Datasets:

ywchoi
/

OpenMedText

+ Pandemic diseases of this century have differentially targeted healthcare workers globally. These infections include Severe Acute Respiratory Syndrome SARS, the Middle East respiratory syndrome coronavirus Middle East respiratory syndrome coronavirus (MERS-CoV) and Ebola. The COVID-19 pandemic has continued this pattern, putting healthcare workers at extreme risk. Just as healthcare workers have historically been committed to the service of their patients, providing needed care, termed their “duty of care”, so too do healthcare employers have a similar ethical duty to provide care toward their employees arising from historical common law requirements. This paper reports on results of a narrative review performed to assess COVID-19 exposure and disease development in healthcare workers as a function of employer duty of care program elements adopted in the workplace. Significant duty of care deficiencies reported early in the pandemic most commonly involved lack of personal protective equipment (PPE) availability. Beyond worker safety, we also provide evidence that an additional benefit of employer duty of care actions is a greater sense of employee well-being, thus aiding in the prevention of healthcare worker burnout.Respiratory infectious diseases of this century, beginning with the severe acute respiratory syndrome (SARS) outbreak of 2003, have differentially targeted healthcare workers globally. Among the 8000 affected with SARS in 29 countries, more than twenty percent were healthcare workers [1]. This included both clinically assigned workers as well as those providing non-clinical support functions [2].Likewise, the Middle East respiratory syndrome coronavirus (MERS-CoV) emergence in 2012 was associated with high numbers of healthcare worker infections with many cases associated with ‘super-spreader’ events [3]. Now numbering more than 2400 cases, an additional 219 were reported by the World Health Organization (WHO) in 2019, of which 52 were linked to hospital transmission, with half of those being infections in healthcare workers [4].Prior to these emerging respiratory infections, the importance of nosocomial transmission of tuberculosis and its threat to worker safety was well known [5]. Most recently, the bitter lessons of Ebola, which also targeted healthcare workers differentially, at more than a ten-fold higher rate than community members [6], suggested early in the COVID-19 outbreak that the healthcare workforce was likely at extreme risk. This awareness warrants an examination of worker protections in place to prevent exposure and minimize harm to workers’ health and safety.Although statute-specific legal requirements of employers toward their workers’ safety have only been instituted in their present form in well-resourced countries over the past 50 years [7,8], the broader employer responsibility toward providing “reasonable care” toward the workforce existed much earlier under common law requirements [9]. Specific expectations of employers here include the prevention of illness, injury and death of workers, providing safe systems of work, safety equipment and generally assuring the health, safety and welfare of workers. These previously existing common law responsibilities have been formalized and acknowledged and are now being enforced under specific, legally binding statues in many countries globally.The British Health and Safety at Work Act of 1974 specified in a provision titled: General duties of employers to their employees, that “It shall be the duty of every employer to secure, so far as is reasonably practicable, the health, safety and welfare at work of all his employees” [8]. In this legal framework, the employer is the ‘duty holder’ of this safety and health responsibility. While this duty is somewhat broad, the general approach to worker safety and health protections depends on performing risk assessments for hazards and establishing policies and procedures to minimize harm [10]. There are also specific provisions for different economic sectors.The International Labor Office (ILO) Occupational Safety and Health Convention (C155) of 1981 had earlier challenged member states, in collaboration with employer and worker representatives, to “formulate, implement and periodically review a coherent national policy on occupational safety, occupational health and the working environment” [11]. Indeed, the British Health and Safety at Work Act, described above, took its language from the ILO Convention which states “It shall be the duty of every employer to ensure, so far as is reasonably practicable, the health, safety and welfare at work of all his employees”. It further states that employers must ensure preventive measures are taken for the protection of life, and physical and mental health of workers. This convention and a related 2002 protocol have been ratified by 69 countries [12].The ethical and legal responsibilities to protect the health and safety of employees are also binding in the healthcare sector and even in emergency settings with evolving risk scenarios. The employer’s duty has come to the fore in all the recent respiratory public health emergencies described above.The employer’s duty mirrors that of healthcare workers, who have historically, in professional ethics codes, committed to the service of their patients, providing protections from risk of harm, termed one’s “duty of care” [13]. This has also been expected during pandemics and other emergencies [14]. In the SARS outbreak of 2003, with the heavy toll of illness and death among healthcare workers, some authors suggested a reciprocal obligation of employers toward their employees to make their added risk-taking as safe as possible [15,16]. This could be achieved, it was proposed, through worker training, performing risk assessments and providing the personal protective equipment (PPE) workers needed. All these actions are the elements of the common law and formal statutes describing the employer’s ‘duty’ to provide for the health and safety of their employees.Ethical responsibilities of the employer under their duty of care include clearly communicating expectations to staff during a pandemic and providing psychosocial support, sufficient skills training and required resources, especially PPE. Additional prevention services, such as vaccines when available and medical care if illness results, should also be provided [17].The WHO/ILO document “Occupational Safety and Health in Public Health Emergencies” lays out similar employer responsibilities, which include providing:Safe working conditions (performing risk assessment and management planning) andInformation for workers (skills training regarding PPE).Safe working conditions (performing risk assessment and management planning) andInformation for workers (skills training regarding PPE).In addition, employers are bound to report occupational illness and injury statistics to government agencies. Worker responsibilities include participating in training, following OSH (occupational safety and health) guidelines and reporting hazards to the employer [18]. Of interest, workers also have the right to refuse unsafe work.WHO also has COVID-specific resources that add detail to the above general approaches to employer and worker responsibilities [19]. These include calling for “healthy, safe and decent” working conditions for workers caring for COVID-19 patients [20] and the basic elements of a comprehensive safety program including infection control (IC) and other safe work practices (PPE use and psychosocial support).Ideally, there should be a balance between employer and healthcare worker duty of care to ensure optimal patient safety and worker protections. The objectives of this paper are to describe the effect of employer duty of care and occupational safety and health (OSH) responsibilities on COVID-19 disease outcomes in the healthcare workforce. Repeating themes identifying the presence of specific duty of care domains of responsibility or deficiencies in these domains will be reported where possible, related to healthcare worker outcomes. Finally, the synergies between a robust duty of care culture and worker well-being will also discussed.A narrative review was performed to assess COVID-19 exposure and disease development in healthcare workers as a function of employer duty of care program elements deployed in the workplace. Peer-reviewed articles as well as public news sources were examined using search terms including “employer duty of care in COVID-19” or “duty of care in the COVID pandemic”, in both PubMed and Google search engines. These articles and sources were culled to address only health sector employers including acute, long-term care and home care settings. Specific duty of care elements provided by, or which failed to be provided by the employer were identified. Search dates included January 2020 through January 2021. An additional query was also performed in PubMed to examine the literature by identifying “risk factors for COVID development in healthcare workers” with various spellings of COVID or COVID-19 included.The narrative review was restricted to publications from January 2020 through January 2021, and thus primarily illustrates the first wave of the pandemic crisis and its immediate aftermath through to the early rise of the second wave in the fall of 2020. This review required two different search queries to obtain sufficient output to assess employer duty of care in the context of healthcare worker disease outcome.The first query specifically used the term ‘duty of care’ and yielded 128 papers, of which only ten were specific to the healthcare sector and sufficiently detailed to contribute examples. The second queried for ‘risk factors for health worker COVID infection’. This resulted in 46 citations, of which 12 were contributory to assessing facility-based employer duty of care safety program elements, which could then be aligned with duty of care domains. All of these papers were from the peer-reviewed literature.Two additional non-peer-reviewed references included a longitudinal report by a health news service on the COVID crisis in the healthcare workforce, largely covering the US and UK experience, and a serially updated report from a global non-governmental organization (NGO) covering the same topic internationally.The narrative literature review query using the ‘risk factors’ term yielded a richer description of working conditions and duty of care domains present in healthcare facilities than did the query using the term ‘duty of care’. The latter query yielded more general discussions of employer legal or ethical requirements.Early in the pandemic, comprehensive reporting of healthcare worker COVID-19 illness and death rates was not easily available. It was not until September 2020, six months after the onset of the pandemic, that the WHO reported that healthcare workers represented 14% of the COVID-19 disease burden globally [21,22]. This was in line with the experience of hard-hit countries during the outbreak’s first wave, such as Spain, Italy and the US The US reported that 18% of cases occurred in healthcare workers early in the pandemic [23]. Additionally, early on, between 28 February 2020, and 23 April 2020, Spain reported that 20.4% of its cases had occurred in healthcare workers, and during the same time period Italy reported a 10.7% rate of infection among healthcare workers [24,25].To assess comparative risk to the healthcare workforce, a prospective, observational cohort study in the UK and the US enrolled healthcare workers and the general community during a thirty-day period between late March and late April 2020, using the COVID-19 Symptom Study smartphone application. The investigators found that healthcare workers were at an increased risk of reporting a positive COVID-19 test of at least 3 fold compared to community reporting (adjusted HR = 3.4, 95% CI 3.37–3.43) [26]. The authors noted that the excess was especially high among Black, Asian and minority ethnic healthcare workers and in workers reporting direct patient contact, lack of PPE, or in those who were required to reuse PPE. Although this study was based on self-report, the authors noted that any misclassification bias would have been non-differential and that the findings were sufficiently robust in subsequent sensitivity and secondary analyses.In a study from the Minnesota Department of Health of over 21,000 COVID-19 exposures in 17,000 healthcare workers reported to a state surveillance program, 66% involved patient care events and 34% involved non-patient events, although much of this latter group included co-workers exposed by an infectious colleague at work [27]. For higher-risk exposures, including aerosol-generating procedures and proximity or longer exposure durations to a positive person, PPE was statistically less likely to have been used by workers in long-term care or congregant settings, as opposed to acute care settings. Mask use for source control by positive patients or residents was also less likely in long-term care and congregant settings. Staff was also more likely to work while experiencing symptoms when in the long-term care setting [27]. Note was made of some deviations from state guidance in risk assessments of exposure events performed in some facilities (an administrative policy or training issue). Limits in infection prevention capacity and training in long-term care, which has been reported elsewhere in these settings, were also mentioned [28].In September of 2020, using multiple sources (memorial pages, government figures, lists compiled by national medical associations, and lists and obituaries published in media around the world), Amnesty International, through their international network, reported that over 7000 healthcare workers had died from COVID-19. As of that date, they reported at least 1320 healthcare worker deaths in Mexico, the highest among countries at the time, followed by the US (1077) and Brazil (643). Also listed were concerning numbers from South Africa (240) and India (573), reflecting recent raised infection rates in those countries. The agency compared this September 2020 figure to their previously reported 3000 healthcare worker deaths through mid-July [29].This account also highlighted many duty of care deficiencies found in COVID-19 death investigations, with workers raising safety concerns. These included inadequate provision of PPE and lack of psychosocial support. They also noted that contracted workers were at increased risk due to less protection observed through employment outsourcing, which may blur lines of which employer has ‘duty holder’ responsibility [29].In a systematic review by Gholani and colleagues [30], risk factors for COVID-19 infection in healthcare workers included the lack and re-use of PPE, but also other work practice deficiencies, including sub-optimal hand hygiene by workers (a training issue), failure to place a mask on a suspect patient and workers not wearing masks even when provided. Also associated with healthcare worker COVID-19 disease was workplace setting (locations where aerosol procedures occur and areas where intensity of exposure to COVID-19 patients is high), including nursing homes [26].The Kaiser Health Network, which has followed this story since the beginning of the pandemic, together with The Guardian, reported in October 2020 the deaths of 2900 US healthcare workers. These included both acute care, nursing home (long-term care) and home care workers. Sources of the report were a review of governmental and public data sources, labor union websites, posted obituaries, as well as required reporting to the government of worker illness in long-term care and home care. In more than 300 interviews with family and others who knew the deceased, concerns about lack of PPE were identified in a third of the cases [31].While the major focus on healthcare worker COVID-19 illness has been on acute care facilities, as the Kaiser report reflects, other environments also threaten the health of workers providing both health and personal care, principally in long-term care (LTC) and home care settings. Surprisingly, while the tragedy of COVID-19 illness and death among elders in LTC has been reported globally, in the US, workers account for close to half (46%) of COVID-19 nursing home infections but only about 10% of deaths according to the US Centers for Medicare and Medicaid Services [32].The high-risk environment in LTC has been linked to several factors, including lack of access to testing early in the pandemic and contact tracing constraints, but also to more long-standing issues of staffing and PPE shortages. Chronic staffing shortages have led to the use of healthcare personnel from staffing agencies, or shared staff between co-owned facilities, permitting infection to be shared. Additionally, PPE shortages are common and recurring. A recent study from a national COVID-19 nursing home database of over 15,000 facilities found about 20% reported severe PPE shortages during the pandemic, with a similar share reporting significant staff shortages in nursing and other staff [33]. The authors point out that both these types of shortages threaten the health of residents and staff alike.One could also argue that such shortages are preventable and suggest a duty of care deficiency. In the US, both Centers for Disease Control and Prevention (CDC) and other governmental organizations have issued guidance to assist in the prevention and management of outbreaks among LTC residents and staff [34]. Many of these recommendations were not innovative but based on long-standing IC practices and reasonable employment strategies, such as providing sick leave, so ill workers could stay home when sick [35]. These examples suggest structural deficiencies that could be remedied to prevent COVID-19 spread.While the LTC worker is far from the relatively more robust safety structures of fixed-site, acute care organizations, the home care worker is even farther away. These healthcare workers provide both personal care, such as bathing and dressing, and medical care, such as taking vital signs and providing wound care, to community-living clients who require assistance [36]. The unique risks to this group include caring for multiple clients as well as their longer duration of ‘exposure’ to the client given their job duties of personal care, compared to acute care. While some authors use the term ‘vulnerable’ for these workers [37], indeed their working conditions are closer to precarious work, that is, non-standard, insecure, unprotected and outside a typical employment relationship [38], when factoring in other social determinants of health commonly encountered in this setting. These include low-wage workers of color, generally working for small companies and providing care for home-bound frail elderly and sick clients. It must be noted, however, that during the pandemic, these workers were the protective bulwark of hard-hit areas in the US, such as New York and New Jersey, keeping their clients out of already critically crowded emergency departments in hospitals. Nevertheless, in a survey of 33 home health workers, a familiar pattern of PPE shortage was found, as well as sometimes inadequate training [36].A summary of duty of care deficiencies reported in healthcare worker COVID-19 exposure and infection studies derived from the narrative review, including those discussed above, are displayed in Table 1. These reports are illustrative of five repeating themes recounted in the literature that threatened worker health and safety.The first major heading in the table discusses the employer’s responsibility to plan for safe work. This was described in the introduction as a key responsibility of the safety duty holder. As seen in the table, there was a widespread lack of planning even for IC protocols which should have been a more well-recognized employer responsibility given the ubiquity of IC activities even in conventional healthcare delivery.The overwhelming deficiency reported was the failure to provide material resources for patient care and worker safety, specifically the lack of PPE. This led to unsafe re-use or use of lesser forms of protection. Another major deficiency commonly reported was the lack of training for both safe PPE use and infection prevention. Examples include failure to train on masking a source patient and on optimal hand hygiene.Additionally, the ability to provide adequate staffing and manageable workloads, which has been repeatedly raised as a determinant of positive working conditions in the context of conventional care, was even more challenged during the COVID patient surge. Finally, related to positive working conditions was the workers’ perception of psychosocial support received from the organization. Each of these five broad areas of responsibilities, displayed in the table, are basic expectations of employer duty of care. Each of these five broad areas of responsibilities, displayed in the table, are basic expectations of employer duty of care, and were repeatedly cited as lacking in the articles reviewed.The basic elements of an employer’s duty of care toward the healthcare workforce are certainly present in their IC and broader safety obligations and so are not unknown to the healthcare sector employer. However, often the focus of such IC and safety provisions patient safety driven and does not encompass the safety of the workforce, the larger organization or the system [63]. Moreover, such safety programs typically address these obligations during conventional rather than contingency or crisis levels of care. Even though much has been written recently about emergency preparedness generally and pandemic readiness specifically, resource constraints compete with investments in harm avoidance, the benefits of which are in the future and perhaps too vague to appreciate.One recent example of the benefits of an employer’s commitment to duty of care taken from the recent Ebola outbreak of 2014–2015 may prove instructive here. The largely WHO-led response to the outbreak in West Africa resulted in approximately 815 worker deaths, and a disease incidence rate of 30–44 per 1000 healthcare worker responders [6]. This is compared to the Ebola incidence rate of Médecins Sans Frontiéres (MSF) healthcare workers of about 4.3 per 1000 [64].A significant difference between the two groups’ experience was MSF’s agency-wide commitment toward employer duty of care. Operationalized through new site risk assessments and structured policies beyond IC, MSF attempts to prevent exposure through safety and emergency procedures design, training and risk communication with workers and with follow-up of worker injury and illness [65]. This commitment to safety likely contributed to the almost ten-fold lower worker incidence rate in the MSF teams compared to those of WHO. The elements of the MSF program include duty of care guidelines and legislation internationally accepted.Thus, beyond a high-minded ethical notion of justice, the employer’s duty of care can demonstrate tangible results in the form of life-saving differences in infection and death rates. This protects not only the healthcare work force but a facility or healthcare system’s response mission during pandemic emergencies.While the employer’s duty of care focus is on preventing foreseeable risk of harm, beyond mere harm avoidance, another benefit of a duty of care disposition arises from the healthcare worker burnout prevention experience.The poor state of healthcare workers’ mental health has been widely reported in the literature and the popular press recently with much of it attributed to worker burnout [66,67]. Burnout here is defined as a state of exhaustion, cynicism and inefficiency driven by factors such as workload and job demands, resource constraints, and misalignment of organizational culture and values [68]. Burnout has also been linked to physician turnover, and to declines in care quality and patient safety [67,69].Burnout is not new and has been widely reported, long before the pandemic [70], with many of its drivers being attributed to raised expectations of healthcare worker productivity, efforts limiting costs of care and excessive documentation requirements of the electronic health record. As many current reports illustrate, these same factors are evident and amplified in the COVID-19 pandemic, with the surge of patient demands for care in unprepared healthcare systems.Using a questionnaire to explore the core burnout domain of emotional exhaustion, a cross-sectional study of healthcare workers caring for COVID-19 patients found that about half of 2700 participants from 60 countries reported burnout. Factors associated with burnout included ‘work impacting household activities’ (time pressures), feeling pushed beyond training, exposure to COVID-19 patients and making ‘life-or-death’ decisions. One factor found to be protective against burnout was adequate PPE [42]. In a British study, symptoms of moderate to severe burnout were reported more frequently by healthcare worker participants if they were younger, female, redeployed from a usual assignment (possibly outside their scope of training or familiarity) and if working on a COVID-19 unit [55].Examples of factors driving burnout in COVID-19 align broadly with domains of burnout risk in general, including organizational culture and values misalignment; workload and job demands; and lack of control, inefficiency and resource constraints. Many of these domains are employer duty of care responsibilities.The similarities in origins of burnout suggest there may be shared solutions from the burnout prevention literature applicable to the COVID-19 context. First, burnout is increasingly recognized not as a diagnosis of the individual worker, but of a work organization [71]. Hence, if it derives from the way work is organized, then the solution must come from the organization.Actions by organizational leadership to address burnout include realigning organizational culture and values and promoting supportive communities at work [68]. Specifically, this must include organizational adjustment of productivity expectations; use of support staff to maximize clinician efficiency; an examination and commitment to culture, values, safety and equity; and enhanced flexibility of schedules. These same approaches contribute significantly toward meeting an employer’s duty of care. Organizational management and psycho-social support of staff and the visibility and accessibility of leadership were repeatedly reported as promoting staff well-being in the COVID context [45,47,52,56,57].The most visible evidence of an employer’s active support of staff physical well-being during COVID-19 was providing availability of adequate PPE [72]. For example, Morgantini reported that the provision of PPE was protective against burnout in her large study of COVID responders [51]. However, Table 1 shows that the most commonly reported and concrete employer duty of care failure was lack of PPE.Not only is providing PPE a highly visible demonstration of an employer’s organizational commitment to safety [63,73], but it has also been found to be an influential determinant of safety culture confidence and of healthcare workers’ willingness to report for duty in a pandemic [73]. However, the mere availability of PPE is not sufficient protection without adequate training on its use.Addressing the healthcare sector employer’s duty of care obligations presents challenges, even in conventional circumstances of healthcare operation, requiring focused effort on safe work policies, providing material resources for patient care and staff protection, on-going safety training, adequate staffing, pay and benefits and attention to staff psycho-social needs. Investing in such efforts, however, provides significant benefits realized not only with respect to staff safety improvements but also in worker well-being, which assists in preventing burnout. Such a duty of care approach establishes an invaluable foundation upon which to build critical response functions during public health emergencies, preserving the healthcare system’s infrastructure and mission while sustaining resiliency in the healthcare workforce.Conceptualization, M.M.; methodology, M.M,; analysis, M.M., J.G.; investigation, M.M.; original draft preparation, M.M., M.C., J.G.; review, M.M., M.C.; editing, M.M., M.C., J.G.; supervision, M.M. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.No new data were created or analyzed in this study. Data sharing is not applicable to this article.The authors declare no conflict of interest.Selected examples of duty of care failures in COVID protections.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06016.txt ADDED Viewed

	@@ -0,0 +1,8 @@

+Classifications of impacted teeth allow defining the type and degree of retention, as well as assessing the degree of difficulty of the procedure. The aim of this study was to conduct retrospective analysis of the degree of retention and difficulty in the surgical removal of impacted mandibular third molars in the clinical material of the Department of Oral Surgery in 2013–2018. This study included 1585 dental panoramic radiographs of patients of the Department of Oral Surgery, who reported in 2013–2018, in order to perform surgical removal of the impacted mandibular third molar. Based on dental panoramic radiographs, the degree of retention was determined based on classifications according to Winter, according to Pell and Gregory, according to Tetsch and Wagner, and according to Asanami and Kasazaki. The difficulty of the procedure was also assessed based on the Pederson index. The most common types of lower wisdom tooth impaction are as follows: in Winter’s classification, mesial-angular impaction; in Tetsch and Wagner’s classification, oblique medial-angular impaction; in Pell and Gregory’s classification, impaction grade 2A; and in Asanami and Kasazaki’s classification, 3A and anterior inclination. In most cases of surgical removal of an impacted tooth, the anticipated difficulty of the procedure was rated as very difficult.Tooth impaction is one of the most common abnormalities of tooth position [1]. An impacted tooth (dens retens) is a tooth with a fully formed root, with complete development, which is partially or completely covered by hard and/or soft tissues, being outside the physiological period of eruption [2]. The procedure for surgical removal of impacted wisdom teeth is routinely performed in a dental surgery. Published studies have reported prevalences of impacted teeth that have ranged from 6.9 to 76.6% [3,4]. The most common of all teeth to be impacted are third molars, especially in the mandible [5,6]. They are among the most commonly impacted teeth next to the maxillary canine and mandibular second premolar [2].The causes of third molar impaction can be divided into general and local [7,8]. In recent decades, an increase in the phenomenon of impaction has been noted. The explanation for this trend seems to be an increasing level of hygiene, as well as less frequent tooth loss and the influence of the lack of physiological tooth attrition due to changes in dietary habits [8]. The genetic etiology of third molar impaction plays an important role in the process of odontogenesis.The MSX1 and AXIN2 genes may be considered as components of the genetic background, characterized by variable expressivity. As such, they promote tooth impaction in circumstances where specific environmental factors coexist and depending on the presence of other modulating genetic factors, additionally increasing the risk and stimulating the onset of abnormalities and/or increasing the degree of phenotypic manifestation of symptoms—such as the number of impacted teeth in the carrier [9]. Two-dimensional panoramic radiographs found that the initial mineralization time of mandibular third molar germs is 8–9 years old. The time of initial mineralization is individually variable [10]. The time of eruption of mandibular third molars is most often 17–20 years old [11]. The surgical removal of impacted wisdom teeth is a difficult procedure due to their different spatial locations and relationships with the surrounding anatomical structures. It is associated in most cases with the occurrence of perioperative complications, as well as with a significant reduction in the patient’s quality of life in the postoperative period [12]. Adequate assessment of the spatial location of the wisdom tooth and the difficulty of the procedure, as well as knowledge of the potential surgical complications associated with intervention into the maxillary bone tissues, is essential in planning the procedure.The most commonly used method of radiographic examination in preoperative diagnosis to determine the position of the wisdom teeth is the panoramic radiograph [10,13]. X-ray diagnostics allows a proper diagnosis to be made, as well as establishing the methodology for clinical management. Due to the limitations of panoramic radiography, cone beam computed tomography is increasingly used [14].Classifications of impacted third molars allow us to determine the degree of impaction and determine the best methodology for the surgical procedure. Planning the procedure based on the subject, physical examination, and additional investigations such as radiographs makes it possible to reduce the risk of complications. The authors of this publication use various classifications. Each of them is characterized by certain limitations. In the literature and clinical practice, authors usually use the following classifications: Winter, Tetsch and Wagner, Pell and Gregory, Asanami and Kasazaki [15,16,17,18].There are many studies available in the literature presenting the problem of impacted mandibular third molars. The description of their impaction in most studies deals only with one classification, without considering the influence of the spatial position in relation to the occlusal plane line and in relation to the anterior margin of the mandibular ramus on the difficulty of the surgical procedure.The inspiration to research the issues in this paper was the lack of studies regarding the assessment of the type of impaction and the degree of difficulty of the procedure discussing the problem among an East Baltic population. This topic is important due to the increasing prevalence of surgical removal of impacted third molars in the mandible. Classifications of the impacted teeth allow defining the type and degree of retention, as well as assessing the degree of difficulty of the procedure. Together with the operator’s experience, this affects the risk of complications associated with the procedure that may be minimized by targeted elective procedures of surgical removal of the impacted wisdom tooth. The aims of this study were as follows:To assess the pattern of impacted mandibular third molars in an East Baltic population;To assess the difficulty of surgical removal of impacted mandibular third molars in an East Baltic population.To assess the pattern of impacted mandibular third molars in an East Baltic population;To assess the difficulty of surgical removal of impacted mandibular third molars in an East Baltic population.The analysis of radiographic images of impacted third molars in the mandible was based on pantographic radiographs taken digitally with a Cranex3Dx camera (Soredex, Tuusula, Finland) with a magnification ratio of 1:1.19. The analysis of digital pantographic images was performed using the Scanora 5.2.6 computer program (Soredex, Tuusula, Finland). The long axes of the second molar, the impacted third molar, the occlusal plane, the greatest width of the crown of the impacted molar, the tangent line to the anterior edge of the mandibular ramus, and the distal surface of the second lower molar were determined on the pantographic images. Based on the above data, the impaction degree of the impacted tooth was determined according to the classifications by Winter, Tetsch and Wagner, Pell and Gregory, and Asanami and Kasazaki, and the difficulty of the procedure was determined according to Pederson.The following lines were plotted to determine the impaction of the impacted third molar:Long axis of the tooth (a);Occlusal plane (A);Tangent to the anterior margin of the mandibular ramus (B).Long axis of the tooth (a);Occlusal plane (A);Tangent to the anterior margin of the mandibular ramus (B).The occlusal plane line (A) is a straight line passing through the buccal cusp of the first mandibular premolar and the proximal buccal cusp of the second lower molar [13] (Figure 1).The mandibular ramus anterior margin line (B) is a line tangent to, and parallel to, the greatest concavity of the mandibular ramus anterior margin [14] (Figure 1).To determine the long axis of the tooth (a), two points were determined: one at the midpoint of the greatest width of the tooth crown (b), and the other at the midpoint of the width of the tooth neck (c). A straight line drawn through the above points determined the long axis of the tooth [19] (Figure 2).G. B. Winter documented impaction types based on angulation—the inclination of the crown of an impacted third molar—concerning the angle formed between the long axes of the second and third lower molars [15,20].
+          Vertical impaction—the long axes of the second molar and the impacted third molar are parallel;Mesioangular impaction—the long axes of the second molar and the impacted third molar are coincident docoronally;Distal-angular impaction—the long axes of the second molar and impacted third molar are convergent apically;Horizontal impaction—the long axes of the second molar and impacted third molar are at right angles;Buccolingual impaction—each tooth is oriented in a buccolingual direction;Inverted impaction;Other orientation.Vertical impaction—the long axes of the second molar and the impacted third molar are parallel;Mesioangular impaction—the long axes of the second molar and the impacted third molar are coincident docoronally;Distal-angular impaction—the long axes of the second molar and impacted third molar are convergent apically;Horizontal impaction—the long axes of the second molar and impacted third molar are at right angles;Buccolingual impaction—each tooth is oriented in a buccolingual direction;Inverted impaction;Other orientation.This classification determines the degree of impaction of the third molar in the vertical and horizontal dimensions. It states the degree of impaction concerning the occlusal plane: A, B, C (vertical dimension), and the mandibular ramus: 1, 2, 3 (horizontal dimension) [16,21].Position concerning the occlusal plane:A: The occlusal surface of the third lower molar is either above or at the level of the occlusal plane;B: The occlusal surface of the third lower molar is between the occlusal plane and the neck of the second molar;C: The occlusal surface of the third lower molar is below the neck of the second molar.Position concerning the anterior margin of the mandibular ramus:Class 1: The distance between the distal surface of the second molar and the anterior margin of the mandibular ramus is greater than the anteroposterior dimension of the crown of the third lower molar;Class 2: The distance between the distal surface of the second molar and the anterior margin of the mandibular ramus is less than the anteroposterior dimension of the crown of the third lower molar;Class 3: The complete absence of space between the distal surface of the second molar and the anterior margin of the mandibular ramus.This classification determines the form of impaction of the third lower molar based on the angle of the long axis of this tooth to the occlusal plane [17].
+          Vertical impaction—the tooth is aligned parallel to the adjacent molars, at an angle of approximately 90° to the occlusal plane.Horizontal impaction—the tooth is aligned parallel to the occlusal plane, at an angle of approximately 0°; depending on the direction of the position of the crown of the tooth, an additional distinction is made.Sagittal impaction:Medial-angular (the impacted third molar faces the occlusal surface to the second molar);Distal-angular (the impacted third molar faces the occlusal surface to the anterior edge of the mandibular ramus).Cross-impaction:Buccal-angular (impacted third molar faces the occlusal surface to the buccal side);Lingual-angular (impacted third molar faces the occlusal surface toward the lingual side).Oblique impaction—the tooth is inclined concerning the occlusal plane in different variants, depending on the course of the long axis of the tooth concerning the occlusal plane from 0 degrees to 90 degrees.Medial-angular;Distal-angular;Lingual-angular;Bucco-angular;Displacement impaction.Vertical impaction—the tooth is aligned parallel to the adjacent molars, at an angle of approximately 90° to the occlusal plane.Horizontal impaction—the tooth is aligned parallel to the occlusal plane, at an angle of approximately 0°; depending on the direction of the position of the crown of the tooth, an additional distinction is made.Sagittal impaction:Medial-angular (the impacted third molar faces the occlusal surface to the second molar);Distal-angular (the impacted third molar faces the occlusal surface to the anterior edge of the mandibular ramus).Medial-angular (the impacted third molar faces the occlusal surface to the second molar);Distal-angular (the impacted third molar faces the occlusal surface to the anterior edge of the mandibular ramus).Cross-impaction:Buccal-angular (impacted third molar faces the occlusal surface to the buccal side);Lingual-angular (impacted third molar faces the occlusal surface toward the lingual side).Buccal-angular (impacted third molar faces the occlusal surface to the buccal side);Lingual-angular (impacted third molar faces the occlusal surface toward the lingual side).Oblique impaction—the tooth is inclined concerning the occlusal plane in different variants, depending on the course of the long axis of the tooth concerning the occlusal plane from 0 degrees to 90 degrees.Medial-angular;Distal-angular;Lingual-angular;Bucco-angular;Medial-angular;Distal-angular;Lingual-angular;Bucco-angular;Displacement impaction.This classification describes the degree of impaction of the third lower molar in both the vertical and horizontal dimensions concerning the angulation of this tooth. The degree of impaction of the tooth in the vertical dimension is determined by A, B, C, the horizontal dimension, and the inclination of the long axis of the third lower molar relative to the long axis of the second lower molar (1–3, 1–4) [18].
+          Vertical position:Vertical position:A—Minor impaction;B—Average impaction;C—Deep impaction.1—The distance between the distal surface of the second lower molar and the anterior margin of the mandibular ramus is greater than the horizontal dimension of the crown of the impacted tooth;2—The distance between the distal surface of the second lower molar and the anterior edge of the mandibular ramus is equal to the horizontal dimension of the crown of the impacted tooth;3—The distance between the distal surface of the second lower molar and the anterior edge of the mandibular ramus is less than the horizontal dimension of the crown of the impacted tooth.
+          2.Distal inclination:Distal inclination:A—Minor impaction;B—Average impaction;C—Deep impaction.1—The distance between the distal surface of the second lower molar and the anterior margin of the mandibular ramus is greater than the horizontal dimension of the crown of the impacted tooth;2—The distance between the distal surface of the second lower molar and the anterior edge of the mandibular ramus is equal to the horizontal dimension of the crown of the impacted tooth;3—The distance between the distal surface of the second lower molar and the anterior edge of the mandibular ramus is less than the horizontal dimension of the crown of the impacted tooth.
+          3.Anterior inclination:Anterior inclination:A—Minor impaction;B—Average impaction;C—Deep impaction.1–4—The degree of inclination of the long axis of the impacted tooth of the third molar relative to the long axis of the second lower molar, depending on the value of the acute angle, is classified in degrees from 1 to 4. The authors Asanami and Kasazaki do not provide a reference value of the angle that determines the degree of inclination of the preceding in each point from 1 to 4.
+          4.Horizontal position:Horizontal position:A—Minor impaction;B—Average impaction;C—Deep impaction.1—The distance between the distal surface of the second lower molar and the anterior margin of the mandibular ramus is greater than the horizontal dimension of the crown of the impacted tooth;2—The distance between the distal surface of the second lower molar and the anterior edge of the mandibular ramus is equal to the horizontal dimension of the crown of the impacted tooth;3—The distance between the distal surface of the second lower molar and the anterior edge of the mandibular ramus is less than the horizontal dimension of the crown of the impacted tooth.
+          5.Horizontal lingual position;6.Lingual inclination;7.Horizontal buccal position;8.Buccal inclination;9.Inverted position.Horizontal lingual position;Lingual inclination;Horizontal buccal position;Buccal inclination;Inverted position.The evaluation of the difficulty of surgical removal of an impacted third molar in the mandible was performed using the Pederson index [22,23]. Based on the impaction according to Winter, the spatial position of the impacted teeth was defined as mesial-angular alignment, horizontal/reversed alignment, vertical alignment, and distal-angular alignment. Concerning the occlusal plane (A, B, C) and the distance of the distal surface of the second molar from the tangent line to the mandibular ramus (1, 2, 3), the position of the impacted tooth was determined according to Pell and Gregory’s classification. Each parameter was assigned a score: mesioangular alignment—1 point, horizontal/reversed alignment—2 points, vertical alignment—3 points, distal-angular alignment—4 points, A—1 point, B—2 points, C—3 points, 1—1 point, 2—2 points, 3—3 points. The difficulty index was determined based on the sum of the scores obtained in the analysis of the pantographic examination. Accordingly, the degree of difficulty of surgical removal of the lower third molar was defined as slightly difficult for scores of 3–4 points, moderately difficult for 5–6 points, and very difficult for scores of 7–10 points.Statistical analysis was performed using the statistical package R-version 3.5.2 [24]. Quantitative variables were described using standard measures of variability and location such as the arithmetic mean with standard deviation, quartiles: Q1, Q3, and median: Q2.We analyzed 1583 pantographic images of patients qualified for surgical removal of an impacted third molar in the mandible with local anesthesia (n = 1583). A total of 63.04% were female (n = 998), and 36.96% were male (n = 585). The mean age of the subjects was 26.95 years (±8.48), and the age groups were selected so that the number of subjects in each group was similar. A total of 50.16% (n = 794) were left and 49.84% (n = 789) were right lower wisdom teeth. Complete impaction was diagnosed in 56.79% (n = 899) of cases, asymptomatic partial impaction in 30.45% (n = 482), and symptomatic impaction in 12.76% (n = 202). In most of the surgical removal procedures of the impacted teeth, the anticipated difficulty of the procedure was rated as very difficult (39.54%). Detailed results are shown in Table 1 and Table 2.In Winter’s impaction classification, the most numerous group was third wisdom teeth in the mesial-angular position (n = 832), and a slightly less numerous group was third molars in the distal-angular position (n = 618). Inverted impaction (n = 7) and impaction described as other (n = 1) were the least frequent. A detailed summary of the degrees of impaction according to Winter is shown in Table 3.The most common position of impacted mandibular third molars according to the amount of space between the anterior margin of the mandibular ramus and the second lower molar (1, 2, 3) was distance 2 (70.44%), and the least common was position 3 (10.68%). In the evaluation of the impaction depth (A, B, C), as many as 50.6% of cases showed depth A and only 9.67% showed depth C. Grade 2A was the most abundant impaction grade according to Pell and Gregory (36.26%), with 1C (2.21%) and 3C (1.83%) being the least abundant. A detailed comparison of impaction grades in the study group is shown in Table 4.In most cases, impacted wisdom teeth occurred in the oblique medial-angular position (n = 833). Inverted impaction (0.44%) and impaction described as other (0.06%) were observed the least. The results of the analysis of the number of different degrees of impaction according to Tetsch and Wagner are shown in Table 5.In this classification, anterior inclination was observed most frequently (52.56%), and inverted (0.44%) and impaction described as “other” (0.06%) were observed least frequently. Depth (A, B, C) was found most frequently with A (50.6%), and depth was found least frequently with C (9.67%). The distance of the anterior edge of the mandibular ramus from the distal surface of the second lower molar (1, 2, 3) was evaluated for all impaction types except for anterior inclination, inverted position, and other impaction (n = 840). Among the other impaction types, the most common distance was 3 (39.86%). The position characteristics of lower wisdom teeth are shown in Table 6.Classifications of the spatial location of impacted lower third molars allow us to determine the degree of impaction of the tooth, which allows preoperative determination of the degree of difficulty of the procedure and the best methodology for the surgical removal procedure [25]. In the literature, authors most commonly use the classifications of Winter, Tetsch and Wagner, Pell and Gregory, and Asanami and Kasazaki.Winter’s classification is the most commonly chosen method for spatial assessment of impacted teeth in the literature because of its simplicity of use. It does not require the use of additional measuring instruments, which influences its widespread use in clinical practice.Our study showed that the most common type of impaction, at 52.56%, according to Winter, is mesial-angle impaction. Padhye et al. presented an analysis of pantographic radiography of 1200 subjects, in which 33.33% of the subjects showed mesial-angle alignment according to Winter [26]. Kumar et al. observed the prevalence of mesial-angle alignment in 52.89% of cases in Eritrean residents [27]. These reports were confirmed in 2016 by Nagaraj and co-authors [28] presenting mesial-angle impaction in 47.1% of patients, as well as many other researchers [29,30,31,32]. In the studies of Al-Dajani et al. [33] and Yilmaz et al. [34], vertical impaction was found to be the most common position. The first team showed the occurrence of this impaction in 40.7% and mesioangular impaction in only 7.1% of patients; the second team showed vertical impaction in 53% and mesioangular impaction in 29% of patients. The differences in results may be due to the adoption of an incorrect modification of Winter’s index in the studies of Al-Dajani et al. [33] and Yilmaz et al. [34]. The researchers determined the long axes of the second and third molars to determine the angulation of the impacted molar. The reference point in the measurements was the axis of the second tooth. The angle of deviation of the third tooth axis from the second tooth axis was measured. If the deviation from the reference line was 10° to either side, tooth impaction was defined as vertical. Winter’s [15] original classification, by design, does not give a value for the angle of tooth deviation and the possibility of deviation from the long axis of the second tooth, but only the spatial relationship of the long axes. It may seem that this modification is identical to the angles adopted in the Tetsch and Wagner classification; however, Al-Dajani and Yilmaz measured the amplitude from the axis of the second tooth, contrary to the Tetsch and Wagner classification, which measures the angle between the occlusal plane and the axis of the impacted tooth. Researchers from Honk Kong [4] presented horizontal impaction as the most common type. In their study of 7486 patients, with 42.45% having impacted lower wisdom teeth, they showed the presence of horizontal impaction in 47.45% of the patients analyzed. In our study, the second most common impaction (39.04%) was distal-angular alignment. The results of this study confirm the findings of Goyal et al. [35], who presented the same order of prevalence of each impaction type according to Winter. In a sample size of 700 subjects, of which 40.7% were impacted mandibular third molars, 53.89% of the teeth were in the mesial-angular position, and 20.56% were in the distal-angular position. The same pattern of prevalence of individual impactions was presented by Al-Anqudi et al. [30]. The predominant second most common impaction in the literature is the vertical position [36,37]. The variability in results regarding the incidence of vertical and distal-angular impaction may be related to the often slight degree of deviation of the crown of the impacted tooth from the long axis of the second molar, which is misinterpreted by the authors of the publications as vertical alignment. In our study, impaction was assumed to be defined as vertical only when the long axes of the mandibular second and third molars were parallel, with no possible deviation of their axes. In the literature, authors of publications unanimously present the occurrence of inverted or other impaction as the least frequent, which was confirmed in our study [4,29,30,31,32,36,37].Analogous to Winter’s classification, impacted third molars in the mandible are graded based on molar angulation according to Tetsch and Wagner. Unfortunately, one publication was found in which the above classification was present. A Polish team [10] presented an evaluation of the structure and impaction type of impacted third molars based on the above classification. The study evaluated 100 pantographic radiographs of patients of the Silesian Medical University in Poland considering both upper and lower wisdom teeth. The researchers presented only teeth with vertical impaction, mesial-angular impaction, distal-angular impaction, and displacement. The frequency of impaction degrees was 16%, 40%, 13%, and 1%, respectively. Many authors incorrectly use a pie chart based on the Tetsch and Wagner classification in assessing the degree of impaction according to Winter [34,38].Winter’s and Tetsch and Wagner’s classifications consider only the angulation of the impacted molar. They do not take into account the amount of space in the arch available for the eruption of the wisdom tooth. Pell and Gregory determined the degree of impaction based on two variables—the depth of impaction of the tooth and the space available between the anterior margin of the mandibular ramus and the distal surface of the lower second molar. Assessment of the spatial location according to Pell and Gregory is often presented by authors of previous publications [27,31,34,37,39,40,41]. Abbas and co-authors presented a study on 358 impacted molars in the mandible. They evaluated the spatial position according to Pell and Gregory, and the occurrence of pathologies associated with the impacted tooth. Position A impaction depth was the most common with 43.02%. A total of 65.95% of the teeth were classified as impaction 2 according to Pell and Gregory. The researchers also determined molar angulation; however, they did not provide the methodology of the study or the classification by which it was assessed. Eshghpour et al. evaluated the impaction rate of impacted mandibular third molars in a northern Iranian population [31]. Based on 1397 cases of impacted mandibular third molars, they identified the most common types of impaction. Concerning the occlusal plane, the most common impaction type was B (n = 892; 64.85%). Concerning the mandibular ramus, 2 was the most common (n = 677; 48.46%). Researchers from Pakistan determined the degree of impaction of 100 wisdom teeth in the mandible during a year-long study [37]. The authors used letter and numerical terms interchangeably, which may confuse interpreting their results. They used the letter designation to refer to the position of the tooth relative to the mandibular ramus, and the number designation to refer to the depth of impaction. The most common type of impaction relative to the mandibular ramus was impaction A with 39% (impaction 1 in the original Pell and Gregory classification and in our study); however, the incidence of impaction B (impaction 2 in the Pell and Gregory classification and in our study) was slightly lower with 35%. Concerning the occlusal plane, the most common impaction was class 2—45% (impaction B in the original classification by Pell and Gregory and in our study); however, the difference between classes 2 and 1 was small (42%). The rarest was class 3—13% (impaction C in the original classification according to Pell and Gregory and in our study). The same order of frequency of impaction groups was presented by Prajapati and co-authors [41]. They conducted a study to evaluate the spatial position of impacted third molars in the mandible using Winter’s and Pell and Gregory’s classifications and also the relationship of the occurrence of caries of the second molar adjacent to the impacted tooth. The authors observed the presence of class 1 in the majority of cases (62.60%). The researchers did not observe the occurrence of class 3 impaction, which may be related to the small group (n = 200). The most common depth of impaction in the subjects was B impaction (47.82%). Obiechina et al. in a study based on 473 impacted mandibular third molars presented 54.55% of cases in the A position, 31.92% in the B position, and 13.53% in the C position. Regarding the mandibular ramus, 22.62% of the teeth were in position 1, 60.89% in position 2, and 16.49% in position 3. The study was limited to patients aged 16–45 years [39].The authors divided the impacted teeth into prophylactic and symptomatic teeth. The only parameter for inclusion in the symptomatic teeth group was the presence of pain. As many as 68.29% of the cases were symptomatic teeth. The connection between the degree of impaction and the occurrence of pain was also evaluated. Different results were reported in a study on 732 impacted lower wisdom teeth by Yilmaz et al. [34]. In up to 61%, the impaction depth, according to Pell and Gregory, was level C. The researchers did not provide a class breakdown of the relationship of the tooth position to the mandibular ramus; they only calculated the size of the sinus space in millimeters. The alveolar space is the distance of the anterior edge of the mandibular ramus from the distal surface of the second molar. Unfortunately, this magnitude has not been related to the anteroposterior dimension of the crown of the impacted tooth. An analysis of the current scientific literature leads to the conclusion that the preponderance of research teams only considers the frequency of occurrence of each degree of impaction depth and reference to the anterior edge of the mandibular ramus, without considering the relationship between depth and distance indices. Kumar and co-authors conducted a retrospective study on an Eritrean population (n = 552) in which they evaluated the impaction degree of impacted mandibular third molars based on Winter’s and Pell and Gregory’s classifications [27]. The most common classification was 1A (59.78%), the least common classifications were 3A (0.36%) and 3B (0.36%), and classification 3C did not occur. Similarly, impaction grades according to Pell and Gregory were classified by researchers from Iran [20]. In their study, they evaluated 1165 impacted third molars, of which 64.4% were teeth in the mandible. The most common impaction types in the study were 2A (38.93%) and 2B (17.67%), and the least common was 3A (1.87%).In our study, the most common grade of impaction depth was grade A (50.60%). Concerning the anterior margin of the mandibular ramus, the impacted molar was most common in position 2 (70.44%). Additionally, the association between depth and distance was also considered. The most common impaction grades were 2A (36.26%) and 2B (28.55%). Type 3C was the least common (1.83%).Due to the limitations of the classifications described above, Asanami and Kasazaki in 1990 created an impaction division that takes into account both angulation and the spatial position of the molar relative to the occlusal plane and concerning the anterior margin of the mandibular ramus. Unfortunately, Asanami and Kasazaki’s classification is used very rarely in the literature. Only one publication on antibiotic therapy in the prevention of postoperative complications in third molar surgery by a research team from Krakow [42] classified lower third molars according to Asanami and Kasazaki.In our study, the results of the classification of impacted wisdom teeth in the mandible using the Asanami and Kasazaki divisions were presented. The most common grades of impaction were grade 3 distance (84.92%), depth A (50.60%), and anterior inclination (52.56%).Despite the analogy of classification according to Asanami and Kasazaki and according to Pell and Gregory, the frequency of each impaction distance class was different. This is because the interpretation of numerical designations 1–3 according to Asanami and Kasazaki is different from Pell and Gregory. The frequency of individual distances also varies because the above classification does not evaluate distances for teeth in anterior inclination.The degree of difficulty according to Pederson takes into account both the angulation of the molar and its spatial position relative to the mandibular ramus and the occlusal plane. The analysis of impaction degrees in the above-mentioned classifications already allows a preliminary assessment of the difficulty of the procedure. The numerical value of the Pederson index is the highest for the distal-angle position according to Winter (4 points), depth C (3 points), and distance 3 (3 points). With the presence of only one of the parameters in the highest value, extraction will be at least moderately difficult. In our study, we were able to draw preliminary conclusions already from the analysis of impaction degrees that extraction of third molars is more difficult in women. They had significantly more distal-angular impactions than men. It was confirmed by statistical analysis that the procedures were significantly more difficult in women than in men.Park [43] evaluated the difficulty of surgical removal of third lower molars using the Pederson scale and the subjective feeling of the operator. They evaluated 762 impacted teeth scheduled for surgical removal between 2009 and 2014. Molar angulation according to Winter and impaction type according to Pell and Gregory were assessed, and the results were converted to a numerical score on the Pederson scale. A total of 45.53% of women participated in the study. The procedures were more difficult in men. There was no significant difference in the predicted difficulty of the procedure between the sexes. The authors of the study noted the differences in procedure difficulty as assessed by the operator and the Pederson index value. The study was performed on a group almost 2.5 times smaller than in our study, which may have influenced the different results. According to the author of the study, the most difficult procedures were in the group over 50 years old. The treatments were easiest in the youngest patients, between 10 and 19 years of age [43]. In the literature, authors have reported a relationship between the position of the impacted mandibular third molar and mandibular fractures. Kumar et al. demonstrated in a study of 64 subjects that a deeper position of the impacted tooth led to an increased risk of angle fracture and decreased risk of condylar fracture [44]. Constantinides et al. conducted a study on the effect of the type of anesthesia on postoperative complications concerning IAN. Nowadays, most mandibular third molar removal procedures are performed under local anesthesia. Only atypical position, extreme difficulty of the procedure, location in spaces that are difficult to access intraorally, and mental illness or anxiety are indications to perform the procedure under general anesthesia. In our study, all patients met the criteria for having the surgical procedure under local anesthesia [45].This study evaluated the predicted difficulty of surgical removal of impacted third molars in the mandible by age group. It was shown that the procedures were significantly easier in subjects under 20 years of age than in others (p < 0.001). Our study shows that the degree of difficulty increases with age until the age of 30 years (46.52% of very difficult procedures in the group of 26–30 years) and then slightly decreases in patients over 30 years of age (43.14%). It should be noted that statistically, the difficulty of the procedure decreases after the age of 30, but this is only based on the spatial location of the impacted tooth. Other factors such as the density of the bone matrix around the impacted third molar, its degree of ankylosis, and the patient’s weight also influence the clinical judgment, which is based on the operator’s experience [46,47,48]. This study has some limitations. Assessment of the position of the tooth based on a 2D panoramic image is not as accurate as that based on a 3D examination. In the next stages of this study, we plan to determine the dependence of gender and age on the position of the impacted tooth as well as comparing the position on 2D and 3D images. The development of 3D diagnostics, 3D printing, and computer methods may allow us to create computer programs that, after marking the measurement points, will determine the type of retention. Additionally, planning based on 3D imaging may allow creating surgical templates to help and allow for precise planning of the surgical procedure.The most common types of lower wisdom tooth impaction are as follows: in Winter’s classification, mesial-angular impaction; in Tetsch and Wagner’s classification, oblique medial-angular impaction; in Pell and Gregory’s classification, distance from the anterior edge of the mandibular ramus 2, impaction depth A, and impaction grade 2A; and in Asanami and Kasazaki’s classification, distance of the mandibular ramus from the distal surface of the second lower molar 3, impaction depth A, and anterior inclination. In most cases of surgical removal of an impacted tooth, the anticipated difficulty of the procedure was rated as very difficult.Conceptualization, G.T. and A.J.; methodology, G.T. and A.J.; software, G.T. and A.J.; validation, G.T. and A.J.; formal analysis, G.T. and A.J.; investigation, G.T. and A.J.; resources, G.T. and A.J.; data curation, G.T. and A.J.; writing—original draft preparation, G.T. and A.J.; writing—review and editing, G.T. and A.J.; visualization, G.T. and A.J.; supervision, G.T. and A.J.; project administration, G.T. and A.J. All authors have read and agreed to the published version of the manuscript.This research received no external funding.This study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of NAME OF Pomeranian Medical University in Szczecin, Poland (KB-0012/270/09/18).Not applicable.Data are available on request: kzchstom@pum.edu.pl.The authors declare no conflict of interest.The occlusal plane line (A) and the mandibular anterior ramus line (B).The long axis of the tooth (a), greatest width of the tooth crown (b), tooth neck (c).Characteristics of the study group.SD—standard deviation. n—number of patients.Characteristics of impacted third molars in the mandible.n—number of patients.Frequency of impaction degrees among impacted mandibular third molars according to Winter in the study group.n—number of patients.Frequency of impaction degrees among impacted mandibular third molars according to Pell and Gregory in the study group.n—number of patients.Frequency of impaction degrees among impacted mandibular third molars according to Tetsch and Wagner in the study group.n—number of patients.Frequency of impaction degrees among impacted mandibular third molars according to Asanami and Kasazaki in the study group.n—number of patients.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06017.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: A subgroup of patients recovering from COVID-19 experience persistent symptoms, decreased quality of life, increased dependency on others for personal care and impaired performance of activities of daily living. However, the long-term effects of COVID-19 on physical activity (PA) in this subgroup of patients with persistent symptoms remain unclear. Methods: Demographics, self-reported average time spent walking per week, as well as participation in activities pre-COVID-19 and after three and six months of follow-up were assessed in members of online long-COVID-19 peer support groups. Results: Two hundred thirty-nine patients with a confirmed COVID-19 diagnosis were included (83% women, median (IQR) age: 50 (39–56) years). Patients reported a significantly decreased weekly walking time after three months of follow-up (three months: 60 (15–120) min. vs. pre-COVID-19: 120 (60–240) min./week; p < 0.05). Six months after the onset of symptoms walking time was still significantly lower compared to pre-COVID-19 but significantly increased compared to three months of follow-up (three months: 60 (15–120) min. vs. six months: 90 (30–150) min.; p < 0.05). Conclusions: Patients who experience persistent symptoms after COVID-19 may still demonstrate a significantly decreased walking time six months after the onset of symptoms. More research is needed to investigate long-term consequences and possible treatment options to guide patients during the recovery fromCOVID-19.At the time of writing, the number of people with a confirmed diagnosis of coronavirus disease 2019 (COVID-19) has risen to 167 million globally, with over 1.6 million confirmed cases in the Netherlands and over 1 million cases in Belgium [1]. As the pandemic continues, many patients have passed through the acute phase of the disease, but may still face difficulties in resuming their daily routine.It has already been demonstrated that hospitalized patients with COVID-19 present with low physical functioning and impaired performance in activities of daily living (ADLs) immediately after discharge [2]. Several months after the infection, a subgroup of patients with COVID-19 still report ongoing symptoms such as fatigue, dyspnoea and muscle weakness, as well as impaired quality of life and increased dependency on others for personal care and the performance of ADLs [3,4,5,6,7,8,9,10,11]. This suggests the presence of a post-COVID-19 syndrome [4,12,13], which was recently referred to as post-acute sequelae of SARS-CoV-2 infection (PASC) by the National Institutes of Health (NIH) [14]. Moreover, a qualitative study indicated that the experience of prolonged symptoms hampers patients in resuming or maintaining physical activity [15]. Similar to other countries, the Dutch and Belgian governments decided to close sport clubs and requested residents’ home confinement as a measure to limit the rate of infections (Netherlands closed sport clubs on 15 March 2020 and requested home confinement on 23 March 2020; Belgium closed sport clubs and requested home confinement on 13 March 2020) [16,17]. Previous studies have already demonstrated that those governmental measures led to a decrease in physical activity (PA) among other unhealthy lifestyle behaviours in the general population [18,19]. However, the long-term effects in patients with COVID-19 on PA levels remain unclear.Therefore, the aim of this study was to assess the impact of COVID-19 on the level of self-reported PA (time spent walking per week and leisure-time sports activities) in patients with post-COVID-19 syndrome (i.e., long COVID, long-haul COVID or PASC). We hypothesized a significant decrease in PA approximately three months after the infection, and at least a partial recovery of PA six months after the infection.As part of a large longitudinal study conducted in the Netherlands and Flanders (Belgium), an online questionnaire was made available between June 4 and June 11 of 2020 (T1) to all members of two long COVID Facebook groups [20,21] and to an online COVID-19 panel (www.coronalongplein.nl) (accessed on 4 June 2020). In total, 1556 participants who agreed to take part in a follow-up of this study were asked to complete a second survey between August 31 and September 8 of 2020 (T2). The medical ethics committee of the Maastricht University stated that the Medical Research Involving Human Subjects Act (WMO) did not apply to this study and that an official approval of this study by the committee was not required (METC2020-1978, METC2020-2554). The medical ethics committee of Hasselt University formally judged and approved the study (MEC2020/041). All adult respondents (aged 18 years or older) gave digital informed consent at the start of the questionnaires. Data about symptoms, care dependency, information needs, functional status, work productivity and quality of life have been published before [4,22,23,24,25,26].The survey contained questions regarding demographics (sex (male/female/other), age (years), body mass index (BMI) (kg/m2), marital status (married or living with partner: yes/no), education level (low/medium/high)), self-reported pre-existing comorbidities, COVID-19 diagnosis (based on reverse transcription polymerase chain reaction (RT-PCR) test and/or computed tomography (CT) scan of the thorax; symptom-based medical diagnosis; no test/medical diagnosis), received care (no care needed/physiotherapy/rehabilitation), symptoms and admission to hospital.In addition, participants were asked about the average time they spent walking in the previous seven days and which sports/activities they performed before COVID-19 (retrospectively) and at the time of completing the two questionnaires (approximately three and six months after symptom onset, respectively) (A summary of the questionnaire is added in the online supplement). The World Health Organization (WHO) recommends a minimum of 150 min per week of moderate-intensity aerobic physical activity [27]. Therefore, participants were divided into two groups: average walking time of less than 150 min per week or an average walking time of 150 min or more per week.Initial analyses were performed in participants with an RT-PCR or CT-confirmed diagnosis. Participants with a presumed COVID-19 diagnosis (n = 766; data presented in the online supplement) were excluded from the primary analyses.Statistical analyses and visualizations were performed using SPSS v25.0 (IBM Corp., Armonk, NY, USA), SankeyMATIC (http://sankeymatic.com/build/) (accessed on 4 June 2020) and GraphPad Prism 8.3.5. (GraphPad Software, La Jolla, CA, USA). Data were presented as mean and standard deviation (SD), median and interquartile range (IQR) or frequency and proportion, as appropriate. Data were tested for normality with a Kolmogorov–Smirnov test. Within-group comparisons were performed using the Friedman test, McNemar’s test or standard Cochran’s Q test (with Bonferroni corrected post-hoc test). Between-group comparisons were performed using a Mann–Whitney U test or Fisher’s exact test. A priori, the level of significance was set at p < 0.05.Of the initial 1556 participants that consented to be approached for a second questionnaire, 1005 participants completed the online questionnaires at approximately six months. See Supplementary Figure S1 for all details. Data from 239 participants with an RT-PCR or CT-confirmed diagnosis (82.8% women; median age: 50 (39–56) years) were used for the primary analyses, of which 62 (26%) were hospitalized (but not admitted to the intensive care unit (ICU)). The average time between the onset of symptoms and filling out the questionnaires was 10.4 ± 2.4 weeks (T1) and 22.6 ± 2.4 weeks (T2).All characteristics of the 239 participants with a confirmed COVID-19 diagnosis are presented in Table 1. Results stratified for hospital admission are presented in the online supplement (Supplementary Tables S2 and S3 and Supplementary Figure S3).Results regarding self-reported time spent walking per week at the three time points (pre-COVID-19 (T0) and three (T1) and six months (T2) of follow-up) are presented in Figure 1. After three months of follow-up, walking time in the previous week was significantly reduced compared to pre-COVID-19 (three months: 60 (15–120) min. vs. pre-COVID-19: 120 (60–240) min./week; p < 0.05). Although there was a recovery in walking time between three months and six months of follow-up (from 60 (15–120) min. to 90 (30–150) min.; p < 0.05), walking time was still significantly lower compared to pre-COVID-19.The proportion of participants reporting walking ≥150 min. per week at the different time points is presented in Figure 2. Approximately 41% of the participants reported walking ≥150 min. per week before COVID-19. This was significantly lower after three (17%) and six months (28%) of follow-up. Only 9% of participants reported walking ≥150 min. per week at all three time points. Approximately 20% of participants reported walking ≥150 min. per week pre-COVID-19 but did not achieve this at three and six months of follow-up, while 11% decreased their walking time per week (from ≥150 min. to <150 min. per week) at three months of follow-up but restored it to ≥150 min. at six months of follow-up. There were no significant differences in baseline characteristics between participants walking ≥150 min. per week and participants walking <150 min. per week, with the exception of the proportion of patients undergoing rehabilitation between three and six months of follow-up (<150 min.: 8%; ≥150 min.: 17%; p = 0.04). Additional analyses of walking time stratified for sex and number of symptoms are reported in the online supplements (Supplementary Tables S4 and S5)Participants reported a wide variety of activities, with walking (pre-COVID-19: 53.1%; three months: 41.1%; six months: 68.2%), outdoor cycling (pre-COVID-19: 35.1%; three months: 21.3%; six months: 42.3%) and (physio)fitness/exercise groups (pre-COVID-19: 30.1%; three months: 10.0%; six months: 38.5%) as the three most reported activities (Table 2).At three months of follow-up, participants reported performing fewer activities compared to pre-COVID-19 and almost 44% of the participants were not able to be physically active or perform sports or activities due to COVID-19. From three months to six months of follow-up the proportion of participants unable to be physically active significantly decreased (from 44% to 12%; p < 0.05) and the proportion of participants reporting walking, cycling outdoors/indoors, participating in (physio)fitness/exercise groups and running significantly increased. The proportion of participants that reported walking and cycling indoors at six months was significantly higher compared to pre-COVID-19.This study aimed to assess the impact of COVID-19 on the level of self-reported PA (time spent walking per week and leisure-time sports activities) in patients with post-COVID-19 syndrome (i.e., long COVID, long-haul COVID or PASC). As hypothesized, a significant decrease in walking time approximately three months after the infection and a partial recovery of walking time six months after the infection was found.These results indicate a possible recovery pattern similar to that which has been established in influenza A, acute respiratory distress syndrome survivors and severe acute respiratory syndrome, where patients experience impaired health-related quality of life, functional disability, psychological problems and impaired exercise capacity after up to two years of follow-up [28,29,30,31].Since the Dutch healthcare system was completely overwhelmed by the COVID-19 pandemic, the primary focus was the treatment of patients with the most life-threatening symptoms. People without the need for ICU admission, were understandably not prioritized during these hectic times, and therefore not intensively monitored or guided in their recovery. Complications like myocarditis or thromboembolic problems might have been missed and therefore might have hindered the recovery of PA in this subgroup of patients [32,33].Furthermore, since the effects of COVID-19 on the human body are still not completely clear, it is important to consider that patients might recover along different trajectories. Indeed, Gandotra et al. have investigated recovery in patients with respiratory failure due to several causes and identified different recovery trajectories and characteristics [34]. It is feasible that similar trajectories are present in patients recovering from COVID-19. Additionally, a study by Sallis et al. has shown that the level of PA pre-COVID-19 is associated with the severity of COVID-19. In fact, physical inactivity was the strongest risk factor for hospital admission, ICU admission and death, exceeding well-known risk factors like smoking, obesity, diabetes, hypertension, cardiovascular disease and cancer [35]. Since PA pre-COVID-19 seems to have a large impact on three important outcomes after COVID-19 (hospital admission, ICU admission and death), we recommend further investigation of the association between PA and outcomes after COVID-19 in a population of patients with persistent symptoms.Although encouraging patients to return to performing daily activities and to start low/moderate-intensity exercise at home is currently recommended for patients recovering from COVID-19 [36,37], Humphreys et al. have described that patients experienced a lack of clear and consistent advice with regard to PA and refraining from PA after suffering a relapse in symptoms after PA or after seeing others relapse after PA [15]. In patients with persistent symptoms, it is recommended to perform a post-COVID-19 assessment and referring them to specialists or pulmonary rehabilitation based on the clinical findings [3]. A patient-tailored approach is needed in order to achieve an optimal recovery after an infection with COVID-19 [15].While interpreting the abovementioned results, some strengths and limitations of this study need to be considered.One of the strengths of this study is the fact that this is the first longitudinal study investigating PA in patients with post-COVID-19 syndrome. We were able to follow up with a significant number of participants during the course of six months. A team of scientists, methodologists and patients worked together during the preparation of this study in order to create a study that provided a complete overview of relevant parameters in this specific population.However, since participants were recruited through platforms that targeted patients with persistent symptoms, there is the possibility of selection bias and the external validity of these results might be limited. It is possible that participants that recovered after three months did not feel the need to fill in the questionnaire at six months and therefore are underrepresented. However, the additional analyses demonstrate that non-responders had a significantly lower walking time three months after the onset of symptoms, compared to the responders (median (IQR) 40 (10–95) vs. 59 (15–105) min. respectively; p = 0.046) while walking time before COVID-19 was similar. This indicates that non-responders showed even less recovery of PA compared to responders.Additionally, females are overrepresented in this sample, possibly due to the higher proportion of women that are part of online long COVID support groups [38]. This is consistent with the gender distribution in previous studies [8,38,39,40]. Additional analyses on walking times stratified for gender demonstrated that female participants had a significantly lower walking duration compared to males at T1 and T2, while no differences were found at T0. This could indicate that the recovery of walking time differs between males and females. Taking the previously mentioned limitations into consideration, there are limitations in extrapolating these results to the general population.Due to the national regulations that were in place during the first wave of COVID-19 infections, the possibilities for sports and activity were limited. In both the Netherlands and Belgium, a lockdown was proclaimed, people were asked to stay at home as much as possible and sport facilities were closed. Therefore, the decrease in activities that were not possible during a large part of the three months after symptom onset (e.g., swimming, team sports) is probably caused by the regulations that were in place at that time. However, at the same time, many sports clubs initiated alternative options like online dance and exercise classes to provide the possibility to exercise at home or outdoors. The regulations that were in place at the time also do not explain the difference in walking time. However, the increase in the proportion of participants who reported walking as an activity might be related to the regulation.Since the pandemic had a huge impact on healthcare resources, medical treatments were focused on the most severe cases of COVID-19 and patients with “milder” symptoms were often not tested. As a result, many patients who presented with milder symptoms were not admitted to the hospital or even officially diagnosed with COVID-19 and remained under the medical radar. The NICE guidelines have indicated that having a positive RT-PCR test or hospitalization is not a requirement for a COVID-19 diagnosis and healthcare should also focus on patients with suspected COVID-19 [13]. In line with these guidelines, we believe that the group of study participants with presumed COVID-19 provides valuable information and needs further attention. Results of participants with presumed COVID-19 are mostly comparable with the results of the participants with a confirmed diagnosis, and are presented in the online supplement (Supplementary Table S1 and Supplementary Figures S2 and S3).Participants were asked to retrospectively assess their physical activity by filling in questionnaires. Therefore, the effect of recall bias cannot be excluded. However, since participants were asked questions that referred to the previous three months or the previous seven days, we expect that the influence of recall bias is limited. Besides recall bias in general, Dyrstad et al. have shown that a self-reported assessment of PA tends to overestimate vigorous activity and underestimate sedentary time [41]. This effect might have led to an overestimation of walking time in our results. Additionally, the questionnaire on sports and activities other than walking provided no information on frequency, intensity or duration, but only whether or not the activity was performed. This limitation should be kept in mind when interpreting these results. At the time of inclusion, due to the acute outbreak of COVID-19, it was impossible to provide participants with accelerometers and perform an objective pre-and post-COVID-19 PA assessment, but in future research objectively measured PA could be used to gain a more complete picture of the PA level in this population.Participants with persistent symptoms while recovering from COVID-19 who were all members of online long COVID support groups still demonstrated a significantly decreased self-reported walking time six months after the onset of symptoms. In contrast, the proportion of participants that reported walking or cycling indoors increased over the course of six months after the onset of symptoms. More research is needed to investigate the long-term consequences and possible treatment options to guide patients during the recovery from COVID-19.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116017/s1. Figure S1: Flowchart of included and excluded participants; Figure S2: Walking time in previous week pre-COVID-19 (T0), at three months of follow-up (T1) and six months (T2) of follow-up of participants with suspected COVID-19; Figure S3: Walking time in previous week pre-COVID-19 (T0), at three months (T1) and six months (T2) of follow-up after stratification into hospitalized and non-hospitalized; Table S1: General characteristics of participants with suspected COVID-19; Table S2: General characteristics of participants stratified into hospitalized and non-hospitalized; Table S3: Activities performed by participants before (T0) and three (T1) and six months (T2) of follow-up stratified into hospitalized and non-hospitalized; Table S4: Walking duration stratified for sex; Table S5: Walking duration stratified for number of symptoms.J.M.D., A.W.V., F.V.C.M., Y.M.J.G., M.V.H., R.M. and S.H.-W. were responsible for the data collection. M.A.S. is the principal investigator of this trial. J.M.D. and A.W.V. drafted the manuscript. J.M.D., F.V.C.M., Y.M.J.G., M.V.H., R.M., S.H.-W, C.B., F.M.E.F., Y.S., H.V., A.J.v.H., D.J.A.J., M.A.S. and A.W.V. reviewed and edited the manuscript. All authors have read and agreed to the published version of the manuscript.The scientific work of Y.M.J.G. is financially supported by Lung Foundation Netherlands grant 4.1.16.085; F.V.C.M. is financially supported by ZonMw (ERACoSysMed #90030355) and R.M. is financially supported by Lung Foundation Netherlands grant 5.1.18.232.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the medical ethics committee of Hasselt University (Diepenbeek, Belgium) (MEC2020/04103-06-2020). Ethical review and approval in the Netherlands were waived for this study, as the medical ethics committee of Maastricht University stated that the Medical Research Involving Human Subjects Act (WMO) did not apply for this study (METC2020-1978 and METC2020-2554).Informed consent was obtained from all subjects involved in the study.The data presented in this study are available on reasonable request from the corresponding author, and only after approval by the medical ethical committee. The data are not publicly available due to ethical restrictions.The research team acknowledges the valuable input from the patient representatives to develop the survey, and the technical support by Martijn Briejers and Oscar Wagemakers (ASolutions, Capelle aan den Ijssel, The Netherlands).Franssen reports grants and personal fees from AstraZeneca, personal fees from Boehringer Ingelheim, Chiesi, GSK and TEVA outside the submitted work. Janssen reports personal fees from AstraZeneca, Boehringer-Ingelheim and Novartis outside the submitted work. Spruit reports grants from Lung Foundation Netherlands, grants from Stichting Astma Bestrijding, grants and personal fees from AstraZeneca and grants and personal fees from Boehringer Ingelheim outside the submitted work.Walking time in previous week pre-COVID-19 (T0), at 3 months (T1) and 6 months (T2) of follow-up; * significant difference vs. pre-COVID-19, p < 0.05; #: significant difference vs. 3 months, p < 0.05. Data presented as median, IQR and 10–90%CI.Proportion of participants divided based on achieving 150 min. walking per week before (T0) and at 3 months (T1) and 6 months (T2) of follow-up; * significant difference p < 0.05.General characteristics of participants.#p < 0.05 vs. 3 months of follow-up.Activities performed by participants before (T0) and at three (T1) and six months (T2) of follow-up.*: significant difference vs. pre-COVID-19, p < 0.05; #: significant difference vs. three months, p < 0.05.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06018.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: Dry needling (DN) is often used for the treatment of muscle pain among physiotherapists. However, little is known about the mechanisms of action by which its effects are generated. The aim of this randomized controlled trial was to determine if the use of DN in healthy subjects activates the sympathetic nervous system, thus resulting in a decrease in pain caused by stress. Methods: Sixty-five healthy volunteer subjects were recruited from the University of Alcala, Madrid, Spain, with an age of 27.78 (SD = 8.41) years. The participants were randomly assigned to participate in a group with deep DN in the adductor pollicis muscle or a placebo needling group. The autonomic nervous system was evaluated, in addition to local and remote mechanical hyperalgesia. Results: In a comparison of the moment at which the needling intervention was carried out with the baseline, the heart rate of the dry needling group significantly increased by 20.60% (SE = 2.88), whereas that of the placebo group increased by 5.33% (SE = 2.32) (p = 0.001, d = 1.02). The pressure pain threshold showed significant differences between both groups, being significantly higher in the needling group (adductor muscle p = 0.001; d = 0.85; anterior tibialis muscle p = 0.022, d = 0.58). Conclusions: This work appears to indicate that dry needling produces an immediate activation in the sympathetic nervous system, improving local and distant mechanical hyperalgesia.Dry needling (DN) can be defined as “a technique in which a fine needle is used to penetrate the skin, subcutaneous tissues, and muscle, with the intent to mechanically disrupt tissue without the use of an anesthetic” [1]. DN is a safe and minimally invasive technique [2] typically used for the treatment of an assortment of neuromusculoskeletal pain syndromes [3,4,5]. In recent decades, DN has been widely used to treat muscle pain and, more specifically, the treatment of the myofascial trigger point (MTrP). The MTrP is a “hyperirritable spot within a taut band of skeletal muscle that is painful on compression, stretch, overload, or contraction of the tissue which usually responds with a referred pain that is perceived distant from the spot” [6]. MTrPs are classified as being active (in the event of spontaneous pain) or latent (absence of spontaneous pain) [7]. The presence of MTrPs in the skeletal muscle has been associated with an impaired range of motion, muscle weakness, loss of coordination, pain, and autonomic reactions [7,8]. DN may have been shown to be effective in the management of pain, improving range of motion, muscle strength, and coordination [9]. The possible explanations found in the literature for the decrease in pain include the effects of dry needling at the local level (producing an interruption of spontaneous electrical activity on the taut band or local vasodilation), activation of the peripheral segmental pain inhibition (explained through Gate Control Theory), or activation of the descending pathways of pain inhibition at the central nervous system level (serotonergic and noradrenergic endogenous opioid release and conditioned modulation of pain) [8]. However, these underlying mechanisms have not yet been fully clarified and are not understood [1,9].Another possible DN mechanism for the modulation of pain is stress-induced analgesia (SIA) [10]. SIA has been described as “a reduced nociceptive response after stress exposure, which is mediated by descending inhibitory opioid and nonopioid brain circuits” [11]. SIA is influenced by activity of the autonomous nervous system (ANS) and the hypothalamic–pituitary–adrenal (HPA) axis [12,13]. The ANS, prior to a stressful stimulus, quickly induces physiological changes through synaptic transmissions via two branches, the sympathetic and parasympathetic nervous systems, resulting in an increase in sympathetic nervous system (SNS) activity [13]. SNS activity is usually determined by measuring skin conductance [14], heart rate, and respiratory rate values [15]. Other therapeutic procedures have been shown to produce sympathoexcitatory changes that had therapeutic benefits for patients [16]. HPA axis activity is measured by determining the cortisol level in saliva. Cortisol is an anti-inflammatory hormone regulated by the HPA axis via feedforward and feedback loops, which is related to the modulation of nociception and stress-induced analgesia [17]. It appears that the cortisol level in saliva increases with a stressful event which, in turn, appears to be related to the SIA process [18]. Therefore, the activation of the neuroendocrine system SNS–HPA axis maintains homeostasis and produces an analgesic effect [19].Given the nature of DN [20], which is a technique that can be considered stressful, SNS–HPA axis activity could be among the possible physiological mechanisms that explain its analgesic effect [9]. Nevertheless, to our knowledge, no research has been undertaken that explores whether the SNS–HPA axis is involved in the response to DN.The purpose of this study was to determine if the application of a DN technique results in changes in skin conductance, heart rate, temperature, breathing rate, or cortisol levels in saliva between different measurements, in addition to assessing improvements in the pressure pain threshold. We hypothesized that a DN technique would result in an activation of the SNS and HPA axis, which plays a crucial role in pain modulation.A randomized controlled clinical trial of parallel groups was carried out to compare a deep DN treatment with a placebo treatment, with the aim of evaluating the effects produced on the ANS and on pain. The study was performed following the CONSORT 2010 (Consolidated Standards of Reporting Trials) [21,22] directives and the STRICTA (Revised Standards for Reporting Interventions in Clinical Trials of Acupuncture) [23] criteria. It was approved by the Committee of Research Ethics and Animal Experimentation of the University of Alcalá (CEIT/HU/2015/06 of 23 November 2015) and registered in the Australian New Zealand Clinical Trials Registry at http://www.anzctr.org.au/ (ACTRN12616000881437) (accessed on 29 May 2021).The sample comprised healthy volunteers from the student body of different degrees and the administrative staff of the University of Alcalá. These were selected using convenience non-probability sampling.The following inclusion criteria were established: (1) 18 to 65 years of age, (2) pain-free, and (3) latent MTrP in the adductor muscle of the left thumb. Participants were excluded if: (1) acute illness was present at the time of the study; (2) any of the following were present: fibromyalgia, diabetes, cardiopathy, essential arterial hypertension, hemophilia, a neurological disease, cognitive decline, a compromised immune system (HIV, cancer, hepatitis, acute immune diseases), left upper limb lymphedema; (3) taking blood thinners; (4) pregnancy; (5) fear of needles; (6) allergic to metals (nickel or chrome); (7) had participated in a dry needling/acupuncture study in the past 6 months.The subjects who met the selection criteria were informed of the study procedure by an information sheet, and they signed a data release document and provided informed consent, according to the standards of the Declaration of Helsinki.Each participant was assigned a code. The subjects were also randomly assigned to receive a deep DN technique or a placebo needling technique. Those not included in the study performed a concealed randomization with the program Epidat 4.2, using a 1:1 allocation ratio through simple randomization.Both the examiner who obtained all of the outcome measures and the statistician who examined the data were blinded to the subject randomization. The subjects participating in the study were blinded to the intervention, because they were not informed about the existence of a placebo group. The physiotherapist assessed the MTrPs before knowing the intervention group to which each subject belonged.Interventions were carried out at the School of Nursing and Physiotherapy of the University of Alcalá (Madrid, Spain). All of the subjects received one session of deep dry needling or placebo needling. Measurements were performed between 9:00 a.m. and 11:00 a.m. [24]. The temperature of the room was maintained in the range of 24–25 °C, and the noise level was kept to a minimum. The subjects could neither ingest alcohol or caffeine, nor perform vigorous physical activity, on the day of the study. They were not allowed to smoke during the 2 h before the study. Moreover, they could not brush their teeth, ingest any liquids, eat solid food, or chew gum 30 min before the study.The subjects lay supine on a stretcher, with their forearms free and legs stretched out. A professional physiotherapist with experience (about 15 years) in the palpation, diagnosis, and treatment of MTrPs and myofascial pain syndrome was responsible for locating and marking the MTrPs. Thus, he was considered a qualified examiner, with a good reproducibility index (k = 0.63) [25]. Subjects were instructed to remain calm and quiet, but completely awake. A ten-minute period was predetermined for the subject to acclimate to the room conditions before beginning the recording of the physiological variables [26,27].The physiotherapist carried out the interventions in both groups, on a latent MTrP in the adductor muscle of the left-hand thumb. The same aseptic measures pertinent to the technique were applied for all subjects in both groups. Due to the supine position on the stretcher, the vision of the subjects of the DN technique was blocked. The procedure followed in the two groups is described below.Deep DN was performed with disposable needles (0.25 × 0.25 mm; AGU-A1038P; Agu-Punt S.L, Barcelona, Spain) [28]. The “fast-in and fast-out” technique asserted by Hong [29,30] was employed. The needle was moved up and down in multiple directions (vertically, without rotations), at approximately 1 Hz frequency for 10 s, to look for LTRs [28].A non-penetrating, simulated DN technique with placebo needles was applied to the subjects who were randomized into this group, with a modification of the protocol developed by Tough et al. [31]. Disposable sterile needles (0.25 × 0.40 mm; DB100-2540; DongBand, AcuPrime®, Exeter, UK) were used. These needles have a red tab and cannot be distinguished from those used in the Deep Dry Needling Group (Figure 1). They also have a spring handle that can be glided up and down, imitating the movement in and out of the skin, without genuine penetration. To achieve an effective blinding [32], all of the needles were held in the box used for the Deep Dry Needling Group. The professional pinched the MTrP of the adductor muscle of the thumb, placed the guide-tube with the needle exerting some pressure, and jabbed the needle against the skin, simulating an insertion. Next, he withdrew the guide-tube, pressing the tip of the needle with his thumb, making sure the needle did not move (to ensure it did not penetrate the skin). The needle stayed in contact with the skin each time. Then, he moved the spring handle up and down 10 times at a speed of 1 Hz, in a “sparrow pecking” movement. Each time the handle moved up and down, the pressure sensation increased, replicating the feeling of a puncture.Before beginning the intervention, participants completed different psychometric tests to assess their initial levels of depression (the Beck Depression Inventory II (BDI-II)) [33], anxiety (the State-Trait Anxiety Inventory (STAI)) [34], and pain catastrophizing (Pain Catastrophizing Scale (PCS)) [35]. This testing was undertaken because high levels in these variables can affect ANS activity and HPA activity.To measure the ANS response when applying a DN technique, multiparametric biofeedback equipment NeXus 10 MK-II was used (Mind Media BV; Herten, the Netherlands) [36,37,38]. Data were processed using the Biotrace software, version V2015B (Mind Media BV). The skin conductance (SC) and peripheral temperature of the skin (Temp) were registered with the sensors placed as shown in Figure 2; heart rate (HR) was registered with EKG sensors, and breathing rate (BR) with a sensor placed on the sternum with an elastic band.Measurements for the physiological variables were collected at different moments (Figure 3): Baseline (average of the 5 min before the intervention); Dry needling (average of the 10 s of the needling technique); Post-1 (average of 1 min immediately after ending the intervention); Post-2 (average of 1 min, 9 min after ending the intervention).Free cortisol levels in saliva were determined using the Cortisol ELISA® kit from IBL International laboratories (Hamburg, Germany) [39,40]. Compared to cortisol determination in plasma, this is a simple, painless, and non-invasive method. It is also less costly, no specialized medical staffing is required, and does not produce stress when performing the vein puncture [41,42]. In addition, the correlation coefficient between the two methods is r > 0.9 [39]. The saliva samples were collected before the intervention and three minutes after it ended [24] (Figure 3). Cortisol levels in the saliva were analyzed in the laboratories of the Research Foundation of the University Hospital Príncipe de Asturias in Alcalá de Henares.The Numeric Rating Scale for Pain (NRS) was used to measure the subjective pain perceived by the participant during the needling technique performance. This is a valid and reliable measuring tool to assess pain intensity during a treatment and/or intervention [43], and presents a good correlation with the Visual Analog Scale (VAS), with an overall intraclass correlation coefficient (ICC) of >0.7 [44]. To implement this scale, the subjects were asked after the intervention (to avoid talking during the recording of the physiological measures) about the maximum pain experienced during the performed needling technique, selecting a whole number between 0 and 10 that best reflected their pain intensity (0: no pain and 10: worst imaginable pain).A Wagner Force Dial™ model FDK 20 (Greenwich, CT, USA) manual algometer was used to measure PPT before and after performing the intervention (Figure 3) on the two points previously marked: on the latent MTrP of the adductor muscle of the left thumb (point selected to examine the implicated area) and on the most painful point on palpation of the anterior tibialis muscle, approximately at 2.5 cm lateral and at 5 cm distal to the anterior tibial tuberosity [45] (point selected as an unrelated segment area to assess descending pain inhibitory mechanisms). For this purpose, the tool was placed perpendicular to the point previously marked, and through an approximate gradual increase in speed of 1 kg/s, pressure was increased until the subject started feeling pain or discomfort (without ever reaching the maximum bearable pressure); at that point, pressure was stopped. Three measurements were taken on the same point, with a 30-s time interval between each measurement, and the average of the three measurements was analyzed [16,28,46]. This method presents high reliability (ICC = 0.91) [46].A prior pilot study was conducted with 20 subjects: 10 in the experimental group and 10 in the control group. In this research, the main variable was skin conductance (SC), measured with the biofeedback Nexus 10 MK-II equipment. A repeated measures contrast was used. The size effect was 0.143; a 0.05 alpha level and a power of 0.95% were assumed, plus 15% possible loss. These assumptions generated a simple size of 65 participants in total. The statistical analysis program G*Power 3.1.9.4 was used.Data were analyzed with the statistical package SPSS for Windows, version 26.0 (SPSS Science, Chicago, IL, USA). To study the homogeneity of the groups at baseline, Student’s t-test was used for independent samples in the quantitative variables, and the Pearson’s chi-Squared test was used for the qualitative variables.To normalize the differences between participants in the variables analyzed, the data of each time period were evaluated in terms of percentage change (% Change), using the formula employed by Perry and Green [47].Regarding the primary outcomes of the physiological variables, a separate 2-by-4 mixed model analysis of the variance was employed to assess the effects of the intervention, for which group (deep dry needling or control) was the between-subjects variable, and time (the different measurements) the within-subjects variable. An a priori alpha level of 0.05 was set. The hypothesis of interest was the group-by-time interaction. In addition, the effect size was estimated by calculating the partial Eta2 coefficient (ηp2). The difference between the two groups in the percentage change in all measurements was compared using Student’s t-test for independent samples, or, alternatively, Welch’s t-test. Bonferroni type adjustment of significance was used.For the analysis of cortisol and PPT, a separate 2-by-2 mixed model analysis of variance was employed to assess the effects of the intervention, for which group (deep dry needling or control) was the between-subjects variable, and time (the different measurements) the within-subjects variable. An a priori alpha level of 0.05 was set. The hypothesis of interest was the group-by-time interaction. Effect size was estimated by calculating ηp2. The difference between both groups in % Change at baseline and at post-test was compared using Student’s t-test for independent samples, or, alternatively, Welch’s t-test. In addition, effect size was estimated using Cohen’s d, considering “small effect size” to be between 0.2 to 0.5, “medium effect size” 0.5 to 0.8, and “large effect size” greater than 0.8. In all statistical tests, the level of significance was set at 95% (p < 0.05; two-tailed test).Finally, different correlations using Pearson’s correlation coefficient were studied. First, the values obtained from the Numeric Rating Scale for Pain (NRS) for each group are shown, in addition to the existing correlations in this scale regarding % Change between the baseline–dry needling measurement of the physiological variables and regarding % Change of cortisol. In addition, correlations between % Change of the Physiological Variables and Cortisol are shown in the different measures with the Psychological Factors.A total of 65 subjects, with an average age of 27.78 years (SD ± 8.41 years), of whom 33 (50.8%) were men and 32 (49.2%) were women, met the selection criteria and participated in the study between November 2016 and February 2017. All of the data were collected for analysis, as shown in the flow diagram in Figure 4. Table 1 and Table 2 contain information about the characteristics at baseline of all of the subjects in each treatment group, confirming no statistically significant differences between the groups. It can also be observed in Table 1 that the subjects did not present signs of depression, anxiety, or pain catastrophizing. Table 2 also shows the values obtained regarding pain intensity perceived during the intervention. The NRS results were significantly higher in DN group than in the placebo group.The mixed model analysis of the variance of the variable HR indicated a statistically significant group-by-time interaction (F = 9.99, p < 0.001, ηp2 = 0.137), and a time effect (Dry Needling Group F = 50.53, p < 0.001, ηp2 = 0.612; Placebo Group F = 12.57, p < 0.001, ηp2 = 0.288). In this variable, no effect was found in the group factor (F = 0.04, p = 0.845, ηp2 = 0.001). As shown in Table 3, comparing the Baseline with the Needling, the % Change of the HR significantly increased in the dry needling group compared to the placebo group. Subsequently, the % Change decreased more in the placebo group at Post-1, whereas the % Change from Post-1 to Post-2 was significantly greater for the HR of the dry needling group.In the variables SC, Temp, and BR, the mixed model analysis of variance indicates no statistically significant group-by-time interaction (SC F = 0.91, p = 0.380, ηp2 = 0.014; Temp F = 1.80, p = 0.173, ηp2 = 0.028; BR F = 0.52, p = 0.667, ηp2 = 0.008), but does indicate a time effect (SC: Dry Needling Group F = 71.88 p < 0.001, ηp2 = 0.692; Placebo Group F = 72.16, p < 0.001 ηp2 = 0.700/Temp: Dry Needling Group F = 3.52, p = 0.042, ηp2 = 0.099; Placebo Group F = 6.84, p = 0.003, ηp2 = 0.181/BR: Dry Needling Group F = 28.33, p < 0.001, ηp2 = 0.470; Placebo Group F = 25.64, p < 0.001, ηp2 = 0.453). No effect was found in the group factor (SC F = 2.08, p = 0.154, ηp2 = 0.032; Temp F = 0.75, p = 0.389, ηp2 = 0.012; BR: F = 2.50, p = 0.119, ηp2 = 0.038).A considerable increase in SC was observed after needling compared to the baseline measurement, although no significant differences existed between the two groups. However, a greater decrease (p < 0.001) was produced in the SC values of the Placebo group (−21.50%, SE = 1.99) in comparison with the Dry Needling Group (−10.36%, SE = 1.96) in the Post-1 measurement after needling (Table 3). On the contrary, no differences were found in the variables BR and temperature between groups in the percentage change of any of the measurements performed (Table 3).Regarding cortisol, no significant group-by-time interaction was found (F = 2.07, p = 0.155, ηp2 = 0.032), nor significant differences between groups (F = 0.002, p = 0.969, ηp2 < 0.001). Notwithstanding the data, an increase of 11.27% (SE = 4.76) was produced in the Dry Needling group, with a difference existing in the time factor (F = 6.84, p = 0.013, ηp2 = 0.176). On the contrary, cortisol increased only 1.51% (SE = 3.08) in the Placebo group, and no difference was recorded in the time factor (F = 0.64, p = 0.428, ηp2 = 0.02).As shown in Table 2, the variance mixed model analysis indicated a significant interaction between the intervention group and the time (different measures) in the algometry thumb adductor (F = 77.88, p < 0.001, ηp2 = 0.553) and algometry of anterior tibialis muscle (F = 50.24, p < 0.001, ηp2 = 0.444), and a time effect (Algometry thumb adductor: Dry Needling Group F = 232.92, p < 0.001, ηp2 = 0.879; Placebo Group F = 58.06, p < 0.001, ηp2 = 0.652/Algometry of anterior tibialis muscle: Dry Needling F = 139.19, p < 0.001, ηp2 = 0.813; Placebo F = 47.54, p < 0.001, ηp2 = 0.605). However, no effect was found in the group factor (Algometry thumb adductor F = 2.23, p = 0.14,1 ηp2 = 0.034; Algometry of anterior tibialis muscle F = 0.84, p = 0.364, ηp2 = 0.013). Additionally, a statistically significant difference was observed in the percentage change (Table 3) between the measurement Baseline and the Post-test in PPT, both in the adductor of the thumb and in the anterior tibialis. In both cases, this percentage was higher in the Dry Needling group compared to the Placebo group, and the effect size was “large” (adductor of the thumb Cohen’s d = 1.87; anterior tibialis Cohen’s d = 1.61).Regarding the analyzed correlations between the Numeric Rating Scale for Pain, and the physiological variables and cortisol, no significant correlations were found (Table 4). The correlations between the physiological variables and cortisol, in addition to the Psychological Factors, are shown in Table 5 for the Dry Needling Group. A modest positive correlation (p = 0.046, r = 0.350) was found in % Change of the temperature between Needling and Post-1 with State Anxiety; and in the BR variable in the % Change Post-1 to Post-2 with the Total PCS (p = 0.024, r = 0.393). The results for the Placebo Group are shown in Table 6, and indicate a significant negative correlation for the BR in the % Change between Post-1 and Post-2 with State Anxiety (p = 0.047, r = −0.354), and between % Change Cortisol and the BDI-II (p = 0.021, r = −0.407), State Anxiety (p = 0.018, r = −0.416), and Trait Anxiety (p = 0.037, r = −0.371).The aim of this study was to evaluate the effects of DN on the ANS and nociceptive processing, by applying DN using the traditional method of fast-in and fast-out technique. The results of this study demonstrate that DN has a neurophysiological effect on the ANS and pain processing, showing an increased heart rate and an increased pressure pain threshold, both locally and at remote sites, compared to the placebo. However, only changes in heart rate were found, and no changes were found in the remainder of the measured parameters of the ANS. The heart rate undergoes innervation from the ANS but not exclusively [48]. These results are supported by a recently published study in which DN was evaluated on the cervical paravertebral muscles. It was found that patients experimented ANS changes, which were detected by measuring pupillometry [49]. In addition, this research found that leaving the needles in the muscle for 21 min caused an activation of the ANS that lasted 18 min, before returning to the basal state.These outcomes are consistent with those of several previous studies [50,51]. Haker et al. [50] investigated the effects of acupuncture applied over the thenar muscle on the ANS in healthy subjects. An increase in parasympathetic and sympathetic activity was found via changes in heart rate variability after the acupuncture stimulation with an increase in LF (low frequency) both during and after the intervention. The results of this study support the postulate that dry puncturing has a simultaneous influence on the SNS and pain processing. In another study, the application of DN resulted in changes in the ANS, with subjects experiencing an increase in blood flow, both in skin and the muscle [51]. In contrast, another study evaluating the effects of DN on skin sympathetic activity found no significant changes after DN in a group of healthy subjects, but did find changes in skin sympathetic activity in the patient group [52]. In a recent study of healthy subjects in which acupuncture to tendons was applied, a change in local blood flow was observed, which was controlled by the SNS, but this was not related to heart rate [53]. In addition, when acupuncture was applied to trigger points in muscles such as the tibialis anterior, changes in heart rate were observed [54]. Contrary to the observations in our study, we are not aware of any previous studies in which puncturing was performed and an increase in heart rate was found. The study also suggests that the parasympathetic nervous system is involved in the relief mechanisms of myofascial pain through acupuncture stimulation.In relation to mechanical hypoalgesia, other studies have shown an improvement in the PPT, and therefore are consistent with the results obtained in this work [28,29,55,56]. In contrast to these studies, it was concluded in a recent meta-analysis that DN shows low to moderate evidence of greater effects versus a placebo or control group in terms of improvement of pain and the pressure pain threshold [1].However, these results contradict those previously found in other conducted studies, in which mechanical hyperalgesia occurs immediately after the DN is applied to healthy subjects, lasting up to 48 h [57,58,59]. It is possible that different pain processing mechanisms are activated in patients when DN is applied to active MTrPs, as shown by immediate increases found in previous studies. Dry needling performed in patients with musculoskeletal pain was found to improve mechanical hyperalgesia after applying different doses [28,29,60,61]. In another study, in which only a single DN session was applied, a percentage change of 54.85% was observed in PPT, with a difference of more than 5 percentage points between pre- and immediately post-treatment [62]. The mechanisms that explain these beneficial effects may occur due to an elimination of pronociceptive substances when DN is performed [63,64]. Alternatively, delta Aδ are stimulated, which activate the descending pain inhibitory systems as a counter-irritation mechanism [9,65]. Finally, the remote hypoalgesic effects obtained in this study contradict those observed by Sterling et al. [61] in patients with whiplash, in which local mechanical hyperalgesia improves, but this effect does not occur over a greater distance.In a crossover clinical trial in which a dry needling-like therapy such as acupuncture was applied to whiplash patients, no relationship was found between pain improvement and changes in the SNS [36]. These results support those obtained in our study in the real DN group, but not in the placebo group, in which a positive correlation existed between the improvement in mechanical hyperalgesia at remote sites and the increment in the skin temperature. In the earlier research, the authors found that acupuncture produced a slight decrease in heart rate and an increase in skin conductance [36].Importantly, salivary cortisol must be noted because it is considered to be the major indicator of stressful stimuli [17]. Cortisol works via circadian rhythms, reaching its maximum level approximately 30 min after waking. In our sample, we observed that the waking time of subjects in both groups was homogeneous. By comparison, the salivary cortisol values found in the literature that correspond to healthy adults in the same time slot as the measurements taken in this study are 0.43 μg/dL [17], which is similar to the basal values obtained in our sample.In this study, no significant differences were found between the groups, although in the DN group a significant increase of 11.27% was found. The fact that no statistically significant differences were found between the two groups may be partly because saliva samples were not collected at the most opportune time. In the existing literature, no studies were found that observed the behavior over time of cortisol after a needle stimulus. Based on the collection of saliva samples in a study by Takai et al. [24], it was observed that after applying a stressful stimulus to psychologically healthy subjects, the highest salivary cortisol levels were obtained 3 min after the intervention.The results obtained in the current research are supported by the study of Knardahl et al. [66], in which they measured plasma cortisol after electroacupuncture in healthy subjects, observing an immediate increase in cortisol in the intervention group and a decrease in the placebo group. No studies that measure salivary cortisol after applying DN were found in the literature; thus, more studies are needed to investigate the effect of this variable.By comparison, several studies have described the fear of medical procedures, such as injections or dental care, as factors that produce fear and pain [67]. In the literature, studies have researched how pain induction produces changes in the ANS [68,69]. In the present study, the pain perception differed during the needling of both groups, and no relationship between the perceived pain during needling, and either the response of the ANS or changes in the PPT, was found. Therefore, pain might not be responsible for the changes observed in the subjects, and the results may be attributed to the effect of the technique.Acupuncture is a procedure that consists of penetrating the skin with a needle, which can stimulate the primary nociceptor and induce pain. Lee et al. [70] performed a study to investigate the effects of acupuncture stimulation on ANS and its relationship with the fear of acupuncture. They found that skin conductance significantly increased after acupuncture stimulation and the fear of acupuncture-induced pain was associated with an enhanced physiological response. These data are consistent with those obtained in our study, in which we found a positive correlation between temperature changes and the state of anxiety, and between the heart rate and the level of catastrophism, in the group that received the real dry needling.Dry needling is a technique used in the management of musculoskeletal pain; however, its specific mechanisms for modulating pain remain unknown. This study examined healthy subjects who were free of depression, anxiety, and pain catastrophism, and is the first step in understanding the effects of dry needling on the autonomic nervous system and nociceptive processing. The results of this study show a short-term improvement in PPT, not only locally, but also remotely, showing that dry needling does not produce changes only in the tissue, but also involves changes at the central level. Furthermore, it does not appear that the pain perceived by the subjects is the trigger for the changes produced at the physiological level. Finally, further studies are needed on the behavior of cortisol following needle stimulation. On the basis of these findings, further studies of subjects experiencing pain, with longer follow up, are needed to further investigate the effects of this technique at the central level.One limitation of this study is that the subjects may have previously experienced a DN treatment, and thus may have prior expectations about the intervention. In addition, the subjects were aware of the invasive treatment using a needle, and anticipation of the treatment may have influenced SNA activity. In future studies, a control group should be included to control for this fact. Moreover, the subjects were not aware of the existence of the placebo group, and whether the subjects had identified the group to which they belonged after the treatment was not monitored. In future studies, it will be necessary to include a blinding index to ensure the success of blinding.The menstrual cycle in women should be monitored because of the influence it could have on the results.Another possible limitation is the lack of close control of the state of wakefulness and sleep of the subjects, because this factor directly influences the responses of the SNS. Finally, only healthy subjects who were not experiencing pain participated in this study; thus, additional studies are necessary to evaluate the response of the ANS in acute and chronic pain processes, considering that the processes that modulate pain are different.The results of this work showed that dry needling applied in healthy subjects immediately produced an increase in heart rate and a decrease in mechanical hyperalgesia, both locally and at remote sites, greater than that of the placebo intervention. Although the skin conductance, temperature, breathing rate, and cortisol levels also increased, no difference was found between the needling group and the placebo group. These results appear to indicate that dry needling produces an immediate activation in the sympathetic nervous system that are related to stress-induced analgesia mechanisms. Further studies are needed to clarify the possible implication of these underlying specific mechanisms.Conceptualization, I.L.-N., D.P.-M.; methodology, I.L.-N., D.P.-M.; software, C.L.-S.-A.; formal analysis, M.J.N.-S.; investigation, I.L.-N. and M.J.N.-S.; resources, D.P.-M. and T.G.-I.; data curation, J.J.J.-R.; writing—original draft preparation, I.L.-N., D.P.-M. and J.F.-C.; writing—review and editing, J.J.J.-R. and C.L.-S.-A.; visualization, I.L.-N.; supervision, T.G.-I. and J.F.-C. All authors have read and agreed to the published version of the manuscript.This research received no external funding.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the University of Alcalá (CEIT/HU/2015/06).Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper.The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical condition.The authors declare no conflict of interest.The needles used with the Deep Dry Needling Group are shown on the left; the needles used with the Placebo Needling Group are shown on the right.Placement of electrodes for SC and Temp on the right hand.Temporal chronogram of the study.Flow diagram of patients throughout the course of the study.Demographic characteristics of each treatment group.* Comparison needling group versus placebo group using Student’s t-test. † Absolute frequency and category percentage n (%) are shown. ‡ Pearson’s chi-squared test was used. BMI: Body Mass Index; BDI-II: Beck Depression Inventory II; STAI: State-Trait Anxiety Inventory; PCS: Pain Catastrophizing Scale.Primary and secondary outcomes.Mean ± standard error (SE) * Comparison using Student’s t-test for independent samples. Bonferroni type adjustment was used. SC: Skin conductance; HR: Heart rate; BR: Breathing rate; PPT: Pressure pain threshold; NRS: Numeric Rating Scale for Pain.% Change in primary outcomes and pressure pain threshold.Mean ± standard error (SE). * Comparison using Student’s t-test for independent samples. † Welch’s t-test was used. Bonferroni type adjustment was used. SC: skin conductance; HR: heart rate; BR: breathing rate; PPT: pressure pain threshold.Correlations of Numeric Rating Scale for Pain with % Change of the Physiological variables and Cortisol.Correlations using Pearson’s coefficient. SC: skin conductance; HR: heart rate; BR: breathing rate; NRS: Numeric Rating Scale for Pain.Correlations between % Change of the Physiological Variables and Cortisol with Psychological Factors in the Dry Needling Group (n = 33).Correlations using Pearson’s coefficient (r) and p-value (p) SC: skin conductance; HR: heart rate; BR: breathing rate; BDI-II: Beck Depression Inventory II; PCS: Pain Catastrophizing Scale.Correlations between % Change of the Physiological Variables and Cortisol with Psychological Factors in the Placebo Group (n = 32).Correlations using Pearson’s coefficient (r) and p-value (p) SC: skin conductance; HR: heart rate; BR: breathing rate; BDI-II: Beck Depression Inventory II; PCS: Pain Catastrophizing Scale.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06019.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: Complaints about medical malpractice have increased over time in Italy, as well as other countries around the world. This scenario, perceived by some as a “malpractice crisis”, is a subject of debate in health law and medical law. The costs arising from medical liability lawsuits weigh not only on individual professionals but also on the budgets of healthcare facilities, many of which in Italy are supported by public funds. A full understanding of the phenomenon of medical malpractice appears necessary in order to manage this spreading issue and possibly to reduce the health liability costs. Methods: The retrospective review concerned all the judgments drawn up by the Judges of the Civil Court of Rome, XIII Chamber (competent and specialized section for professional liability trials) published between January 2018 and February 2019. Results: The analysis of data concerning the involved parties showed that in 84.6% of the judgments taken into account, one or more health facilities were sued, while in 58.2% of cases, one or more health workers were present among the defendants. When healthcare providers are the only ones to be summoned, it is dentists and aesthetic doctors/plastic surgeons who undergo most of the claims. In the overall period analyzed, the amount paid was 23,489,254.08 EUR with an average of 163,119.82 EUR. Conclusion: The evidence provided by the reported data is a useful tool to understand medical malpractice in Italy, especially with regard to the occurrence of the phenomenon at a legal level, an aspect still hardly mentioned by existing literature.The concept of medical tort has ancient origins. In Roman law, if the Lex Cornelia settled a series of crimes by physicians, the Lex Aquilia (hence, the so-called “Aquilian” liability takes its etymological origin), introduced a penalty grading scale, including also further criminal hypotheses such as abandonment and experimentation as well as the possibility to compensate for the damage caused.An article of the Corpus Iuris Civilis introduced the punishability of physicians for their unskillful conduct.After the fall of the Roman Empire and the rise of the Goths, the issue was solved in a much more rough way. Visigoths asked physicians to pay a deposit before treating the patient, the Ostrogoths, quite simply, left doctors in the hands of patients’ relatives if a treatment-related fatal event occurred.In the modern era, the general concept of professional malpractice appears in English legal doctrine since the early 17th century. In 1768, Sir William Blackstone, in his famous work, Commentaries on the Laws of England, introduced the concept of mala praxis (hence the term malpractice, currently used), “Injuries … by the neglect or unskillful [sic] management of [a person’s] physician, surgeon, or apothecary … because it breaks the trust which the party had placed in his physician, and tends to the patient’s destruction” [1].In light of the above, it is doubtless that from the ancient times until today, despite the deep changes that have occurred, the ambivalent feeling underlying the patient–doctor relationship has remained substantially and justifiably unchanged and, perhaps, the resentment resulting from the disappointment in the event of failure to heal seems to have also strengthened as a consequence of the enormous scientific and technical progress achieved in the medical field.With regard to the patient–doctor relationship, in fact, the so-called “paternalistic” approach has changed, moving from the idea of physicians as hieratic entities, nearly endowed with magical powers, to healthcare services providers to turn to and from whom to expect an outcome.In modern times, especially in the last 2 decades, medical liability has received increasing attention both in the Italian [2] and international [3,4] medico–legal context.The increasing attention that forensic scientists devote to medical liability is encouraged not only by the captivating issues arising from it but also by the significant rise in litigations resulting from MedMal (Medical Malpractice). This contingency is widely perceived as a real “malpractice crisis” [5].An idea of the economic relevance of the issue comes from the reading of the report for the years 2016–2017 drafted by the Association of Italian Insurance (ANIA), which states that the amount relating to the premiums paid by public healthcare facilities amounted to 343.5 million EUR, compared to just over 87 million paid by private facilities. On the contrary, the amount of the insurance premiums paid by every single healthcare professional was ~208 million EUR. It was also reported that in 2016, insurance companies received approximately 14,803 claims, 6,884 of which were from healthcare facilities and 7919 from every single healthcare professional. The average cost per claim with regard to facilities was ~70,000 EUR, whereas for each professional involved it was ~40,000 EUR [6].This situation, in less than a decade, has repeatedly required the intervention of the Legislator on a subject that, until the entry into force of Balduzzi Law (Legislative Decree no. 158/2012), had never been evenly discussed, since the case-law had been charged to draw up its features.Over time, the development of a fruitful legal debate has also led judges to change their way of thinking and rethinking modern medicine, moving from the idea of medicine as an intellectual activity to that of a medical act performed by a healthcare provider.All the above has led to considerable concern in the medical professional community, whose members, being involved in expensive trials with unfavorable rules of evidence, have significantly changed their way of thinking and experiencing their medical profession, being caught in a vicious circle resulting from a widespread mistrust towards patients, on one hand, and the rise of the well-known phenomenon of defensive medicine, on the other.This debate culminated with the entry into force of Law 24/2017: “Provisions on patient care safety and professional liability of healthcare providers”, the so-called “Gelli-Bianco” Law.One of the key targets of the aforementioned law was to standardize and align medical liability both in the civil and criminal context. In particular, with regard to the civil framework, the dividing line between the contractual liability by the healthcare facility, on the one side (Article 1218 of the Italian Civil Code—Liability of debtor: “The debtor who does not exactly render due performance is liable for damages unless he proves that the non-performance or delay was due to impossibility of performance for a cause not imputable to him”.), and the tortious liability by the individual professional, on the other (Article 2043 of the Italian Civil Code—Compensation for unlawful acts: “Any intentional or negligent act that causes an unjustified injury to another obliges the person who has committed the act to pay damages”.), has been definitively set out by Law 24/2017, a distinction having significant effects both on the burden of proof and on the limitation period. In Italy, in fact, in the civil field, the patient who believes to have suffered damage due to medical responsibility can request compensation directly from the healthcare professional or the healthcare facility, public or private, where the same operate. This setting, from a strictly legal point of view, could appear more favorable to the physician involved than to the healthcare facility; in the area of contractual liability (to which, in the light of Law 24/2017, the healthcare facility is subject) the principle of the presumption of fault applies, with the creditor (the damaged patient) only having the burden of proof of non-performance and the extent of the damage, while, conversely, the debtor (healthcare facility) will have to demonstrate the supervening impossibility of the performance for reasons not attributable to him in order to escape the obligation to pay compensation. On the other hand, if only the healthcare professional is called upon, the field is, except for particular cases, that of tort liability, in which the damaged party must prove all the constituent elements of the illicit act and, therefore, both the damage and the violation [7].Despite the social and also economic relevance of the issue relating to litigations arising from medical liability, there is currently no public body charged to gather and examine data resulting from this phenomenon. Additionally, the continued absence of a national shared table for the judicial compensation for permanent disability from dynamic-relational damage makes it even harder to identify the exact correspondence between damage and amount of compensation, forcing case-law, on the issue and the merits, to a continuous search for uniform parameters and criteria.The difficulty in obtaining a precise estimate has already been raised by various authors who, while acknowledging the value of many projects aiming to this objective, have observed that the available data are still not homogeneous and not performing a complete and global picture of the Italian MedMal phenomenon. The source of the data analyzed and published so far, both at the national and international level, is mainly provided by the insurance area, a fact that entails a clear underestimation of the phenomenon. Indeed, in Italy, a considerable portion of claims is directly managed by the Healthcare Facilities, failing to disclose to the companies insuring the facilities themselves.In this respect, the lack of a “control room” able to integrate the data gathered by insurance companies, brokers, and organizations for the protection of patients and courts does not currently allow a precise estimate of disputes arising from Italian medical malpractice.The purpose of this paper is to report the legal data relating to the phenomenon of liability arising from Med Mal by the analysis of a whole year of judgments of the Civil Court of Rome, the main Court at the national level by the number of litigations, representing about 20% of all national disputes [8].The retrospective review concerned all the judgments drawn up by the Judges of the Civil Court of Rome, XIII Chamber, published between January 2018 and February 2019. The XIII Chamber of the Civil Court of Rome is the competent and specialized section for professional liability trials, including the medical sector. The judgments were provided by the Court pursuant to an agreement signed by the same and “Tor Vergata University of Rome”. With the exception of some replicated judgments, the documents were saved in PDF format and anonymized in order to preserve litigants’ personal identities and any connection between the tort in question and specific individuals or institutions. After the gathering and anonymization phase, 290 documents were analyzed. The actual analysis of the judgments was performed by three different auditors, experts in the field of medical liability. For the analysis, a working grid was prepared using the EXCEL program (office 365) in order to systematize the data mining. Furthermore, to reduce the risk of human error arising from the inter-subject variability between the three auditors, some locked fields with a drop-down list were added to the grid, so as to exclude the risk of inserting different definitions in overlapping fields. The grid is composed of the following items:-Judgment no.;-Occurrence year: the year when the event filing the lawsuit occurred;-Publication year of the judgment;-Difference between the registration and publication: a field to be filled with the difference, in terms of number of years, between the year of registration on the docket and the publication year of the judgment;-Difference between the event and claim: field to be filled with the difference, in terms of number of years, between the year of the tort bringing the action (if available in the text of the judgment) and the year of registration on the docket;-Medical specialty involved: a locked field has been set for this item with a drop-down list including all medical specialties acknowledged by Italian law;-Defendants/facilities: a locked field with a drop-down list to enter the type of facilities involved in the proceedings, if any, (available options: public health facility, private health facility, more public facilities, more private facilities, public and private facilities, no facility mentioned);-Defendants/persons: a locked field with a drop-down list to enter the type of professional(s) involved, should natural persons have been involved in the proceedings (available options: individual physician, several physicians together, non-physicians practicing a health profession, physician and non-physician health professionals, persons left unmentioned);-Provision of a court-appointed expert (CTU in Italy): a field with YES/NO locked options;-Inclusion of the CTU in the judgment: a locked field with a drop-down list to enter whether the provided technical report has or has not been fully or in part admitted by the judge (available options: yes/no/partially);-Civil conviction: a locked field relating to the outcome of the judgment with the possibility to select “yes” (if the defendant is guilty) or “no” (if the defendant is innocent);-Claimed damage: a locked field with a drop-down list to enter the type of damage claimed by the plaintiff (available options: injuries, death, consent, injuries and consent, death and consent);-Damage from consent: a locked field with a drop-down list to enter a possible dispute on the consent during the procedure and whether this dispute was deemed worthy of compensation by the judge (available options: claimed and paid, claimed and not paid, not claimed);-Damage for loss of chance: a locked field with a drop-down list to enter whether a damage from loss of chance was claimed or otherwise paid by a plaintiff during the proceeding (available options: claimed and paid, claimed and not paid, not claimed and paid, not claimed);-Compensation paid out;-Compensation claimed;-Compensation for consent: a field to enter the amount of compensation for damage from consent specifically if liquidated by the judge and mentioned in the judgment.Judgment no.;Occurrence year: the year when the event filing the lawsuit occurred;Publication year of the judgment;Difference between the registration and publication: a field to be filled with the difference, in terms of number of years, between the year of registration on the docket and the publication year of the judgment;Difference between the event and claim: field to be filled with the difference, in terms of number of years, between the year of the tort bringing the action (if available in the text of the judgment) and the year of registration on the docket;Medical specialty involved: a locked field has been set for this item with a drop-down list including all medical specialties acknowledged by Italian law;Defendants/facilities: a locked field with a drop-down list to enter the type of facilities involved in the proceedings, if any, (available options: public health facility, private health facility, more public facilities, more private facilities, public and private facilities, no facility mentioned);Defendants/persons: a locked field with a drop-down list to enter the type of professional(s) involved, should natural persons have been involved in the proceedings (available options: individual physician, several physicians together, non-physicians practicing a health profession, physician and non-physician health professionals, persons left unmentioned);Provision of a court-appointed expert (CTU in Italy): a field with YES/NO locked options;Inclusion of the CTU in the judgment: a locked field with a drop-down list to enter whether the provided technical report has or has not been fully or in part admitted by the judge (available options: yes/no/partially);Civil conviction: a locked field relating to the outcome of the judgment with the possibility to select “yes” (if the defendant is guilty) or “no” (if the defendant is innocent);Claimed damage: a locked field with a drop-down list to enter the type of damage claimed by the plaintiff (available options: injuries, death, consent, injuries and consent, death and consent);Damage from consent: a locked field with a drop-down list to enter a possible dispute on the consent during the procedure and whether this dispute was deemed worthy of compensation by the judge (available options: claimed and paid, claimed and not paid, not claimed);Damage for loss of chance: a locked field with a drop-down list to enter whether a damage from loss of chance was claimed or otherwise paid by a plaintiff during the proceeding (available options: claimed and paid, claimed and not paid, not claimed and paid, not claimed);Compensation paid out;Compensation claimed;Compensation for consent: a field to enter the amount of compensation for damage from consent specifically if liquidated by the judge and mentioned in the judgment.To complete the filling of the grid, the data were analyzed and checked by a fourth auditor different from those who dealt with the previous data mining from the judgments in order to exclude any inconsistencies and errors which, if any, would have required a re-evaluation of the judgments. During the analysis, it was found that 10 judgments did not deal with cases concerning medical professional liability lawsuits, but they were rather related to motor T.P.L. (RCA in Italy) and veterinary professional liability. It is for this reason that the latter judgments were not taken into account during the data mining phase.Finally, it should be noted that during the data mining it was not possible to fill in every single field per judgment, since the judgments were drafted in an open form that might not contain all the same information. The next processing phase was carried out through a data-cross match using the EXCEL program filter function (Windows Office 365).With regard to the proceeding timeline, analysis of the data showed an average period of 5.3 years (DS ± 3.87) between the claimed event and the start of the actual dispute, while the interval between the registration on the docket (that is its start) and the publication of the judgment was about 4.3 years (DS ± 1.84).However, it should be considered that in a few cases analyzed, a previous proceeding had already been filed and settled; in eight judgments the argument referred to a previous criminal conviction and in another three judgments an appeal was previously filed pursuant to art. 696 bis. Nevertheless, the final data concerning the proceedings timeline are to be considered reliable by virtue of the numerical scarcity of the aforementioned occurrences. Most of the judgments analyzed concerned proceedings registered from 2010 to 2017; of these, approximately 54% had started between 2013 and 2015.The analysis of data concerning the involved parties showed that in 84.6% of the judgments taken into account, one or more health facilities were sued, while in 58.2% of cases, one or more health workers were present among the defendants. On the 43 judgments (15.4% of the total) in which only physicians were involved, 25 (or 58.1%) concerned dentists, while a smaller part (six judgments) were about cases related to plastic surgery, three orthopedics, two ophthalmology, and another two oncology, all different in the remaining five. It is important to highlight that in 74% of the judgments concerning dentistry malpractice, dentists were singly summoned before civil courts, whereas aesthetic surgeons/doctors were singly summoned in 27% of cases.In 51% of cases (144 judgments out of 280 analyzed), some medical malpractice profiles were detected; in 46% of cases (128 judgments), liability was excluded, whereas in 3% (eight judgments) no judgment was settled, as a result of the reaching of a settlement agreement or the ineligibility of the claim itself (Figure 1).In 280 judgments analyzed, the judge provided the execution of a court expert’s report in 93% of cases (260 lawsuits); in the remaining 7% (20 judgments) any type of technical investigation was required by the judge. In these cases, proceedings concerned either the ineligibility of the claim or a previous appeal pursuant to Article 696 bis of the Italian Code of Civile Procedure or the case could be solved without the need of a court expert. The acknowledgment by the judges of the court-appointed experts’ outcomes occurred in 92.7% of cases (241 judgments out of 260 in which an expert witness was appointed by the Court). A partial acknowledgment occurred in 3.5% of cases, while in 3.8% the conclusions by the court-appointed experts were entirely rejected by judges.With reference to the data relating to the type of the most claimed damage, it is clear that physical injuries were claimed by the injured parties in 80.6% of cases. Deaths were significantly less and affected only the remaining 19.4% of disputes. However, only in one case, with a conviction of the defendants, improper informed consent was complained as a damage (Figure 2).During the analysis, the authors differentiated those cases claiming both injuries or death and the prejudice to the right to self-determination arising from an omitted or poor consent; the latter prejudice was explicitly claimed by the plaintiff in about 24.9% of cases. This prejudice, despite having been claimed in about one out of four cases, was admitted only in 42% of the cases in which it was complained, with an average compensation of 8648 EUR.With regard to the damage claimed, the 278 judgments taken into consideration showed that in 22 of these, approximately 8%, the plaintiff requested to the judge damages from loss of chance. Furthermore, it should be noted that in two judgments such damage was liquidated despite the lack of an explicit claim by the party, or at least, the reading did not show a request to that effect.For each medical field, the number of proceedings along with the judgments dealt with facilities liability and medical professionals’ malpractice were extrapolated. The data obtained were then compared as seen in Figure 3.It should be noted that in 280 judgments analyzed, eight were left undefined (termination of the dispute subject-matter and reaching of a settlement agreement, etc.) and so counted as non-conviction.As many as 172 out of 280 judgments involved only six medical specialties, thus representing 61.4% of the cases, with a total number of 107 convictions, or 74.3% of the total convictions. Table 1 shows the ratio of lawsuits and convictions for the six most involved specialties.In the overall period analyzed, the amount paid was 23,489,254.08 EUR (except for any legal interest or court fees) with an average of 163,119.82 EUR. The highest compensation paid for a judgment was 4,741,34.43 EUR and concerned a neonatology lawsuit. Table 2 shows the average compensation paid for each of the six most involved specialties.In 117 out of 280 judgments, it was also possible to extrapolate the data relating to the average sum of compensation claimed, amounting to 473,694.63 EUR, or about three times the average value of the compensation actually paid.As reported in the opening section, the overall analysis of the data relating to researches carried out on medical disputes shows a clearly growing trend, so that in some states, legislators have provided regulatory measures, reducing its excessive increase [9], especially in case of disputes initiated in the absence of any liability or of any sort of damage occurrence.The importance and relevance of the phenomenon are suffered by the civil society too, with an impact at an economic and social level.With regard to the economic framework, in the United States, a recent estimate of the annual costs of this phenomenon showed that they amount to 55.6 billion USD, or 2.4% of total healthcare spending [10].In Italy, as mentioned in the statistical bulletin published by the Insurance Authority (IVASS) [11], the insurance premiums collected from civil liability for medical malpractice-related risks (both for healthcare facilities and individual professionals) amounted to 612 million EUR in 2018, up compared to 2017 (+3.7%).The average premium covering a public healthcare facility costs was 456.000 EUR (+25.5% compared to 2017). The value is 24 times higher than that for a private healthcare facility.The economic burdens weighing on public facilities are not only represented by insurance premiums, but also, of course, by compensation paid using the facilities’ own funds, being lower than the expected deductible or SIR (Self Insured Retention) pursuant to the insurance contract. It should also be noted that an increasing portion of Italian health facilities has opted to self-insure (self-risk retention), as an exclusive form of protection. The document drawn up by IVASS reported that the allocations made in 2017 amounted to 592.4 million EUR (+16% compared to the previous year) [11].If, on one hand, the approach represented by self-risk retention seems to be potentially virtuous, on the other, it requires the implementation by the healthcare facility of adequate internal structures made up of healthcare professional experts in the field, able to settle claims quickly and ensure a proper defense during proceedings.The economic burden weighs directly on every single healthcare provider, especially if he/she is a physician; the IVASS statistical bulletin shows that the average premium paid by medical providers is 1001 EUR, against the 183 EUR paid, on average, by non-medical health professionals.As mentioned at the beginning of this section, medical malpractice costs are also social.Numerous authors have identified a growing use by the medical community of the so-called “defensive medicine” as an indirect consequence arising from health litigations [12], which, however, do not lead to an improvement in quality of care and outcomes for patients [13].Furthermore, it was highlighted that to be involved in a legal action may be stressful for the healthcare professional (also due to the length of the judicial process), who may suffer damage to his/her reputation and a loss of trust in the physicians’ own performance by other patients [14].With respect to the proceeding timeline, the analysis carried out made it possible to highlight two important facts: on average, around 5.3 years elapse between the contested event and the start of the litigation, whilst the time elapsing between the registration of the proceedings (coinciding with the start of the same) and the publication of the judgment (coinciding with its end) is around 4.3 years. From the collected data it is clear that, in Italy, the proceedings relating to medical professional liability lawsuits are characterized by a long latency period between the damaging event and the reach of judgment by the injured party or his/her heirs. This discrepancy can be explained by patients’ delayed comprehension of the damage suffered as a result of both inadequate medical assistance and the need to find, under the Italian civil code, the liable persons and the “abstractly eligible” breaches occurred before reaching the judgment.This period of time, certainly important, has a great impact on the economic management of the capital placed in reserve by health companies and insurance companies, as well as on the cost of the legal fees needed for the management of each proceeding.It should also be considered that the analysis focused only on actual judicial actions. It is, therefore, very likely that in some cases, the possibility of an out-of-court resolution of the dispute was considered prior to the settlement of the judgment. n this regard, it should be noted that Article 8, Law 24/2017, provides that the submission of an appeal pursuant to Art 696 bis of the Italian Code of Civil Procedure or, alternatively, a mediation process attempt, is a prerequisite of admissibility for compensation.The reported data confirm what has already been highlighted by various authors on the most involved specialties in lawsuits and convictions [15,16,17].The reported analysis has shown that most of the disputes involved health facilities, especially public, compared to 15.4% in which health professionals were directly and exclusively summoned. Among the latter, physicians were the most involved (this has indirectly been confirmed by the significantly higher average cost of medical insurance policies than those intended for other health professionals). In this regard, it has been estimated that, for consultants specializing in specialties at higher risk, the possibility of being involved in litigation within the age of 45 is 88% [18].By our analysis, it can be concluded that when healthcare providers are the only ones to be summoned, it is dentists and aesthetic doctors/plastic surgeons who undergo most of the claims (74%).This figure was entirely expected, since the dentist and aesthetic doctor medical practice, mainly carried out on a private basis, exposes healthcare professionals themselves as a result of the signing with the patient of a specific “contract” under the ex-art. 1218 of the Italian Civil Code. It should also be noted that for these two specialties, Italian law aims to admit a substantial result obligation [19].This investigation documented that in 51% of judgments, the defendant party was held liable. The comparison of this result with the data reported by ORME (Observatory on Medical Liability), referring to less than a decade ago, reveals that there is a difference arising from the decrease of data concerning the acknowledgment of liability against the defendants; in fact, ORME analysis revealed a liability rate exceeding 60% [20].The data concerned are, in the authors’ opinion, extremely meaningful, and would require a regular updating in order to find out the impacts resulting from legislative changes, doctrine, and especially case-law, with their potential implications also in the practice, claims management, and insurance market focused on malpractice.With regard to the type of damage claimed by patients, data suggest that most cases dealt with physical injuries, although compensation was requested by heirs for the patients’ death only in about 20% of cases.In addition, in one out of four cases, injuries resulting from an infringement of the right to self-determination were also claimed, and this results, practically, in a lack of informed consent to the medical act.The concept of informed consent, used with this meaning for the first time in 1957, has been recently introduced in Italian legislation with the Law n.219/2017, which is the first law on this issue. It can be defined as a real communication path between doctor and patient, by which the doctor makes the patient competent about all medical information needed in order to make him/her aware and able to voluntarily and consciously accept or refuse the planned medical treatment. The three fundamental criteria that are needed for informed consent are that the patient must be competent, adequately informed, and not coerced [21].For years, the international scientific literature has shown that effective communication with the patient is an undeniable advantage for the doctor too, leading to higher patient compliance with the planned treatment as well as better tolerance in the event of complications [22].In Italy, the importance of the informed consent concept in modern medical practice has been strengthened after the entry into force of Law 219/17, “Regulations in the area of informed consent and advance treatment directives” [23]. It is assumed that the introduction of this new law, by which new and clear directives have been set out imposing both on the doctor and health facility a huge attention in the collection and documentation of the consent, could lead the civil judge to an increase in interest with respect to this type of prejudice. Furthermore, a possible rise in new cases worthy of protection and therefore of compensation following the new regulatory provisions on the shared care plan should not be underestimated [23].Given that, under the Italian Civil Law, the amount of compensation for damage arising from omitted or poor consent is equitably liquidated by the court, as set out in Article 1226 of the Italian Civil Code (Art. 1226 of the Italian Civil Code: “If damages cannot be proved in their exact amount, they are equitably liquidated by the court.”), it was observed that this prejudice, explicitly mentioned in a few judgments, was liquidated in a range fluctuating between 2000 EUR and 20,000 EUR, with an average of 8648 EUR. Given the variability arising from the economic assessment of injuries, it might be helpful to draw up a compensation grading scale allowing the various figures involved to predict, albeit roughly, the economic risk in the event of conviction, allowing an equitable and standardized, as much as possible, “monetization “of the prejudice.Particularly worthy of mention are data concerning paid compensation for damage from loss of chance. This case is actually a relative novelty in the field of medical liability that is still not fully accepted and on the assessment of which there are still no shared guidelines either in the legal medical context or in the case-law, uncertainty that is also reflected on the verification of the causal link between conduct and event, on which there is not yet a universally shared orientation [24].The analysis carried out has ascertained that the damage from loss of chance is still a reality hardly admitted by the court, having been confirmed only in 3% of the analyzed judgments; given the shortage of the sample collected, it might be useful, in a further analysis, to focus attention on the jurisprudential implications from this particular type of damage.The evidence provided by the reported data are useful tools to understand medical malpractice in Italy, especially with regard to the occurrence of the phenomenon at a legal level, an aspect still hardly mentioned by literature.Some of the findings, especially those concerning the most affected specialties and the type of damage, confirm what has been already outlined by several authors at an international level. Conversely, with regard to the case of loss of chance, damage arising from the breach of the right to self-determination and the amount of compensation, the analysis carried out has given rise to some new and interesting issues, with potential doctrinal and practical implications. In this sense, the information provided could be used not only by forensic scientists working in the field, but also by health facilities that, due to the growth of the MedMal phenomenon, have to deal with a considerable number of disputes with significant disbursements of money.A new study carried out, taking into consideration a longer period, would allow researchers to obtain a larger sample and, consequently, higher statistical reliability. In addition, it might be interesting to consider in a future analysis the outcome of the judgments following the registration on the docket after the entry into force of Law 24/2017, in order to understand the effect of the recent regulatory intervention.It should be noted that during the data mining it was not possible to fill in all fields for every single document analyzed, since the judgments, being drafted in an open form, might not contain the same amount of information. For example, in some cases, it was not possible to deduce from the judgments the exact claim by the plaintiff, while in many cases the economic quantum requested by the plaintiff was not mentioned. However, this is not believed to have significantly influenced the validity of the data collected. It could also be useful to obtain the entire case file for each judgment analyzed. Unfortunately, despite the introduction in Italy of the Telematic civil proceedings, actually, it is not possible to access Telematic civil trials.Conceptualization, L.T.M. and A.M.C.; methodology, M.T. and J.G.; formal analysis, M.T., P.P., and J.G.; data curation, M.P. and J.G.; writing—original draft preparation, M.P. and L.D.L.; writing—review and editing, M.T.; supervision S.M.; project administration, L.T.M. and A.M.C. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.Data available on request due to privacy restrictions. The data presented in this study are available on request from the corresponding author. The authors declare no conflict of interest.Judgment outcomes.Categories of damages claimed by patients in court.Cases and convictions divided by medical specialty. The ordinates show the branches involved and the abscissae show the number of total cases and the number of convictions.Most involved specialties and risk of recognizing liability in court.Average compensation for the most involved specialties.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06020.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Members of the SHINE Trial team are listed at https://doi.org/10.1093/cid/civ844.Background: With millions of people experiencing malnutrition and inadequate water access, FI and WI remain topics of vital importance to global health. Existing unidimensional FI and WI metrics do not all capture similar multidimensional aspects, thus restricting our ability to assess and address food- and water-related issues. Methods: Using the Sanitation, Hygiene and Infant Nutrition Efficacy (SHINE) trial data, our study conceptualizes household FI (N = 3551) and WI (N = 3311) separately in a way that captures their key dimensions. We developed measures of FI and WI for rural Zimbabwean households based on multiple correspondence analysis (MCA) for categorical data. Results: Three FI dimensions were retained: ‘poor food access’, ‘household shocks’ and ‘low food quality and availability’, as were three WI dimensions: ‘poor water access’, ‘poor water quality’, and ‘low water reliability’. Internal validity of the multidimensional models was assessed using confirmatory factor analysis (CFA) with test samples at baseline and 18 months. The dimension scores were associated with a group of exogenous variables (SES, HIV-status, season, depression, perceived health, food aid, water collection), additionally indicating predictive, convergent and discriminant validities. Conclusions: FI and WI dimensions are sufficiently distinct to be characterized via separate indicators. These indicators are critical for identifying specific problematic insecurity aspects and for finding new targets to improve health and nutrition interventions.Globally, over two billion people do not have regular access to safe, nutritious and sufficient food [1], while about four billion are exposed to water stress at least once a month [2]. However, the definitions of both food insecurity (FI) and water insecurity (WI) go beyond only inadequate access. The most widely accepted definition of food security is “when all people at all times have physical, social and economic access to sufficient, safe ad nutritious food to meet their dietary needs and food preferences for an active and healthy life” [3]. Water security is analogous to food security and refers to “safe and reliable access to adequate quantity and quality of water for consumption, economic production and cleanliness” [4,5]. Both these definitions identify multiple dimensions of FI and WI, like availability, access, quality and safety, and reliability of supply [3,6,7]. Nevertheless, inconsistencies exist between these internationally recognized definitions and the ways in which the concepts of FI and WI are applied in research and policy [8,9].Decades of work on FI have produced a diverse range of metrics at the household-level [8]. Many of these existing metrics are often used interchangeably even though they capture different combinations of FI dimensions [8,10]. For example, the commonly used Household Food Insecurity Access Scale (HFIAS) [11] and Food Insecurity Experience Scale (FIES) [12] capture economic food access and food sufficiency, the Household Dietary Diversity Score (HDDS) captures diet quality [13] and the Coping Strategy Index (CSI) captures food access experiences in emergency settings [14]. These metrics, due to their respective conceptualization and intended use, are correlated, but not equivalent [10,15,16,17]. In empirical comparisons of some of these household measures, different estimates of FI prevalence were reported for the same population at the same time point [16]. For example, CSI in Ethiopia identified 58% food insecure households, but HFIAS detected 66% [16]; in Bangladesh and India, 34% of households with adequate caloric intake were classified as food insecure based on CSI or HFIAS [17]. As a result of these challenges, the Food and Agriculture Organization (FAO) recommends the development and application of multiple indicators to distinguish FI dimensions, to improve accuracy and specificity of FI assessments, and to prevent misclassification based on dimensions of FI [18].Assessment of household-level WI is less established than FI, and few cross-culturally valid metrics exist. The most common measures of WI developed by the World Health Organization (WHO) categorize water access based on fetching time, and water sufficiency based on quantity of water available per person [19]. Another measure from the Joint Monitoring Programme (JMP) for Water Supply and Sanitation approach distinguishes “improved” water access based on the type of the main water source [20]. Population- and context-specific WI scales have also been developed in Kenya [21,22], rural Ethiopia [23,24], urban Nepal (Household Water Insecurity Scale (HWIS)) [25], slums of India (Water Insecurity Experience Scale (WIES)) [26], Uganda (Household Water Insecurity Access Scale (HWIAS) [27], Bolivia [28], Brazil [29], Jamaica (Water Accessibility Index (WAI)) [30], and the colonias along the US-Mexico border [31]. These metrics vary in the WI dimensions they attempt to capture. For instance, the Household Water Insecurity Experiences (HWISE) Scale, a unidimensional 12-item cross-cultural scale, captures emotions and behaviors predominantly in response to inadequate water supply [5]. The WAI captures only water access using seven components including affordability, source and collection time [30]. Although these existing WI measures are instrumental in identifying water insecure households, further distinguishing between WI dimensions will allow targeted interventions and policy decisions.FI and WI have been shown to contribute directly and indirectly to undernutrition [1,7], psycho-social stress [25,28,32,33,34], and increased risk of infectious and chronic diseases [1,7]. They often chronically co-exist within the same households [35]. The high prevalence and co-occurrence of FI and WI may have synergistic effects on adverse health outcomes [36]. This may have significant implications for achieving sustainable development goals (SDG) two, three and six to “end hunger and all forms of malnutrition”, “ensure healthy lives and promote well-being for all” and “ensure availability and sustainable management of water and sanitation for all” [37], respectively. However, few studies have been able to explore FI and WI concurrently as environmental stressors in causal pathways to specific health and nutrition outcomes [36,38,39,40,41]. This is in part because of the complexity surrounding the conceptualization and the measurement challenges of FI and WI as described above [8,42,43].More than one third of the world’s food insecure and water insecure people live in Sub-Saharan Africa (SSA) [44,45]. In Zimbabwe, economic crises, recurrent droughts, and depletion of ground water are causing severe food and water shortages [46]. Approximately 30% of the rural Zimbabwean population is undernourished [47] and obtains water from unprotected sources [48]. A prior study in rural Zimbabwe demonstrated the value of using multiple FI indicators in designing and evaluating interventions, when CSI and HDDS gave different prevalence of food insecure households [49]. The same is likely true for WI, where the application of distinct indicators for each dimension will be more valuable. Our objectives were therefore to develop separate multidimensional measures for household FI and WI, and to test their internal, predictive, convergent and discriminant validities in the context of rural Zimbabwe.Data for the development of the household measures of FI and WI were obtained from the Sanitation, Hygiene and Infant Nutrition Efficacy (SHINE) trial. The trial’s primary objectives were to test the independent and combined effects of an improved water, sanitation and hygiene (WASH) intervention, and an improved infant and young child complementary feeding (IYCF) intervention on stunting and anemia among rural Zimbabwean children. The design, protocol, and primary outcomes have been published elsewhere [50,51,52]. Briefly, SHINE was a four-arm cluster-randomized community-based 2 × 2 factorial trial conducted in two rural districts in Zimbabwe: Shurugwi and Chirumanzu. The two districts were divided into 212 clusters which were then randomly allocated to one of the four trial arms: (1) Standard of Care (SOC), (2) SOC + IYCF, (3) SOC + WASH and (4) IYCF + WASH. Recruitment occurred between 22 November 2012 and 27 March 2015. Village health workers (VHWs) employed by the Zimbabwe Ministry of Health and Child Care prospectively identified and referred eligible women for the trial. Only women residing permanently in a cluster and who were pregnant at the time of recruitment were enrolled. Written informed consent, in the language of their choice (English, Ndebele, or Shona), was obtained prior to data collection. SHINE was approved by the Medical Research Council of Zimbabwe and the Johns Hopkins University Bloomberg School of Public Health Institutional Review Board.SHINE included an extensive structured questionnaire to collect detailed information on household, maternal and child characteristics. Baseline data collection spanned the recruitment period mentioned above. A few weeks after obtaining consent, research nurses made home visits for face-to-face interviews with the women. Additional home visits for subsequent data collection were also made at one, three, six, 12- and 18-months post-partum, until the end of the study in July 2017. The questionnaire and data collection protocol are available on OSF at https://osf.io/w93hy/ (accessed on 17 May 2021).From the 5280 pregnant women who were recruited, 4675 took part in the baseline interview. For the following analysis, the sample was restricted to households with complete information on the selected food (N = 3551) and water (N = 3311) variables. Figure 1 illustrates participant inclusion.The creation of FI and WI measures were carried out in a stepwise manner, starting with item variable selection for inclusion in the quantitative analyses. The next steps included descriptive analyses, item reduction, multiple correspondence analysis (MCA) with extraction and rotation of dimensions, and validity assessments.The starting point for item selection was the internationally accepted definitions and dimensions of FI [3] and WI [7]. The FAO [3] and Action Contre la Faim (ACF) [53] provide some recommendations for indicators of FI dimensions, while WaterAid [7], Global Water Partnership (GWP) [54] and JMP [20] suggest items for WI dimensions. Indicators relevant to rural Zimbabwe and available from SHINE were then selected. Table 1 provides detailed descriptions of all variables selected to represent each FI and WI dimension. Brief justifications are also provided below for the choice of item variables:A.1. Food availability refers to the food supply aspect of food security [3]. This dimension considers whether food is actually present for the population [55]. At the national-level, this has historically been addressed via the use of food balance sheets of food production and imports. At the rural household-level, food availability may be captured by considering food stocks, presence of markets and ability to produce food. We used three variables to operationalize this dimension: (1) number of days of staple food stocks available for household members to eat according to their needs, (2) availability of a garden where the household grows fruits and vegetables, and (3) the availability of left-over food from the last cooking occasion.A.2. Food access concerns economic, physical and social resources that enable acquisition of sufficient, nutritious and preferred foods in a dignified manner [3]. Physical food access is linked to infrastructure and at the household-level can be captured by considering time spent, distance travelled and transportation to safe food sources. Economic access depends on the ability of households to purchase or barter resources to obtain food [55]. Social access concerns food preferences in terms of taste, health requirements and religious restrictions. It also implies that food is obtained in socially acceptable ways. The following seven household-level variables were considered for this dimension: (1) access to preferred food, (2) food sufficiency for all household members, (3) help required from family and/or friends to obtain food, (4) purchasing or borrowing food on credit, (5) selling assets for food, (6) time from home to food market, and (7) method of transportation to food market.A.3. Food utilization reflects differences in the intra-household allocation of food, nutritional quality of food, and food safety in terms of preparation, handling, and storage [8,55]. Within SHINE, four variables were available as proxies for food utilization: (1) household dietary diversity, (2) handwashing behavior prior to handling food, (3) whether food containers were covered, and (4) food storage location. No information was available as proxy for intra-household allocation, which also depends on age, work load, and other factors.A.4. Food stability covers the barriers and promotors of food security dimensions [8,55]. At the household-level, this can be captured by considering exposures to risks, shocks or vulnerabilities that influence the ability of household to consistently acquire food [55]. The variables most appropriate to represent this dimension from SHINE were household experiences of social, economic, agriculture and health shocks.B.1. Water availability depends on the physical presence of water resources or infrastructure that makes it available in sufficient quantity to households [56]. Sufficient quantities of water must be available for drinking to prevent dehydration (≥5 L per person/day) and for cooking, bathing, hygiene and sanitation (>100 L per person per day) [19]. Within SHINE, two variables were considered: (1) volume of water, calculated from storage capacity of water containers and water collection frequencies, and (2) whether the households had access to water for irrigation purposes.B.2. Water access refers to physical delivery and economic access to water. Methods for assessing water access include the distance to water points, fetching time, and water expenditures [19,57]. Water access is inadequate if households have to travel >1 km or >30 min (return journey) to collect water [19,58]. Water is affordable if households spend <3–5% of their total income on it [59]. Five variables were considered to assess water access: (1) whether the household purchases water, (2) drinking water collection time, (3) distance to drinking water point, (4) non-drinking water fetching time, and (5) distance to non-drinking water point.B.3. Water utilization is meant to reflect the quality and safety of water for drinking and other purposes. Physical quality can be measured by considering the color, smell and taste of the water. Chemical quality and microbiological safety are determined by testing turbidity, total dissolved solids, chlorine levels and the presence of bacterial coliforms in the water. In low-income settings, types of water sources are used as proxy for water quality and safety [19]. For instance, protected sources such as piped water, boreholes and wells are considered microbiologically and chemically safer compared to surface water from rivers or streams. To capture this dimension, three SHINE variables were used: (1) reported satisfaction with the water smell, color and taste, (2) water source for drinking, and (3) water source for non-drinking purposes.B.4. Water reliability refers to whether water supply is consistent or intermittent. Whether water is piped into dwellings or available off premises, it may be periodically or seasonally inaccessible [2]. To assess the reliability of water supply among SHINE households, two variables were considered: (1) whether drinking source and (2) non-drinking source ran dry over the past year.Separate multiple correspondence analysis (MCA) were conducted on the selected item variables to develop FI and WI measures. MCA for categorical variables is equivalent to exploratory factor analysis (EFA) or principal component analysis (PCA) designed for continuous variables [60]. Analyses were conducted as explained below using Stata Version 16 (StataCorp LLP, College Station, TX, USA) for descriptives, ‘FactoMineR’ [61] and ‘PCAmix’ [62] packages from the software R Version 4.0.2 for MCA and factor rotation, and MPlus Version 8.4 (Muthén & Muthén, Los Angeles, CA, USA) for validity tests.First, we looked at the distributions of participants across the categories of each item variable using frequencies and percentages. Variables with categories reporting frequencies of ≤5% or ≥95% were excluded.Second, we ran polychoric correlations on all variables. Items indicating negative correlations and those without adequate variance (<0.1) were dropped. We also used the Kaiser-Meyer-Olkin (KMO) measure for sampling adequacy and Barlett’s test of sphericity to ensure robustness of our approach. We then carried out MCA on the remaining variables. Scree plots were used to decide the number of dimensions for extraction. We investigated factor extraction using oblique (geomin) and orthogonal (varimax) rotations. Since correlations among the extracted factors were small (<0.5), we report varimax-rotated loadings in our results. Dimensions extracted were interpreted and named based on the variables that loaded on them from the theoretical framework (Table 1). We report the squared correlation ratios between each item variable and dimension, eigenvalues and percentage explained variances. Squared correlation ratios <0.20 were not considered relevant in explaining a dimension. We then used post-estimation commands in R to obtain standardized dimension scores for individual households.Validity refers to the extent to which certain measures are acceptable indicators for what they are intended to capture [63]. We tested four types of validity for our FI and WI measures: internal, predictive, convergent and discriminant. These are briefly described in Table 2 with an explanation of their purpose and statistical methods used. For internal validity, we assessed multidimensional model fit via confirmatory factor analysis (CFA) in two groups: (1) a sub-sample of the baseline participants constituting 60% of the dataset, and (2) the same baseline households more than 18 months after the baseline interview. We used model fit statistics such as root mean square error of approximation (RMSEA), standardized root mean square residual (SRMR), comparative fit index (CFI), and Tucker Lewis index (TLI). Satisfactory fit was determined using recommended arbitrary cut-offs of RMSEA ≤ 0.05, SRMR ≤ 0.08, CFA ≥ 0.95 and TLI ≥ 0.95 [64,65]. CFA was performed in MPlus using geomin rotation with diagonally weighted least squares estimator (WLSMV).For predictive, discriminant and convergent validity, we used a group of exogenous variables, also obtained from the SHINE trial. Self-reported perceived health status of women was measured using an adapted version of the RAND Health Survey [66]. Scores for perceived health status ranged from 0 to 5 units, with 0 indicating least healthy and 5 most healthy [67]. The Zimbabwe-validated version of the 10-question Edinburgh Postnatal Depression Scale (EPDS) was used to assess depression among the women [68]; those with a score ≥12 out of 30 were classified as clinically depressed. Household receiving food aid over the past 12 months from government or other organizations (yes/no) was self-reported by women. Usual frequency of water collection was reported as daily, weekly or monthly. HIV-status of the participating women was determined via rapid blood tests performed by trained nurses [50]. Household socio-economic status (SES) was based on a household wealth index [69]. Seasonality was determined based on the date of interview; hungry season was from January through March and rainy season was from November through March. These variables were used as predictors in simple regressions to estimate associations with FI and WI dimension scores from MCA.We tested the robustness of the MCA results after accounting for missingness in the selected items. Almost all variables had <10% missing values (Table S1). Lower SES, HIV-status and interview months were found to influence missingness (Table S2). To account for missing data uncertainty, we imputed missing variables using the multiple imputation by chained equations (MICE) method via the ‘MICE’ function from the ‘missMDA’ package in R [70]. We then re-ran MCA by including the additional households with imputed variables. Only households with less than three imputed variables were used for sensitivity analysis. The sample size increased considerably (N = 4622 for FI and N = 4575 for WI).Table 3 summarizes the distribution of households and participating women according to socio-demographic and food- and water-related characteristics. The two samples used to generate FI and WI measures were similar for all variables. The average age of the respondents was 26 ± 7 years. Approximately the same proportion of participants were randomized in the four SHINE trial arms, 15% were living with HIV, 44% were interviewed during the rainy season and 28% during the hungry season, more than 40% of the participants had completed secondary school, more than half had children, and 45% were of Apostolic faith. At least 6% of women were clinically depressed. More than 90% were partnered and were not generating income outside the home.Descriptive and item reduction analyses suggested the removal of three food variables (transportation, leftover food and food container) and two water variables (water purchase and irrigation water) from the initial list (Table 1). The KMO statistic was above 0.74 for food and 0.48 for water. Barlett’s sphericity tests were significant for both food and water samples (p < 0.01). These two indicators suggest that the final 15-item food dataset is adequate to further explore underlying latent constructs. Adequacy of the 12-item water dataset is poor according to the KMO statistic. However, Barlett’s sphericity statistic suggests substantial inter-item correlation between the water variables, indicating that factor analysis may still be useful.A visualization of the scree plots for both food and water MCA showed that the eigenvalues plateaued after the third dimension for FI (Figure 2). For WI, there appears to be multiple inflection points, at dimensions three and five. However, we consider only the first three dimensions, since beyond this, the water items were cross-loaded at lower factor loadings.We therefore chose three dimensions for FI. They are named as follows based on the items that loaded on them (Table 4): (1) “poor food access” (food not preferred, insufficient food, food help and food on credit), (2) “household shocks” (economic shocks, agriculture shocks and health shocks), and (3) “low food availability and quality” (stock of staple food, garden and household diet diversity). These dimensions accounted for a cumulative variance of 20.12%. Similarly, the WI dimensions are named according to the characteristics they capture (Table 4): (1) “poor water access” (time to drinking source, distance to drinking source, time to non-drinking source, distance to non-drinking source), (2) “poor water quality” (drinking source, non-drinking source, water satisfaction), and (3) “low water reliability” (whether drinking and non-drinking sources ran dry). Together, these dimensions accounted for a cumulative variance of 31.36%.The median dimension scores indicated low insecurity in our population with wide interquartile ranges. Nevertheless, the minimum and maximum values ranged from −1.04 to 3.93 units suggesting variations in FI and WI across households. Poor water access was significantly correlated with all FI dimensions; poor water quality was correlated with poor food access and low food quality and availability, and low water reliability was correlated with poor food access and household shocks. However, none of the correlations were very high (r < 0.15 for all pairwise associations) (Table S3).Table 5 summarizes model fit statistics for internal validity of the multidimensional FI and WI measures. There was strong support for both FI and WI measures, with at least two of the four indices indicating satisfactory cut-offs.Almost all assessments related to predictive, discriminant and convergent validities were in the expected direction, although all were not statistically significant (Table 6). Higher perceived health status was associated with lower FI and WI scores. Depression, lower SES and living with HIV were associated with higher FI and WI scores. Households interviewed from January through March (hungry season) scored higher on FI dimensions, while those interviewed from April through October (dry season) had poorer water access. Households receiving food aid had higher scores on shocks. Less frequent water collection was associated with poorer water access.Our sensitivity analyses, after imputation of missing values for the relevant food and water item variables, confirmed the robustness of the results presented. The number of dimensions identified, items loading on each dimension, correlation ratios and eigenvalues were similar to the complete case analysis (Table S4).The goal of this study was to develop new measures of FI and WI that are cognizant of their multidimensionality to advance the discussion on impactful nutrition and health interventions for vulnerable populations. We used rigorous analytical procedures to develop these measures among rural Zimbabwean households. Each of the FI and WI measures obtained consist of three dimensions. The multidimensionality observed through the development process is in part consistent with the definitions of both FI [3] and WI [7]. The distinction between the dimensions from our household measures of FI and WI provide additional depth that may complement existing FI and WI metrics. The quantification of the dimensions will advance our understanding of their prevalence and consequences for health and well-being. We named our measures the multidimensional household food insecurity (MHFI) and the multidimensional household water insecurity (MHWI), respectively.MCA with 15 food-related variables resulted in the identification of multiple dimensions of FI as theorized previously (Table 1). The first dimension refers to “poor food access” through quantity, affordability and food preference. Access to food includes the social, physical and economic aspects [3,55]. However, this first dimension captures only the social and economic access to food. The variables ‘time to market’ and ‘mode of transportation to market’, that represent physical access to food, did not load on any dimension. The second FI dimension, “household shocks”, describes the reliability component of food supply and includes households’ experiences of economic shocks through loss of employment or assets; agricultural shocks through loss of crops and livestock; and health shocks through death, disease and injury of household members. Social shocks such as conflict or legal problems were not retained in this dimension. The third dimension, “low food availability and quality”, includes poor household dietary diversity, low stock of staple food and lack of household garden. This dimension partially encompasses several theoretical components of FI: utilization (dietary diversity) and availability (stock of staple food, having a garden). This may be because after the item reduction step, only two variables were left to represent availability (stock of staple food and having a garden) and two to represent utilization (handwashing and food storage location) in the MCA model.In contrast to existing FI scales (HFIAS, FIES, CSI) whose internal consistency arise from assessing similar constructs, our FI dimensions reflect different conceptual constructs due to the use of variables with disparate measurement approaches and recall periods (Table 1). These differences may impede our ability to interpret and compare the FI dimensions to each other, and to existing scales. Nevertheless, our three-dimensional measure of FI was found to be valid within a test sample of the SHINE population and across time (Table 5). Moreover, it is possible for households with similar scores on one of the existing FI metrics to have different characteristics on individual FI dimensions [10,43,49]. This is an important limitation for exploring impact pathways or to identify relevant intervention targets, because composite scores do not inform on which aspects of availability, access, utilization or stability to modify. Therefore, the three FI dimensions identified in our population contribute to addressing the need for multiple indicators to improve the identification of food insecure households [3,18,43,71].In our study, 12 water-related variables loaded on three dimensions (Table 4). The first dimension refers to “poor water access” and includes time and distance variables. The second dimension, “poor water quality”, includes types of water sources and degree of satisfaction with water quality. Finally, the third dimension refers to “low water reliability” and includes information on whether water for drinking and non-drinking purposes was unavailable at any point. Of the four hypothesized dimensions of WI (Table 1), availability in terms of water quantity was the only dimension not identified, likely due to the low number of variables in SHINE to represent this dimension thoroughly. Interestingly, neither Stevenson et al. (2012) in Ethiopia [24] nor Tsai et al. (2016) in Uganda [27] reported any correlations between their experience-based WI scores and water quantity. However, HWISE-4 (shorter version of the HWISE questionnaire) does capture the experience of adequate quantity of drinking water for consumption [72]. Although our WI model was found to be valid at baseline, it was not completely supported at 18 months, suggesting that additional water-related information may be needed to better represent WI over time.Our study, like some others [5,21,24,27], acknowledges the importance of water for both non-drinking (e.g., laundry, bathing, cooking, irrigation, etc.) and drinking purposes. Recent efforts to develop WI metrics have mostly focused on composite scales that capture at least one, but not all, WI dimensions (HWISE, WIES, HWIAS, WAI) [5,21,26,27,30,31]. Like for FI, using unidimensional composite scales masks the contributions of the multiple WI dimensions. For example, the initial HWISE questionnaire included questions on the taste, smell, color, treatment, reliability, and usability of water, among others. However, the focus was on how people experienced those things and only some aspects were retained in the final scale [5]. This is in contrast to our analyses, where water quality was identified as a distinct dimension. Our water quality dimension has the added advantage of including types of water sources, also considered reliable and factual by the JMP [20]. Similarly, for water access, the WHO uses water fetching time as a simple measure [19], while the HWISE scale was correlated with time to water source [5]. Our water access dimension includes distance to water point, and as such, captures additional access information. Distance travelled and time taken for water collection form part of the WAI [30]. Whereas in our analyses, reliability of water supply was a distinct dimension, it was an integral part of the WAI. This may be explained by the difference in variable measurement: WAI reported time period when water supply is cut-off on any given day while we report any disruption in water supply over the year prior to the interview.Validity of our FI and WI measures was supported in a number of ways (Table 5 and Table 6). Firstly, internal validity was confirmed in two separate groups: a test sample of the baseline population and the SHINE households more than 18 months later. The consistent results produced in both instances indicate structural adequacy of the FI and WI dimensions within this population. Secondly, higher perceived health status was associated with lower FI and WI, while maternal depression was associated with higher FI and WI. These associations, although small, suggest good predictive validity. The results are consistent with findings in similar populations between health conditions and experiential measures of FI [34,73,74,75] and WI [5,21,25,33]. These validity analyses add to the existing literature by considering additional dimensions of FI and WI.Convergent validity was examined by assessing the relationship between dimension scores with food aid for FI and frequency of water collection for WI. Households having experienced shocks were more likely to receive food aid, which suggests convergent validity. Although the association with poor food access and low quality and availability were not significant, they were in the expected direction. Lower frequency of water collection was associated with poorer water access. Upon further exploration within SHINE, we found that the longer women spent on water collection at any one time, the less often they collected water.Finally, we evaluated discriminant validity via the associations between FI and WI dimension scores and season of interview, SES and HIV-status. As expected and consistent with international reports, households with lower SES, or those interviewed during the hungry season, and women living with HIV had poorer food access and lower food quality and availability. In the months between the last season’s food stores and current season’s harvesting of crops, food supplies in farming communities run low, and people often use food aid and other coping strategies [53]. Similarly, households at lower SES, especially subsistence farmers, may be less able than those at higher SES to access sufficient and nutritious food [1]. The association between HIV-status and FI is considered bidirectional [76]. People living with HIV have a greater need for adequate food to ensure the success of antiretroviral therapy (ART). At the same time, because they are weakened by their disease, they may be unable to procure food or resources needed to obtain food [77,78]. This may be true of other health outcomes, such as physical health status and depression used for predictive validity. Since we are primarily concerned with associations rather than causation in these validity analyses, although relevant, bidirectionality does not affect our interpretation.Water access was found to be poorer among households interviewed during the dry season compared to the rainy season. During the dry season, households may need to travel longer distances and spend more time on finding water because their usual water source ran dry, whereas in the rainy season, water sources may be more abundant [27]. SHINE households with low to middle SES had poorer water quality compared to those of high SES, because they were less likely to have access to piped water or protected water sources. The association with HIV status was not significant for any of the WI dimensions. However, a prior study measuring WI on the experiential scale reported that women living with HIV in Kenya were more water insecure [21].The strengths of this study make it a worthwhile exploration of new FI and WI measures in an underserved population. First, we were able to use information from over three thousand households to develop and test the structural validity of these new measures. The households are representative of the rural population in Zimbabwe, implying that the FI and WI measures may be valid in other similar rural areas [52]. Second, sensitivity analyses accounting for missing data further increased the sample size and enhanced our confidence in the robustness of the multiple dimensions identified (Table S4). Third, unlike previous scales, our measures reflect additional theoretical multidimensionality in both FI and WI separately, and allow distinction between key dimensions. We were able to show via the statistical method of MCA that it is possible to come up with and quantify different aspects of FI and WI. We expect that future research will consider using similar methods to distinguish between FI and WI dimensions in other settings. The Gallup World Poll implemented the FIES questionnaire in 2019 for FI monitoring and recently decided to include HWISE for WI. Our method and potentially our MHFI and MHWI measures may serve as supplements for identifying and addressing specific food and water problems in vulnerable populations. Fourth, our study is unique in that it considers the simultaneous development of both FI and WI measures in a rigorous randomized controlled trial like SHINE, which included training of data collectors and quality control that ensured information accuracy. Finally, we recently implemented the questions identified in our study in another survey with a separate group of Zimbabwean households, and found that it took less than 15 min for trained data collectors to obtain the required FI and WI information. For all questions, higher literacy levels shortened the interview time. Since these questions are easy to add and are of low time burden, we hope that this paper further encourages global health scientists and policy makers to think about the individual components that define FI and WI when designing health and nutrition interventions.Although SHINE’s IYCF intervention showed a reduction in stunting prevalence, the improvement was modest; while SHINE’s WASH intervention showed no improvement [50]. Similar modest improvements have been reported in other IYCF and WASH interventions, in countries such as Bangladesh and Kenya [79,80,81]. It is hypothesized that the impact on stunting reduction could be larger if underlying determinants (e.g., FI and WI) as per the UNICEF’s framework for undernutrition are addressed [82]. In most instances, efforts to mitigate FI and WI are complicated by interactions with social, environmental, and physical processes [83,84]. Therefore, our deconstructed measures of FI and WI provide an opportunity to explore specific aspects of FI and WI on undernutrition and related interventions for better targets. The contribution of our WI measures will be particularly important for the transformative WASH movement, which calls for the radical reduction of fecal contamination in the household environment in LMICs [85,86]. Despite intensive implementation and uptake of low-cost household-based WASH interventions in recent trials [50,80,81], environmental fecal contamination remains pervasive [85]. Our WI measures can be used to assess convenient and adequate access to uncontaminated water, which is central to transformative WASH. This concept is also critical for the utilization dimension of food security, which requires safe water for food handling and preparation.This study is innovative in its approach for finding multidimensional measures of FI and WI, but it is not without limitations. First, the dimensions explained a small percentage of the cumulative variance between items. This may be a concern, although the percentage variance is generally smaller in MCA than in PCA, because individuals are located in a high K-J dimensional space which gets larger as the number of categories increase [87]. The low variances are not unlike what have been observed in other categorical or binary factor analyses [69,87]. Second, these measures were developed in rural Zimbabwe and among households with pregnant women. Although the methods and overall dimensions are potentially transferable, generalizability outside of this population may not be appropriate. When used in other settings, modifications will be required to the items included for the MCA. For example, water purchase was not relevant in rural Zimbabwe because the majority of its rural population does not buy water. Similarly, water purchase and affordability were not retained in HWISE [5]. However, this is an important access consideration for some populations, like those residing in colonias on the US-Mexico border and in low-income communities in Jamaica [30,88]. For FI, transportation and time to markets may be important among households that do not engage in subsistence farming, e.g., in areas with food scarcity in the USA [89].Third, the dimensions retained are limited to the variables available from the SHINE questionnaire. We could not capture the experiences of thirst, hunger or emotional distress (e.g., anger, frustration, shame) associated with FI and WI because such information was not collected. Therefore, our household FI and WI measures may be used to complement experience scales such as FIES and HWISE for an all-encompassing view of these resource insecurities. Moreover, to better capture FI utilization, supplementary information suggested by the FAO, like food consumption scores (to determine equitable intra-household food distribution), and utilization of clean utensils for cooking and eating, would have been useful [3]. Similarly, additional information on objective microbiological and physico-chemical assessments of water at point of use, details on intermittent availability of water from all sources, quantification of the amount of water used per household member to determine need-based equitable access, and water sufficiency for purposes other than drinking, would have improved our WI dimensions. We also caution on the ‘poor water reliability’ dimension, created with only two variables, making it prone to quantitative estimation problems [65]. Nevertheless, a two-item dimension is considered reliable if they are highly correlated with each other, as is our case (r = 0.75), but fairly uncorrelated with other variables [90,91].Finally, another concern may be seasonality, due to the dependence of FI and WI on environmental conditions in a land-locked country like Zimbabwe and elsewhere. However, the SHINE baseline data was collected over three calendar years, from 2012 to 2015. Therefore, all months are represented in our analyses since households were interviewed year-round. In addition to the validity analyses that looked at differences in scores by season (Table 5), we also ran multiple sensitivity analyses of separate factor analyses with groups of households that were interviewed by calendar quarter, by dry season and rainy season, and by hungry season and plenty season. In all instances, the same items were loaded on the same dimensions as our main analyses. Furthermore, we had another sample of the households at a different time point. As shown in Table 6, when we ran factor analysis at the 18-month mark, the results we report in this paper remained consistent in terms of dimensions and item variables. All these analyses greatly strengthen our confidence in the structural validity of our measures and the dimensions identified for FI and WI.We developed new and culturally-specific measures for each FI and WI among rural Zimbabwean households. In accordance with the theoretical definitions of FI and WI, each measure was multidimensional, with three distinct dimensions retained. FI was characterized by ‘poor food access’, ‘household shocks’ and ‘low food quality and availability’. WI was characterized by ‘poor water access’, ‘poor water quality’, and ‘low water reliability’. The application of such multidimensional measures will make it possible to pinpoint the components of FI and WI that impact health, and may facilitate the provision of better interventions to households in need of specific food and water support. These measures will also contribute to transformative WASH actions, and transdisciplinary FI and WI mitigation efforts.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116020/s1, Table S1. Description of missing item variables (N = 4675); Table S2. Comparison of baseline characteristics of participants with complete vs. missing values; Table S3. Correlation (rho) between the scores of food insecurity dimensions and water insecurity dimensions; Table S4. Sensitivity analysis of multiple correspondence analyses with imputed variables.Conceptualization, N.K.; methodology, N.K., L.E.S., C.-S.L., K.K.; validation, N.K.; formal analysis, N.K.; resources, L.E.S., J.H.H.; data curation, N.K., R.N., B.C.; writing—original draft preparation, N.K.; writing—review and editing, N.K., A.D.J., R.S., G.O.B., M.N.N.M., L.E.S., K.K., C.-S.L., J.H.H.; visualization, N.K.; supervision, L.E.S., K.K., C.-S.L. All authors have read and agreed to the published version of the manuscript.The SHINE trial is funded by the Bill & Melinda Gates Foundation (OPP1021542 and OPP113707); Department for International Development (DFID), UK; Wellcome Trust, UK (093768/Z/10/Z, 108065/Z/15/Z and 203905/Z/16/Z); Swiss Agency for Development and Cooperation (SDC); National Institutes of Health, USA (2R01HD060338-06); and UNICEF (PCA2017-0002). The study funders approved the trial design, but were not involved in data collection, analysis, interpretation, manuscript preparation, nor decisions related to publication. The lead author and corresponding author had full access to all study data and ultimate responsibility for the decision to submit for publication.SHINE was approved by the Medical Research Council of Zimbabwe and the Johns Hopkins University Bloomberg School of Public Health Institutional Review Board.Written informed consent from participants, in the language of their choice (English, Ndebele, or Shona), was obtained prior to data collection.The authors thank all the mothers, infants and their families who participated the SHINE trial. The authors gratefully acknowledge the leadership and staff of the Ministry of Health and Child Care in Chirumanzu and Shurugwi districts and Midlands Province (especially environmental health, nursing and nutrition) for their roles in the operationalization of SHINE’s procedures. The authors acknowledge the Ministry of Local Government officials in each district who supported and facilitated field operations. We would also like to thank our colleagues at Zvitambo Institute for Maternal and Child Health Research, especially Joe Piper and Ceri Evans. We acknowledge funding support for NK during this project, from the Department of Epidemiology and Environmental Health and from the Community of Global Health Equity at the University at Buffalo, The State University of New York. We also thank all members of the SHINE Trial team (listed at: https://doi.org/10.1093/cid/civ844).The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.Sample selection for FI and WI factor analyses.Scree plots of food and water dimensions.Complete set of item variables from the Sanitation Hygiene and Infant Nutrition Efficacy (SHINE) trial considered for each dimension of household food insecurity and water insecurity, collected at baseline from November 2012 to March 2015.* All item variables were either dichotomous or ordered categorical, and reverse coded so that insecurity scored higher; ** Parameterization of variables as used in the subsequent quantitative analyses in this study; a Variables excluded in the subsequent steps of factor analysis if categories were too small (≤5%) or too common (≥95%).Validity assessments for dimension scores of food insecurity and water insecurity.Socio-demographic characteristics of households included in analyses.* All socio-demographic variables had <5% missing data unless otherwise stated. ** All variables represent complete data as n(%) unless otherwise stated. 1 Missing data >5%. - Empty cells imply that the variables were not described for that sample.Squared correlation ratios, eigenvalues, percentage variances and descriptive statistics of multiple correspondence analysis for final dimensions in each food insecurity and water insecurity measures.Items in bold are retained as relevant (squared correlation ratio ≥ 0.2) and further indicate which dimension they load on. * Scores are standard dimension scores obtained from post-estimation commands. Higher positive scores on dimensions are indicative of higher insecurity.Internal validity of multidimensional food insecurity and water insecurity measures.1 Baseline test sample refers to a sub-sample of the complete case used to confirm the measures. We used 40% of the initial baseline sample as a training dataset and ran exploratory factor analysis; the remaining mutually exclusive 60% of the sample was then used as the testing dataset to confirm the exploratory findings via confirmatory factor analysis. 2 The 18-month time point refers to 18 months after the pregnant woman gave birth. Therefore, this sample could be between 19–24 months post baseline interview. Bolded values represent satisfactory model fit statistic. RMSEA= root mean square error of approximation; CFI = comparative fit index; TLI = Tucker Lewis index; SRMR = Standardized Root Mean Square Residual.Predictive, discriminant and convergent validity of food insecurity and water insecurity measures.Values in bold represent statistically significant (at p < 0.05) associations in the expected directions from linear regressions. a Information self-reported by mothers for their status and activity; b Based on date of interview recorded by data collector; c Household-level information; d Mother’s information obtained from rapid blood test.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06021.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background and Objectives: Stroke is a strong risk factor for recurrent cardiovascular disease (CVD) incidents. The risk of post-stroke CVD incidents can be reduced by eliminating the most relevant risk factors. The aim of the study was to compare the incidence of recurrent CVD events and to determine the quantitative and qualitative differences in CVD risk factors over the 5-year follow-up period in patients with ischemic stroke (IS) and haemorrhagic stroke (ICH) with the use of ICF classification categories to present these differences. Materials and Methods: The study was retrospective. The study groups included 55 post-IS patients and 47 post-ICH patients. The results were translated into the categories from the International Classification of Functioning, Disability and Health (ICF) classification. Results: As compared to post-ICH patients, post-IS patients were significantly more frequently observed to have recurrent CVD incidents (p < 0.001), including fatal CVD incidents (p = 0.003). More risk factors in total were identified in both post-IS patients (p = 0.031) and post-ICH patients (p = 0.002) who had a recurrent CVD incident. Post-IS patients were more often found to have arterial blood pressure higher than 140/90 mmHg (p = 0.045). On the other hand, post-ICH patients were more frequently observed to have carotid artery stenosis in the range of 50–69% (p = 0.028) and an eGFR of <15 mL/min/1.73 m2 (p = 0.001). Conclusions: The type of primary stroke determines the type and incidence of risk factors as well as the recurrence rate of CVD incidents over a 5-year follow-up period. Patients after IS have a higher risk of recurrence of CVD events, including fatal ones in the 5-year follow-up compared to patients after ICH. In addition, post-IS patients who have a recurrent CVD event over a 5-year follow-up have more risk factors for a CVD event than ICH. The ICF classification can be useful for assessing and analysing risk factors for recurrent CVD incidents, which can help to improve the effectiveness of secondary prevention.Stroke is a serious social problem because it is the second leading cause of death in the world and the leading cause of long-term disability [1]. The risk of death in patients with recurrent stroke is approximately 50% [2]. Mortality due to recurrent stroke is twice as high as in the first stroke event [3]. Moreover, post-stroke patients are at high risk of recurrence of cardiovascular diseases (CVD) [4]. The cumulative risk of a recurrent CVD event ranges from 10% to 12% in the first year and from 30% to 40% over a five-year period. Recurrent stroke, which most often has the same etiopathogenesis as the primary stroke, accounts for more than 75% of the secondary sequelae of CVD. Interestingly, more than 40% of post-haemorrhagic stroke (ICH) patients have ischemic stroke (IS) and only 5% of post-IS patients have ICH as a recurrent CVD incident [2].Based on analyses carried out by experts of the World Health Organization (WHO), it can be concluded that about 80% of recurrent CVD incidents could be avoided if the most relevant risk factors were eliminated [5]. The generally recognised modifiable risk factors for CVD related to secondary prevention include arterial hypertension, atrial fibrillation, diabetes, dyslipidaemia, abnormal body mass index (BMI), carotid artery disease, depression, insomnia, smoking, and alcohol abuse [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20]. The coexistence of the above risk factors increases the likelihood of a recurrent CVD incident [21]; therefore, it is reasonable to monitor patients for all these risk factors simultaneously.The International Classification of Functioning, Disability and Health (ICF) makes it possible to compare and identify differences in the assessment of impairments which affect functions and body structures as well as the activity and participation. This assessment is particularly important in the management of post-stroke patients due to the complexity of their psychosomatic dysfunctions. The ICF classification has been approved by the WHO as an international standard for describing health status and health-related conditions [22]. This makes it possible to organise the clinical data contained in medical records and to present them in a clear graphical form. Translating clinical data into the categories from the ICF classification increases the reliability of a complex assessment of a patient’s health status [23]. The usual clinical practice is to use the ICF classification mostly for post-stroke patients at the time of initial diagnosis and during the acute treatment period [24]. The available medical literature contains only a few publications relating to the use of the ICF classification for assessing post-stroke patients in the course of secondary prevention [25].To compare the incidence of recurrent CVD incidents over a 5-year follow-up period in post-IS and post-ICH patients;To determine quantitative and qualitative differences relating to risk factors for CVD in post-IS and post-ICH patients according to the recurrence of CVD incidents over a 5-year follow-up period;To use the categories of the ICF classification as a tool for presenting differences in the incidence of risk factors for recurrent CVD incidents according to previous IS or ICH and the recurrence of CVD incidents over a 5-year follow-up period.The study was retrospective and was conducted in the Neurological Rehabilitation Unit of the Rehabilitation Department at Wiktor Dega Orthopedic and Rehabilitation Teaching Hospital attached to K. Marcinkowski Poznań University of Medical Science in the period from 1 January 2015 to 31 December 2016. In this period in the Neurological Rehabilitation Unit were hospitalized 258 patients after stroke, including 169 after IS and 89 after ICH. The study was divided into 2 stages. The first stage was to analyse the frequency of risk factors for a recurrent CVD event based on the patient’s medical history. After the inclusion and exclusion criteria were applied, 189 patients were enrolled in this stage, including 109 after IS and 80 after ICH. The second stage was to analysed the frequency of recurrent CVD events within five years of the first stroke episode based on contact with the patient or a close family member authorized to provide information about the patient’s health status. Finally, 55 patients after IS and 47 patients after ICH were qualified for the second stage. The results were translated into categories of the ICF classification and are presented in graphical form. A detailed study diagram is shown in Figure 1.The study was approved by the Ethics Committee attached to K. Marcinkowski Poznań University of Medical Sciences (Approval No. 174/21 of 11 March 2021). The study was conducted in accordance with the ethical principles for biomedical research as stated in the Declaration of Helsinki. The study was registered in the Clinical Trial Registry: NCT04590287 https://clinicaltrials.gov/ct2/show/NCT04590287 (accessed on 19 October 2020).The study inclusion criteria were as follows: (1) patients with stroke as confirmed by medical imaging, (2) patients hospitalised within 14 days of the stroke in the neurological rehabilitation department, (3) full medical records containing detailed information on all possible risk factors, (4) current medical records to confirm whether there was a recurrent CVD incident over the 5-year follow-up period, and (5) use of the prescribed pharmacotherapy of chronic diseases by patients.The study exclusion criteria were (1) patients hospitalised more than 14 days after the stroke for neurological rehabilitation in the neurological rehabilitation department, (2) patients who could not be followed up or whose current medical records could not be obtained to confirm whether there was a recurrent CVD incident over the 5-year follow-up period, and (3) not use of the prescribed pharmacotherapy of chronic diseases by patients.The individual categories of the risk factors included in the analysis were assigned relevant code numbers and qualifiers as per the ICF classification.The effect of depressive disorders on the risk of a recurrent CVD event was assessed using ICF category b152: emotional functions. The following Beck Depression Inventory (BDI) scores were used to measure the severity of depression [26]: qualifier 0: BDI total score 0 to 11—no depression; qualifier 2: BDI total score 12 to 19—mild depression; qualifier 3: BDI total score 20 to 25—moderate depression; qualifier 4: BDI total score 26 to 63—severe depression.The effect of sleep disturbance on the risk of a recurrent CVD event was assessed using ICF category b134: sleep functions. The following criteria were used to measure the severity of insomnia [7]: qualifier 0—no sleep disturbance (sleep time 6–9 h); qualifier 4—sleep disturbance (sleep time < 6 or > 9 h).The increased risk of CVD related to heart rate (HR) was estimated using ICF category b4100: heart rate. The following criteria were used to quantify heart rate disorders [8]: qualifier 0—HR < 80/min; qualifier 4—HR >80/min. Heart rhythm disorders were encoded as ICF category b4101: heart rhythm. The following criteria were used [9]: qualifier 0—normal sinus rhythm; qualifier 4—atrial fibrillation.The effect of carotid artery stenosis on the risk of a recurrent CVD event was assessed using ICF category b4150: functions of arteries. The following criteria were used [10]: qualifier 0—<50% carotid stenosis; qualifier 3—50% to 69% carotid stenosis; qualifier 4—>70% carotid stenosis.The effect of increased blood pressure (BP) on the risk of a recurrent CVD event was assessed using ICF category b4200: increased blood pressure. The following BP values were used [27]: qualifier 0—BP < 130/80 mm/Hg; qualifier 1—BP > 130/80 mm/Hg; qualifier 2—BP > 140/90 mm/Hg; qualifier 3—BP > 160/90 mm/Hg; qualifier 4—BP > 180/110 mm/Hg.The effect of liver and renal impairment on the risk of a recurrent CVD event was assessed using ICF category b4301: metabolite-carrying functions of the blood. The following criteria were used to classify renal impairment [28]: qualifier 0—estimated glomerular filtration (eGFR) > 90 mL/min/1.73 m2; qualifier 1—eGFR 60–89 mL/min/1.73 m2; qualifier 2—eGFR 30–59 mL/min/1.73 m2; qualifier 3—eGFR 15–29 mL/min/1.73 m2; qualifier 4—eGFR < 15 mL/min/1.73 m2, and liver impairment [13]: qualifier 0—bilirubin level < 2x the upper limit of normal (ULN) and ALT (alanine transaminase)/AST (aspartate transaminase)/ALP (alkaline phosphatase) < 3x ULN; qualifier 4—bilirubin level > 2x ULN and ALT/AST/ALP > 3x ULN.Patients receiving anticoagulants have got increased risk of bleeding [14]. This parameter was encoded as ICF category b4302: functions related to the coagulation of blood. If taking VKA (vitamin K antagonist) following values were used: qualifier 0—NO; qualifier 4—YES. If taking NOAC (non-vitamin K antagonist) following values were used: qualifier 0—NO; qualifier 4—YES.The effect of impaired glycaemic control on the risk of recurrent CVD event was assessed using ICF category b5401: carbohydrate metabolism. The following HbA1c (glycated haemoglobin 1c) values were used [29]: qualifier 0—HbA1c < 7%; qualifier 4—HbA1c > 7%.The effect of LDL-C (low-density lipoprotein cholesterol) levels on the risk of a recurrent CVD event was assessed using ICF category b7302, lipid metabolism. The following LDL-C values were used [30]: qualifier 0—LDL-C < 55 mg/dL; qualifier 2—LDL-C 55 mg/dL−70 mg/dL; qualifier 3—LDL-C 71 mg/dL−115 mg/dL; qualifier 4—LDL-C > 116 mg/dL.Alcohol consumption is an additional risk factor associated with increased risk of a recurrent CVD event. This risk factor was assessed using ICF category e1100, food: alcohol consumption. The following criteria were used [31]: qualifier 0—alcohol intake per day < 10 g; qualifier 4—alcohol intake per day > 10 g.The increased risk of CVD related to NSAID (nonsteroidal anti-inflammatory drugs) [18] and to smoking [19] was estimated using ICF categories e1101: drugs, and e1109: products or substances for personal consumption, respectively. The following criteria were used: qualifier 0—NO; qualifier 4—YES.In the following stage, to better highlight any differences which might be present, the graphical summary included a percentage distribution of the qualifiers of risk factor categories as per the ICF classification: qualifier 0, dark green—no risk factors if the value of the percentage distribution was in the range of 0% to 4%; qualifier 1, light green—low risk factor if the value of the percentage distribution was between 5% and 24%; qualifier 2, yellow—moderate risk factor if the value of the percentage distribution was 25–49%; qualifier 3, orange—high risk factor if the value of the percentage distribution was 50–95%; qualifier 4, red—extremely high risk factor if the value of the percentage distribution was 96–100%.The data analysis was carried out using Statistica v. 13.1. The parameters of descriptive statistics are reported as mean values with standard deviations (SD) and median, minimum, and maximum levels. The categorical variables are presented as counts and frequencies. The Shapiro–Wilk test was used to assess the normality of the distribution of test scores. Non-parametric analyses were used when the data were found not to meet the assumptions defined for parametric analysis. The significance of differences between results or both groups was evaluated based on the parametric Student’s t-test for independent variables or the non-parametric Mann–Whitney test. The chi-squared test was used to compare differences between groups in terms of categorical variables. p-values less than 0.05 were considered to be statistically significant.The group after IS was consisted of 55 patients, the group after ICH was consisted of 47 patients. The study groups were significantly different in terms of the age of patients at the time of stroke (p < 0.002). The average age of post-IS patients was 69.3 years (SD ± 12.5), whilst the average age of post-ICH patients was 61.3 years (SD ± 12.6). The majority of post-IS patients were women (50.90%) whilst the majority of post-ICH patients were men (57.40%). As compared to post-ICH patients, post-IS patients were significantly more frequently observed to have had recurrent CVD incidents (p < 0.001), including recurrent IS (p = 0.031) and myocardial infarction (p < 0.019). Additionally, recurrent CVD incidents were more often fatal (p = 0.003) in this patient group. Detailed characteristics of the study groups are shown in Table 1.Table 2 presents the results of the analysis of the incidence of recurrent CVD incidents over a 5-year follow-up period according to the clinical type of previous stroke, taking into account age and gender. Fatal CVD incidents were significantly more frequently observed among male patients in the post-IS group than male patients in the post-ICH group (p = 0.030). Additionally, post-ICH patients over 65 years of age (p = 0.016) and male patients (p = 0.015) were more frequently observed to not have any recurrent CVD incidents over a 5-year follow-up period than post-IS patients.Table 3 shows the incidence of recurrent CVD incidents in post-IS or post-ICH patients, taking into account the identified risk factors.Post-IS patients who suffered recurrent CVD incidents were significantly more frequently observed to have atrial fibrillation (p = 0.004), abnormal glycosylated haemoglobin levels (p = 0.018), and LDL levels above 116 mg/dL (p < 0.001), as well as to more often use NSAIDs (p < 0.001) as compared to post-ICH patients, who were significantly more frequently observed to have LDL levels in the range of 55–70 mg/dL (p < 0.001).Post-IS patients who did not suffer a recurrent CVD incident were significantly more frequently observed to have atrial fibrillation (p = 0.046) and LDL levels above 116 mg/dL (p = 0.008), as well as to more often use NSAIDs (p < 0.001) as compared to post-ICH patients, who were more frequently observed to have LDL levels in the range of 55–70 mg/dL (p = 0.014).Post-IS patients who suffered a recurrent CVD incident were significantly more frequently observed to have abnormal arterial blood pressure above 140/90 mmHg than patients who suffered no recurrent CVD incidents (p = 0.045).Post-ICH patients who suffered a recurrent CVD incident were significantly more frequently observed to have carotid artery stenosis in the range of 50–69% (p = 0.028) and glomerular filtration rates below 15 mL/min/1.73 m2 (p < 0.001) as compared to patients who had no recurrent CVD incidents.For patients with extremely high risk factors, post-IS patients who suffered a recurrent CVD incident were more frequently observed to have depression, insomnia, abnormal heart rate and rhythm, carotid artery stenosis over 70%, elevated arterial blood pressure (above 180/110 mmHg), glomerular filtration rate below 15 mL/min/1.73 m2, abnormal glycosylated haemoglobin levels, and LDL levels higher than 116 mg/dL and to smoke and abuse alcohol, as compared to patients with no recurrent CVD incidents.For those with a significant risk factor, carotid artery stenosis in the range of 50% to 69%, glomerular filtration rates in the range of 15 to 29 mL/min/1.73 m2, and LDL levels in the range of 71 mg/dL to 115 mg/dL were observed more frequently.In terms of a moderate risk factor, glomerular filtration rates in the range of 30 to 59 mL/min/1.73 m2 and elevated arterial blood pressure (above 140/90 mmHg) were more frequently observed.For patients with extremely high risk factors, post-ICH patients who suffered a recurrent CVD incident were more frequently observed to have depression, abnormal heart rhythm, carotid artery stenosis over 70%, elevated arterial blood pressure (above 180/110 mmHg), and abnormal glycosylated haemoglobin levels and to smoke and use NOACs, VKAs, and NSAIDs, as compared to patients with no recurrent CVD incidents.In terms of significant risk factors, carotid artery stenosis ranging from 50% to 69% and LDL levels in the range of 71 to 115 mg/dL were observed more frequently.For those with moderate risk factors, elevated arterial blood pressure (above 140/80 mmHg), glomerular filtration rate in the range of 30 to 59 mL/min/1.73 m2, and LDL levels ranging from 55 to 70 mg/dL were more frequently observed.For patients with extremely high risk factors, post-IS patients who suffered a recurrent CVD incident were more frequently observed to have abnormal heart rate and rhythm, carotid artery stenosis over 70%, glomerular filtration rates below 15 mL/min/1.73 m2, LDL levels higher than 116 mg/dL, and abnormal glycosylated haemoglobin levels and liver function test results and to smoke and use NSAIDs and NOACs as compared to post-ICH patients. On the other hand, post-ICH patients were significantly more frequently observed to suffer from depression, insomnia, elevated arterial blood pressure (above 180/110 mmHg), and alcohol abuse. The frequency of the use of VKAs was found to be similar in both groups.For those with significant risk factors, carotid artery stenosis was observed to be in the range of 50% to 69%, in both the post-IS and the post-ICH groups. On the other hand, post-IS patients were more often observed to have glomerular filtration rates in the range of 15 to 29 mL/min/1.73 m2 and LDL levels ranging from 71 to 115 mg/dL. Post-IS patients, however, were more frequently observed to have elevated arterial blood pressure (exceeding 160/90 mmHg).In terms of a moderate risk factor, post-IS patients were observed more often than post-ICH patients to have glomerular filtration rates in the range of 30 to 59 mL/min/1.73 m2. On the other hand, post-ICH patients were more frequently observed to have elevated arterial blood pressure (above 140/80 mmHg) and LDL levels in the range of 55 to 70 mg/dL.Table 4 shows a list of categories of the ICF classification and the percentage distribution of risk factors for CVD in secondary prevention according to the recurrence of CVD incidents and the clinical type of stroke.Table 5 shows a comparison of the cumulative risk of CVD in the study groups. In both post-IS patients (p = 0.031) and post-ICH patients (p = 0.002) who had a recurrent CVD incident, significantly more risk factors for CVD were identified than for patients who had no recurrent CVD incidents.The values of distribution of the total risk factors for CVD were observed to be higher for the post-IS group than for the post-ICH group (Figure 2). In both groups, the greatest values of distribution of risk factors for CVD were observed for patients following a fatal recurrent CVD incident, whilst the lowest values were for patients who had no recurrent CVD incidents.Sequelae of cardiovascular disease are a major cause of death around the world and more than half of the patients with a history of stroke are at an increased risk of recurrent CVD incidents, including recurrent stroke in particular [1]. In our study (Table 1), as many as 76% of post-IS patients and only 40% of post-ICH patients (p < 0.001) had recurrent CVD incidents over the 5-year follow-up period. Our results are consistent with those reported by Vickrey et al. [3], where the risk of recurrent stroke over a 5-year follow-up period was greater than 40%. According to Yamamoto et al. [2], recurrent stroke incidents in post-IS patients most often have the same etiopathogenesis and the risk of future myocardial infarction is 15%. Only 5% of all incidents are classified as IS. On the other hand, 42% of post-ICH patients are observed to have recurrent IS. In the present study (see Table 1), recurrent IS episodes were observed in more than 60% of patients with a history of primary IS and in 40% of patients with a history of primary ICH (p = 0.031). No recurrent IS was observed in either group. Additionally, more than 14% of the post-IS patients had myocardial infarction, whilst no cases of myocardial infarction were identified in the post-ICH group (p < 0.019). Yamamoto et al. [2] demonstrated that more than half of the patients who suffered a recurrent stroke are at risk of death [2]. This is consistent with our results, where recurrent CVD incidents in the post-IS group were fatal in more than 54% of patients. In this study, a significantly lower mortality rate (p = 0.003) and a lower incidence of recurrent CVD incidents (p < 0.001) were observed in patients with a history of primary ICH. In our opinion, these differences are due to the pathogenesis and the more severe clinical presentation of IS as compared to ICH [32]. Moreover, there was no difference in the occurrence of a recurrent CVD incident and the mortality due to a recurrent CVD incident in a 5-year follow-up depending on the conservative or surgical treatment in both IS and ICH. The obtained results are consistent with the studies by McCarthy et al. [33], who proved that treatment with mechanical thrombectomy in IS does not reduce long-term mortality compared to conservative treatment. Moreover, Hemphill et al. [34] proved that surgical treatment in patients after ICH did not show any clear benefits compared to conservative treatment. In terms of age and sex (Table 2), fatal recurrent CVD events were significantly more frequently observed in men with a history of primary IS (p = 0.030). As proved by Zhang et al. [32], the incidence of IS in men aged over 55 was more than two times higher than in the case of ICH, which may result in higher mortality in men after IS in this age group compared to ICH. The absence of recurrent cardiovascular events was significantly more often observed after ICH in men (p = 0.015) and over 65 (p = 0.016). The research results correspond to the results obtained by Zhang et al. [35].As argued by Adams et al., the more risk factors that are identified, the greater the likelihood of a recurrent CVD incident [21]. As compared to patients with no recurrent CVD incidents, significantly more coexisting risk factors for CVD were observed both in post-IS patients (p = 0.031) and post-ICH patients (p = 0.002) who had a recurrent CVD incident (Table 5). Additionally, in all three options included in the analysis, higher values of distribution of the total risk factors for CVD were observed for post-IS patients than for post-ICH patients (Figure 2). This is due to the fact that post-IS patients have more coexisting diseases [36]. Therefore, for the purposes of secondary prevention, it is reasonable to monitor patients for various risk factors simultaneously. In our study (Table 3), post-IS patients who had a recurrent CVD incident were significantly more frequently observed to have atrial fibrillation than the post-ICH patients (p < 0.004). This is consistent with the results obtained by Lip et al. [9], who determined that post-IS atrial fibrillation was associated with a high risk of recurrent CVD incidents. Another risk factor included in the analysis was abnormal glycosylated haemoglobin level. As evidenced by Wu et al., it is significantly more frequently associated with the recurrence of CVD incidents in post-IS patients than in post-ICH patients [15]. Our findings were similar (p = 0.018). Post-IS patients who suffered a recurrent CVD incident were also significantly more frequently observed to have abnormal LDL levels (above 116 mg/dL; p < 0.001), and were more rarely observed to have LDL levels in the range of 55–70 mg/dL than post-ICH patients (p < 0.001). High LDL levels are a strong predictor of the recurrence of CVD incidents. The relationship between dyslipidaemia and the recurrence of CVD incidents due to atherosclerosis is well evidenced [37]. Patients in the post-IS group were also observed to more frequently use NSAIDs (p < 0.001), which are associated with a higher risk of intracerebral haemorrhage [18]. In the post-IS group, patients who suffered a recurrent CVD incident were significantly more frequently observed to have elevated arterial blood pressure (p = 0.045). Lower arterial blood pressure in post-stroke patients reduces the risk of recurrent CVD incidents [11]. Post-ICH patients who suffered a recurrent CVD incident were significantly more frequently observed to have carotid artery stenosis in the range of 50–69% (p = 0.028) and glomerular filtration rates below 15 mL/min/1.73 m2 (p < 0.001). Significant carotid artery stenosis and an abnormal glomerular filtration rate both increase the risk of recurrent CVD incidents [10,12].When using the ICF classification in secondary prevention of recurrent CVD incidents, specific attention should be paid to the risk factors associated with extremely high incidence (qualifier 4), which are marked in red (Table 4). Common factors in post-IS and post-ICH patients which are of particular relevance to the recurrence of CVD incidents include depression, abnormal heart rate and rhythm, carotid artery stenosis above 70%, arterial blood pressure higher than 180/110 mmHg, abnormal glycosylated haemoglobin levels, and smoking. Additional risk factors in post-IS patients include insomnia, a glomerular filtration rate below 15 mL/min/1.73 m2, LDL levels higher than 116 mg/dL, and alcohol abuse. In post-ICH patients, the use of NOACs, VKAs, and NSAIDs is an additional risk factor of relevance to the recurrence of CVD incidents.The percentage distribution of risk factors for recurrent CVD incidents according to the categories of the ICF classification, as shown in Table 4, provides information on risk factors and their incidence based on the type of previous stroke. The use of a single tool for monitoring various risk factors for CVD—in the form of the ICF questionnaire—could help to increase the effectiveness of secondary prevention and thus to reduce the risk of recurrent CVD incidents. The simultaneous presentation of several categories of risk factors in the form of ‘dynamic graphs’ makes it possible to analyse them in a legible and concise manner, which can help in taking appropriate clinical decisions.The limitations of our study were retrospective character of the study and relatively small number of study groups. Additionally, as concerns the analysis of individual risk factors, the study groups were composed of different numbers of patients. We have not also included BMI as a known risk factor for recurrent CVD incidents due to the lack of data in the available medical records from which BMIs could be calculated.The type of primary stroke determines the type and frequency of risk factors and the frequency of recurrence of CVD events over the 5-year follow-up period. Patients after IS have a higher risk of recurrence of CVD events, including fatal ones in the 5-year follow-up compared to patients after ICH. In addition, post-IS patients who have a recurrent CVD event over a 5-year follow-up have more risk factors for a CVD event than ICH. ICF can be useful in assessing and analysing risk factors for recurrent CVD events, which can help improve the effectiveness of secondary prevention.Conceptualization, M.L., E.C. and P.L.; Investigation, M.L., E.C. and A.W.; Methodology, M.L., E.C., A.W. and P.L.; Writing—original draft, M.L.; Writing—review and editing, M.L. and P.L. All authors have read and agreed to the published version of the manuscript.The author(s) received no financial support for the research, authorship, and/or publication of this article.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Karol Marcinkowski Memorial. Medical University in Poznań (protocol code 174/21 of 11 March 2021). The study was registered in the Clinical Trial Registry: NCT04590287 https://clinicaltrials.gov/ct2/show/NCT04590287 (accessed on 19 October 2020).Informed consent was obtained from all subjects involved inthe study.The data presented in this study are available on request from the first author. The data are not publicly available due to ethical restrictions.The author(s) declared no potential conflict of interest with respect to the research, authorship, and/or publication of this article.Study design. CVD—cardiovascular disease; ICF—International Classification of Functioning, Disability and Health.Distribution of the total of CVD risk factors. CVD—cardiovascular disease; ICF—International Classification of Functioning, Disability and Health; IS—ischemic stroke.Characteristics of the study groups.CVD—cardiovascular disease; n: size of sample; IS—ischemic stroke; ICH—haemorrhagic stroke; SD—standard deviation; a chi-squared test; b Student’s t-test for independent variables; c Mann–Whitney test.Analysis of the recurrence of CVD incidents, according to age and gender.CVD—cardiovascular disease; F—female, M—male; IS—ischemic stroke; ICH—haemorrhagic stroke; a chi-squared test.Analysis of the incidence of risk factors for CVD in secondary prevention, according to the clinical type of the previous stroke.ALT—alanine aminotransferase; AST—aspartate aminotransferase; ALP—alkaline phosphatase; BP—blood pressure; CVD—cardiovascular disease; eGFR—estimated glomerular filtration rate; ICH—intracerebral haemorrhage; ICF—International Classification of Functioning, Disability and Health; IS—ischemic stroke; HbA1c—glycated haemoglobin 1c; HR—heart rate; LDL-C—low-density lipoprotein cholesterol; n—size of the sample; NOAC—nonvitamin K antagonist oral anticoagulants; NSAIDs—nonsteroidal anti-inflammatory drugs; ULN—upper limit of normal; VKA—vitamin K antagonist.Profile of CVD risk factors as per categories of the ICF classification.ALT—alanine aminotransferase; AST—aspartate aminotransferase; ALP—alkaline phosphatase; BP—blood pressure; CVD—cardiovascular disease; eGFR—estimated glomerular filtration rate; ICH—intracerebral haemorrhage; ICF—International Classification of Functioning, Disability and Health; IS—ischemic stroke; HbA1c—glycated haemoglobin 1c; HR—heart rate; LDL-C—low-density lipoprotein cholesterol; n—size of the sample; NOAC—nonvitamin K antagonist oral anticoagulants; NSAIDs—nonsteroidal anti-inflammatory drugs; ULN—upper limit of normal; VKA—vitamin K antagonist; Red color—extreme problems; Orange color—significant problems; Yellow color—moderate problems; Light green color—minor problems; Dark green color—no problems.Comparison of the total incidence of CVD risk factors.CVD—cardiovascular disease; IS—ischemic stroke; ICH—haemorrhagic stroke; SD—standard deviation; a Mann–Whitney test; b Student’s t-test for independent variables; *—Wilcoxon test.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06022.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Acute febrile illnesses occur frequently in Guinea. Acute fever itself is not a unique, hallmark indication (pathognomonic sign) of any one illness or disease. In the infectious disease context, fever’s underlying cause can be a wide range of viral or bacterial pathogens, including the Ebola virus. In this study, molecular and serological methods were used to analyze samples from patients hospitalized with acute febrile illness in various regions of Guinea. This analysis was undertaken with the goal of accomplishing differential diagnosis (determination of causative pathogen) in such cases. As a result, a number of pathogens, both viral and bacterial, were identified in Guinea as causative agents behind acute febrile illness. In approximately 60% of the studied samples, however, a definitive determination could not be made.Reliable and detailed epidemiological data are essential for both low-income and high-income countries. The best way to get this kind of data is likely through international collaboration. This paper presents the results of a cooperative study involving specialists from Guinean University and five leading Russian and Belarusian epidemiological research institutions.The Republic of Guinea is a West African country with an area of 245,857 km2. According to the World Health Organization (WHO), the country’s population (2016) is approximately 12.4 million people [1]. Guinea is one of the poorest countries in the African region and, as a result, a low level of public health continues to be a problem, as reflected by several indicators. According to the WHO, the average life expectancy in Guinea (2016) is 59.8 years [2]. The infant mortality rate (for children aged 5 and below) is 91.7 per 1000 thousand live births [3].The top causes of illness and death in children include malaria, respiratory illness, intestinal diseases, tuberculosis, HIV infection, measles, parasitic illnesses, and infectious diseases of unknown etiology. Among the adult Guinean population, the greatest health burdens are HIV/AIDS, malaria and tuberculosis (a combined incidence of approximately 1500 per 100,000 population), acute respiratory illnesses (approximately 1000 per 100,000), and other infectious diseases (approximately 1400 per 100,000) [1,4]. It should also be noted that the above data are likely underestimates. In most cases, it is not possible to establish a precise diagnosis due to a lack of qualified medical personnel and necessary diagnostic laboratory facilities, particularly in remote areas of the country.In addition, the situation in Guinea is complicated due to numerous, widespread zoonotic pathogens (both viral and bacterial), including Yellow Fever Virus (YFV); Dengue Virus (DENV); West Nile Virus (WNV); Chikungunya Virus (CHIKV); Crimean-Congo Hemorrhagic Fever Virus (CCHFV); Rift Valley Fever Virus (RVFV); Lassa Virus (LASV); spotted fever group rickettsias (SFG rickettsias, including Rickettsia africae, R. aeschlimannii, R. massiliae); relapsing fever Borrelia spp.; Bartonella spp.; Anaplasmataceae bacteria; Coxiella burnetii; and poorly studied viruses such as O’nyong-nyong virus (ONNV) and Tahyna virus (TAHV), etc. These cause diseases of varying severity with similar clinical symptoms [5,6,7,8,9].Timely detection and diagnosis of the above diseases are important in themselves, but it is also important that they are able to masquerade the early manifestations of filovirus fever epidemics, which are characterized by high mortality and rapid spreading. For example, the largest recorded Ebola outbreak in history (between 2014 and 2016), accounting for 11,323 deaths among 28,646 confirmed cases, can be cited [10]. Notably, the first cases associated with that epidemic are now known to have occurred in 2013 [11]. From an epidemiological point of view, infection patterns have changed, and the vast majority of cases occurred through person-to-person contact [12]. The disease spread widely, including urban population involvement, due to a lack of diagnostic tools, similarities between Ebola virus disease (EVD) and other diseases which could clinically present with hemorrhagic syndromes and/or fever, as well as a lack of vigilance by health authorities [13].Moreover, the gaps in our understanding of epidemic threats in Africa is not solely an African problem. Numerous cases of importation of African zoonotic infections to Europe or the Americas have been well documented. In the worst case, they may lead to autochthonous outbreaks or even permanent establishment of an infection in a new area [14,15,16]. As cases of imported infection are rare and their initial symptoms are often unspecific, such diseases can be easily missed during a patient’s differential diagnostics.It would be inaccurate to say that infectious disease incidence in Guinea has not been studied at all. Indeed, the medical literature does reflect numerous publications devoted to various aspects of infectious disease in the region. These works, however, have been narrow in focus. They have mainly centered on the study of infectious agent biological properties, organization and conduct of specific prevention, or the clinical course and treatment of West African diseases. We utilized molecular and serological methods to analyze numerous samples from acute fever patients with the goals of determining which infections are most relevant to Guinea and which require improved differential diagnosis relative to hemorrhagic fevers, including EVD.Serum samples were collected from December 2016 to April 2017 from patients unvaccinated against YFV displaying symptoms of an acute febrile illness. The patients were hospitalized in different regions of the country, and their samples were sent for study to the Virology Laboratory of Hemorrhagic Fevers Research Project, located in Gamal Abdel Nasser University (Conakry, Guinea) in 2016–2017.The samples were collected in 25 prefectures representing lower, middle, upper, and forested Guinea (see Figure 1). Serum samples were collected on the first or second day of hospitalization, and malaria was excluded at the point of care using a rapid diagnostic test. Onset of illness, however, was between 4 and 11 days prior to sampling. Venous blood samples (5 mL) were collected in VACUETTE Serum Fast Separator tubes (Greiner Bio One, Kremsmünster, Austria), kept at room temperature for 10 min, and centrifuged at 3000× g for 5 min. Serum samples were then placed in 1.5 mL tubes and delivered to the laboratory within two days after sampling; portable containers featuring frozen-block cooling were used (temperature not higher than 8 °C). Delivered samples were stored at −70 °C until analysis (storage from one month up to one year). Repeated freeze–thaw cycles were avoided during the transportation and testing process. To ensure patient confidentiality, each clinical record was assigned an individual number, which was subsequently used to label tubes containing serum for serological study.Viral RNA was extracted from 140 µL of undiluted sera using a QIAamp viral RNA kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions.Molecular methods were used to study all of the samples in terms of Ebola virus disease (EVD), Marburg fever disease (MFD), Crimean-Congo hemorrhagic fever disease (CCHFD), Lassa fever disease (LFD), Denge fever disease (DFD), and West Nile fever disease (WNFD) (Table 1).All samples were studied for EVD, MFD, CCHFD, DFD, and WNFD using the following commercial kits: CCHFV-FL Kit (Amplisens®, Moscow, Russia); Dengue virus-FL (Amplisens®, Moscow, Russia); WNV-Fl Kit (Amplisens®, Moscow, Russia); and FiloA-screen-FL and FiloB-screen-FL Kits (Amplisens®, Moscow, Russia) for the diagnosis of all pathogenic filoviruses and non-pathogenic Reston Ebolavirus [17]. All analyses were performed according to the manufacturers’ instructions. Analysis for LFD was conducted using Qiagen One-Step RT-PCR Kit reagents (Qiagen, Germany) as described by Ölschläger et al. [18] with minor modifications. In brief, GPC gene fragments (300 bp) were amplified in a 25 μL reaction containing 5 μL of viral RNA, 0.6 μM of S36 [19] and 0.6 μM of LVS-339-rev (GTTCTTTGTGCAGGAMAGGGGCATKGTCAT) [20] primers, 0.4 mM dNTPs, 5 μL of 5x OneStep buffer, and 1.0 μL of enzyme mix. Thermal cycling parameters were as follows: 50 °C for 30 min, 95 °C for 15 min, followed by 45 cycles of amplification (95 °C for 30 s, 52 °C for 30 s, and 72 °C for 30 s). A final elongation was performed at 72 °C for 5 min. Following electrophoretic separation on a 2% agarose gel, amplified PCR products were visualized with ethidium bromide using UV illumination.ELISA was used for yellow fever diagnosis and Lassa fever confirmation. Detection of serum anti-YFV IgM was performed using the MAC-ELISA IgM capture assay, developed by the Center for Disease Control and Prevention (CDC, Atlanta, GA, USA), as recommended for WHO Yellow Fever network laboratories [20]. Specifically, wells of MaxiSorp 96-well flat-bottomed strips (Nunc, Roskilde, Denmark) were coated overnight at 4 °C with 75 μL/well of goat anti-human IgM (Sigma-Aldrich, St. Louis, MO, USA) diluted at 1:2000 in carbonate buffer (pH 9.6), followed by 4 washes with Wash Buffer (PBS with 0.05% Tween 20) and a 1 h blocking step using 200 μL/well of Blocking Buffer (10% horse serum in PBS/0.05% Tween 20/1% nonfat milk). After 4 plate washes with Wash Buffer, 75 μL/well of patient serum, the positive control (CDC, USA), or the negative control (CDC, USA) was added to wells and incubated for 1 h at 37 °C; all samples and controls were diluted at 1:400 in Dilution Buffer (PBS/0.05% Tween 20/1% nonfat milk).Each diluted sample was added to two separate wells for future incubation with the yellow fever antigen and control antigen. In parallel, the yellow fever antigen or control antigen (CDC, USA) were reconstituted in Dilution Buffer (1:40). After 4 plate washes, the yellow fever antigen and control antigen were added to their separate partner wells (75 μL/well), followed by a 1 h incubation at 37 °C. After 4 plate washes, the pan-flavivirus 6B6C-1 HRP-conjugated mAb (CDC, USA), pre-diluted at 1:6000 in Dilution Buffer, was added and incubated for 1 h at 37 °C. After 5 plate washes, 10 μL/well of Enhanced K-Blue TMB substrate (Neogen Corp., Lexington, KY, USA) was added. Plates were incubated for 10 min at room temperature, after which the reaction was stopped by addition (50 μL/well) of 1 M H2SO4. Optical absorbance was measured at 450 nm. Detection of serum anti-LASV IgM was performed using the ReLASV® Pan-Lassa IgG/IgM ELISA Test Kit based on a GP-linked protein (Zalgen Labs LLC, Germantown, MD, USA), according to the manufacturer’s instructions.The presence of specific IgM antibodies in patient sera (anti-DENV, anti-ZIKV, anti-WNFV, anti-CCHFV, anti-CHIKV, anti-RVFV, anti-ZEBOV (Zaire Ebola virus), anti-MARV (Marburg virus), anti-SFG rickettsia, and anti-Borrelia spp.) was determined using a microarray containing the following recombinant antigens: DENV type 1 E protein; DENV type 1 NS1 protein; DENV type 3 E protein; DENV type 3 NS1 protein; ZIKV E protein; ZIKV NS1 protein; CHIKV E1-protein (Meridian Life Science, Memphis, TN, USA); ZEBOV NP; MARV NP (The Research and Practical Center for Epidemiology and Microbiology, Minsk, Belorussia) [21]; CCHFV fragments of glycoprotein G1 (AA 1451–1469, 1451–1469, and 1613–1631) and fragment of L protein (AA 859–873); for RVFV, the NP protein, its NPsh fragment (AA 121–201), and a G2 glycoprotein fragment (AA 522–535); CHIKV E2 protein fragment (AA 1–264); WNFV NS1 protein (all from the Central Research Institute for Epidemiology (CRIE), Moscow, Russia) [22]; B. afzelii and B. garinii proteins p100, p41, p39, p58, BBK32, OspC, p17, and the antigenic fragment of the VlsE protein; B. miyamotoi GlpQ, Vsp1, Vlp15/16, Vlp18, and Vlp5 proteins (all from CRIE, Moscow, Russia); and R. sibirica GroEl, OmpA (AA 1256–1734), and OmpB (AA 1210–1654) proteins (all from CRIE, Moscow, Russia) [23,24].Antigens in predetermined concentrations (from 35 to 200 µg/mL) and human IgM control (Jackson ImmunoResearch, Chester County, PA, USA) solutions (in concentrations of 5, 10, and 50 µg/mL) were spotted, in triplicate, on the surface of aldehyde-activated VALS glass slides (CEL Associates, Los Angeles, CA, USA) using a sciFLEXARRAYER SX (Scienion AG, Berlin, Germany) to produce an ELISA-like multiplex assay in a microarray format. PBS (1x) was used as a printing buffer and negative control, while bovine serum albumin (Sigma-Aldrich Inc., USA) labeled with Cy3 NHS ester (Sigma-Aldrich Inc., USA) was used as an internal positive control (array border marker, ABM). After printing and overnight incubation in a humid chamber, slides were blocked for 1 h at 37 °C with a 0.5% BSA (Sigma-Aldrich Inc., USA) solution in PBS. Slides were stored at −20 °C. Immediately before use, microarrays were washed with PBST (1x PBS containing 0.01% of Tween 20) for 2 min at 37 °C with shaking at 500 rpm. Sixteen-well Fast Frames and Fast Slide Holders (Sigma-Aldrich Inc., USA) were used for well-formation. Serum samples (diluted 1:10 in PBS solution with 2% BSA) were added to each array and incubated for 30 min (37 °C with 500 rpm shaking). Next, all liquids were aspirated, and the wells were washed with PBST for 2 min (37 °C, 500 rpm shaking), followed by removal of all wash solutions (by aspiration) from all wells. After those wash steps, Cy3-conjugated goat anti-human IgM antibodies (50 ng/mL) (Jackson ImmunoResearch, USA), diluted in assay buffer, were added to each well and incubated for 30 min at 37 °C. After aspiration of liquids, a washing step was performed as described above. Slides were then removed from their frames and holders, washed with ultra-pure water (Milli-Q), and dried. The resultant fluorescent signals were measured on a MArS laser microarray scanner (Ditabis, Pforzheim, Germany).Images were quantified using SpotScout software (Ditabis, Pforzheim, Germany) in accordance with its user manual. The obtained raw numeric data were processed as follows: human IgM dose-response calibration curves were fitted using a 3-parameter curve-fit for each array, and concentrations of IgM specific to recombinant antigens were interpolated from the human IgM calibration curves using ImStar software (CRIE, Moscow, Russia) for each array. Immunoglobulin M levels were calculated as micrograms per unit volume (µg/mL). Specific IgM levels to each antigen were considered significant if they exceeded 5 µg/mL [25,26,27]. This value was used as the permanent cut off for all antigens used in the microarray. Unification of the cut off value was achieved by varying antigen concentrations and sorption conditions during microarray development.For all pathogens, a determination was made as to what criterion defined a positive result (Table 2).Samples were considered positive for the presence of anti-Borrelia spp. IgM if antibodies were found against (1) at least one OspC antigen, in the absence or presence of antibodies against any other Borrelia spp. antigen; (2) at least two antigens belonging to different antigen groups (p41, p17, VlsE); or (3) against GlpQ antigen, in the presence of Abs against at least one antigen in a set (p39, p41, VlsE, Vsp1, Vlp5, Vlp15/16, Vlp18). All other samples were considered negative with respect to the presence of anti-Borrelia IgM.Samples were considered positive for the presence of anti-SFG rickettsia antibodies if Abs were found against OmpA or OmpB antigens in the absence or presence of antibodies to GroEl. All other samples were considered negative with respect to the presence of anti-SFG rickettsia IgM.Samples were considered positive for the presence of anti-DENV IgM if Abs were found against at least against one DENV NS1 antigen in the absence or presence of antibodies against E protein. Samples were considered positive for the presence of anti-ZIKV IgM or anti-WNFV IgM if Abs were found against ZIKV NS1 or WNFV NS1 antigen, respectively, in the absence or presence of antibodies against E protein.Samples were considered positive for the presence of IgM against unspecified flaviviruses if antibodies were found against at least one DENV or ZIKV E antigen; samples were considered negative for the presence of IgM against flaviviruses if no antibodies against any of the antigens (DENV or ZIKV) were found.Samples were considered positive for the presence of anti-CHIKV IgM if antibodies were found against the E1 or E2 antigens. In all other cases, samples were considered anti-CHIKV IgM negative.Samples were considered positive for the presence of anti-CCHFV IgM if antibodies were found against (1) the NP and/or NPsh antigens in the absence or presence of antibodies against any of the other antigens or (2) against any number of G-antigens in the presence of antibodies to L-protein. In all other cases, samples were considered anti-CCHFV IgM negative.Samples were considered positive for the presence of anti-RVFV IgM if antibodies were found against one of the NP or NPsh antigens in the absence or presence of antibodies against G2 antigen. All of the other samples were considered anti-RVFV IgM negative. For ZEBOV and MARV, samples were considered positive for the presence of anti-virus IgM if antibodies were found against their corresponding NP antigens.Some previous evaluation work has been performed with the microarray [22,23,28,29,30,31,32]. Negative control samples were taken from Russian healthy donors. Presumptively positive samples were taken from our collection of PCR-confirmed cases of diseases either endemic in Russia (borrelioses, ricketttsioses, CCHFD, WNFD), imported to Russia (DFD, ZIKFD, CHIKD), or from Guinean patients (EVD). Positive serum samples were not available for RVFV and MARV. The array sensitivity and specificity are shown in Table 3. No cross-reactivity was observed between serum samples of patients with diseases caused by Borrelia spp., SFG rickettsia, CHIKV, CCHFV, or ZEBOV in relation to off-target pathogen antigens immobilized on the array. Due to the well-known phenomenon of high cross-reactivity between antibodies against different flaviviruses, interpretation criteria were adapted to minimize such effects. Only antibodies against their corresponding NS1 proteins were considered to be a marker of DENV or ZIKV infection (Table 2). Regarding ZIKV cross-reactivity, one anti-DENV IgM+ sample (of 33) and three anti-WNV IgM+ samples (of 48) cross-reacted with ZIKV proteins.The 95% confidence interval for a proportion was calculated according to refinements made by R. Newcombe on the procedure outlined by E. Wilson [33,34], using the calculator at https://epitools.ausvet.com.au/ciproportion (accessed on 16 April 2021). The significance of the difference between nominal variables (proportion of positive samples in different groups and subgroups) was estimated using Fisher’s exact test. The significance of differences between numeric scale variables (age of patients and interval between the date of disease onset and the date of sampling in different groups and subgroups) was estimated using nonparametric Mann–Whitney test and exact test (2-tailed) in IBM SPSS Statistics 19. The magnitude of differences between groups (effect size) was estimated for scale variables (Cohen’s d, Glass’s Δ) and nominal variables (odds ratio, Phi, Cramer’s V) [35].In total, 164 serum samples from 25 Guinean prefectures were analyzed (Table 4). Patient ages ranged from 2 to 75 years (median age 19 years, interquartile range 9–21). Women accounted for 44.5% ± 0.5% of the specimens and men for 55.5% ± 0.5%.Using molecular methods (PCR), seven Lassa+ samples (4.3% ± 1.5% of the samples studied) were identified. All Lassa+ samples were confirmed using the ReLASV® Pan-Lassa IgG/IgM ELISA Test based on GP-linked protein (Zalgen Labs LLC, Germantown, MD, USA). No other pathogens were detected during the PCR-based study of the samples. Using serological methods, detectable levels of specific serum IgM to antigens of one pathogen were observed in 45 of the studied samples (n = 157, excluding the 7 Lassa+ samples), as follows: IgM against YFV was found in 20 patients (12.2% ± 0.9% of all samples); IgM against DENV in 1 patient (0.6% ± 0.5% of all samples); IgM against ZIKV in 2 patients (1.2% ± 0.4% of all samples); IgM against CCHFV in 1 patient (0.6% ± 0.5% of all samples); IgM against SFG rickettsia in 8 patients (4.9% ± 1.2% of all samples); and IgM against Borrelia spp. in 13 patients (7.9% ± 1.3% of all samples). IgM against ZEBOV or MARV was not found.In addition, we identified numerous samples with IgM against two or three pathogens together, as follows: IgM against SFG rickettsia and relapsing fever Borrelia spp. in 9 patients (5.5% ± 1.7% of all samples); IgM against DENV and CHIKV in 2 patients (1.2% ± 0.4% of all samples); IgM against Borrelia spp. and RVFV in 2 patients (1.2% ± 0.4% of all samples); and lastly, IgM against WNFV, SFG rickettsia, and relapsing fever Borrelia spp. in 2 patients (1.2% ± 0.4% of all samples). A total of 97 samples (59.1% ± 0.8% of all samples) were found to be negative using both molecular and serological methods (Figure 2).The proportion of positive findings of laboratory diagnostics and the type of detected pathogens did not depend significantly on the place and time of sampling, gender, and age of patients (p > 0.1 in all comparisons). Even though differences between groups were present, their magnitude (effect size) should be considered “small” [35].All 164 clinical samples were obtained from patients unvaccinated against YFV with acute severe fever as their main clinical sign. Severe acute fever itself is not a unique, hallmark indication (pathognomonic sign) of any one illness or disease. As such, patient samples representing any number of viral or bacterial pathogens, including the Ebola, may have been present. Diagnosis was achieved in slightly over 40% of the studied samples; viral and bacterial pathogens were both identified. Similar data was obtained during an investigation of causative agents of acute fever in samples collected in the neighboring West African country of Mali: evidence of viral or bacterial infection was found in 39.9% of samples (14.4% Leptospira spp., 7.7% DENV, 5.3% CHIKV, 0.27% WNFV, 7.2% hantaviruses, 0.27% LASV, and 4.8% CCHFV IgM-positive, respectively) [36].Among viral pathogens, most were attributable to YFD (12.2%) and LFD (4.3%), which is not surprising given that both are endemic to the territories of Guinea and neighboring Sierra Leone [5,37,38]. In 2008, Guinea reported about six confirmed cases of YFD (two in each of the Faranah, N’zérékoré, and Kankan health districts) and about 41 suspected YFD cases (21, 14, and 6 in Faranah, N’zérékoré, and Kankan health districts, respectively), four of whom died due to fever and jaundice [39]. YFD and LFD cases were not linked epidemiologically. Natural foci of YFV and LASFV exist in Guinea, and these contribute to a sporadic background incidence. This background makes outbreaks and epidemics of these viruses (YFD and LFD) continuously possible. We considered the LFD+ and YFD+ case determinations to be reliable based on methodology. Lassa+ samples were confirmed using both PCR and ELISA, while YFD+ samples were diagnosed using the WHO-recommended ELISA method.Furthermore, CCHFD+ (0.6%), ZIKFD+ (1.2%), and DFD+ (0.6%) cases were identified. Therefore, these viral infections were demonstrably present among humans in Guinea. This fact should be taken into account during differential diagnosis of acute febrile illnesses. Ideally, field results would be further confirmed with high confidence using a plaque reduction neutralization test (PRNT). It is hardly feasible, however, in the low-resource settings in which studies are being carried out in Guinea. Notably, the first human WND case in Korea was imported from Guinea and confirmed by PRNT in a Korean laboratory [40]. In addition, there was evidence of ZIKV circulation in Senegalese vectors and the population, suggesting the possibility of such circulation in Guinea due to similarity in climate and habitat type [41]. The seroprevalence of DENV, CHIKV, and ZIKV in surrounding West African countries was found to be 10–30%, 30–40%, and 3–5%, respectively (the possible cross-reactivity of detected IgG antibodies within flaviviruses and alphaviruses was not considered) [42].Obviously, all of these infections may be imported from Africa. In Europe, the most affected are southern countries where importation may lead to autochthonous outbreaks [14]: Spain, France, Croatia, Greece, and Italy, in particular. From 2008–2011, 109 imported cases of DENV infection and 21 imported cases of CHIKV infection were reported to the Italian National Institute of Health [15]. When the National Plan on Human Surveillance of Vector-borne Diseases was implemented, the Italian National Reference Laboratory for Arboviruses diagnosed 68 laboratory-confirmed imported cases of DENV infection, 35 imported cases of CHIKV infection, along with the detection of the first four confirmed ZIKV cases, in the period from July 2014 to October 2015 [16]. The number of DVD cases in Russia is even higher. Since its first detection in the country, more than 1500 clinical DVD cases have been officially registered (2012–2019), with a maximum of 415 cases in 2019; all of them were imported [43,44,45]. According to estimates by Napoli et al. [15], the number of DENV-exposed travelers may be about 20-fold higher.Although most imported cases of vector-borne infection come from popular recreational areas in Thailand, Maldives, Vietnam, etc., in absolute numbers [15,16,43,44,45], the relative risk of travelers contracting zoonoses may be higher in West Africa. In addition to illnesses of viral etiology, cases of SFG rickettsiosis (4.9%) and relapsing fever caused by Borrelia spp. (7.9%) were identified. Mixed infections, mainly Borrelia spp. with SFG rickettsiosis (5.5%), were seen. Other mixed infections were also seen: DENV and CHIKV; RVFV and Borrelia spp.; and three together (WNV, SFG rickettsia, and relapsing fever Borrelia spp.). These findings suggest that Guineans are likely attacked by both mosquitoes and ticks, and that these attacks may be occurring simultaneously or over short time frames. In total, 15.9% of acute febrile illnesses had bacterial etiology in our study. Similarly, DNA from at least one pathogenic bacterium were identified in 80/440 (18.2%) of the samples from febrile patients in Senegal (35, 30, 23, 2, and 1 cases for Borellia crocidurae, Rickettsia felis, Bartonella spp., Coxiella. burnetii, and Tropheryma whipplei identification, respectively [46].Bacterial infections can also be imported. Evidence of B. crocidurae infection has been noted in travelers returning to France and Italy from Senegal and Mali [47,48,49]. Spotted fever group rickettsioses, first of all, R. africae infection, have been diagnosed in travelers returning to the U.S. from Liberia, Gambia, and other African countries [50]. R. typhi infection was found in a traveler returning to Spain from Senegal [51].The geographic distribution of positive samples was uneven (Table 4). Most positive samples (all types) were obtained from lower Guinea (29 positive samples of 85, or 37%), while 14 positive samples were obtained from middle Guinea (44% positive of 32 samples) and upper Guinea (61% positive of 23 samples) each. In addition, seven positive samples (29% positive of 24 samples) were obtained from forested Guinea. Although noticeable, differences in the proportion of positive samples in different geographic areas did not reach statistical significance (p = 0.12, Fisher’s exact test), and effect size was small (Cramer’s V = 0.19). For some pairwise comparisons, odds ratios were rather high: 3.8 (p = 0.04) and 2.7 (p = 0.06) when comparing upper Guinea versus forested Guinea or lower Guinea, respectively.No significant socio-demographic differences among positive samples were seen.In a sizable number of cases (approximately 60%), a diagnosis was not determined. Several factors may have been contributing to this. One potential reason may have been sample handling issues, such as sub-optimal sample collection time frames, improper storage conditions, or deficiencies related to delivery of materials to the laboratory. It is also possible that the set of infectious agents (of various etiologies) causing fever was broader than the pathogen panel used to analyze patient samples. For example, Leptospira spp., Bartonella spp., hantaviruses, as well as Bombali virus can also cause febrile illness [36,46,52]. Perhaps it is appropriate here to quote verbatim the results of a long-term, large-scale study of the Soviet era [53]: “In 1978–1991, the USSR–Guinea Virological and Microbiological Laboratory functioned in Kindia, the Republic of Guinea. … About 74,000 mosquitoes, 100,000 Ixodidae ticks, 1500 wild birds, 2700 bats, 106 monkeys, 308 other mammals, and 927 blood samples collected from febrile patients were examined in 1978–1989, using inoculation of new-born white mice. As a result of this work, 127 strains of the following arboviruses were isolated: Chikungunia (one strain), Dengue 2 (four), Saboya (seven), Wesselsbron (one), Bunyamwera (four), M’Poko (five), Rift Valley Fever (six), CHF-Congo (nine), Dugbe (22), Bhanja (six), Forecariah (two), Jos (26), Abadina (15), Kindia (two), Ark 6956 (one), Fomede (two), Bluetongue (nine), Mossuril (two), AnK 6009 (one), and Kolente (two). Dengue 2, Wesselsbron, Bunyamwera, M’Poko, Kindia, and Mossuril viruses were isolated from mosquitoes. Ixodidae ticks were sources for isolation of Chikungunia, Saboya, CCHF, Dugbe, Bhanja, Forecaciah, Jos, Abadina, Kindia, Ark 6956, Fomede, Bluetongue, and Kolente viruses. Saboya, RVF, Fomede, Kolente, and AnK 6909 were isolated from bats (Chiroptera); Saboya, Abadina, and Bluetongue viruses were isolated from birds. One strain of Dugbe virus was originated from the brain of Cercopithecus patas. Bunyamwera and Abadina viruses were isolated from the blood of two febrile patients. Serological identification of many strains was kindly conducted at the Pasteur Institute, Dakar (J. P.Digoutte) and some at the YARU, USA (R. Shope)”.Not all of these virus species are pathogenic to humans, of course, but some of them might potentially be responsible for undiagnosed febrile illness cases in our study or other studies [5,38]. T. Pierson and M. Diamond have considered African Wesselsbron arbovirus and Zika-like Spondweni virus to have potential as newly emerging flaviviruses [54].Another problem is that the currently available nucleotide sequence data on African pathogenic strains are scarce. For example, of the 125 DENV and CHIKV “reference” nucleotide sequences used to clarify the origin of DENV or CHIKV causing human cases imported into Italy [16], only eight “reference” isolates originated from Africa. Therefore, unidentified target-proximal genetic variability can hinder PCR diagnostics, leading to negative results, even for diseases caused by known pathogens. Possibly a Pan-Degenerate Amplification and Adaptation (PANDAA) approach [55] could be useful in this situation.The main limitation of our study was its modest statistical power; we were not authorized to collect enough clinical samples by ourselves and used additional samples donated by the collection of the Hemorrhagic Fevers Research Project in Guinea. As a result, some additional pathogens might not have been found (by random chance), or conversely, the frequency of other pathogens may have been overestimated. This study did, however, reveal a number of possible febrile illness agents and their relative importance in Guinea. As such, it outlined the way for further research. Such assessments can identify regions where needs and provisions do not align. These areas should be targeted for future strengthening and support of public health, as has been encouraged by a huge team of experts in a multistage analysis [56].Therefore, determining exactly which infectious pathogens are most relevant to Guinea is an extremely important step in terms of improving the local health care system and facilitating differential diagnosis of acute fevers. Clarifying data, from this work and future research, will be useful in the event of new Ebola Virus Disease outbreaks. Knowledge of the spectrum of zoonoses endemic in Guinea is also necessary to improve the quality of differential laboratory diagnostics for rapid identification of imported cases among people who have arrived from West Africa to Europe. In addition, such knowledge could be necessary, both for assessing the risks of people traveling to Guinea and for planning preventive measures, such as vaccination.We characterized, in Guinea, a number of infectious diseases that present with severe fever, including YFD, LFD, CCHFD, DENFD, SFG rickettsiosis, and relapsing fevers caused by Borrelia spp. In addition, several co-infections were identified by microarray (DENV-CHIKV, RVFV-Borrelia spp., and WNFV-SFG rickettsia-Borrelia spp.). Although CHIKV and RVFV were only identified in co-infections (CHIKV+other or RVFV+other), it was clear that both etiologic agents could exist separately. The immobilized protein microarray presented here was determined to have quite acceptable sensitivity and specificity, and successful express identification of several highly dangerous pathogens was demonstrated. Still, the methods used did not identify a pathogen in a number of severe fever cases. Therefore, further work is needed to establish a list of the most important etiologic agents relevant to Guinea to improve existing methods and to develop new diagnostic tools.Conceptualization, V.G.D.; Data curation, N.M.; Formal analysis, V.A.S., M.V.S. and A.E.P.; Investigation, O.A.S., J.C. and B.S.; Methodology, V.G.D., M.V.S. and P.S.; Writing—review & editing, V.G.D. All authors have read and agreed to the published version of the manuscript.This article was funded by RSF grant N. 20-64-46014.Participation was voluntary, and all patients agreed to participate. The study was evaluated and approved by local Ethics Committees of the Pasteur Institute, Saint Petersburg, Russia (No. 058-03) and Comité National D’Ethique pour la Recherche en Santé, Guinea (No. 129/CNERS/16). All parties involved in sample/data collection and handling met all requirements and followed all regulations on personal information data protection.Consent for child or adolescent inclusion was obtained from the parent or authorized representative.Data collection and data handling procedures strictly complied with the EU’s General Data Protection Regulation (GDPR) and the Helsinki Declaration.The authors declare that they have no competing interests.Map of Guinea. Districts where clinical samples were collected are marked with red stars.Distribution of zoonotic diseases in the study group, in percentages.Methods used to diagnose infectious diseases.* Explanation of abbreviations and details of the methods used are given in the text (Section 2.3, Section 2.4 and Section 2.5). YFD (yellow fever disease), ZIKFD (Zika fever disease), RVFD (Rift Valley fever disease), CHIKF (Chikungunya fever), SFG rick. (spotted fever group rickettsias), Bor. spp. (relapsing fever Borrelia spp.).Algorithms for the interpretation of microarray data.* Explanation of abbreviations and details of the method used are given in the text (Section 2.5).Sensitivity and specificity of the protein microarray assay.Analysis results from acute febrile illness patient samples.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06023.txt ADDED Viewed

	@@ -0,0 +1 @@

+ These authors share senior authorship.Mental health problems are relatively common during university and adversely affect academic outcomes. Evidence suggests that mindfulness can support the mental health and wellbeing of university students. We explored the acceptability and effectiveness of an 8-week instructor-led mindfulness-based course (“Mindfulness: Finding Peace in a Frantic World”; Williams and Penman, 2011) on improving wellbeing and mental health (self-reported distress), orientation and motivation towards academic goals, and the mechanisms driving these changes. Eighty-six undergraduate and post-graduate students (>18 years) participated. Students engaged well with the course, with 36 (48.0%) completing the whole programme, 52 (69.3%) attending 7 out of 8 sessions, and 71 (94.7%) completing at least half. Significant improvements in wellbeing and mental health were found post-intervention and at 6-week follow-up. Improvements in wellbeing were mediated by mindfulness, self-compassion, and resilience. Improvements in mental health were mediated by improvements in mindfulness and resilience but not self-compassion. Significant improvements in students’ orientation to their academic goal, measured by “commitment” to, “likelihood” of achieving, and feeling more equipped with the “skills and resources” needed, were found at post-intervention and at 6-week follow-up. Whilst exploratory, the results suggest that this mindfulness intervention is acceptable and effective for university students and can support academic study.A growing proportion of the population is benefitting from the social, occupational, and academic opportunities offered by higher education. In recent years, there has been a three percent increase in higher education enrolments across both undergraduate and post-graduate university students, with larger numbers of ethnic minority students and students registering with a disability [1]. For many, higher education is a major transition met with increasing social, academic, and financial demands [2,3]. Rates of mental health difficulties amongst university students are noteworthy. A global survey of 13,984 students found 35% reported at least one DSM-IV mental disorder [4]. Rates of self-reported depression and anxiety in the United Kingdom are greater than the age-matched United Kingdom general population [5]. A growing body of research suggests that mental health difficulties worsen throughout the degree programme [6,7,8]. Whilst an increase in symptoms may not be caused by higher education itself, it is frequently suggested that the daily stressors associated with university life are a significant contributing factor [7]. Post-graduate students are burdened with additional daily stressors, including difficulties in the supervisory relationship and isolation [9].The negative impact of having a mental health problem during university is broad, impacting the quality of life. Specifically, the presence of depressive symptoms in university students has been associated with role limitations due to physical health problems; while anxiety symptoms have been related to bodily pain; and both depressive and anxiety symptoms have been associated with reductions in general health, energy/fatigue, social functioning, as well as psychological distress, and lower psychological wellbeing [10]. It has also been observed that mental health problems during university affect academic performance [11] and the likelihood of dropping out [12]. Recently, UK Universities called for universities to transform institutions into “mentally healthy universities” that place the mental health and wellbeing of staff and students as foundational for all aspects of the university system [13]. The strategic framework developed by UK Universities outlines university-wide systematic changes which can be made across different domains (e.g., learning, support, etc.). For example, under the ‘learn’ domain, universities must make sure assessments “stretch and test” learning without imposing unnecessary stress. The ‘support’ domain focusses on implementing, in consultation with staff and students, safe and effective mental health interventions that are regularly audited for safety, quality and effectiveness [13]. Whilst cognitive, behavioural and mindfulness-based interventions have been found to effectively reduce symptoms of anxiety and depression in university students [14], the implementation of such interventions should be conducted in collaboration with university students to ensure that there is a demand for such programmes and that interventions are acceptable and accessible for all [13].Over the last decades, there has been a growing interest both nationally and internationally in mindfulness-based programmes (MBPs) in higher education institutions [15]. Mindfulness is a natural, trainable capacity, which encourages people to approach experiences with attitudes of curiosity and care, and in ways that support overall wellbeing, and general functioning [16,17]. By training ‘mindful awareness’, MBPs aim to promote a conscious shift away from automatic, habitual responses that may increase distress towards greater self-regulation [17]. As well as cultivating mindfulness skills, the psycho-educational content within MBPs can be tailored to the target population whilst promoting an understanding of psychological processes at the core of distress and wellbeing [18]. A meta-analytic review of MBPs implemented in universities found overall improvements in depression, anxiety and wellbeing for students post-intervention, with lasting effects (>3 months) on distress [19]. MBPs should be further tested for their appropriateness with a student population, ensuring people across the spectrum of wellbeing can make use of them and at a level of intensity that balances accessibility with potency [20].One example of an MBP is the “Mindfulness: Finding Peace in a Frantic World” course (M-FP) [21], which was adapted from Mindfulness-Based Cognitive Therapy (MBCT) [22] to be an accessible universal MBP. The course is developed for a non-clinical population and participants learn skills they can use in their daily lives to break the cycle of anxiety, stress, and exhaustion, as well as promote mental health and wellbeing [21]. M-FP themes include psycho-education, waking up from autopilot, how to respond to negative thoughts and difficult feelings, and self-care. Delivered in a university context, M-FP aims to support students to develop sustained awareness of the different aspects of daily living. Students are then encouraged to apply mindfulness skills to manage emotions, as well as academic and social pressures that may occur in day-to-day university life.A self-guided version of the M-FP course in a university sample had positive effects on depression, anxiety, stress, satisfaction with life, mindfulness and self-compassion, with good acceptability [23]. A pragmatic randomised control trial using the instructor-led version of the M-FP course for university students alongside ‘mental health support as usual’ evidenced significant improvements in mental health difficulties in comparison to a control group [24].Students also report both direct and indirect improvements in study skills and behaviour, including greater analytical thinking and memory capacity, increased enjoyment of studying, and reduced procrastination [25] following the M-FP course. Students reporting higher levels of mindfulness engage more with autonomous academic goals that are intrinsically motivated than students with lower levels of mindfulness [26], which is a consistent predictor of academic achievement across different educational contexts [27]. These findings are particularly relevant considering the disproportionate number of students with poor mental health who drop out of university or perform less well academically compared to those without mental health problems [11,12]. By providing students with internal resources to support wellbeing, mindfulness programmes could alleviate some of the disadvantages related to studying with poor mental health whilst potentially promoting positive academic behaviours. Trait mindfulness and behaving in accordance with intrinsic values are also related to greater wellbeing [28]. This is in line with Self-determination theory, which suggests that the relationship between mindfulness, intrinsic motivation and wellbeing is driven by greater awareness of self, leading to the choice of behaviours that are consistent with individual interests, values, and desires. It follows that awareness, cultivated through MBPs, may improve academic outcomes. Exploring the effects of the MF-P on students’ motivations towards academic goals is an aim of the current study.Whilst the effects of MBPs on mental health and wellbeing are widely researched [29,30], the mechanisms through which these improvements are facilitated remain unclear. There is growing evidence to suggest that both mindfulness and self-compassion may be important mechanisms of change [31,32]. For example, in a study of an instructor-led M-FP tailored for secondary school teachers, there was reduced stress and increased rates of wellbeing, mindfulness and self-compassion [33]. Additionally, mindfulness and self-compassion have been shown to mediate improvements in wellbeing [34], stress, burnout and mental health [35].The relationship between mindfulness and self-compassion is complex. Initially, it was conceptualised as a bidirectional relationship with each enhancing the other [36], but emerging research suggests that mindfulness and self-compassion improve mental health and wellbeing independently [37]. The delivery of the MBP may also impact these mechanisms for change, with the instructor-led delivery of the M-FP program indirectly improving wellbeing, mental health and burnout in secondary teachers by significantly enhancing mindfulness and self-compassion compared with the self-taught M-FP program [20]. Such mechanisms are also yet to be tested in a university student population.Resilience is another key mechanism of change in MBPs [38]. Resilience refers to positive adaptation in the face of stress or trauma [39], a skill which is of particular importance to university students, who are faced with a number of novel challenging experiences (e.g., increased academic and financial demands). Drawing upon clinical models depicting mechanisms of change, the cultivation of mindfulness skills is thought to enhance self-regulation skills, promoting the psychological resilience that supports mental health and wellbeing [40]. This is supported by cross-sectional and experimental research, revealing a partially mediating serial relationship between mindfulness, resilience and subjective wellbeing in community, and university samples [38,41]. In a recent cross-sectional study of general population participants, significant direct effects of mindfulness, self-compassion and resilience on anxiety and depression symptoms were observed, and also indirect effects of mindfulness and self-compassion through resilience on depression symptoms were found [42].An improved understanding of the mechanisms of change for MBPs would enable further refinement of these programmes, thereby optimising individual outcomes. Mindfulness, self-compassion, and resilience are likely to be important mechanisms in MBPs for improving wellbeing and mental health outcomes in students. Based on the literature presented, Figure 1 shows a proposed model whereby there is a sequential relationship between mindfulness, self-compassion, resilience, mental health, and academic outcomes. Testing this model is beyond the analytic and methodological capacity of this paper but forms the theoretical basis of the analyses conducted.The present study explores engagement with acceptability and effectiveness of the instructor-led M-FP course [21] in improving mental health and wellbeing in university students. Second, the impact of the M-FP course on the students’ orientation and motivation towards their academic goals is also explored. Finally, three mechanisms of change were independently explored in relation to mental health and wellbeing: (1) mindfulness, (2) self-compassion, and (3) resilience.The study was conducted at the University of Oxford and utilised an open pre-post-test intervention design with a 6-week post-course follow-up. Participants gave informed consent and completed self-report questionnaires online of the outcomes and mechanisms. The questionnaires were administered using Qualtrics software (Qualtrics, Provo, UT, USA). Acceptability was assessed through both bespoke questions as well as engagement with the programme monitored by the M-FP teachers.Participants were undergraduate and post-graduate students (aged ≥18 years) who signed up to an 8-week group-based mindfulness course offered by the Oxford Mindfulness Centre and volunteered to participate in the study. Prior to course enrolment, participants self-assessed the suitability of the mindfulness course based on information supplied by the Oxford Mindfulness Centre [43], including a statement on contraindications (such as serious mental or physical health concerns or recent bereavement) to the mindfulness course. Expression of interest for the present study occurred between October 2017 and April 2019, and therefore participants were consecutively recruited during this time. The research team contacted those who expressed an interest, provided them with information about the study and an opportunity to ask questions. Participants were then sent a link to consent to the study via an online form. During the study period, 200 and 96 students took part across 10 mindfulness groups, each group following the 8-week mindfulness-based programme. Of these, 86 students (29%) consented to participate in the study.Data was collected pre-intervention, one week prior to the first-course session (T0), post-intervention, at the end of the 8-week course (T1), and at a 6-week follow-up after completion (T2). Participants received £20 after completing the first 2 time points and an additional £10 for continued participation at T2 (a maximum of £30 in total) as a reimbursement for their time. The payment was not conditional on attendance at the mindfulness course.A risk and safeguarding protocol was implemented to protect participants’ safety. Those reporting significant distress were provided with information on sources of support. Additionally, participants who reported suicidal ideation were contacted for further assessment and, when appropriate, the student’s college nurse was informed. The University of Oxford Research Ethics Committee approved the study (R52786/RE004).The MBP used for the intervention was the M-FP [21]. The course is an adaptation of Mindfulness-Based Cognitive Therapy (MBCT) [44] and includes 8 in-person weekly 90-min sessions of reduced-intensity MBCT with a focus on reducing general distress and improving wellbeing [21]. The main themes include “waking up to the autopilot,” “keeping the body in mind,” “the mouse in the maze,” “moving beyond the rumour mill,” “turning towards difficulties—from reacting to responding,” “practicing kindness,” “when did you stop dancing?” and “your wild and precious life.” Students paid £65 to take part in the MBP, which was the standard price of the course, a cost unrelated to the study. Each of the 10 mindfulness groups that did the 8-week programme consisted of up to 30 participants (M = 29.6, SD = 0.97). Students were recommended to engage in daily home practice from the course books audio files and activities, such as mindful walking and eating, amongst others. All groups were taught by a qualified mindfulness teacher from the Oxford Mindfulness Centre, who met the good practice guidelines developed by UK Network for Mindfulness-Based Teacher Training Organizations (http://mindfulnessteachersuk.org.uk/#guidelines, accessed on 2 June 2021). The mindfulness teachers were not involved in the research study and were unaware of which students were study participants.At baseline (T0), participants were asked to provide information on their age, gender (Male, Female, Other, Prefer not to say), ethnicity (White British; White Irish; Any other White background; White and Black Caribbean; White and Black African; White and Asian; Any other mixed/multiple ethnic backgrounds; Indian; Pakistani; Bangladeshi; Chinese; Any other Asian background; Caribbean; African; Any other Black/African/Caribbean background; Arab; Any other ethnic origin group; Prefer not to say) and education (Bachelor’s degree, Master’s degree, Doctorate degree, Other). Participants were additionally asked whether they were currently suffering from a mental health disorder or had previously received a diagnosis of a mental health disorder and were asked to specify their diagnosis. Furthermore, we collected information on the students’ previous experiences with meditation in the format of an open answer question.At T0, participants were asked to report how much they expected to benefit from the mindfulness course on a Likert scale [‘not at all’ (0) to ‘very much’ (10)]. Following the mindfulness intervention (T1), participants were asked to report on the number of mindfulness sessions they had attended (0–8) and the number of days and minutes they had practiced mindfulness at home during the eight-week course. Course acceptability was assessed using four statements, which the participants rated from 0 to 10: “How much do you feel that you benefitted from the course?”; “Please rate the quality of teaching,” “Mindfulness courses should be made widely available to students at the University of Oxford,” and “How likely are you to use mindfulness in the future?” A mean of these four statements was used to calculate a “total acceptability score.” The internal consistency of these four statements was good in the current sample (alpha = 0.78).At all three time points (T0, T1 and T2), the Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS) [45] was administered. The WEMWBS is a psychometrically robust 14-item measure, assessing wellbeing over the past two weeks. Upon being presented with a statement (e.g., I have been feeling good about myself), participants were asked to score each item on a 5-point scale [“none of the time” (1) to “all of the time” (5)], yielding a total score ranging from 14 to 70. A higher total score indicated greater wellbeing [45]. Internal consistency in the current sample was good (T0 alpha = 0.86; T1 alpha = 0.87; T2 alpha = 0.90).The Clinical Outcomes Routine Evaluation-10 (CORE-10) [46] is derived from the 34-item Clinical Outcomes Routine Evaluation-Outcome Measure [47] and was used as a measure of psychological distress at all data collection time points (T0, T1 and T2). This questionnaire consists of 10 statements about thoughts and feelings (e.g., “I have felt tense, anxious, or nervous”). Participants were asked how often they felt this way over the past week, using a 5-point Likert-type scale, [“not at all” (0) to “most or all of the time” (4)]. The total score ranges from 0 to 40; a score of 11 or higher is considered clinically significant. The CORE-10 has shown good validity, internal consistency, and sensitivity to change [46]. The internal consistency in the current sample was good (T0: alpha = 0.84, T1: alpha = 0.81, T2: alpha = 0.82).At all data collection time points (T0, T1 and T2), participants completed the Five-Facet Mindfulness Questionnaire-Short Form (FFMQ-SF) [48], which is a shortened, 15-item version of the Five-Facet Mindfulness Questionnaire (FFMQ) [49]. Participants rated statements (e.g., “I’m good at finding words to describe my feelings”) on a 5-point scale [“never or rarely true” (1) to “very often or always true” (5)]. As per recommendations for pre–post research [50], the “observation” subscale was excluded from the total score calculations, leading to a range of 12–60. The internal consistency in the current sample was appropriate (T0: alpha = 0.81, T1: alpha = 0.82, T2: alpha = 0.75).The present study used a short-form version of the Self-Compassion Scale (SCS-SF) [51] to measure self-compassion at all 3 time points (T0, T1 and T2). With 12 items rather than 26 [51,52], this measure has good convergent validity and reliability [51]. Participants were presented with a statement (e.g., “When I’m going through a very hard time, I give myself the caring and tenderness I need”), which they rated in terms of the frequency of this experience, using a five-point scale [“almost never” (1) to “almost always” (5)]. We used a total score of self-compassion [53] that ranged from 1 to 5, with higher values indicating greater levels of self-compassion. Internal consistency in the current sample was good (T0 alpha = 0.84; T1 alpha = 0.81; T2 alpha = 0.82).The Connor–Davidson Resilience Scale (CD-RISC) [54] is a 25-item measure of resilience and coping. At all three time points (T0, T1 and T2), resilience was assessed using a shorter 10-item version (CD-RISC-10) [39], which has a stable factor structure, good reliability and validity [39]. Participants rated each item (e.g., “I believe I can achieve my goals, even if there are obstacles”) for their “truthfulness” over the past month on a 5-point scale [‘not true at all’ (0) to ‘true nearly all the time’ (4)]. The total score can range between 0 and 40, with a higher score indicating greater resilience [39]. Internal consistency in the current sample was good (T0: alpha = 0.88, T1: alpha = 0.89, T2: alpha = 0.90).The ‘Measure to elicit positive future goals and plans’ (MEPGAP) [55] has been adapted in previous research to measure conditional goal setting [56,57]. For the purpose of the present study, we further adapted this measure to assess students’ academic goals. At each time point, participants were asked to generate their most important “academic goal,” defined as “something that they would like to happen or to be true of their academic life in the future.” Subsequently, participants were asked to rate (1) the likelihood of achieving this goal, (2) the extent to which they felt that they have the skills and resources necessary to obtain this goal and (3) how committed do they felt to attaining this goal on a scale of 0–10. Intrinsic and extrinsic motivation towards achieving this goal was measured by asking participants to rate the following two statements (1) “I am pursuing this goal because someone else wants me to, or because I will get something from somebody if I do. I probably wouldn’t pursue it if I didn’t get some kind of reward, praise, or approval for it” and (2) “I am pursuing this goal because I really believe that it is an important goal I have. I endorse it freely and value it wholeheartedly.” from (“not at all because of this reason” 0 to “completely because of this reason” 10). Each subscale (i.e., likelihood, skills and resources, commitment, intrinsic motivation and extrinsic motivation) was treated as a separate measure of academic goal pursuit, with a higher subscale score indicating greater orientation towards academic goals on that specific domain.Baseline socio-demographic characteristics were described using means (SDs) for continuous data and frequencies (percentages) for categorical data. Associations between socio-demographic characteristics and research attrition were analysed at post-test and follow-up, using odds ratios obtained from bivariate logistic regression analyses.We used the mean (SD) to describe expectations of benefit from the mindfulness course. The engagement was presented in terms of the number of sessions attended, using frequencies (percentages), and the rates of home practice during the programme, using medians and interquartile ranges (IQRs). Participant ratings of acceptability were described using means (SDs). Using bivariate linear regression analyses and standardised beta coefficients (β), we explored the possible relationships between baseline characteristics and the number of sessions attended, as well as between course acceptability and pre–post differential scores in outcomes.The primary analysis was carried out for the main outcomes of wellbeing and mental health at post-test (T1) and follow-up (T2), using intention-to-treat analysis (ITT). We used mixed linear regression analyses in which time was entered as the independent variable, and within-person variance was captured by the random part of the model. We used the restricted maximum likelihood algorithm (REML) to obtain less biased estimates of parameters given the small sample size [58]. We calculated unstandardised regression coefficients (B) from complete cases for our primary analyses. Secondarily, we developed sensitivity analyses carrying out adjusted—controlling for previous experience of mindfulness or meditation and expectancy at baseline—and imputed models. Imputed models used linear multiple imputations of 20 datasets based on chained equations to address missing data at post-test and follow-up in the main outcomes (i.e., distress and wellbeing), considering: (a) those variables included in the primary outcome analysis; (b) variables significantly related to, or potentially related to, non-response (see supplementary materials Table S1); (c) variables that explained a significant amount of variance in the main outcomes (see supplementary materials Tables S2 and S3 for a list of the included variables). Additionally, we calculated the possible “time × current diagnosis of mental health problems” and the “time × previous diagnosis of mental health problems” interactions on distress. We also estimated the effect of time on the mechanistic variables of mindfulness, self-compassion, and resilience, and the effect of time on the secondary outcomes (i.e., the student’s orientation towards their academic goals). We reported within-group effect sizes (ESs) from the marginal means by correcting for the dependence of the repeated measures [59]. For the proposed interactions, we computed ESs for pairwise comparisons, using the pooled pre-test standard deviation to weight the differences in the pre‒post marginal means and to correct for the population estimate [60]. Standardized ESs of d ≤ 0.20 are considered small, d = 0.50 as medium, and d ≥ 0.80 or more as large [61].Using Spearman’s correlations, we explored the possible relationships between the number of mindfulness sessions attended as well as the amount of practice during the programme—calculated as the multiplication of the number of days practiced and the estimated mean practice duration per day—and the pre–post changes in each outcome. Additionally, we also completed these same analyses on a subset of participants with current mental health difficulties.To test the possible mediating effects of mindfulness, self-compassion, and resilience, we used an exploratory within-participant path-analytic framework [62,63]. Firstly, we developed a post hoc power analysis for participants with complete data. Assuming a partial mediation scenario, we estimated the product of the tests of paths “a” and “b,” which closely estimates the power of bootstrap estimates [64,65] for a large effect in both paths “a” and “b” (i.e., a standardised path-value of 0.40 each) and an intermediate effect in path “c’” (i.e., the direct effect after conditioning on the indirect effects, with a path-value of around 0.30) of the mediation models. Secondly, we evaluated the correlations between pre–post-intervention changes in mindfulness, self-compassion, and resilience, with pre-intervention to follow-up changes in the primary outcomes of wellbeing and distress. Finally, we explored the relationships between time (i.e., the independent variable), the pre‒post-intervention changes in the mediators (i.e., mindfulness, self-compassion, and resilience), and the pre-intervention to follow-up changes in the main outcomes (i.e., distress and wellbeing), using ordinary least squares (OLS) analyses. For these analyses, we used unstandardised path estimates from the regression coefficients and developed independent models for each tested mediator (i.e., mindfulness, self-compassion, and resilience) and main outcome (i.e., distress and wellbeing). Thus, we entered ‘time’ as the independent variable (X), the mindfulness (or self-compassion or resilience) pre‒post-intervention change scores as a possible mediator (M), and the pre-intervention to follow-up change scores for wellbeing (or distress) as a dependent variable (Y). The simple within-group path mediational model followed is graphically depicted in Figure 2. We calculated the unstandardised regression coefficients and the corresponding standard errors for bootstrapped indirect effects. We present the 95% bias-corrected confidence interval based on 10,000 bootstrap samples to overcome possible problems of asymmetry in the distribution of the indirect effects. Indirect effects are considered statistically significant when their 95% confidence interval does not include zero [66]. Finally, we used the multiple determination coefficient (R2) to calculate the ESs for the mediation models (i.e., values of 0.00 = null effect, 0.14 = small effects, 0.39 = medium effects, and 0.59 or more = large effects) [67].An overall 2-sided α level of 0.05 was used. We did not correct for multiple testing, given the exploratory nature of the present study [68]. Analyses were performed using STATA v17.0 (StataCorp. College Station, TX, USA) and IBM SPSS v27.0 (IBM Corp. Armonk, NY, USA).Table 1 provides an overview of the participant characteristics at baseline, and a flowchart of participants can be seen in Figure 3. Research attrition was predicted by age, the current presence of a diagnosed mental health problem, as well as baseline levels of mindfulness, self-compassion, distress, and the secondary outcome of commitment towards academic goals (see supplementary materials Table S1). No other variables were involved in the missing data pattern.The expectation of benefit was M = 7.09 (SD = 1.64). Of the 74 participants who completed T1 measures, 36 (48.0%) participants attended the whole M-FP programme, and including these, 52 (69.3%) attended seven out of eight sessions, with 71 (94.7%) attending half or more of the programme (Figure 3). The median number of days participants reported practising mindfulness meditation during the programme was 5 days per week (IQR: 3–7), and on these days, they reported engaging in mindfulness practices for a median of 10–15 min per day (IQR: 5–10 to 15–20). Participant ratings of acceptability, following the intervention, were as follows: “the intervention was beneficial”, M = 6.93 (SD = 1.90); “the intervention should be made widely available”, M = 9.12 (SD = 1.21); “quality level of the teaching,” M = 8.05 (SD = 1.97); “likelihood of using mindfulness in the future,” M = 8.12 (SD = 1.99). The total acceptability score had a mean value of M = 8.06 (SD = 1.40). Expectations of benefit from the mindfulness course was not a predictor of completion of the programme (β = 0.17; p = 0.148). However, participants classified as ‘White,’ including ‘White British, ‘White Irish, any ‘Other White’ ethnicity, were significantly associated with a higher number of sessions attended compared with other ethnic groups (β = 0.28; p = 0.019), and thus, ethnicity was a predictor of completion of the programme. The expectations of benefit from the mindfulness course was significantly related to the total acceptability score (β = 0.50; p < 0.001), while the total acceptability score was significantly related to pre‒post improvements in mindfulness (β = 0.29; p = 0.012), self-compassion (β = 0.25; p = 0.030), resilience (β = 0.39; p = 0.001) and wellbeing (β = 0.40; p < 0.001). No other significant relationships were found between expectations of benefit, total acceptability score and pre–post differential scores in other outcomes.Table 2 shows that there were significant improvements in the main outcomes of wellbeing and distress, following the intervention at T1 (WEMWBS: B = 2.08; p < 0.001; CORE: B = −2.63; p < 0.001) with moderate effects, and at T2 (WEMWBS: B = 2.09; p < 0.001; CORE: B = −2.39; p = 0.001), with moderate and low-moderate effects, respectively. Adjusted models and analyses using imputed data reinforced these results, with significant but reduced slopes (see supplementary materials Table S2).As can be seen in Table 3, there was an interaction effect between time and current mental health problems on distress at post-intervention and follow-up, with large effects. Furthermore, the analyses revealed a significant interaction effect between time and previously diagnosed mental health problems on distress at post-intervention and follow-up, with moderate to large effects. Those participants suffering from a current or previous history of mental health problems showed significantly more improvements in distress than those without a current or a previous history of mental health problems. These results were replicated when using adjusted models and imputations, although ESs were reduced when using imputations (see supplementary materials Table S3).There were significant pre–post-intervention and pre-intervention to follow-up improvements in the mediators, including mindfulness, self-compassion, and resilience, with moderate effects (Table 4). Furthermore, the analyses revealed significant pre−post-intervention and pre-intervention to follow-up improvements in the secondary outcomes, including the domains of likelihood, skills and resources, and commitment towards achieving one’s academic goals, with moderate to low effects. Intrinsic and extrinsic motivation did not show significant changes at any time point (Table 5).The number of mindfulness sessions attended during the programme was not significantly related to improvements between time points across all measures. However, the amount of home practice during the programme (calculated as the multiplication of the number of practice days and the estimated mean time practiced each day) was significantly related to pre‒post-intervention improvements in wellbeing (R = 0.28; p = 0.017), mindfulness (R = 0.23; p = 0.045) and self-compassion (R = 0.27; p = 0.021). Although the amount of home practice during the programme was not significantly related to pre‒post-intervention improvements in distress (R = −0.04; p = 0.716) and resilience (R = 0.12; p = 0.327), intermediate effects were observed in a sub-group of students, who reported current mental health problems (n = 16; distress: R = −0.36, p = 0.169; resilience: R = 0.33, p = 0.207).Firstly, we computed bivariate correlations between pre‒post-intervention differences in mindfulness, self-compassion, and resilience and pre-intervention to follow-up differences in wellbeing, and distress (see supplementary materials Table S4). The shared variance between pre‒post-intervention changes in mindfulness and self-compassion was 36%; between mindfulness and resilience, the shared variance was 30%; between self-compassion and resilience, it was 19%, and it was 29% between pre-intervention to follow-up changes in wellbeing and distress. The path analysis results are presented in Table 6, Table 7 and Table 8 and illustrated in Figure 2.We observed that participants reported significant improvements in mindfulness at post-intervention (a = 3.80; p < 0.001) and that these improvements predicted changes in wellbeing (b = 0.29; p < 0.001) and distress (b = −0.42; p < 0.001) at follow-up (see Table 6). The 95% bias-corrected bootstrap confidence intervals for the interaction effects on wellbeing (0.38−2.05) and distress (−2.65 to −0.69) did not cross zero, indicating a possible mediating effect of mindfulness on both wellbeing and distress. This mediating effect explained 56% of total effects in wellbeing and 75% of total effects in distress, with small to medium ESs (Table 6). Furthermore, participants reported significant improvements in self-compassion at post-intervention (a = 0.38; p < 0.001), and these changes predicted improvements in wellbeing (b = 2.57; p = 0.005) at follow up (see Table 7). The 95% bias-corrected bootstrap confidence intervals for the interaction effects on wellbeing (0.22 to 1.98) did not cross zero, indicating a possible mediating effect of self-compassion on wellbeing, explaining a 51% of total effects, with a small ES. The possible mediating effect of self-compassion on distress was not significant (Table 7). Finally, participants reported significant improvements in resilience following the intervention (a = 2.07; p = 0.004), and these improvements predicted changes in both wellbeing (b = 0.25; p = 0.013) and distress (b = −0.33; p = 0.033) at follow−up. The 95% bias-corrected bootstrap confidence interval for the interaction effects on wellbeing (0.12–1.10) and distress (−1.59 to −0.07) did not cross zero, indicating a possible mediating effect of resilience on wellbeing and distress, explaining a 27% and 33% of total effects respectively, with small ESs (Table 8).This study explored the acceptability and effectiveness of the ‘Mindfulness: Finding Peace in a Frantic World’ course [21] in a university student population. Exploratory investigations were also conducted into changes in attitudes and motivation towards a self-selected academic goal. Preliminary explorations were conducted into the mediating role of mindfulness, self-compassion, and resilience for wellbeing and mental health (self-reported distress) independently. Results suggest that university students engaged with the M-FP course and reported high rates of acceptability. In addition, there were significant improvements in wellbeing and decreased mental health difficulties following the intervention and at 6-week follow-up, despite just under half of the sample attending all eight sessions. Improvements in wellbeing were significantly mediated by mindfulness, self-compassion, and resilience, whilst reductions in mental health problems were mediated by improvements in mindfulness and resilience, but not self-compassion. Further exploration revealed significant improvement in perceived “commitment” to, “likelihood” of achieving, and feeling more equipped with the “skills and resources” required to accomplish a self-selected academic goal at post-intervention and at 6-week follow-up. No improvements were revealed for intrinsic or extrinsic motivation towards academic goals. The M-FP course had a moderate effect in improving wellbeing and decreasing mental health problems at post-intervention, with moderate-low effects at 6-week follow-up, which reflects similar findings from recent meta-analyses and systematic reviews exploring university student samples [19,69]. Consistent with previous research, improvements in wellbeing were associated with more time engaging in home practices [70]. Engagement, by M-FP course attendance, was higher than alternative MBPs offered to university students [23], and slightly lower in comparison to a secondary teacher sample [20]. Engagement by home practice was lower than the recommended practice duration within the 8-week M-FP course (20 min per day) [21], but it was within the previously reported total range of home practice for MBPs in university student samples [23,24]. Thus, universities might be implicated in improving student well-being by enhancing mindfulness home practice (e.g., making available tools that facilitate practice such as online platforms with tutorials, audios, videos, etc.). Participants perceived the course to be beneficial and useful, and many had intentions to continue using mindfulness in the future. Ethnicity was a significant predictor of completion of the programme, and those who categorised themselves as being of white ethnicity attended significantly more sessions than other ethnic groups, although we recognise this is a small sample (Table 1). This raises an important perspective on the inclusivity and accessibility of mindfulness interventions for individuals from minority ethnic backgrounds. Similar studies in the area [23,24] have not reported the relationship between ethnicity and attendance but have similar percentages of ethnic diversity in their participant groups. In a systematic review of Mindfulness and Meditation-Based Interventions (MMBI), only 24 out of 12,265 studies were identified as ‘diversity-focused’ [71]. Research efforts, therefore, need to be made to ensure such courses are accessible, acceptable and effective to people across a range of ethnic backgrounds at the stage of both design and delivery.The present student sample reported moderate (26%) rates of current mental health problems in line with recent student population estimates [4,10,72]. Commonly, mindfulness practitioners advise against participation in MBPs if a potential participant is experiencing severe psychological distress, current depression, mania, or recent bereavement. In the present study, participants who reported experiencing current or past mental health problems reported significantly greater reductions in distress compared to those without current or past mental health problems. This reflects similar evidence from Galante et al. [23], which found that baseline levels of distress moderated the benefit obtained from an MBP in a university student population. These findings could have important implications for the implementation and benefits of MBPs with vulnerable students and should be explored in further research.There is preliminary evidence to suggest that MBPs implemented in a student population can impact academic outcomes and behaviour [23,25]. Whilst academic outcomes were not measured in the present study, we found an effect of time on students’ orientation towards their most important academic goal, suggesting the M-FP may have facilitated a more positive orientation towards academic goals. Intrinsic motivation is widely considered an important determinant for academic success [27] and is associated with trait mindfulness [73]. However, the present study did not find any significant changes in intrinsic and extrinsic motivation towards obtaining an academic goal following participation in the M-FP course. This finding is likely the result of ceiling effects, as much of the study sample reported very high levels of intrinsic motivation and very low levels of extrinsic motivation at baseline. This may arguably be characteristic of the present sample, comprising of students who have met the high entry requirements for studying at Oxford University [74]. Future research would benefit from a more diverse sample representative of a broad university student population. However, this result could also be derived from the instrument that was used, which only included one single item to measure each academic goal domain. Nevertheless, these preliminary findings contribute to our understanding of the effectiveness of the M-FP course on improving students’ orientation towards their academic goals and may form the basis of future investigations.In line with systematic reviews and meta-analyses exploring mechanisms of change in MBPs [31,32], mindfulness, resilience, and self-compassion were found to be significant mediators for pre-post-intervention changes in wellbeing. Furthermore, whilst improvements in distress were mediated by resilience and mindfulness, self-compassion was not found to significantly mediate improvements in distress in the present sample. This finding is consistent with previous findings from a study with a sample of secondary school teachers that found mindfulness, but not self-compassion, mediated changes between the frequency of mindfulness practice during an MBP and mental health outcomes [20], suggesting that improvement in mental health symptoms and wellbeing may be driven by different pathways of change. Recent cross-sectional evidence from general population participants also found significant direct effects of mindfulness, self-compassion, and resilience on anxiety and depression symptoms, with indirect effects of mindfulness and self-compassion through resilience on depression symptoms [42]. Thus, in order to optimise outcomes and delivery of MBPs, the disparity between mechanisms underpinning improvements in wellbeing and mental health symptoms, including distress, should be investigated further.The findings are interpreted in the light of several limitations. Given the exploratory nature of our study, we did not control for multiple comparisons, possibly leading to false-positive findings. The study is characterised by a small sample size and large rates of attrition. Although statistical techniques were employed to address these characteristics, it may not have sufficient power to fully investigate all the questions proposed with the potential to lead to false-negative findings. Therefore, it is important that these findings are considered preliminary and that future research aims to formally test these observations in larger university student samples. The absence of a control condition means that it is unclear whether improvements in wellbeing, distress, mediators and the students’ orientation towards their academic goals can be attributed to their participation in the M-FP course or the passage of time. Students’ orientation towards their academic goals is a new line of enquiry. Thus, it is currently unknown how academic goal orientation (i.e., commitment to, likelihood of achieving, and having the skills and resources to achieve the academic goal) may fluctuate during the academic year. The present study did not explore the mechanisms through which the observed changes in academic goal orientation were facilitated to avoid over-testing. Future research may wish to explore these longitudinal changes by combining self-report measures of academic goal orientation with objective measures of academic achievement and study behaviour. As mentioned, the academic goal findings may also be derived from the instrument that was used, which only included one single item to measure each academic goal domain, and future research should improve the quality of the measures used in this regard. In addition, whilst the M-FP course was found to be particularly effective for students with mental health problems, these findings can also be explained by the observation of regression to the mean, as students reported high levels of mental health problems at study entry [75]. Finally, opportunity sampling was employed, whereby students, who were already considering partaking in the M-FP course, were recruited. Hence, participants may present with characteristics (e.g., financial resources to afford course participation), which differ from the general student population at large, limiting the generalisability of the present findings. In this sense, we observed that ethnicity predicted completion of the programme, and that raises questions of inclusivity. For example, marginalized identities (rather racial, gender, gender identity, sexual orientation, migrant, etc.) may play a role in the outcomes and would require special consideration, as findings may be differential based on students’ identity.Whilst the findings are exploratory in nature and must be considered preliminary, they highlight possible avenues for future investigations. Considering the limitations, the M-FP course was found to be acceptable and effective in the university student population. Students who undertook the course showed improvements in wellbeing and distress over the study period. Whilst this may also be the result of regression to the mean, mindfulness, self-compassion, and resilience were found to mediate changes in wellbeing, while changes in distress were mediated by mindfulness and resilience. Such mediation effects would be expected of active intervention. Further, to our knowledge, this study is the first to suggest that participation in the M-FP course may improve students’ orientation towards their academic goals (i.e., the perceived likelihood, commitment, and skills and resources of achieving their goal). Given the exploratory nature of this study, future research should aim to formally test these observations in larger student samples, using randomised controlled trial designs and combining more robust self-report measures of academic goal orientation with objective measures of academic achievement, and study behaviour. In light of the corresponding relationship between ethnicity and completion of the intervention, it is pertinent that barriers to attendance and engagement for students from ethnic minority backgrounds are explored. To optimise outcomes of the M-FP course for university students, the disparity between mechanisms underpinning improvements in wellbeing and distress should be investigated further. The results merit future investigation whilst having implications for public health and calls for universities to further support and promote positive mental health in their students.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116023/s1, Table S1: Baseline variables predicting completion at post-test and follow-up. Table S2: Adjusted and imputed analysis of primary outcomes. Table S3: Adjusted and imputed analysis of the interaction effect between the time and mental health problems (current or previous) on distress (CORE-10) scores. Table S4: Correlations between the pre‒post-treatment changes in the proposed mediators and pre-intervention to follow-up changes in the main outcomes (n = 60).Conceptualization, C.C., A.T. and W.K.; methodology, C.C., J.M.-M., A.T. and W.K.; software, J.M.-M.; validation, E.M., A.P., V.H. and A.T.; formal analysis, J.M.-M.; investigation, A.T.; resources, A.T. and W.K.; data curation, E.M., A.P., V.H. and A.T.; writing—original draft preparation, E.M., A.P. and J.M.-M.; writing—review and editing, E.M., A.P., C.C., V.H., L.T., A.T., J.M.-M. and W.K.; visualization, A.P.; supervision, C.C. and W.K.; project administration, W.K.; funding acquisition, W.K. All authors have read and agreed to the published version of the manuscript.This research was funded in whole, or in part, by the Wellcome Trust WT104908/Z/14/Z and WT107496/Z/15/Z and supported by the NIHR Oxford Health Biomedical Research Centre. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission. The Wellcome Trust and NIHR Oxford Health Biomedical Research Centre had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the University of Oxford Research Ethics Committee (protocol reference code: R52786/RE004; date of approval: 16 October 2018).Written Informed consent was obtained from all subjects involved in the study. Participants could withdraw from the study at any time after they had filled out the questionnaires on request.Following the International Committee of Medical Journal Editors (ICMJE), all of the individual anonymized and completely de-identified participant data are available for any analytical purpose that is related to achieve aims in the present study upon reasonable request to researchers (a) who provide a methodologically sound proposal and (b) whose proposed use of the data has been approved by an independent ethical review committee. The data and codebook will be provided by the corresponding author (willem.kuyken@psych.ox.ac.uk) to interested researchers that meet the aforementioned criteria.We would like to thank all of the participating University students for giving their time so generously to participate in this project. Special thanks to the Oxford Mindfulness Centre for their support of this project.Willem Kuyken is the Director of the Oxford Mindfulness Centre. Catherine Crane, Verena Hinze, Laura Taylor, Jesus Montero-Marin, Alice Philips, Emma Medlicott and Alice Tickell were affiliated with the Oxford Mindfulness Centre. The views expressed in this publication are those of the author(s) and not necessarily those of the National Institute for Health Research or the Department of Health and Social Care.The proposed model of change following participation in a MBP for university students.The independent variable is the repeated-measures factor (e.g., time: X). M is the pre‒post difference in the corresponding mechanistic variable. The dependent variable is the pre-follow-up difference in the corresponding main outcome (Y). “a*b” = indirect effect through the mediator. “c’” = direct effect after adjusting for the mediating effects.Flowchart of participants.Baseline characteristics of the study sample.a e.g., PGCE. b Prefer not to say: n = 1; c Prefer not to say: n = 2; d Missing data: n = 11; e e.g., Yoga.Complete cases analysis of primary outcomes.WEMWBS: Warwick Edinburgh Mental Wellbeing scale; CORE-10: Clinical Outcomes Routine Evaluation-10. d: Cohen’s d effect size from adjusted means. B: unstandardised regression coefficient using mixed models with subjects as random effects. 95% CI: 95% confidence interval.The interaction effect between the time and mental health problems (current or previous diagnoses) on distress (CORE-10) scores using complete cases.Current case: currently experiencing mental health problems. Previous case: previous diagnoses of mental health problems. d: Cohen’s d effect size from adjusted means. B: unstandardised regression coefficient using mixed models with subjects as random effects. 95% CI: 95% confidence interval.Complete cases analysis on the proposed mediators.FFMQ-SF: Five-Facet Mindfulness Questionnaire-Short Form. SCS-SF: Self-Compassion Scale-Short Form. CDRISC: Connor–Davidson Resilience Scale. d: Cohen’s d effect size from adjusted means. B: unstandardised regression coefficient using mixed models with subjects as random effects. 95% CI: 95% confidence interval.Complete cases analysis of student’s orientation towards their academic goals.Analyses of secondary outcomes. d: Cohen’s d effect size from adjusted means. B: unstandardised regression coefficient using mixed models with subjects as random effects. 95% CI: 95% confidence interval.The mediating role of mindfulness on main outcomes.WEMWBS: Warwick Edinburgh Mental Wellbeing scale. CORE−10: Clinical Outcomes Routine Evaluation-10. R2: multiple determination coefficient as an effect size measure. Coef: unstandardised regression coefficient. Boot: bootstrapped unstandardised regression coefficient. SE: standard error. 95% CI († 95% confidence interval; ‡ 95% bias-corrected bootstrap confidence interval for the indirect effect using 10,000 samples). “a*b” = indirect effects through mindfulness (see Figure 2). Path “c” refers to the unadjusted direct effects of X on Y. *** p < 0.001. * p < 0.05.The mediating role of self-compassion on main outcomes.WEMWBS: Warwick Edinburgh Mental Wellbeing scale. CORE-10: Clinical Outcomes Routine Evaluation-10. R2: multiple determination coefficient as an effect size measure. Coef: unstandardised regression coefficient. Boot: bootstrapped unstandardised regression coefficient. SE: standard error. 95% CI († 95% confidence interval; ‡ 95% bias-corrected bootstrap confidence interval for the indirect effect using 10,000 samples). “a*b” = indirect effects through mindfulness (see Figure 2). Path “c” refers to the unadjusted direct effects of X on Y. *** p < 0.001. ** p < 0.01. * p < 0.05.The mediating role of resilience on main outcomes.WEMWBS: Warwick Edinburgh Mental Wellbeing scale. CORE-10: Clinical Outcomes Routine Evaluation-10. R2: multiple determination coefficient as an effect size measure. Coef: unstandardised regression coefficient. Boot: bootstrapped unstandardised regression coefficient. SE: standard error. 95% CI († 95% confidence interval; ‡ 95% bias-corrected bootstrap confidence interval for the indirect effect using 10,000 samples). “a*b” = indirect effects through mindfulness (see Figure 2). Path “c” refers to the unadjusted direct effects of X on Y. *** p < 0.001. ** p < 0.01. * p < 0.05.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06024.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Counselling helplines or hotlines are key support services for young people with mental health concerns or in suicide and self-harm crises. We aimed to describe young peoples’ use of a national youth helpline (Kids Helpline, Australia, KHL) to understand how usage changed over time. A descriptive analysis was conducted on 1,415,228 answered contacts between 2012–2018. We described the trend of service usage over the observed period, the types of youth who used the service, and the problems young people contacted the service about. Phone (APC = −9.1, KHL: −10.4 to −7.8, p < 0.001) and email (APC = −13.7, 95%CI: −17.1 to −10.2, p < 0.001) contacts decreased over time whereas webchat contacts increased (APC = 16.7, 95%CI: 11.7 to 22.0, p < 0.001). With this increase in webchat contacts, there was an associated increase in total webchat contact duration. Concerns raised in contacts to the service were primarily related to emotional wellbeing and mental health concerns (53.2% phone, 57.3% webchat, 58.2% email) followed by social relationship issues (20.4% phone, 20.3% webchat, 16.8% email) and family relationships (19.4% phone, 17.2% webchat, 21.8% email). The increased preference for online text-based information and counselling services can help inform development of services for young people and allocation of staff/service training and resources. Many mental health conditions first present at a young age (e.g., anxiety) [1]. A recent Australian study reported that nearly one in seven (13.9%) young people aged 4–17 years were assessed as having a mental health disorder in the previous 12 months [2]. Moreover, suicide mortality increases through the teenage years, with suicide remaining a leading cause of death in children and young adults [3,4,5]. Despite this, help-seeking and treatment utilization is often low [6,7]. According to a recent systematic review of young peoples’ perceived barriers to help-seeking for emotional well-being and mental health concerns, barriers were related to individual factors (e.g., mental health and services literacy), social factors (e.g., perceived stigma), relationship factors (e.g., fears surrounding confidentiality), and systemic/structural factors (e.g., access/time/transport and costs) [6].Helplines can be a critical first point of contact in the prevention and care for young people with psychosocial concerns and those at risk of suicide. They provide confidential information and emotional support for people with various psychosocial concerns (e.g., emotional wellbeing, violence and abuse, social relationships [8]) and crisis intervention for individuals experiencing suicidal crisis [9]. By providing support through telephone, email, SMS, or online chat [10], helplines offer multiple advantages for users compared to face-to-face services, potentially addressing several barriers to help-seeking. For example, the confidentiality/anonymity of helplines may overcome perceived stigma, and helplines are also highly accessible and cost-effective. Furthermore, the introduction of online and text-based helpline modalities may also enhance help-seeking in young people [11,12].Regarding the effectiveness of youth helplines around the world, a recent systematic review described the state of the literature and, despite noting a lack of empirical studies, suggested that youth helplines may effectively reduce immediate distress [13]. In addition, Mathieu et al. [13] concluded that relatively limited previous research has explored the reasons young people contact youth helplines, described up-to-date demographic features of those who use the services, or explored the use and effectiveness of different service contact media (e.g., phone vs. webchat). Identifying who, how, and why youth use helplines would help to tailor services to the needs and preferences of young people and contribute to improving their experience with and outcomes from engaging with helplines. The current study aimed to analyse changes in help-seeking behaviour through the Kids Helpline (KHL), a nationwide youth service in Australia from 2012–2018. We explored service utilisation, including the medium of contact, the reasons children, and young people (5–25 years) contacted the service and how use of contact media changed over time. Specifically, our research questions were: (1)How many contacts did the service respond to through different contact media and by contact type (information and referral, counselling) and how has this changed over time?(2)What are the differences by gender and age across media and by contact type?(3)What is the duration of the different contact types and has it changed over time?(4)What are the concerns of Australian children, do they differ across media, and did this change over time?How many contacts did the service respond to through different contact media and by contact type (information and referral, counselling) and how has this changed over time?What are the differences by gender and age across media and by contact type?What is the duration of the different contact types and has it changed over time?What are the concerns of Australian children, do they differ across media, and did this change over time?Established in 1991, KHL provides young people aged 5–25 years with free 24/7 information and counselling accessible through telephone, webchat, and email. The KHL service model uses a stepped-intervention framework, providing a range of support options from ‘universal care’, in which young people are provided with information or short-term support for a simple problem, to ‘complex care’ for those with acute, chronic, and often multiple presenting issues such as complex mental ill health or risk of harm to self or others. Support is calibrated to respond to the needs of each individual and includes provision of information, referral to other services, psychoeducation, counselling, case planning, risk of harm assessment, and for those with the most complex needs, wrap-around care with other services. Counsellors are degree-qualified in psychology, social work, counselling, human services or similar, with at least one year of experience when recruited. Comprehensive in–house training provides additional understanding of specific issues and evidence–informed responses relevant to KHL. For the data collection period, all counsellors received one full day of training exclusively in the coding system and recording of data, this included approximately three hours on coding the ‘problem’ or concern raised during the KHL contact whereby counsellors discussed case examples in small groups before large group consensus.The de-identified contact data analysed in this study were recorded by KHL counsellors between January 1, 2012 and December 31, 2018 as part of routine service delivery. The justification for the 2012 start date is that the coding system was revised at this time and there were differences in the data collected prior. More recent data were also not included in the current analyses, as the unprecedented impact of the global novel coronavirus disease (COVID-19) pandemic warranted a more granular analysis of weekly and monthly changes in demand (as opposed to annual), which have been reported elsewhere [14]. Furthermore, there were increases in web-chat service delivery from late 2018 onwards that may have unduly impacted the results. For each contact, counsellors record the client’s demographic details (e.g., age, identified gender, living arrangement) and other information about the contact (e.g., concerns raised by the young person) in an online platform. For reasons of confidentiality, counsellors do not specifically request demographic details for the purpose of data collection. Age is generally disclosed by the young person as part of a counselling conversation or inferred through other indicators (e.g., school grade). Gender (male, female, transgender, or gender diverse) is either inferred based on the interaction or asked at an appropriate time. Other details (e.g., living arrangement) are only recorded if the information arises in conversation. Consequently, age and gender are more likely to have been recorded for repeat than single callers. For the current study, young people have been separated into age groups, with ‘children’ aged 5–12 years, ‘teens’ aged 13–17 years, and ‘young adults’ aged 18–25 years.Contact type is distinguished as either ‘counselling’ (i.e., contact focused on a specific concern), ‘information and referral’ (e.g., a request for information or contact details of a specific service), or ‘other ways of engaging’ (i.e., non-conversational, roleplaying a problem, and nonsensical, aggressive, or sexual ‘prank’ contacts). Counsellors record up to four concerns for each counselling contact using the KHL Concern Classification System [15], which is a list of 49 concerns (e.g., mental health concerns, bullying, suicidal thoughts, sexual abuse) grouped into 11 broader concern groups (e.g., family relationships, identity and self-concept, violence and abuse). Concerns reflect reasons for contact, that is, the issues the young person wished to speak about, not the counsellor’s perception of the young person’s most important problems.The study was approved by the Griffith University’s Human Research Ethics Committee (reference number: 2020/505).Dependent variables included the total number of contacts, contacts per medium (phone, email, webchat), concern type, and the number of contacts across gender (male, female, transgender or gender diverse) and age group (children, teens, young adults). The duration of webchat and phone contacts (seconds, minutes) was also analysed, but duration of email contacts was not, as email exchanges may occur over multiple days. Time trend analyses of dependent variables were analysed from 2012 to 2018.Time trends were analysed using Joinpoint regression, which identifies time points where a statistically significant change in trend occurs. Joinpoint provides an estimate of the average annual percentage change (APC), with 95% confidence intervals (95% CI). Chi-square analyses were used to compare categorical data. IBM SPSS Statistics for Windows version 26 (IBM Corp, Armonk, NY, USA) and Joinpoint Regression Program version 4.8.0.1 (National Cancer Institute, Bethesda, MD, USA) were used to analyse the data.Between 1 January 2012 and 31 December 2018, 1,415,228 contacts to the KHL service were recorded. For further analyses some data were excluded. First, ‘Other’ contacts (35.9%, n = 507,648) were excluded as these did not inform the aims of the study. Second, contacts made by people outside the target age range (i.e., 5–25 years) were excluded (2.1%, n = 19,004), as were contacts where age could not be determined conclusively (20.6%, n = 182,424). Third, phone counselling contacts lasting less than 60 seconds were excluded (0.2%, n = 1462) following discussion with KHL staff, which indicated these do not represent meaningful counselling conversations. ‘Information and referral’ contacts of all lengths were included as the study team deemed these contacts were relevant, regardless of duration. The following analysis includes 704,690 contacts. KHL allows young people to remain anonymous; nevertheless, many of those who contact multiple times identify themselves with a name or pseudonym. Based on these names, and matching of phone numbers, IP addresses and email addresses, approximately 35% (n = 24,745) of contacts were from a young person who made a single contact and approximately 65% (n = 457,445) were from a person who made two or more contacts. The total number of contacts decreased over the study period, with the APC showing a decline of 6.5% per year on average (95%CI: −8.3, −14.8, p = 0.005) between 2012 and 2016, followed by a non-significant decline from 2016–2018 (Figure 1). This primarily reflected a reduction in information seeking rather than counselling contact types.Most contacts (66.4%, n = 467,626) were by phone, with 19.9% (n = 139,999) by webchat and 13.7% (n = 97,065) by email. Trends of contacts by medium and contact type from 2012–2018 are provided in Figure 1 and Joinpoint trend analyses in Table 1. In terms of contact medium, the number of phone (APC = −9.1%, 95%CI: −10.4, −7.8, p < 0.001) and email contacts decreased (APC = −13.7%, 95%CI: −17.1, −10.2, p < 0.001), whereas webchat contacts showed a rapid increase (APC = 16.7%, 95%CI: 11.7, 22.0, p < 0.001) over the study period.Each contact’s age, reported gender, and contact type for phone, webchat, and email are presented in Table 2. Contacts by phone were primarily from young adults (53.4%), whereas contacts by webchat and email were primarily from teens (63.8% and 67.4%, respectively). Most phone contacts from young adults were for information and referral purposes, whereas phone contacts from teens and children were primarily for counselling. Across all media, most contacts were from those whose gender was recorded as female. Counselling contacts were the primary contact type across all media (63.8% of all contacts). However, the proportion of contacts that were counselling related (versus information and referral) was greater for webchat (83.3%) compared to the proportion of counselling contacts made by either phone (55.7%) or email (70%).A breakdown of information and referral contacts across media, age and gender are provided in Table 2. A total of 254,876 (36.2%) contacts were requesting information and/or referral. Information and referral contacts decreased over time (APC = −12.6, 95%CI: −14.7, −10.5, p = 0.001). Most were made by phone (81.3%, n = 207,113), followed by email (9.5%, n = 24,322), and webchat (9.2%, n = 23,441). Information and referral contacts made by phone (APC = −15.6, 95%CI: −17.6, −13.4, p <0.001) and email (APC = −12.7, 95%CI: −17.3, −7.8, p = 0.001) decreased between 2012 and 2018, whereas webchat contacts increased for this contact type (APC = 19.1, 95%CI: 8.4, 30.9, p = 0.005; see Figure 1 and Table 1). Information and referral contacts were more likely to be from young adults (56.4%, n = 143,677) than teens (34.6%, n = 88,226) or children (9.0%, n = 22,973) (Table 2). The age breakdown varied for each contact medium. Relatively more children made information and referral contacts by email, whereas relatively more teens made contact by webchat and email, and relatively more young adults made contact by phone. Most information and referral contacts were with girls (72.9%, n = 183,690) and fewer were with boys (25.5%, n = 64,238) or transgender or gender diverse young people (1.6%, n = 4044). The trend of gender breakdown was consistent across all media; however, relatively fewer contacts were made by boys through webchat and email compared girls and transgender or gender diverse young peopleThe median duration of phone information and referral contacts across all years was just over two minutes (Table S1). The change in median durations between 2012 and 2018 was less than one minute (2012 = 1.92 minutes, 2018 = 2.72 minutes), which reflected a non-significant increase in duration (APC = 5.21, 95%CI: −0.1, 10.8, p = 0.052). As a measure of load on KHL services, the total duration of information and referral contact time per year by phone decreased significantly from 346,593 minutes in 2012 to 107,674 minutes in 2018 (APC = −18.48, 95%CI: −24.5, −12.0, p = 0.001). Regarding contact medium, the median duration of information and referral contacts made by webchat across all years was 8 minutes (Table S2), demonstrating a significant decline from 11.5 minutes in 2012 to 6 minutes in 2018 (Table S2; APC = −9.21, 95%CI: −11.7, −6.7, p < 0.001). However, regarding load on KHL services, there was a non-significant increase in total webchat information and referral contact time per year from 28,627 minutes in 2012 to 52,871 minutes in 2018 (APC = 5.99, 95%CI: −2.3, 15.0, p = 0.125). While non-significant, this near double in total load reflects the increase in number of webchat contacts made over time (Table S2).A breakdown of counselling contacts across age, gender, and contact medium are provided in Table 2. There were 449,814 (63.83% of total contacts) counselling contacts made from 2012–2018 that were greater than 60 seconds in length. Counselling contacts decreased slightly over time (APC = −0.9, 95%CI: −1.7, −0.2, p = 0.024; Table 1). Most counselling contacts were by phone (57.9%, n = 260,513) followed by webchat (25.9%, n = 116,558) and email (16.2%, n = 72,743). Across all media, most counselling contacts were made by teens (52.1%, n = 234,559) followed by young adults (35.7%, n = 160,714), with the fewest by children (12.0%, n = 54,541). As shown in Table 2, this differed by medium. Fewer teens (41.2%) made counselling contacts by phone compared to young adults (46.1%), with a greater proportion of teens making counselling contact by webchat (65.2%) compared to young adults (25.6%). Contacts by children were evenly dispersed across media.A greater proportion of counselling contacts were made by girls (79.7%, n = 353,606) compared to those made by boys (18.9%, n = 84,044) and transgender or gender diverse people (1.4%, n = 6290). This breakdown of gender across counselling contacts was consistent across contact media. However, a greater proportion of email contacts were made by girls (87.1%) compared to other media, and a greater proportion of phone counselling contacts were made by boys (23.4%) compared to other media.Table 3 presents counselling contact concerns and concern groups across each medium. When counselling contacts across all media were examined, more than half were related to ‘emotional wellbeing and mental health concerns’, with suicide-related concerns and self-injury/self-harm being raised in 12% and 6% of contacts (respectively). Relationship matters were also common, with social issues (i.e., friend/peer, dating) and family relationships (i.e., child–parent, parenting, other family) being raised in 20.4% and 19.4% of contacts, respectively. As would be expected, comparisons across contact media showed that the phone was used more so to discuss most concern groups; however, when comparing the prevalence of different concern groups across media there were no notable differences (Table 3).Table S3 presents Joinpoint trends of concerns over time, which varied by medium. For phone contacts, all concerns except emotional abuse declined from 2012–2018, consistent with the overall trend of phone counselling contacts over time. For email contacts, a decrease was observed for all concerns other than ‘offending, abusive & violent actions’. In contrast, webchat counselling contacts increased over time for all concerns, with notable increases in contacts that discussed gender/sex identification, abusive and violent actions, and practical and material assistance.Frequencies of counselling contacts by concern groups across medium by gender and age-group are presented in Table S4. ‘Emotional wellbeing and mental health concerns’ were most common for all gender groups, followed by ‘social and family relationship-related’ problems. ‘Identity and self-concept-related’ concerns were present in more than a quarter (28.1%) of counselling contacts with youth who identified as transgender or gender diverse. This varied by contact medium, with 41.1% of email, 30.5% of webchat, and 23.9% of phone counselling contacts related to this concern area. Moreover, compared to contacts whose gender were recorded as male (48.1%) or female (56.8%), those who identified as transgender or gender diverse (63.4%) more frequently raised concerns about ‘emotional wellbeing and mental health concerns’.Across all age groups, ‘emotional wellbeing and mental health concerns’ (>39%), and family and social relationships (>12%) were the most raised concern groups, regardless of contact medium. However, children raised ‘violence and abuse (non–family)’ in 19.2% of phone contacts, whereas this was less frequent for teens (8.9%) and young adults (4.8%).The median duration of a phone counselling contact was 32 minutes (Table S1), which was stable between 2012 and 2018 (APC = −0.17, 95%CI: −1.5,1.2, p = 0.757). In contrast, the total minutes of phone counselling delivered per year decreased significantly from 1,496,924 minutes in 2012 to 1,137,593 minutes in 2018 (APC= −4.19, 95%CI: −5.4, −2.9, p < 0.001). Webchat counselling contacts ranged in duration from less than one minute to 221 minutes. The median contact duration was 56 minutes, which reduced from 63 minutes in 2012 to 53 minutes in 2018 (Table S2). The decreasing median contact duration was significant between 2012 and 2015 (APC = −4.73, 95%CI: −8.0, −1.3, p = 0.030), but non-significant thereafter (APC = −1.24, 95%CI: −4.7, 2.3, p = 0.027). However, given the increased number of webchat counselling contacts, the total minutes delivered per year increased from 552,598 minutes in 2012 to 1,218,537 minutes in 2018 (APC = 12.75, 95%CI:9.5, 16.1, p < 0.001). In combination, the increase in the number of webchat counselling contacts and the longer average duration of a webchat in comparison to a phone contact resulted in the total time spent on counselling contacts increasing from 2,049,521 minutes in 2012 to 15,397,949 minutes in 2018, indicating a more than seven-fold increase in counselling load on KHL services.Young people under the age of 25 are at heightened risk of developing mental health disorders [1,2,16], and responsive mental health services are critical. This study aimed to describe how and why young people use the Australian nationwide KHL over time. Findings may be used to inform future service improvements for the KHL and related helpline services in meeting the ongoing needs of young people.The current analysis of contacts made to KHL from 2012–2018 indicates significant changes in the way young people engaged with the service. Over these years, there was an increase in the use of webchat and a decrease in use of phone and email (although the phone continued to be the most used medium for KHL contacts). Despite the total number of phone and webchat contacts (combined) decreasing per year over time, the service load provided by KHL (through counselling and information and referral) remained constant, with a total of 2,424,741 minutes in 2012 and 2,516,676 minutes in 2018. This is explained firstly by a reduction in the number of information and referral contacts, which tend to be brief, but minimal change in the number of counselling contacts. The reduction in the number of contacts seeking information and referral from KHL may reflect young people’s increasing access to mental health and well-being information via the internet more broadly. For instance, the KHL website was redesigned with informational content added from late 2017 onwards. Secondly, while the number of counselling contacts was stable, the time spent responding to counselling contacts increased, largely reflecting the rise in use of webchat. Indeed, while the increased use of webchat was evident for all contact types, the annual number of counselling webchats more than doubled from 2012 to 2018. This increase, combined with the longer duration of a webchat counselling contact (median 56 minutes) compared to a phone counselling contact (median 32 minutes), led to a greater total load on KHL services from 2012 to 2018 despite a decrease in the number of phone counselling contacts and a decrease of 10 minutes in the median counselling webchat duration. Increased webchat use reflects recent findings that young people prefer text-based counselling services to phone or face-to-face services [17,18]. In an Australian survey of young people (aged 15–25 years) who had contacted the KHL webchat service, the preference for text-based counselling in comparison to face-to-face services was primarily due to a perceived increase in safety and control (i.e., privacy, reduced emotional intensity, increased control and autonomy [18]). Avoidance motivations (i.e., avoiding being overheard, avoiding verbal social interactions, and minimising difficult emotions), increased perceived accessibility (i.e., speed, convenience), and perceived counselling expectations (i.e., help with low-complexity issues, realistic expectations) were also identified [18]. Additionally, a US study reported that a combined 59% of youths preferred one of three text-based services (i.e., text-messaging 25%; online chat 18.7%; social media 15.3%) compared to a phone service (41%) [19]. Emphasizing and providing ease-of-access, timeliness, and confidentiality of services from the privacy of home (or other private place) is clearly important in attracting and engaging young people in helpline services. Furthermore, the use of peer based online forums or message boards as alternative options for mental health support have also been shown to be acceptable to young people [20,21,22,23].While there are clear reasons young people may prefer webchat services, the sector must be confident that this medium can offer similarly effective counselling. There is evidence to suggest the utility of text-based counselling. For example, after a single text-messaging counselling session from a Danish children’s helpline, 35.9% of young people reported feeling better and more than half reported having a plan of action [24]. Similarly, Navarro, Bambling et al. [18] reported young people perceived greater efficacy of text-based counselling with preferences for text-based communication (i.e., feeling heard/understood, catharsis, feeling normalised/validated/supported). Building upon these findings, Navarro and colleagues explored mental health professionals’ perspectives of factors related to higher and lower effectiveness of text-based counselling services [25]. They noted that increased complexity of presenting problems and subsequent assessment, slower response time (text vs. speaking), no non-verbal cues, and connectivity issues were factors believed to decrease service effectiveness. However, text-based communication, the counsellor’s interpersonal skills, and the use of self-management strategies were perceived as factors that would increase service effectiveness.In response to increased use of, and some indication of preference for, text-based services, it is important that counsellors can provide appropriate support and guidance to meet young people’s needs in webchat contacts. Sindahl et al. [24] suggest appropriate text-based support includes discussing emotions, expressing empathy, and encouraging the young person to speak to someone, while Navarro, Bambling et al. [18] reported participants value feeling heard, understood, and supported during text-based contact. Moreover, given there is some indication that help-seeking by young people is low (e.g., [6]), appealing to preferred aspects of text-based counselling (e.g., safety, accessibility) [18], may increase the likelihood that young people engage with mental health services, including helplines. For instance, there may be utility in ensuring there is sufficient information and resources available online to improve awareness of helplines and the multiple avenues for contact [19]. Doing so may help increase awareness, reduce stigma, and ensure services can meet the increasing demand of young people using online services. It is also important to consider the financial impact of text-based services. Given the longer duration of a webchat counselling contact (as compared to telephone), increasing demand for webchat will increase costs and place financial strain on helpline services, unless efficiencies can be achieved. The 10-minute reduction in median counselling webchat duration during the study period may indicate that KHL became more efficient over time as counsellors become increasingly familiar and skilful in the provision of text-based services, but further investigation is needed to support this conclusion. A vast majority of counselling contacts were made by girls, and teens or young adults rather than children or boys and gender diverse youth. These findings align with a similar study from the US showing the greatest use of a youth helpline service was by young girls aged 15–16 years, albeit that service provided peer support rather than the professional counselling provided by KHL [26]. Recent research has specifically explored barriers to help-seeking in young men and found that peers and traditional ‘masculine’ ideals were key to low mental health service utilization and tailored mental health advertising and using specific language (e.g., mental fitness) may be ways to overcome these [27], which could be applied to the helpline field. Given that many mental health disorders manifest during adolescence and young adulthood [16], and the increasing calls for greater early intervention for mental health problems, services such as KHL must be able to respond effectively to the varying severity of needs and cognitive developmental stages of children, teens, and young adults. Consequently, it is important to understand the differences in the extent of service use by different genders and age groups. Whether these differences reflect greater rates of mental health and related concerns in different demographics or differences in the likelihood of help–seeking/service utilisation requires further investigation. Nevertheless, ensuring helpline services are accessible, appealing, and relevant for any young person (regardless of age or gender) is critical. In terms of young people’s concerns, more than half the counselling contacts across all media related to emotional wellbeing and mental health concerns, followed by social and family relationship concerns. This is comparable with previous descriptive analyses of youth helpline services in the USA (e.g., [26]), where anxiety, stress, sadness, and depression were the most reported concerns by young people. However, while such concerns are commonly reported, we must also be mindful of the contrast in reported concerns across gender and age to adequately address those who contact the helpline. For example, in comparison to contacts from girls, contacts from boys were far less likely to relate to emotional wellbeing concerns. Contacts by youth identifying as transgender or gender diverse were more likely to discuss identity and self-concept than contacts from those identifying as male or female. Contacts about violence and abuse (non–family) were more likely to be from children than either teens or young adults. While at face value these differences may be expected, a better understanding is needed regarding why these differences exist, so that online and phone services can be targeted more effectively to specific groups of young people.Nearly a fifth of counselling contacts included either suicide or self-injury/self-harm concerns. In numbers, this reflects 89,173 contacts between 2012 and 2018 (across all media) which equates to (on average) 34 suicide or self-injury/self-harm related contacts per day. This proportion of contacts is comparable to those reported by Kerner et al. [26] for a hotline in the USA between 2010 and 2016, with 14.2% reporting suicidal ideation and 8.2% reporting self-harm. This fits the conceptualization of helplines as a potentially useful crisis service [13]; however, future research is needed to increase understanding of the patterns of help-seeking and type of assistance preferred/required by young people at risk of suicide and/or self-harm.Given the observed increase in webchat use, counsellors may need additional, specific training to maintain effective service provision through this medium, for all concern types. Further, with most counselling contacts considered complex (e.g., mental health, suicide), ongoing skill development of counsellors tailored to these areas may be useful. Similarly, tailoring the service to those using the service will be important (e.g., young women, teenagers), as well as adjusting strategies for less frequent users (e.g., young men) to ensure they are aware of, can access, and are satisfied with the services KHL provides. Interestingly, while the number of information and referral-related contacts have decreased over time, most continue to be made through the phone even though there is an enormous amount of information available online (the KHL website itself being updated toward the end of the study period). Exploring ways to continue to meet the needs of those who prefer the phone medium to source their information may be required. Lastly, the increase in webchat and total duration of counselling contacts places an increased the load on the service and this may result in financial implications for helplines in accessing and maintaining adequate funding to support this. The current analysis is descriptive, which limits the conclusions that can be drawn. Nevertheless, the large nationwide database and analysis of trends over seven years of KHL usage is informative and detailed. Limitations in the data itself also affect the interpretation of findings. Importantly, analysis is of contacts not individuals and the dataset includes both first-time users of the service and young people who have contacted multiple times. This may affect obtained results around problem types and demographic information, particularly for those categories with a relatively smaller sample size. There was also substantial missing data for age and gender of clients in some analyses, and while reflective of the anonymous nature of a helpline service, needs to be considered in drawing conclusions. Interpretation of observed changes in the data may be enhanced with reference to population data to determine rates of service usage in this demographic. It is also important to note that we analyse the response data (i.e., answered contacts) as the demand for the KHL service consistently outstrips capacity to respond (response rate in 2018 was 52%) and our analysis was limited to answered contacts (response), which may have different characteristics to contact attempts (demand). Future research could also examine specific barriers and facilitators to young men accessing helplines, investigate the trajectories and needs of repeat users, and explore the effectiveness of different helpline media (webchat vs. phone). A further important consideration for future research in this field is the use of consistent terminology to differentiate services such as helplines, crisis lines, hotlines, and various online therapies such as e-therapy, web-based CBT, and mHealth, which may or may not be part of a helpline service.The results revealed that from 2012–2018 there was a change in the way young people contacted KHL and the types of help being sought. There was a reduction in information seeking contacts, but overall little change in the number of contacts seeking counselling. There was a substantial increase in webchat contacts with a concomitant decrease in contacts by telephone. These findings have important implications for helpline service providers by providing indications for resource investment and training, particularly in relation to increases in online based help-seeking.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116024/s1, Table S1: Descriptive statistics of the duration (in minutes) of phone counselling (n = 260–513) and information and referral (n = 207,113) contacts; Table S2: Descriptive statistics of the duration (in minutes) of webchat counselling (n = 118,332) and information and referral (n = 24,488) contacts; Table S3: Counselling concern trends across each medium in 2012–2018 (total n = 449,814); Table S4: Number and percentage of contacts reporting at least one problem in each problem group across gender and age groups (n = 449,814).Conceptualization: all authors; methodology: K.K., S.B., V.R., S.H.S., and D.W.; formal analysis: D.W.; resources: S.B. and B.C.; data curation: D.W.; writing—original draft preparation: D.W.; writing—review and editing: all authors; visualization: D.W., S.M., and K.K.; supervision: K.K.; project administration: K.K.; funding acquisition: K.K., S.B., V.R., and S.H.S. All authors have read and agreed to the published version of the manuscript.This research was supported by the Griffith Health–Building Partnerships Grant. The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Griffith University (reference number: 2020/505).Patient consent was waived due to the use of anonymous archival administrative clinical records collected as part of routine practice for Kids Helpline Australia.Restrictions apply to the availability of these data. Data was obtained from yourtown (Kids Helpline) and are available to Dr Kairi Kõlves only with the permission of yourtown.The authors wish to acknowledge and express thanks to yourtown (Kids Helpline).The authors declare no conflict of interest. The funding body had no involvement in study design, data collection, analysis/interpretation, manuscript preparation or submission.Number of contacts to KHL service in 2012–2018: (a) total, (b) counselling contacts, (c) information and referral contacts.Joinpoint trend analyses across 2012–2018 for type and medium of contacts (n = 704,690).Note: APC = annual percentage change; LL = Lower limit of 95% confidence interval; UL = Upper limit of 95% confidence interval.Contacts separated by KHL age groups, gender, and contact type (n = 704,690; 2012–2018).Note: Percentages reflect column totals unless specified. Gender totals reflect only contact where the young person’s gender was recorded.Number and percentage of contacts across each concern type and concern group (n = 449,814).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06025.txt ADDED Viewed

	@@ -0,0 +1 @@

+ There are large socioeconomic inequalities in alcohol-related harm. The alcohol harm paradox (AHP) is the consistent finding that lower socioeconomic groups consume the same or less as higher socioeconomic groups yet experience greater rates of harm. To date, alcohol researchers have predominantly taken an individualised behavioural approach to understand the AHP. This paper calls for a new approach which draws on theories of health inequality, specifically the social determinants of health, fundamental cause theory, political economy of health and eco-social models. These theories consist of several interwoven causal mechanisms, including genetic inheritance, the role of social networks, the unequal availability of wealth and other resources, the psychosocial experience of lower socioeconomic position, and the accumulation of these experiences over time. To date, research exploring the causes of the AHP has often lacked clear theoretical underpinning. Drawing on these theoretical approaches in alcohol research would not only address this gap but would also result in a structured effort to identify the causes of the AHP. Given the present lack of clear evidence in favour of any specific theory, it is difficult to conclude whether one theory should take primacy in future research efforts. However, drawing on any of these theories would shift how we think about the causes of the paradox, from health behaviour in isolation to the wider context of complex interacting mechanisms between individuals and their environment. Meanwhile, computer simulations have the potential to test the competing theoretical perspectives, both in the abstract and empirically via synthesis of the disparate existing evidence base. Overall, making greater use of existing theoretical frameworks in alcohol epidemiology would offer novel insights into the AHP and generate knowledge of how to intervene to mitigate inequalities in alcohol-related harm.Systematic socioeconomic inequalities in health persist and continue to widen across the globe, including in countries ranked highly on indices of economic prosperity and human development [1,2]. Alcohol-related health outcomes are not only an example of health inequality but also contribute to inequalities in both life expectancy and death age between socioeconomic groups [3]. There is a large body of evidence to suggest that, although those of lower socioeconomic position (SEP) tend to drink the same or less on average as those in higher SEPs, they still experience greater rates of alcohol-related harm [4]. One record linkage study found that, despite controlling for alcohol consumption and other risk behaviours, the most deprived group still maintain a three-fold higher risk of alcohol-related harm [5]. This phenomenon, termed the alcohol-harm paradox (AHP), treats alcohol use as a risk factor for health-related harm, although when alcohol use crosses into alcohol dependence the social/health state of dependence that arises is viewed as a harm outcome [6]. The AHP is found consistently across several outcomes, including alcohol dependence [7], alcohol-related morbidity [8] and mortality [4]. Yet the causal mechanisms remain unclear.Despite socioeconomic inequalities in alcohol-related health outcomes, health behaviour has been central to research investigating the AHP [9]. This reflects wider public health trends, as for decades epidemiological research has been criticised for its emphasis on using individual-level proximal risk factors to predict population-level health [10,11]. Arguably this has led to the most affluent reaping the health benefits due to their increased access and uptake of behaviour change interventions [2].Cross-sectional research has demonstrated that low SEP groups tend to drink on fewer occasions but drink more heavily per occasion compared to high SEP groups [12,13]. They are also more likely to engage in multiple health-risk behaviours (e.g., smoking, poor diet) [12]. However, these studies do not measure harm outcomes.Conversely, two record-linkage studies found that behavioural factors, including drinking pattern, smoker status and BMI, could not fully explain the paradox [5,14]. These factors attenuated inequalities, but low SEP groups still had a persistently higher risk of alcohol-related harm [ibid.]. These findings were confirmed by a recent meta-analysis, which found that quantity of alcohol consumed and drinking patterns could not explain socioeconomic inequalities in the relative risk of both all-cause and alcohol-attributable mortality [4]. This suggests health behaviours are unable to fully explain the AHP.While empirical research on the AHP has been limited in exploring other factors associated with socioeconomic circumstances [9], there is an increasing appetite to draw on explanations used to understand health inequalities. A report summarising the AHP discusses access to healthcare and material resources as potential explanations [15]. However, at present, there is a lack of theoretical structure to research investigating the AHP. Our understanding of the causes of the paradox remains stagnant due to a continual focus on individual behaviour. This is reflected in recent calls for exploration of contextual factors (e.g., characteristics of drinking environment) and how they not only influence health behaviour but may also directly impact harm [16].The aim of this paper is to address this gap by identifying alternative approaches rooted in health inequality theory which could be used to design future research on the AHP. To achieve this, we review theories of health inequality and their potential to understand the causes of—and therefore potential solutions to—the AHP. We do not aim to synthesise these theories or recommend any one theory. In the context of the AHP, drawing on any of these approaches would be a novel way to conceptualise the problem or inform research design. In Section 2 we introduce prominent theories including the social determinants of health (SDH), fundamental cause theory (FCT), the political economy approach and the eco-social model and discuss the extent to which these approaches are present in the existing AHP literature. We do so by explicitly drawing on a recent systematic review which presents an overview of the explanations for the AHP [9]. We then examine how these theories could be used to explicitly frame research on the AHP. In Section 3 we discuss the potential use of computer simulations to assess their explanatory value. In Section 4 we discuss what adopting a health inequality lens could mean for the wider alcohol-harm research agenda.Since the publication of the UK Black Report on Inequalities in Health [17], several theories have been developed which seek to explain how SEP drives health outcomes. Most have a common focus: to shift attention away from the individual-level and behavioural factors, and instead take a multi-level approach. In this section we outline four main theoretical approaches: the SDH, FCT, the political economy approach and the eco-social model (see Table 1 for descriptions of each theory), and referring to a recent review [9], discuss how these approaches fit with explanations for the AHP used within scientific literature. The review highlighted which explanations had remained hypothetical, and which were present in the empirical research. We aim to highlight how explicitly drawing on theories of health inequality could support research aiming to identify the causal mechanisms that drive the AHP.The SDH refers to the social and economic factors that shape health at the individual and population level [28,29]. This approach, originated from the Rainbow Model [30], was refined by the WHO in the 2000s [18] and continues to be central to public health research. It attempts to shift the focus from individual-level behaviours as the cause of health inequality to social determinants which themselves determine not only health but also behaviour [31]. Drawing on this theoretical approach to understand the AHP could be the first step to shift from an individual approach.Within the SDH theoretical approach there are four underlying-interrelated-explanations: culture-behaviour, materialist, psychosocial and lifecourse [32].Culture-Behaviour. Norms and cultural practices associated with socioeconomic groups have been hypothesised to impact alcohol-related harm. There has been discussion, but no formal hypothesis testing, for how normative differences in drinking patterns between socioeconomic groups might contribute to AHP [33,34]. Culture and norms may also influence help-seeking and engagement with preventative healthcare services [35]. There is further scope to examine occasion-level risk factors, such as drinking contexts and their association with acute alcohol harm [36].Materialist. The materialist approach is not present in empirical work investigating the AHP. Researchers have hypothesised that some of the mechanisms associated with materialism lead to socioeconomic inequalities in alcohol-related harm without explicitly drawing on theory. This includes individual material deprivation (e.g., housing and employment), which results in individuals having worse health and a lack of resources to protect themselves from a problem or stressful life event [15,37,38]. Additionally, place-based materialist mechanisms, such as a lack of environmental resources (e.g., treatment facilities and preventative services), alcohol outlet density and barriers to accessing healthcare have all been hypothesised contribute to the AHP [39,40].When providing materialist explanations for the AHP, researchers tend to focus on the mechanisms which impact the most deprived in society, without considering the material advantages available to wealthier socioeconomic groups. Additionally, material explanations are typically discussed in isolation in the AHP literature, meaning the link between materialist explanations and societal structures (e.g., the welfare state and benefits system) is missing from the current narrative.Historically, the contribution of individual material factors (car ownership) to health inequalities has been well evidenced [41,42,43]. Subsequent research also includes resources available in the environment (e.g., access to destinations, transportation systems) [44]. Applying these measures to identify material differences within and between socioeconomic groups could reveal the contribution of material mechanisms to the AHP.Psychosocial. The psychosocial approach has yet to be used in research investigating the AHP. Stress-related mechanisms are hypothesised to play a role, particularly lower socioeconomic groups experience a greater number of stressful life events, negative stereotyping, stigma, and social isolation [7,37,45,46]. The lack of social relationships is purported to lead to maladaptive coping strategies, consuming alcohol to cope and a reduced resilience to future negative events [47,48]. Conversely, it is acknowledged that affluent individuals have a beneficial network of social connections and therefore a greater social ‘buffer’ against stressful life events [40,49]. While these hypothesised mechanisms touch on components of the psychosocial approach, the role of social comparison (when lower socioeconomic groups compare themselves with others) and discussion of the biological consequences, both central to the psychosocial approach, are missing from the current AHP literature.Explicitly using the psychosocial approach would reframe the discussion of psychological and social mechanisms to consider how people feel compared to others and the psychological and biological consequences of those feelings which may contribute to the inequalities expressed within the AHP. This concept of relative deprivation is particularly important given the presence of the AHP in high income social welfare state countries [9] where social inequality persists. There is a vast literature on psychosocial pathways, which have been shown to contribute to health inequalities more generally [50], particularly in the form of social capital (capturing both social buffer and potential negative effects of social inequality and exclusion). Future work aiming to understand the AHP could usefully refer to the measures of social capital used in existing studies.Lifecourse. The life-course explanation integrates aspects of several other explanations, allowing different causal mechanisms and processes to explain socioeconomic health inequalities. Risk factors associated with other SDH explanations have been situated in time by some researchers investigating the AHP. This work shows promise, with one study finding cumulative behaviours (those that persist over time) attenuate the link between SEP and all-cause mortality by 38–77% compared to adjusting for proximal behaviours which attenuated the link by only 24–55% [51]. Some literature on the AHP discusses the impact of experiencing material disadvantage at critical time periods (e.g., childhood) and the accumulation of negative events as having prolonged negative health effects [52,53,54].The life-course perspective has been adopted in research using event history analysis and retrospective data, for example in a study investigating the role of cultural capital and cultural health capital during childhood in the uptake of mammography in later life [55]. There is a lack of application of these methods in the context of alcohol-harm, with only one similar example identified in the review which investigated factors associated with the development of a comorbid alcohol and mental health condition [53].The overall SDH approach is, however, subject to criticism. It has been argued that those who adopt it remain focused on the intermediary causes of health inequalities despite the consensus that it is the macro-level structures that result in health inequality [56]. These macro-level structures are viewed as being outside individual control and have become ‘causes of causes’ obscured by more proximate factors (e.g., health behaviour). This has resulted in theoretical and empirical research dedicated to describing the mechanisms that link socioeconomic inequalities to health, as opposed to identifying the source of socioeconomic inequality [56]. One theory developed to address this gap is FCT [22,57].FCT shifts the focus from individual-level causes of health inequalities to looking at the context; what puts people “at risk of risks” [22]. This means acknowledging that risk factors (e.g., alcohol consumption) are generated by social conditions, specifically the socioeconomic stratification of society. Crucially this theory does not deny the role of social determinants but suggests that base mechanisms associated with SEP determine whether individuals can adapt to the introduction of new disease, risks or treatment [58]. Proponents of this theory highlight that SEP should be viewed as the fundamental cause of health inequality and any downstream risk factors rooted within it [58]. From this perspective, neglecting the social conditions which generate risk factors has slowed progress in reducing health inequalities.FCT is not apparent in research investigating the AHP. Using this theory to frame the mechanisms underlying the paradox requires a focus on the societal structures which generate social inequality. Viewing SEP as a fundamental cause of health inequality requires the understanding that disparities are generated through multiple intervening risk-factor mechanisms which alter over time [22]. Key to this is the role of resources (money, knowledge, power, prestige, and access to social connections), closely linked to the materialist approach [22,58]. FCT asserts that health inequalities will remain despite societal and healthcare changes so long as the socioeconomic structure giving access to resources remains stable [58]. Drawing on this perspective to understand the AHP would require acknowledging the existence of this structure and treating SEP as a meta-mechanism responsible for access to resources which could mitigate the effects of other factors associated with the SDH.A comparative case-study using FCT predicted that as lung cancer becomes more preventable, due to knowledge of the link between smoking and the disease, those with greater access to resources disproportionately benefit, thus increasing health inequalities [59]. Contrastingly, for a disease lacking in major prevention or treatment innovation (e.g., pancreatic cancer), there was found to be no mortality advantage associated with socioeconomic group, and this trend was consistent across time [ibid.]. Alcohol-related harms (e.g., liver disease), are largely preventable. Trend analysis could test the role of FCT and investigate whether the introduction of prevention or treatment measures over time has resulted in socioeconomic inequalities in alcohol-related harm.Sitting between the SDH and FCT is the political economy of health approach. The political economy explanation is an attempt to acknowledge the role of upstream factors in generating and distributing risk factors. It argues that the social- and behavioural- determinants of health are themselves shaped by structural determinants: politics, the economy, the (welfare) state, political institutions, the organisation of work and the structure of the labour market [60,61,62] and that population health is shaped by the “social, political and economic structures and relations” that may be, and often are, outside the control of the individuals they affect [25,27].Structural influences within the political economy approach have only been tenuously linked to the AHP. The economic and socio-political conditions, alcohol policy, corporate influence, employment, and power relations are provided as potential explanations for the AHP [45,63,64,65]; however, authors do not clearly articulate the underlying mechanisms. They touch on the commercial determinants of health as key drivers of alcohol-related harm which aligns with recent calls to acknowledge the detrimental role of the private sector on both the environment and health behaviour, which in turn determines health [66]. The political economy perspective clearly defines the role of these structures as influencing the distribution of the other SDH. Drawing on a synthesis of these perspectives in the context of alcohol-related harm would highlight these mechanisms. For example, the social and political attitudes of residents and decision makers influence the investment of public services in deprived areas, which then determines the availability of services [67], a materialist determinant of health.Studies investigating the role of political economy in the generation of health inequalities typically take a cross-national comparative approach. This involves comparing different economic and political systems to understand how these systems contribute to health inequalities, both within and between countries [68]. This approach to research provides the opportunity to identify how the structure of the labour market, employment and welfare systems can prevent or increase health inequalities [ibid.]. There is a current lack of cross-national comparisons in the existing AHP literature.A recent commentary by Bloomfield has called for future research investigating the AHP to draw on the eco-social approach, acknowledging that inequalities in alcohol-related harms cannot be explained by drinking patterns alone [16]. The main distinguishing feature of the eco-social approach is the emphasis it places on biological and ecological analysis [69].Biological mechanisms have been hypothesised to contribute to the AHP. Primarily these have been related to health behaviours and genetic alterations due to the experience of disadvantage [12,33]. For example, engaging in certain patterns of behaviour (e.g., multiple unhealthy behaviours or drinking with meals) has metabolic effects which compound or protect against the effects of alcohol consumption [33]. Biological alteration related to the experience of disadvantage or differences based on ethnicity were also more vaguely linked to the AHP [48].Explicitly using the eco-social approach would shift the focus to how individuals biologically embody their social conditions. Achieving this in empirical research requires access to biological and social data. A recent paper which analysed data from several cohort studies investigated the relationship between social disparity and biology, finding evidence of biological changes in response to the environment [70]. There may be opportunities for alcohol researchers to engage in collaborative projects or gain access to data sets, for example, the UK Biobank [71], which would allow the opportunity to investigate the eco-social model in the context of alcohol harm.Explicitly drawing on existing theories of health inequality may address the gap in identifying and extracting relevant variables and relationships in the pursuit to understand the AHP. However, the methods best placed to test these causal relationships requires further scrutiny.To study these complex relationships, which exist on a multi-level plane (e.g., individual, community, and structural levels) and are dynamic in nature, suitable research methods are required. The “risk factor” approach to epidemiology explores decontextualized and independent relationships between dependent and independent variables and uses linear reductionist models to test these relationships [72]. To capture the features of complexity, a mechanism-based approach is required which explicates the details of how regularities are brought about rather than focusing on statistical regularities between variables [73]. Mechanisms consist of “entities” and the “activities” entities engage in, either as a collective or independently, to bring about a particular outcome [ibid.]. Computer simulation methods are a good candidate to test mechanisms, and complex system models have become increasingly attractive in public health research [74].A review of the use of simulation models in the context of health inequality concluded that they enhance our understanding of socioeconomic health inequalities [75]. Specifically, the class of techniques known as agent-based modelling (ABM) can flexibly model the multilevel, reciprocal, and indirect effects of socioeconomic inequalities [ibid.]. ABMs are computer simulations comprised of agents (e.g., individuals or households) and their interactions within the context of their environment [76]. ABMs provide the opportunity to test mechanisms specified in theory [77]. This ranges from abstract theory testing to more concrete applications which draw on empirical data to inform the properties and environments of agents [78]. Much like other types of simulation model, ABMs enable otherwise fragmented evidence to be synthesised to address research questions and inform decision making [79].One example of an ABM implemented to understand socioeconomic health inequalities explored the role of bounded rational choice mechanisms (individual level) and spatial segregation (structural level) in the emergence of income gradients in healthy eating [80]. This model represented both food stores and households as having agency over decisions to supply and purchase, respectively, healthy, or unhealthy foods. The model equations define the mutual interactions between stores and households, enabling feedback loops to be represented. The model’s findings suggest that differences in diet between socioeconomic groups arises only when high income households and healthy stores are both spatially segregated from low-income households and unhealthy stores. Once established, these diet inequalities could only be overcome when both groups had favourable preferences for healthy foods and when healthy food was relatively cheap [ibid].A similar approach could be taken to investigate the mechanisms specified in health inequality theory. Here, we sketch such a model. In psychosocial theory, one mechanism proposed to result in socioeconomic differentials in health is that high SEP groups have a protective social buffer [40,49]. Hypothetically, this mechanism could be represented in an ABM by simulating individuals as agents and defining a macro-level social network structure with connections based on agent attribute similarity (e.g., age, gender). Agents would possess the capability to give or receive support in the presence of a stressful event. However, this capability would be contingent on their own resources (e.g., income), type of support available to them (e.g., emotional support) and their own stress burden. Individuals who receive support from their network would have a reduced stress burden and therefore reduced risk of harm. The network could also be responsive to changes in relationships (e.g., providing support strengthens ties while refusing support breaks social ties between agents). A simulation such as this would allow in silico experimentation with changes in resources, types of support and stress, to determine how these features impact not only individuals but also, potentially, their social network structure.Recent developments in computer model integration have also demonstrated that ABMs which combine mechanisms from multiple theories can provide an improved explanation for complex phenomena (in terms of parsimony and empirical goodness-of-fit) [81,82]. These integration findings are particularly relevant given that theories of health inequality do not necessarily compete, but rather attempt to explain health inequality from different viewpoints.Computer simulation methods such as ABM have yet to be applied to understand the AHP and would allow us to make best use of the available evidence to test the explanatory value of the mechanisms described in existing theories of health inequality. When we extract the mechanisms from these theories and implement them in an ABM simulation, does the simulation generate inequalities in alcohol-related harm?It is clear research investigating the AHP eschews the use of theory. Many of the mechanisms specified in health inequality theory are touched on as hypothetical explanations for the paradox, mainly on an ad-hoc basis and in the absence of clear theoretical structure. Structure would be provided by drawing on any of these theories explicitly. In the one instance where one of the theories was present in the empirical work on the AHP, this showed promise, as cumulative behaviours across the lifecourse could explain a greater proportion of harm experienced by lower socioeconomic groups [51]. There is a lack of evidence, which makes it difficult to conclude whether one theory over another can best explain the AHP, especially as these theories do not necessarily compete but examine causes of health inequality at different levels and with differing emphasis on certain factors. One thing is clear: the use of these theories will shift how we think about the causes of the paradox from health behaviour in isolation, to the wider context of complex interacting mechanisms between individuals and their environment.Framing alcohol research using health inequality has significant implications for the study of the AHP and the wider alcohol harm research agenda. In the past, behavioural framings have resulted in empirical work underpinned by individual proximal factors, specifically alcohol consumption and other health behaviours. In Section 2, for each theory, we identified research designs implemented in social epidemiology which attempt to understand the causes of health inequality more generally (e.g., new measurements that capture social capital [50], or cross-national comparisons [68]). We can utilise the advances in social epidemiology, for example the introduction environmental resources in the materialist perspective [44] and apply this to the AHP.Taking a behavioural approach has resulted in the implementation of policies which often rely solely on individuals taking action to reduce their alcohol consumption (e.g., educational campaigns), which arguably increase inequalities [15,83]. There have been attempts to reduce inequalities by introducing minimum unit pricing in several countries, including Scotland, Wales, and Australia’s Northern Territory. In theory this policy reduces the consumption of alcohol, particularly for those of a lower SEP, as they typically purchase alcohol at cheaper price points [84]. However, the focus of this policy remains on reducing alcohol consumption, which will not address the underlying causes of inequality.Critically, shifting from this focus on alcohol consumption as the fundamental cause of harm in alcohol research requires researchers to acknowledge the causal processes driving harm are complex and that understanding of these processes requires different methodological perspectives, drawing on ideas from complexity science [85].While the focus of this paper has been on the AHP, a well evidenced phenomenon, it is possible that a harm paradox could exist for other health behaviours. Hypothetically, with the same number of cigarettes smoked, those of a lower SEP may experience greater rates of smoking related harm, and there is evidence to support this hypothesis [86]. This reflects a slight misnomer—the AHP is not particularly paradoxical if it simply reflects wider causes of health inequalities. This concern further reinforces the need to utilise theories of health inequality to understand the complex interactions between health behaviour, the environment and harm and explore why lower socioeconomic groups are more vulnerable to the negative effects of risk behaviour.The existing research on the causes of the AHP lacks theoretical structure and relies heavily on analysing the contribution of health behavioural risk factors. Drawing on health inequality frameworks would result in a more structured effort which gets at the root causes of both alcohol-related harm and alcohol-related health inequalities. Using these multi-level frameworks would allow us to understand the role of other mechanisms, in addition to alcohol consumption, which exist in the wider socioeconomic environment. Simulation methods (e.g., ABMs) allow for the opportunity to meaningfully explore the complexity captured in health inequality theory. Combining these theories with simulation methods has the potential to inform policy, which not only reduces consumption but also reduces harm, and in turn health inequalities, more broadly.Conceptualization, J.B., C.B., R.C.P. and J.H.; investigation, J.B.; writing—original draft preparation, J.B.; writing—review and editing, C.B., R.C.P. and J.H.; supervision, R.C.P. and J.H. All authors have read and agreed to the published version of the manuscript.This research was funded by The Wellcome Trust (108903/B/15/Z), the National Institute on Alcohol Abuse and Alcoholism of the National Institutes of Health (Award Number: R01AA024443) and the UKPRP SPIHER Consortium: Systems science in Public Health and Health Economics Research (MR/S037578/1). For the purpose of Open Access, the author has applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.The authors declare no conflict of interest. The funders had no role in the design of the study; in the writing of the manuscript, or in the decision to publish the results.Health inequality theories with descriptions.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06026.txt ADDED Viewed

	@@ -0,0 +1,32 @@

+The substitution of chemical pesticides by biopesticides is crucial to ensure the quality of agricultural products and to foster environmental sustainability. This study takes the willingness and the behaviors of rice farmers on the application of biopesticides as the research object. The survey questionnaire was designed based on the theory of rational small-scale farmers from three aspects: “individual and family characteristics of farmers”, “cognition of farmers” and “external factors”. The survey was then conducted on 163 rice farmers in seven prefecture-level cities in Jilin Province of China. The logistic model was used to analyze the influencing factors resulting in the deviation of the behaviors of the rice farmers from their initial willingness on the application of biopesticides. The explanatory structure model (ISM) was used to analyze the logical hierarchical relationship among various influencing factors. The results show that: (1) For 45% of the farmers surveyed, there’s a deviation between their willingness and behaviors regarding the application of biopesticides; (2) Among the significant factors leading to the deviation between farmers’ willingness and behaviors concerning the application of biopesticides, the surface-level direct factor is biopesticide awareness. The mid-level indirect factors are agricultural product quality and safety awareness and the deep-level root cause is farmers’ education level. (3) The primary reason for the deviation of the farmers’ behaviors from their willingness is their lack of knowledge about biopesticides and the biopesticides’ incomplete market structure. Based on the comprehensive analysis, it is recommended to improve the professionalization of the farmers, to strengthen the publicity of green production and to accelerate the formulation of the biopesticides market to further promote the usage of biopesticides.The amount of pesticides used globally to control crop pests and diseases is estimated to be around 6 million tons per year. The effective utilization rate of pesticides is less than 30% while the non-effective misusage is as high as 70% [1]. Residual pesticides in the environment spread rapidly under wind, rain and other meteorological conditions, leading to the condition where pesticide residues are trapped in the air, oceans, soil and organisms worldwide [1]. Since 2007, China has ranked no.1 in the world in terms of pesticide production and usage, however there’s also adequate evidence showing that overusage of pesticides is very common in China. According to statistics, the total amount of pesticides utilized in China in 2019 was still as high as 1,392,000 tons and the average dosage of pesticide application reached 8.39 kg/ha, which was higher than the internationally accepted upper limit of 7.5 kg/ha based on safety considerations [2]. The massive use of chemical pesticides will not place heavy pressure on the ecological environment, it will also adversely affect the quality and safety of agricultural products due to the potential presence of pesticide residues [3], Pesticide residues can also pose a serious threat to human being’s health through the food chain and the accumulation of bioconcentration effects [4]. Some member states of the European Union were the first to put forward and implement the concept of reducing pesticide usage to reduce the impact on the agroecological environment. Since then, the use of pesticides in several countries have shown a decreasing trend [5]. Similar policies on the reducing pesticide usage have also been implemented in Korea, Japan and other countries in Asia, where the usage of pesticides has decreased significantly in recent years [6]. However, the usage of pesticides in major countries in the Americas has still been growing rapidly, especially for herbicide usage [7]. To better tackle the problem of limited global agricultural resources and food safety issues, biopesticides have been more frequently used over traditional chemical pesticides because they control through natural substances or living organisms [8]. They also have the advantages of being flexible, less likelihood of producing resistance, harmless to plants, human beings, animals and the environment and they are eco-friendly products [9]. They are also the most important production inputs for organic agriculture and they a crucial role in agricultural sustainability [10]. In order to promote green agriculture and to increase agricultural product quality, it’s necessary to shift farmers’ traditional way of using chemical pesticides into using biopesticides [11]. Environmental protection and sustainable development are nothing new in China and majority of the Chinese farmers do have an expressed willingness to adopt green production techniques [12]. However, due to high production and preservation costs for biopesticides and the farmers’ lack of knowledge, the adoption level of biopesticides in China is still not promising. The market share of biopesticides usage stands at less than 10%, which is 50% lower than the world average [13]. Differences in the behaviors and willingness of farmers to apply biopesticides in actual agricultural productions have been observed [14]. Pray et al. found that more than one-third of agricultural producers in India expressed a willingness to use biopesticides, however only 3% of the farmers in the selected sample had actually used biopesticides in the past year [15]. Some scholars have noticed this phenomenon and it’s referred to as deviation or conflict between willingness and behaviors [16]. It will easily lead to wrong decisions from the government and enterprises on the production and promotion of biopesticides. Therefore, reducing the deviation of farmers’ behaviors from their willingness on the application of biopesticides is of vital importance to effectively promote the application of biopesticides and to realize the green transformation of agriculture.Rice has been the predominant crop feeding 800 million people in China with a massive plantation area [17]. Rice has also played an important role in ensuring food security for the country. China is the largest producer of rice and, at the same time, China is also the largest consumer of rice. Jilin Province, as a major agricultural province in China, is an important rice production area in China [18]. Therefore, rice farmers in Jilin Province of China were chosen as the subjects of this paper. It is important to adopt biopesticides in the sustainable development of agriculture, especially in the current stage of low utilization rate of biopesticides. Through conducting research on rice farmers on their willingness and behaviors of biopesticides application and analyzing the influencing factors and the logical hierarchy among the influencing factors, the weak links in the promotion and the adoption of biopesticides in China at the present stage can be identified. It is also of great significance to alleviate the deviation of farmers’ behaviors from their initial willingness to apply biopesticides and to promote biopesticides more efficiently. It will also help to improve the competitiveness of the rice industry, to reduce environmental pollution and to gradually replace chemical pesticides with biopesticides. The study on the deviation of farmers’ willingness and behaviors to apply biopesticides in China will also be useful for the policy making of biopesticides in other developing countries.Villa-Rodríguez et al. found that Bacillus thuringiensis (Bt), one of the major biopesticides used worldwide currently, was effective against rice leaf borer, stem borer, and stem borer [19]. In a survey conducted in U.S. farms, Wozniak concluded that the more educated one is, the more willing one is to try new things and take the risks they entail, and the more likely one is to embrace new technologies [20]. Paudel et al. concluded that the risk-averse psychology of growers affects the adoption of organic fertilizer technologies and that the degree of risk aversion (preference) of farmers affects their agricultural production and input behaviors [21]. Existing studies under the farmer perspective found that farmers’ personal characteristics, household characteristics [22], risk preference [23], technology perception [24], information ability [25], psychosocial perception [26] and production purpose [27] have significant impacts on farmers’ biopesticides purchase behaviors, their willingness to apply and their application behaviors. The government perspective focuses on the promotion role of policy guidance [28] as well as government publicity and education [29] in the adoption process of biopesticides by farmers.Most of them are based on the Theory of Rational Behavior (TRA) or the Theory of Planned Behavior (TPB) [24,27,30], using methods such as structural equations to analyze the influence of behavioral attitudes [31], perceptual behavioral control [32], subjective norms [33] and other dimensions on the divergence of some behavioral intentions and applied behaviors [31]. The research area is mainly focused on the product consumption of individuals. Relevant studies in the field of agriculture are fewer, with research topics scattered widely. Theoretically, Icek argued that willingness is a condition of the process of achieving the desired behavioral goal and it’s predictive [34]. Meanwhile Newman argued that willingness and behavior can show inconsistencies, either in the form of a blocked conversion of willingness into behaviors or in the form of a deviation of behaviors from the initial willingness due to external interference, and that willingness will not effectively convert behaviors [35]. Waithaka argued that deviation is influenced by internal endogenous drivers and external situational changes [36]. Jeffrey R. found that the theory of planned behavior adds to the individual’s subjective willingness the conditions and ability to perform a specific behavior [37], and since the ability to perform and subjective willingness are collectively referred to as perceived behavioral control, perceived behavioral control can directly influence individuals’ behavioral intentions and applied behaviors. To date, most of the relevant studies on the behavioral analysis of farm household are based on the willingness-behavior deviation perspective focusing on new rural cooperative medical care [38], food security [39] and small-scale water constructions [40].The existing international research findings on biopesticides show that the relevant literature started earlier and the topic has been studied deeply. However, there are few studies on the behaviors of farmers on the application of biopesticides, especially for rice farmer households in Northern China. There is a gap in the research on the deviation of behaviors of rice farmers from their willingness to apply biopesticides and the mechanisms and factors influencing the deviation of biopesticide application decisions need to be analyzed.Therefore, this study uses a logistic regression model to empirically analyze the factors influencing the divergence between the willingness and behaviors of the rice farmers on the application of biopesticides. Theoretical support and practical guidance are pro-vided for better and prompt promotion of biopesticides and improved utilization rate of biopesticides.The data used in this study were obtained from survey questionnaires and interviews conducted by the research team from October to December 2020 among rice farmers in Jilin Province of China. A multi-stage random sampling method was used to select the samples during the actual survey [41]. First, based on the scale of rice cultivation, a total of seven counties and cities were selected in Jilin Province, including Changchun, Jilin, Siping, Liaoyuan, Tonghua, Songyuan, and Baicheng. Then, two townships were randomly selected in each county and city. Finally, three natural villages were randomly selected in each township, and 3–8 rice growing households were randomly selected in each village for our questionnaire survey. Data were obtained on individual characteristics, household characteristics, knowledge of biopesticides, willingness to apply biopesticides in rice production and their application behaviors, together with other related variables. We issued a total of 200 questionnaires, and the focus of this study is on the difference between the willingness and behavior of biopesticide application, i.e., farmers who have the willingness but do not have the behavior, so the questionnaires of farmers without the willingness to apply biopesticides were excluded. After the later research error checking and sorting, questionnaires were excluded that are invalid or farmers who don’t have the willingness, and finally 163 valid and willing farmers’ questionnaires were obtained, with an effective rate of 81.5%. The basic information of the farmers surveyed is shown in Table 1.Descriptive statistics of the samples show that: the age of the farmers surveyed was mainly between 41 and 50 years old (43.6%), with an average age of 45.88 years old; 73.6% of them were men and 26.4% were women; 42.9% of the farmers participated in cooperatives; the education level of farmers were mainly primary and junior high school, accounting for 74.8% of the total number of samples; the average scale of rice cultivation was 18.3 mu; the average annual household income was 118,600 yuan; the proportion of income from rice plantation mainly ranged from 80% to 100% (33.1% of them) with an average value of 78.75%. Therefore, most of the farmers being surveyed were middle-aged males with a higher proportion of household income from planting rice. It’s also worth mentioning that a majority of them have an education level of junior high school and below. Jilin Province is an important rice production area in China and this paper uses stratified random sampling from dispersed geographical locations for the survey to ensure the samples selected are representative in serving the research needs for this study.Based on the theoretical basis of the rational smallholder theory proposed by Schultz [42] in combination with related studies [43], we analyzed the factors leading to the divergence between farmers’ willingness to apply biopesticides and their behaviors from three aspects: farmers’ individual and family characteristics, farmers’ perceptions and external factors. Based on the construction of the theoretical model, “farmers’ willingness and behaviors to apply biopesticides” was set as the dependent variable and “factors influencing the deviation of farmers’ behaviors from willingness to apply biopesticides” as the independent variable, including farmers’ individual and family characteristics, farmers’ perceptions and external factors. Referring to the existing scholars’ measures [30,31,37], this paper defines the deviation between the farmers’ willingness and behaviors to apply biopesticides as a phenomenon in which farmers show willingness to apply biopesticides in the agricultural production process without taking actual actions. In other words, there’s inconsistency shown between their willingness and behaviors. Based on this definition, the samples of this paper should be those farmers who have the initial willingness to apply biopesticides in their agricultural production process. Statistical analysis shows that 163 farmers out of 200 samples have the willingness to apply biopesticides, hence this paper will conduct empirical analysis based on these 163 samples. For those farmers that don’t have biopesticide application behaviors, deviation exists and y = 1; if farmers have biopesticide application behaviors, there’s no deviation and hence y = 0. The details are shown in Table 2.
+                (1)
+                Individual and Family Characteristics
+          Individual and Family CharacteristicsNumerous studies have shown that both individual characteristics of farmers and their family characteristics pose an impact on their deviation between behaviors and willingness. The individual characteristics of farm households mainly refer to gender, age, and education level of the farmers surveyed. Gender differences are reflected in the decision-making process in agricultural production. Different scholars have different opinions on this issue. Some of the scholars’ studies concluded that male farmers usually have higher exposure to the outside world than women with a better understanding of pesticides and they have a better awareness of the associated health risks from using chemical pesticides [3,44]. However, some other scholars believed that women are more concerned about their own safety and health than men in the process of pesticide application [45,46]. Binswanger et al. showed that younger farmers are more inclined to take risks [47], while older farmers, who may have developed an empirical dependence on chemical pesticides during their long-term agricultural practices, are more inclined to choose chemical pesticides [48]. The education level of farmers reflects, to some extent, their ability to obtain in-formation and to acquire skills. Farmers with higher education level are more likely to adopt biopesticides [49].The agricultural households’ business characteristics of farmers mainly include whether they participate in cooperatives, annual household income, percentage of annual income from rice plantation and scale of rice cultivation. Farmer cooperatives are an important part of the agricultural science and technology extension system and they play an crucial role in the promotion of biopesticides [50]. Therefore, farmers who participate in farmers’ cooperatives are generally aware of modern agricultural productions and thus they know the advantages of environmentally friendly agricultural technologies. So, it is more likely that they will choose biopesticides over conventional chemical pesticides [51]. Biopesticides are more expensive compared with chemical pesticides. The higher the income level of farmers’ households, the higher their probability of putting in more investment and adopting biopesticides [52]. The percentage of the farmers’ household’s annual income from rice plantation measures the dependence of farmers on land. By considering opportunity cost, the lower the percentage of their annual income out of rice plantation, the less likely they are to choose biopesticides [53]. Related studies have shown that planting scales have a facilitating effect on rice farmers’ biopesticide application behaviors [54] and large-scale households tend to have more social capital and human capital compared with smaller farmer households hence they have better access to external resources [55]. This implies that farmers with a large cultivation scale have more human and financial resources, pest control expertise as well as broader information access, so they are inclined to choose biopesticides for early prevention and disease control at the right timing [56].Therefore, this paper hypothesized that farmers’ education level, annual household income and rice cultivation scales have a negative effect on the deviation of farmers’ willingness and behaviors to apply biopesticides. While farmers’ age, participation in cooperatives and percentage of income out of rice plantation have a positive effect on the deviation of farmers’ willingness and behaviors to apply biopesticides. The gender of the farmer has an uncertain influence on the deviation of farmers’ willingness and behaviors to apply biopesticides.
+                (2)
+                Farmers’ Awareness
+          Farmers’ AwarenessFarmers’ perceptions are one of the most important factors affecting farmers’ choices of pesticides [48]. Farmers’ perceptions of pesticides mainly include their understanding of the characteristics of biopesticides, the hazards posed on human health and environmental pollution from long-term application of chemical pesticides, their concern about the quality of agricultural products and their confidence level in the efficiency of biopesticides. Farmers are more inclined to choose biopesticides over chemical pesticides if they recognize that long-term application of chemical pesticides will do harm to human beings’ health and bring about environmental problems such as soil acidification, soil caking and nutrient decline [29]. The better the farmers’ awareness of effectiveness of biopesticides and the more serious they are with the quality and safety of agricultural products, the greater the possibility that they will choose biopesticides over chemical pesticides [57]. In contrast, if farmers do not believe in the promotional effects of biopesticides and the more skeptical they are about biopesticides, the more likely they will be driven to abandon the adoption of biopesticides [58].Therefore, this paper hypothesized that the farmers’ knowledge about the characteristics of biopesticides, their awareness about the environmental pollution caused by the long-term application of chemical pesticides and their concern over the quality and safety of agricultural products have a negative influence on the deviation of farmers willingness and behaviors to apply biopesticides. On the contrary, farmers’ skepticism about the effectiveness of biopesticides has a positive influence on the deviation of their behaviors from willingness on the application of biopesticides.
+                (3)
+                External Factors
+          External FactorsDifferent external factors have different impacts on the deviation of willingness and behaviors. Farmers tend to abandon the application of biopesticides themselves when there’s no one around them applying biopesticides [26]. The prerequisite for farmers to apply biopesticides is the availability of adequate biopesticides [59]. In the case of unexpected outbreak of severe pests’ diseases in the fields, farmers will be more likely to choose chemical pesticides that are fast-acting and easily accessible [60]. When purchasing pesticides, farmers will consider the price of the pesticides. Commercial biopesticides are generally more expensive, which may lead to farmers’ reluctance to apply biopesticides [61]. Therefore, this paper hypothesized that peer influences, emergency conditions, market availability and price affordability all have positive effects on the deviation of farmers’ willingness and behaviors to apply biopesticides from the farmers’ perspective.Therefore, the independent variable including 15 factors in 3 areas. The specific variable definitions and their descriptive statistics are shown in Table 3.A deeper analysis of the logical hierarchy among the influencing factors is of great theoretical and practical significance in studying the correlation between farmers’ willingness and behaviors in biopesticide application. Therefore, in this paper, a logistic regression model has been chosen to filter the influencing factors. Moreover, the hierarchical relationships among influencing factors have been analyzed using the ISM model [30,31,32,37,62].This study investigates the factors affecting the deviation of rice farmers’ behaviors from their initial intentions of biopesticide application. The dependent variable is “whether biopesticide application intentions and behaviors deviate from each other”, which is a typical binary decision problem, i.e., “deviation” and “non-deviation”. Therefore, in this paper, a logistic regression model has been chosen to investigate the factors influencing the deviation of farmers’ behaviors from their initial willingness in biopesticide application [63,64,65]. For those farmers that don’t have biopesticide application behaviors, deviation exists and y = 1; if farmers have biopesticide application behaviors, there’s no deviation and hence y = 0. The logistic regression model is as follows:
+		  (1)Pi=Fyi=β0+∑j=1nβjXij=expβ0+∑j=1nβjXij1+expβ0+∑j=1nβjXij
+           where Pi is the probability of deviation between the application intentions and behaviors of farmer i; Fyi is the probability distribution function; β0 is the intercept term; βj is the regression coefficient of the j-th independent variable; n is the number of independent variables; Xij is the value of the j-th variable of the i-th farmer.By taking the logarithm of both sides of Equation (1), the simplified form is obtained as:(2)yi=lnPi1−Pi=β0+∑j=1nβjXijThe factors influencing the divergence between farmers’ willingness to apply biopesticides and their behavior in rice cultivation are both independent and interrelated, and it is important to distinguish the hierarchy of relationships among the factors to identify the key reasons for the divergence between willingness and behavior, and even to solve the problem of biopesticide promotion efficiency [37]. Therefore, this paper further analyzes the correlation and hierarchy between the factors influencing the divergence between farmers’ willingness to apply biopesticides and their behavior by using the ISM model [62,66]. The steps of the ISM model are as follows [67]:Assuming that there are k significant influencing factors; S0 is the deviation of farmers’ intentions to apply biopesticides from their behaviors; SiSj denotes the ij significant influencing factor; the components of the adjacency matrix R are defined by Equation (3):(3)rij=1 Si is related to Sj0 Si is not related to Sj i=0,1,…,k; j=0,1,…,k
+          Determine the reachable matrix M among the factors, which is calculated from Equation (4)
+		  (4)M=R+Iλ+1=R+Iλ≠R+Iλ−1≠…≠R+I2≠R+I
+          where I is the unit matrix, 2≤λ≤k and the Boolean operator is used in the power operation of the matrix.According to Equation (5), the reachable matrix is divided into the reachable set PSi and the antecedent set QSi and both represent the set of all factors in the reachable matrix that can be reached from the factor Si, where both mij and mji represent the factors in the reachable matrix. Equation (6) determines the highest level Li and its influencing factors, as well as the other levels of factors. To do this, we remove the rows and columns of the highest-level factors from the reachable matrix M to form the reachable matrix. By repeating the steps in Equations (5) and (6), the factors at the second level and all other levels can be obtained:
+          (5) PSi=Sjmij=1,QSj=Sjmji=1
+          (6)Li=SiPSi∩QSi=PSi;i=0,1,…,k
+		  Determine the Hierarchical Structure of Each Influencing Factor.Directional arrows have been used to connect factors between adjacent levels and at the same level to obtain a hierarchical structure of all the influencing factors.Before adopting the logistic regression model, the possible multicollinearity in the explanatory variables was firstly diagnosed by the multicollinearity test and the results showed that the variance inflation factors (VIF) were all less than 10, indicating that there’s no multicollinearity among the variables. Based on that, the regression analysis of the sample data was performed using Stata software and the results are shown in Table 4.As seen from the regression results of the logistic model in Table 4, eight significant factors are influencing the deviation of behaviors and willingness of biopesticide application of rice farmers and the systematic composition of the deviation is determined as Si = (S1, S2, …, S8), representing education level, scales of planting, biopesticide awareness, awareness of hazardous effect from chemical pesticides, quality and safety awareness of agricultural products, peer influences, emergency conditions and price affordability, respectively. The willingness to apply biopesticides and behavioral deviations are represented by S0. According to the ISM explanatory structure model, the Matlab matrix operation tool was applied to obtain the hierarchical structure T of the factors influencing the willingness to apply biopesticides and behavioral deviations of rice farmers. The box indicates the same level of factors as shown in Figure 1.The comprehensive regression results show that both education level and scales of rice planting pass the 5% significance test hence both have a significant negative effect on the deviation of willingness and behaviors of biopesticide application. It’s confident to conclude that the higher the education level of the farmers and the larger the scales of rice planting, the less likely there’s deviation between their behaviors and willingness. This is because the farmers with better education level have a deeper understanding of biopesticides, so it’s the easier for them to adopt the application of biopesticides. At the same time, for those farmers having larger scales of rice plantation, agricultural production has become their major work. Since the application of biopesticides can effectively ensure the smooth implementation of agricultural production, their behaviors and willingness to apply biopesticides are less likely to diverge.Biopesticide awareness has a significant negative effect on the deviation of farmers’ willingness and behaviors of biopesticide application and it’s at 10% significance level. Combined with the results from descriptive statistics, there’s lack of knowledge of biopesticides currently for most of the farmers. By taking into consideration their age and low education level, they hardly search for information about biopesticides actively. Although farmers claim that they are willing to apply biopesticides, they have difficulties in appreciating the advantages of biopesticides due to their lack of knowledge and expertise about them and therefore they tend to give up transforming their willingness into concrete behaviors.The awareness of hazardous effects of chemical pesticides and quality and safety awareness of agricultural products are negatively correlated with the deviation of willingness and behaviors of biopesticide application, with both of the factors are statistically significant at 5% level. The reason is that, on one hand, farmers with a higher level of awareness of environmental pollution and the hazards of chemical pesticides are usually better educated and relatively younger. The more farmers willing to pay attention to the protection of rural ecological environment and health, the higher the chance they will adopt the usage of biopesticides. So, the likelihood of deviation of their behaviors from willingness on the application of biopesticides is lower. On the other hand, quality and safety awareness of agricultural products is part of social responsibility and it’s also the farmers’ own psychological initiative to protect rural ecological environment. Therefore, the stronger the awareness towards agricultural products’ quality and safety, the less likely the divergence between farmers’ willingness and behaviors on biopesticides application will take place.Peer influences have a significant positive effect on the deviation of farmers’ willingness and behaviors in biopesticide application and it’s at 1% significance level. It means that farmers tend to rely heavily on neighboring farmers when it comes to the procurement and application of pesticides. Peer influences can also be referred to as social customs or social norms. Farmers are pressured by social norms to conform to the behavioral expectations of others when conducting agricultural productions. In the survey, it was found that many farmers were willing to apply biopesticides at the first place, however the intention was abandoned as the neighboring farmers were still using chemical pesticides.The emergency conditions have a significant positive effect on the divergence between farmers’ willingness and behaviors to apply biopesticides and it has passed the 5% significance test. Farmers often face the dilemma of whether purchasing and applying the highly toxic but fast-reacting chemical pesticides or sticking with environmentally friendly biopesticides. Emergency conditions refer to the temporary anxiety, excitement, and tensions that farmers show in their daily agricultural productions [68]. Although a certain degree of contingency exists, it can still affect farmers’ environmental perceptions by changing their own emotions, which in turn affects the deviation of farmers’ willingness and behaviors in biopesticide application.The price affordability has a significant positive effect on the divergence between farmers’ willingness and behaviors of biopesticide application and it has passed the 1% significance test. It indicates that the price of pesticides is still one of the most crucial factors that farmers consider when purchasing. This is because farmers are still mainly rational people and profit optimization is the goal of conducting agricultural productions. Although most farmers are aware of environment protection and they have showed their willingness to apply biopesticides, the higher price of biopesticides discourages farmers in the end. When farmers’ willingness to protect the environment conflicts with the high purchasing price, majority of them will opt for cheaper options, which is chemical pesticides.Based on the results of logical hierarchical analysis (Figure 2), it is observed that the influencing factors are in different hierarchical structures, which are both independent from and interrelated with each other. Price affordability, emergency conditions, peer influences and biopesticide awareness are the direct influencing factors at the surface level. Awareness of hazardous effects from chemical pesticides and quality and safety awareness of agricultural products are the indirect factors at the mid-level while education level and cultivation scales are the root causes. The logical hierarchy among these factors can be summarized as a “single path with three drivers” model. The reason for this hierarchy to appear is mainly because that the application of biopesticides is determined by farmers but at the same time it’s constrained by realistic situations.Single path: education level, scales of rice planting → awareness of hazardous effects from chemical pesticides, quality and safety awareness of agricultural products → biopesticide awareness → farmers’ willingness to apply biopesticides and behavioral deviation.In this pathway, farmers’ individual and business characteristics such as farmers’ education level and planting scales are the most fundamental driving forces. Their awareness of hazardous effects from chemical pesticides and their awareness of quality and safety of agricultural products are the external manifestations of the root factors. Intermediate factors will further influence farmers’ perception of biopesticides, which in turn directly affects the deviation of farmers’ willingness and behaviors to apply biopesticides. Farmers’ actual behaviors are influenced by farmers’ perceptions, which are derived from their own perceptions of the hazards that chemical pesticides pose and their sense of responsibility to protect the quality and safety of agricultural products. Such perceptions and awareness of the environment protection reflect the personal characteristics of farmers and the operational characteristics of agricultural productions as well.Three drivers: price affordability, emergency conditions, peer influences → deviation of farmers’ willingness and behaviors to apply biopesticides.Farmers, as the most important group of people in agricultural productions, will face a variety of realistic scenarios when choosing pesticides. They need to consider the effectiveness of pesticides and the reaction time; they will also conduct a comparative analysis of various inputs and outputs to choose the most cost-effective production methods. At the same time, they are also heavily relying on their neighboring farmers as for the selection and application of pesticides.The analysis in this study was based on data from sample surveys of rice-growing farmers in seven prefectural-level cities in Jilin Province of China from October to December 2020. Logistic-ISM model has been used to analyze the key factors influencing the deviation of rice growing farmers’ willingness and behaviors to apply biopesticides and the logical hierarchy among the key factors have been analyzed in depth:
+		(1)There are still many farmers using chemical pesticides and there are many deviations between their willingness and behaviors in the application of biopesticides, so it is not promising to promote biopesticides as an alternative to chemical pesticides on a full scale. The divergence between the willingness and behaviors of rice farmers to apply biopesticides is influenced by various factors. In terms of individual and family characteristics, both education level and scales of rice planting have a negative effect on it. In terms of farmers’ awareness, biopesticide awareness, awareness of hazardous effects from chemical pesticides and quality and safety awareness of agricultural products have a negative effect on the deviation. The better the farmers’ awareness towards environment protection, the lower the possibility of the deviation to take place. As for external factors, peer influences, emergency conditions and price affordability have positive effects on the occurrence of deviation.(2)The logical hierarchy of influencing factors can be summarized as a “single path with three drivers” model. Biopesticide awareness is at the surface level, awareness of chemical pesticides’ hazards and awareness of agricultural quality and safety are indirect factors at the mid-level while the farmers’ characteristics such as education level and planting scales are root causes. The three drivers refer to external factors such as price affordability, emergency conditions and peer influences and they are also surface-level direct influencing factors. It’s very important for governing agencies to put focus on these root causes while promoting the application of biopesticides to achieve a promising outcome.(3)Some of the key reasons for the deviation of farmers’ willingness and behaviors are listed below: farmers’ education level is generally low, farmers are not much concerned about the quality and safety of agricultural products, farmers’ lack knowledge and expertise about the characteristics of biopesticides and the hazardous effects from chemical pesticides. Farmers are constrained by economic conditions and their purchasing power is quite limited in terms of biopesticides procurement. In addition, the lack of publicity and incomplete construction of markets for biopesticides have led to farmers having difficulties distinguishing between biopesticides and chemical pesticides.There are still many farmers using chemical pesticides and there are many deviations between their willingness and behaviors in the application of biopesticides, so it is not promising to promote biopesticides as an alternative to chemical pesticides on a full scale. The divergence between the willingness and behaviors of rice farmers to apply biopesticides is influenced by various factors. In terms of individual and family characteristics, both education level and scales of rice planting have a negative effect on it. In terms of farmers’ awareness, biopesticide awareness, awareness of hazardous effects from chemical pesticides and quality and safety awareness of agricultural products have a negative effect on the deviation. The better the farmers’ awareness towards environment protection, the lower the possibility of the deviation to take place. As for external factors, peer influences, emergency conditions and price affordability have positive effects on the occurrence of deviation.The logical hierarchy of influencing factors can be summarized as a “single path with three drivers” model. Biopesticide awareness is at the surface level, awareness of chemical pesticides’ hazards and awareness of agricultural quality and safety are indirect factors at the mid-level while the farmers’ characteristics such as education level and planting scales are root causes. The three drivers refer to external factors such as price affordability, emergency conditions and peer influences and they are also surface-level direct influencing factors. It’s very important for governing agencies to put focus on these root causes while promoting the application of biopesticides to achieve a promising outcome.Some of the key reasons for the deviation of farmers’ willingness and behaviors are listed below: farmers’ education level is generally low, farmers are not much concerned about the quality and safety of agricultural products, farmers’ lack knowledge and expertise about the characteristics of biopesticides and the hazardous effects from chemical pesticides. Farmers are constrained by economic conditions and their purchasing power is quite limited in terms of biopesticides procurement. In addition, the lack of publicity and incomplete construction of markets for biopesticides have led to farmers having difficulties distinguishing between biopesticides and chemical pesticides.Through the formulation of policy and measures on major influencing factors, the conversion of willingness to behaviors can be improved hence reducing the deviation of behaviors from willingness. A few suggestions have been made to relevant departments and local governments based on the results from this study:
+		(1)It is extremely important to improve the expertise level of the farmers, to reduce the constraints of farmers’ resource endowment and to promote the conversion of farmers’ willingness to apply biopesticides into behaviors. It’s also necessary to enhance the education level of rural farmers through face-to-face coaching sessions and education on fields for farmers with low education level. In this approach, a new generation of young professional farmers can be cultivated with better agricultural expertise level. The promotion of biopesticides should also be focused such as the development of differentiated promotion programs for farmers of different planting scales in different regions.(2)It is also recommended to strengthen the publicity of the ideas of green production and to raise the cognition level of farmers towards green production. On one hand, publicity and promotion work for biopesticide popularization through television, Internet and other social medias and face-to-face coaching can strengthen farmers’ understanding of green production and green transformation of agricultural production. On the other hand, it is necessary to deepen the farmers’ perceptions of green agricultural production experiences by carrying out special environmental protection activities such as organizing visits to green production demonstration projects and establishing green production demonstration households. In this way, farmers’ sense of responsibility to protect the environment in agricultural productions can be improved.(3)Finally, it is crucial to speed up the establishment of the biopesticide market and to optimize the policy mechanisms and enforcement of biopesticide use. At present, farmers are facing the problem of selecting from various types of pesticides, which makes it difficult for farmers to distinguish between biopesticides and chemical pesticides. This phenomenon reminds us that attention should be paid to improving the identifiability of biopesticides at pesticide distribution sites hence reducing the extra identification costs for farmers. At the same time, price affordability is also one of the major concerns of farmers. The price of biopesticides need to be regulated to a relatively acceptable range through improved subsidy schemes and promotions. Moreover, subsidy schemes and promotions need to be made known to the public to obtain satisfaction from the farmers in order for them to have confidence in the application of biopesticides.It is extremely important to improve the expertise level of the farmers, to reduce the constraints of farmers’ resource endowment and to promote the conversion of farmers’ willingness to apply biopesticides into behaviors. It’s also necessary to enhance the education level of rural farmers through face-to-face coaching sessions and education on fields for farmers with low education level. In this approach, a new generation of young professional farmers can be cultivated with better agricultural expertise level. The promotion of biopesticides should also be focused such as the development of differentiated promotion programs for farmers of different planting scales in different regions.It is also recommended to strengthen the publicity of the ideas of green production and to raise the cognition level of farmers towards green production. On one hand, publicity and promotion work for biopesticide popularization through television, Internet and other social medias and face-to-face coaching can strengthen farmers’ understanding of green production and green transformation of agricultural production. On the other hand, it is necessary to deepen the farmers’ perceptions of green agricultural production experiences by carrying out special environmental protection activities such as organizing visits to green production demonstration projects and establishing green production demonstration households. In this way, farmers’ sense of responsibility to protect the environment in agricultural productions can be improved.Finally, it is crucial to speed up the establishment of the biopesticide market and to optimize the policy mechanisms and enforcement of biopesticide use. At present, farmers are facing the problem of selecting from various types of pesticides, which makes it difficult for farmers to distinguish between biopesticides and chemical pesticides. This phenomenon reminds us that attention should be paid to improving the identifiability of biopesticides at pesticide distribution sites hence reducing the extra identification costs for farmers. At the same time, price affordability is also one of the major concerns of farmers. The price of biopesticides need to be regulated to a relatively acceptable range through improved subsidy schemes and promotions. Moreover, subsidy schemes and promotions need to be made known to the public to obtain satisfaction from the farmers in order for them to have confidence in the application of biopesticides.Conceptualization, H.G. and C.P.; methodology, H.G. and F.S.; software, H.G. and F.S.; validation, H.G., F.S. and Y.L.; formal analysis, H.G. and B.Y.; investigation, H.G., F.S. and Y.L.; resources, H.G.; data curation, Y.L. and B.Y.; writing—original draft preparation, H.G.; writing—review and editing, F.S. and C.P.; visualization, H.G.; supervision, C.P.; project administration, H.G.; and funding acquisition, H.G. All authors have read and agreed to the published version of the manuscript.This research was funded by Research on Science and Technology Strategy and Planning of Science and Technology Department of Jilin Province, China, grant No.20200101130FG; Key Scientific and Technological Projects of Science and Technological Department of Jilin Province, China, grant No.20190301080NY; MOE (Ministry of Education in China) Project of Humanities and Social Sciences, grant No.18YJC630128; Social Science Fund Project of “the 13th Five-Year” of Education Department of Jilin Province, China, grant No.JJKH20190736SK.Not applicable.Not applicable.Data sharing not applicable.The authors declare no conflict of interest.Driving Factor Hierarchy T Diagram.Interpretative Structural Model of Influencing Factors.Basic Characteristics of the Samples.Pesticide Application by Farmers.Variables of the Model and Descriptive Statistics.Simulation Results of Regression Model.Note: ***, ** and * indicate that the coefficients of the explanatory variables are significant at the 1%, 5%, and 10% levels, respectively.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06027.txt ADDED Viewed

	@@ -0,0 +1,13 @@

+Authors have contributed equally to this work.Polyvinyl alcohol (PVA) is a water-soluble plastic commercially used in laundry and dish detergent pods (LDPs) for which a complete understanding of its fate in the environment and subsequent consequences is lacking. The objective of this study was to estimate the US nationwide emissions of PVA resulting from domestic use of LDPs, corroborated by a nationwide, online consumer survey and a literature review of its fate within conventional wastewater treatment plants (WWTPs). Peer-reviewed publications focusing on the degradation of PVA in critical processes of WWTPs were shortlisted as a part of the literature review, and subsequent degradation data was extracted and applied to a model with a set of assumptions. Survey and model results estimated that approximately 17,200 ± 5000 metric ton units per year (mtu/yr) of PVA are used from LDPs in the US, with 10,500 ± 3000 mtu/yr reaching WWTPs. Literature review data, when incorporated into our model, resulted in ~61% of PVA ending up in the environment via the sludge route and ~15.7% via the aqueous phase. PVA presence in the environment, regardless of its matrix, is a threat to the ecosystem due to the potential mobilization of heavy metals and other hydrophilic contaminants.Plastic pollution has been steadily increasing since the 1950s [1]. Due to an upsurge in public awareness regarding plastic usage and pollution, more “sustainable” alternatives have increased in popularity, and are thus being utilized in higher quantities by the general public [2]. These new materials are often marketed as “biodegradable”, as they are considered to be susceptible to microbial degradation under specific conditions, but this specificity often makes it difficult to understand their ultimate fate in the environment. Polyvinyl alcohol and its corresponding blends (PVOH, PVAI, Polyviol, Alcotex, Covol, Gelvatol, Lemol, Mowiol, Mowiflex, and Rhodoviol) are examples of polymers that have become more popular both in usage and within scientific research (see Figure 1A) due to their water-solubility. Typically, PVA is used as a protective film for laundry and dish detergents; as a sizing and finishing agent in the textile industry [3]; and as a thickening or coating agent for paints, glues, meat packaging, and pharmaceuticals in paper and food industries (see Figure 1A) [3].Up to 650,000 tons of PVA is produced yearly across the globe [4] and this has been expected to increase 4.09% annually from 2018 to 2023 [5]. In 2018, due to its general increase in usage, PVA was considered to be one of the most ubiquitous pollutants in wastewater [4,6,7]. A thorough understanding of its path to and breakdown within the environment is presently lacking. Although water-soluble, its constituents, such as ethylene (a petroleum-based product), can remain intact within the solvent. Studies have shown ethylene to have negative effects on surrounding organisms, such as plants, which naturally produce and utilize ethylene [8]. Similar to table salt and sugar, PVA dissolves in water, and if the water volume is low, a viscous solution will be formed. The high water volume in WWTPs means the texture of the water should remain unchanged. When PVA is discharged into water bodies, it has the ability to foam due to its surface properties [9]. This can inhibit oxygen transfer, causing irreparable harm to aquatic life [10]. Additionally, because of its hydrophilicity, PVA has the potential to adsorb dangerous chemicals or contaminants [11], such as antibiotics [12] or heavy metals [13,14,15], at high concentrations. These can then concentrate up food chains [16], posing a threat to the environment, similar to behavior of traditional polluted plastics. WWTPs are known to contain a variety of dangerous contaminants, creating a higher-risk situation for PVA particles passing through [17].The PVA used in LDPs is composed of PVA polymeric chains with a fraction of polymeric acetate groups (see Figure 1B). This is referred to as partially hydrolyzed PVA, and the percentage of hydrolysis and molecular weights vary with its application [18,19]. The PVA used in LDP blends is typically 88% hydrolyzed [18], while its molecular weight can vary within several ranges, including but not limited to 1000–1,000,000, 10,000–300,000, and 20,000–150,000 Da [19]. Upon contact with water, the presence of polymeric acetate groups enables the water molecules to penetrate the bonds of the outer coating and break it into smaller chains. Once flushed down the drain, partially hydrolyzed PVA chains (See Figure 1B) enter wastewater channels, eventually interacting with WWTPs. The fate of PVA in wastewater treatment systems has been partially explored, with some studies highlighting specific aspects of the wastewater treatment process, but very few studies aimed to establish a complete degradation estimate in conventional WWTP processes from beginning to end.Available studies suggest that the degradation of PVA occurs under a specific set of circumstances, which may not be ubiquitous within WWTPs or the natural environment. Ultimately, PVA degradation is reported to be a slow process, which greatly depends upon the surrounding conditions mentioned in Table S2 [20]. Bacteria utilize enzymes to degrade PVA to its constituent form by attacking specific bonds within the polymeric chain [7,21]. Bacteria oxidizes the tertiary carbon atoms, leading to the endo-cleavage of PVA molecules, one of the main degradation routes, which leads to the creation of byproducts such as hydrolyzable hydroxy ketone and 1,3-diketone [20,22]. Other microorganisms mainly utilize PVA as a carbon source; such is the case with the bacteria Pseudomonas [20], which generate hydrogen peroxide and other byproducts, including a lower molecular weight PVA [23,24]. Many of these processes can take place simultaneously to begin the degradation of the polymer [20]. While several bacterial species have been documented degrading PVA, these are infrequently found within conventional WWTPs or the environment [20], as are the other optimal circumstances necessary for PVA to completely degrade (Table S1). WWTPs are predominantly designed to remove suspended solids, harmful bacteria, and pollutants of emerging concern [25], but the removal of PVA is still in question due to a lack of comprehensive research. Independent studies have been performed demonstrating removal of PVA using different bacteria, enzymes, and chemical processes [20], but a complete fate assessment in a conventional WWTP or the environment is currently lacking.In this study, we present a detailed qualitative, quantitative, and spatial analysis on annual mass emissions of PVA in the US, consisting of three defined segments: (i) an online survey involving 500+ respondents investigating their laundry and dish detergent purchasing habits (ii) a United States Geological Survey (USGS) report on water usage, sources, and population; (iii) and a detailed literature review on the degradation of PVA in wastewater, utilizing the data to conduct a subsequent mass balance analysis. An online survey was conducted with 527 respondents, 80% female and 20% male, ages 24–55 years. The survey was directed at the primary decision-maker responsible for purchasing cleaning products in the household. In total, 60% of the responses were from the top 20 designated market areas (DMAs). Survey questions mainly investigated the frequency of buying laundry and dish detergent and the type of detergent bought (single dose pods or detergent bottle) (see Supplemental Information). State wise data on water use and application for the United States was published in the USGS report titled “Estimated Use of Water in the United States in 2015” [26]. Treatment plants in the US receive wastewater from public facilities, domestic sources, and industrial effluents. Hence, the total water used that ultimately resulted in the generation of wastewater was the sum of public, domestic, and industrial supply (Mgal/d) as shown in Equation (1) which assisted in populating Figure 2.
+        (1)WUse,2015=WP+WD+WIIn this equation, WP, WD, and WI stand for public, domestic, and industrial water supplies, respectively.Not all water used is received by WWTPs as wastewater; there are an estimated 20%–25% in losses [27]. which include but are not limited to leakages of sewers and sanitary sewage outflows. The total wastewater generated was calculated using Equation (2), assuming a 20% loss in volume. In this case, WWG,2015 stands for the wastewater generated for each state.
+        (2)WWG,2015=0.8∗WP+WD+WIRoughly 16,000 US WWTPs treat 34 billion gallons (BGD) of wastewater daily [28]. This represents an average of approximately 2.12 billion gallons treated daily per facility. Wastewater treated for each state can be calculated using Equation (3), where NF and WWT,2015 represent the number of wastewater treatment facilities in each state and amount of wastewater treated, respectively, assuming a complete use of operational treatment capacity.
+        (3)WWT,2015=NF∗2.12 BGDThe untreated wastewater was calculated as the difference between generated and treated wastewater, as shown in Equation (4). Untreated wastewater could be a result of several possibilities, including but not limited to the lack of connectivity between households and WWTPs, improper disposal of wastewater generated, and/or generated wastewater exceeding the operational capacity of a WWTP (current assumption).
+        (4)WWUT,2015=WWG,2015−WWT,2015Subsequently, the percentage of wastewater untreated and treated were calculated as shown in Equations (5) and (6) respectively.
+        (5)WWUT,%=WWUT,2015WWG,2015 ∗ 100
+        (6)WWT,%=WWT,2015WWG,2015 ∗ 100Laundry pods from three brands and dish pods from two brands were drained, air-dried overnight, and weighed on an Ohaus Adventurer weighing scale (AR1530, China). Laundry pods from different brands (n = 9) and dish pods from varying brands (n = 6) were weighed in triplicate. The average weights for laundry (ML,Avg) and dish (MD,Avg) pods were 1.0 ± 0.6 g and 0.5 ± 0.2 g, respectively. Data from the online survey revealed consumption of ~15 billion laundry pods (NL,Avg) and ~12 billion dish pods (ND,Avg) per year by 126 million households in the US. Using 2015 state-wise population numbers from the USGS report [29] (P2015), the number of per capita pods consumed in the US (NPC,Avg) was calculated using Equation (7).
+        (7)NPC,Avg=NL,Avg+ND,AvgP2015The outer dissolvable coating is composed of PVA and other additives in varying proportions. Based on patents and reports, the PVA ratio (by weight) lies between 65% and 99% of the total outer coating weight [30]. An average mass of PVA in laundry and dish pods was calculated using Equation (8), where fL and fD are the fractions of PVA in laundry and dish pods, respectively, in grams.
+        (8)MPVA,Avg=ML,Avg∗fL+MD,Avg∗fDThe number of pods (laundry and dish) used by each state was calculated as per Equation (9).
+        (9)NPod=NPC,Avg∗P2015The mass of discarded PVA from laundry pods (ML,G) and dish pods (MD,G), in each state was expressed as per Equation (10).
+        (10)MPVA,G=NPod∗MPVA,AvgUntreated and treated masses of PVA were calculated as per Equations (11) and (12), respectively.
+        (11)MUT,PVA=MPVA,G∗ WWUT,  %
+        (12)MT,PVA=MPVA,G∗ WWT,  %Applying total degradation percentages (solid + aqueous phase) from the modeled scenario to the PVA treated (MT,PVA) would result in total PVA emissions from WWTPs.Figure 3 was created in Arc GIS pro 2.0.0. Data specific to the US states and environmental emissions of PVA (metric tons/yr) were collected from outside sources. These were then imported into the GIS software program ArcMap. A comprehensive literature review was performed on the presence and degradation of PVA in US wastewater according to PRISMA guidelines. A detailed breakdown of the publication selection is presented in SI Figure S1. Databases accessed included Google Scholar and Scopus, utilizing the search terms polyvinyl alcohol*; PVA*; Polyvinyl alcohol with results further specified using the constraining term (AND): Pollution*; Degradation*; Biodegradation*; Wastewater*; WWTP*; Sludge*; Sewage*; Effluent*; Influent*; Activated sludge*; US. Papers published between 1950 and 2020 were sought out. These studies were specifically sought out as they aimed to assess the breakdown of PVA within a conventional WWTP by microbial attack or other forces, such as UV radiation or chlorine oxidation. From each relevant paper, the degradation mechanism was extracted as well as the percent degradation, WWTP section (if applicable), the microorganism species, and whether or not the species was adapted to the wastewater itself. As many of the papers included several independent studies relating to PVA degradation in wastewater, a range of percentages were then included to encapsulate all reported values within a single manuscript. Papers with incomplete text or a lack of relevance to wastewater were excluded (see Figure S1).The majority of the PVA, generally in the form of greywater from domestic, public, and industrial sources, reaches WWTPs, where it encounters primary treatment, secondary treatment, sludge treatment, and disinfection before leaving the WWTP. Primary treatment consists of initial screening, grit removal, and a primary clarifier, with the objective of removing coarse solids and other large items [31]. PVA has not been studied in the context of its partitioning and removal in primary treatment within a conventional WWTP. Degrading PVA is a challenge and can contribute to the total chemical oxygen demand (COD) [32] from the incoming wastewater. The most influential mechanism of removal in a primary clarifier is the sorption to the suspended solid particles [33]. However, PVA is a hydrophilic polymer and has a greater affinity toward water, tending to stay in the liquid phase as opposed to solid [34]. It is possible that PVA could form gelatinous consistencies with fats and lipids from influent wastewater, which may increase its partition toward solid matter and its removal from the aqueous phase. An empirical study analyzing hydrophilic pharmaceuticals and endocrine-disrupting chemicals in wastewater revealed the low sorption capacity of hydrophilic contaminants to solids, requiring relatively longer retention times (5–11 h) for their efficient removal from the aqueous phase [35]. Current literature evidence is not sufficient to establish a definitive path PVA may take in a primary treatment system, but the COD removal efficiency of the primary clarifiers likely eliminates PVA from the aqueous phase [36]. Secondary treatment is designed to further remove organics and suspended solids [31]. Two main components of secondary treatment are activated sludge processing and the use of the secondary clarifier. The activated sludge process involves the recycling of the bacteria/microbes present in the sludge, which are sent back into the aeration chambers. These microbes are responsible for the degradation of organic waste and chemicals from the influent wastewater. The bulk of PVA degradation is anticipated to occur at this stage. There are several key factors influencing the biodegradation of PVA in an ASP. One important element is the food-to-microorganisms ratio (F:M), defined as the load of substrate applied daily per unit of biomass [37]. A general F:M ratio lies anywhere between 0.25 and 0.45 [38]. However, an F:M ratio better suited for PVA degradation lies between 0.1 and 0.15 [39]; thus, a higher number of microorganisms is required to fully degrade PVA, compared with conventional domestic sludge. Typical microorganisms may not be able to adequately break down PVA, as the presence of PVA-adapted microbes is necessary for thorough biodegradation of PVA in the ASP [22,39,40]. PVA adaption often requires a lag time spanning over several weeks [22,39], and the ASP can only be PVA-adapted in WWTPs receiving a heavy influx from textile industries, allowing sufficient time for the microbes present at the facility to adapt to the incoming COD. After the lag phase is complete, biodegradation thereafter occurs exponentially [39]. The number of PVA-degrading microbes is limited, and their presence is specific to certain environments and environmental conditions. Schonberger et al. discovered that WWTPs consistent with PVA adaption within the ASP achieved 80% degradation in a 7-day period. Alternatively, the unadapted ASP was consistent with only 18% biodegradation within the same time frame [39]. Hoffman et al. observed similar findings when four different blends of PVA were tested in adapted and unadapted sludge. In unadapted sludge, a degradation of 88 ± 9% of incoming PVA was achieved in 187 ± 25 h for four blends, and ~20% was degraded in 25 h [22]. For adapted sludge, 90 ± 5% was degraded in just 29 ± 2 h [22]. The above findings utilized 88% hydrolyzed PVA, the same composition as that used in detergent pod coatings [30,41]. The average hydraulic retention time (HRT) in the ASP is approximately 18–24 h, and the sludge retention time (SRT) is 12–15 days [42]. Due to the hydrophilic nature of PVA, the majority of PVA is expected to be in the water phase, in which the HRT would subsequently play a larger role in its degradation.The decomposed sludge mixture then enters the secondary clarifier, where the solid waste is given time to settle, allowing the liquid to enter the tertiary treatment stage. In the secondary clarifier, three removal mechanisms may occur, which include volatilization, biodegradation, and adsorption to solids [33]. The current literature does not present evidence of volatilization, biodegradation, or adsorption of PVA in secondary clarifiers. However, since the majority of secondary sludge contains fluid, and the density of PVA is ~1.2 g/cm3, a fraction of PVA may settle and exit via secondary sludge. Tertiary treatment typically consists of a disinfection chamber and a filtration unit. Older WWTPs use chlorination as an effective disinfectant, along with trickling or gravity filters, due to their moderate efficiency and low cost [43]. These technologies are evolving, as most modern WWTPs rely on advanced oxidation processes (AOPs) and membrane filtration, according to the United States Environmental Protection Agency (USEPA) [44].In 2017, Ye et al. studied the comparative effects of using UV, chlorine, and a UV-chlorine combination in PVA degradation/oxidation [45]. Their experiments revealed that 20 min of UV irradiation (Is = 2.6 mW/cm2) had a minimal effect on the initial concentration of PVA (50 mg/L). Chlorine had a similar result, with just 1.5% of original PVA degraded in 20 min. Chlorine conditions of 20 mg Cl2/L and a pH of 7 are synchronous with WWTP disinfection conditions in the US [46]. However, a combination of UV-chlorine treatment resulted in 92% degradation within 10 min and 100% in 20 min. Advanced oxidation technologies have been documented to eliminate PVA completely from wastewater in a matter of minutes [15]. These processes can be expensive and thus not cost-effective for the municipality. In 2004, the US EPA reported that chlorination is the most utilized method for wastewater disinfection (EPA primer, 2014). Given the degradation conditions under chlorine alone, only a minimal amount of PVA coming from domestic wastewaters can be expected to be degraded.Sand filtration is one of the most widely used filtration techniques due to its low cost of operation and maintenance [47]. Research on the behavior of PVA in a sand or trickling filter system is lacking. However, in cases of missing data, its fate can be predicted based on the behavior of other commonly used hydrophilic materials, such as pharmaceuticals with an octanol-water partitioning coefficient (log Kow) below 2 (see Table 1). A low log Kow indicates the compound’s affinity toward water and categorizes the compound as hydrophilic in nature. The table below demonstrates the very low removal efficiencies of widely known hydrophilic pharmaceuticals in a sand filtration system. It should be noted that sand filtration had minimal to no effect on the concentrations of the compounds studied, with a mean and standard deviation of −0.08% ± 20%. It is likely that sand filtration would not have a significant effect on PVA concentrations. Wastewater activated sludge coming from secondary clarifier effluent is treated in an anaerobic digester to reduce the overall volume, destroy pathogens, and control odors [50]. Using anaerobic microbes, the organic matter is further broken down, releasing methane and water as byproducts. This treated sludge, termed “biosolids”, is then safe to be used for agricultural purposes, as an example. In the US, anaerobic digesters are typically catered towards mesophilic (30–37 °C) and thermophilic (50–60 °C) bacteria with retention times of 12–25 and 10–12 days, respectively [50]. Matsumura et al. explored PVA degradability using a lab-scale anaerobic digester with activated sludge obtained from a local WWTP in Japan [51]. Two drastically different molecular weights (14,000 and 2000 Da) were selected for the study. In the first 25 days, both PVA blends showed similar biodegradation results of 12.5%. As time progressed, PVA-14,000 degraded at a much higher rate compared with PVA-2000. After 150 days, PVA-14,000 reached 50% degradation, while PVA-2000 was only 37% degraded [51]. Such low biodegradation rates could be attributed to the usage of PVA-unadapted sludge. Different PVA-starch blends were studied by Russo et al. under anaerobic conditions. They studied starch to PVA ratios of 90:10, 75:25, 50:50, and 0:100. Anaerobic conditions of 38 ± 5 °C with microbes sourced from activated sludge processes were maintained under nitrogen for up to 100 h. Because PVA used in detergent pods does not contain any starch additives [30], the results of the starch to PVA blend of 0:100 most closely mimicked detergent-derived PVA behavior. The solubilization study yielded lower results for the 0:100 blend, with only 10% of the PVA able to solubilize under anaerobic conditions in 100 h. Its 90:10 counterpart was solubilized up to 60% in the same time allotment [52]. Anaerobic digestion efficacy is often assessed based on the amount of methane and carbon dioxide produced as a byproduct. Upon isolating methane and CO2 production analysis, the 0:100 blend produced less than 5 mL/g COD, whereas the 90:10 blend produced ~40 mL/g COD [52]. These results highlight the inefficiency of traditional anaerobic digestion methods for the biodegradation of PVA. Another study analyzing the breakdown of PVA-glycerol-starch blends in anaerobic digestors was conducted by Pšeja et al. in 2006 [53]. After 30 days of incubating PVA blends at 35 ± 2 °C with anaerobic bacterial cultures, the percentage of biodegradation was assessed based on the balance of carbon, biogas, as well as the liquid phase [53]. The bacterial inoculum was sourced from unadapted municipal sludge of a local WWTP. A total of 13 blends were studied, including 85%/15% PVA-glycerol, 70%/15% PVA-starch, and a 75%/15%/10% PVA/starch/glycerol blend. The biodegradation percentages varied from 4.1% to 19.8%. The 85%/15% PVA-glycerol blend degraded the least, at 4.1% in 30 days, whereas the 75%/15%/10% PVA/starch/glycerol blend degraded the most, at 19.8% [53]. According to the authors, high degradation rates (high carbon differential) can be attributed to the biodegradation of starch, not the PVA itself. Since detergent pods do not contain starch additives, high biodegradation rates may not be observed during anaerobic digestion. The mass balance presented here is a combination of degradation percentages adopted from Section 3.1, Section 3.2, Section 3.3, Section 3.4, Section 3.5 and Section 3.6, as well as a WWTP scenario assumed by Garrido et al. in 2013 [54]. Additionally, WWTP assumptions similar to those of Garrido et al. regarding primary and secondary clarifier COD removal efficiencies were also adopted. A variety of scenarios can be suggested based on a high number of variations in treatments across the US. However, many US WWTPs are older in age and rely on conventional treatment technologies. A conventional activated sludge facility (completely mixed) with primary treatment (screening, grit removal, and primary clarifier), secondary treatment (aeration basins, secondary clarifier, and activated sludge process), tertiary treatment (disinfection and sand filtration), and anaerobic digester was assumed for this study. The COD removal efficiency of the primary clarifier was fixed at 30% and 75% for the secondary clarifier [54]. PVA was considered to be a part of the total COD in this model, and the efficiencies signifying the percentage of the partitioning of COD within the clarifiers were directly applied to PVA. Table 2 indicates the treatment section and the corresponding degradation percentages for the SRT, HRT, and other process conditions based on the literature review within this study. PVA in the influent wastewater entering primary treatment is considered to be 100%. As PVA passes the primary clarifier, 30% is expected to partition into the solid phase and eventually reach the anaerobic digester. The remaining PVA (70%) in the aqueous phase then enters the activated sludge system (see Table 2). The microbial activity in the aeration basins degrades 20% of the PVA in the presence of unadapted sludge during 18–24 h of HRT. Residual PVA (64%) enters the secondary clarifier, where 75% (48% of total) is partitioned into the sludge phase, and 25% (16% of total) is carried over via the liquid phase to the sand filters. The 48% in the sludge phase is further divided into 21% (10% of total) as return activated sludge (RAS) and 79% (38% of total) as waste activated sludge (WAS). Therefore, 10% of the total PVA is assumed to be in the form of return activated sludge (RAS). RAS values are adopted from the EPA report, in which RAS flow varied between 15% and 127% of the secondary influent flow, with the number eventually settling in the lower 20s [55]. RAS and primary effluent amount to 80% of the total PVA entering the aeration basin.Solid phase effluent from the primary clarifier (30%) and WAS (38%) are treated in an anaerobic digester, where 10% is degraded (6.8% of the total) and 61.2% of the total is left untreated and ready to be landfilled, land applied, or incinerated. The aqueous phase, containing 16% of the total PVA reaching the sand filtration stage, enters unaltered into the disinfection basin, where 1.5% is degraded (0.24% of the total) and 15.76% remains intact within the aqueous phase. A detailed mass balance can be found in Figure 2. References for degradation percentages and respective retention times are listed in Table 2.Referencing the above model, it is estimated that ~61.2% of PVA is emitted via sludge, and ~15.7% is emitted through effluent, with ~77% of the PVA still intact after passing through conventional wastewater treatment (see Figure 2). Once PVA has passed through conventional water treatment, either untreated or within sludge, its journey and fate become important to understand. By accessing data relevant to sewage treatment in US states, we were able to project the amount of intact PVA emitted by each US state. Our research indicates that the Midwest of the US has the lowest amount of untreated PVA, possibly due to lower populations and fewer population-dense areas (see Figure 3, left panel). The Southern and some Western areas of the US have lower to moderate volumes of PVA emissions. However, states such as California, Florida, New York, and Pennsylvania have the highest loadings via untreated wastewater. This may be due to the presence of more metropolitan cities and overworked public facilities as populations grow, causing urban expansion [56].PVA emissions via WWTP effluent demonstrated slightly different data, with certain regions remaining the same (see Figure 3, right panel). Similar to untreated PVA emissions, much of the Midwest had lower emissions, with the South showing slightly higher numbers, and larger states with high treatment capacities, such as Texas, California, New York, and Florida, having the highest loadings via effluent. It is worth noting that most of the states with the highest PVA emissions, either untreated or via effluent, have coasts bordering the Pacific or Atlantic Oceans, suggesting a quicker release of PVA into aquatic or marine ecosystems. Previous research has detected wastewater-derived contaminants in the ocean [57], and similar methods could be used to trace the presence of PVA within surrounding marine or terrestrial ecosystems. As human populations and their LDP usage continue to increase, it is expected that wastewater-derived contaminants will also increase [57].Our data suggest that, on average, only ~10,500 ± 3000 mtu/yr (See Figure 4) of PVA enters treatment infrastructure, and only a fraction of this is biodegraded due to the specificity of conditions required to facilitate complete degradation. Based on the assumed WWTP scenario, 15.76% remains in the aqueous phase (~1600 ± 500 mtu/yr) and 61.2% (6500 ± 1900 mtu/yr) remains in the biosolids exiting the anaerobic digester. Thus, a total of 8100 ± 2400 mtu/yr of PVA is estimated to remain untreated by WWTPs annually in the United States. Of that, 6500 ± 1900 mtu/yr of PVA remains untreated due to lack of treatment capacity or inaccessibility to a functioning WWTP in certain remote communities.Once biosolids leave a WWTP facility, 50–60% are applied to agricultural lands, 20% are sent to be incinerated, and 17% are sent to a landfill [58]. Each of these locations carries environmental risks associated with the distribution of plastics from biosolids. Initial research chronicling the negative impact of environmental PVA and future areas of study are listed in the following section. The pathways of sludge have been well documented, as has the ability of WWTPs to act as sources of contaminants and microplastics entering the environment. These emissions can cause deleterious impacts on surrounding ecosystems and the biota within them. PVA that passes through conventional water treatment can similarly pose a threat to the environment in several ways, once released into the environment or if land applied Our data suggest that around 3500 mtu/yr of PVA is sequestered within agricultural soils in the US. As mentioned previously, ethylene is a byproduct of PVA degradation and is also a hormone utilized by plants. It is unknown if ethylene derived from PVA could affect agricultural yields, but it warrants investigation. The ability for plastic particles to adsorb dangerous contaminants at high concentrations has been documented, but this research is currently lacking as it pertains to PVA. Initial studies revealed that PVA can alter gas exchanges, such as carbon dioxide exchange, affecting aquatic ecosystems [20]. It is also capable of leaching into the groundwater, and it has even been documented to mobilize heavy metals from sediments to water resources [59,60,61]. Hydrophilic compounds, such as biocides, insecticides, herbicides, flame retardants, corrosion inhibitors, personal care products, and pharmaceuticals are present in wastewater and stormwater [62]. Some of these are proven carcinogens [63] with great aqueous phase stability. As the sorption of organic and inorganic pollutants is not limited to hydrophobic compounds but can also occur with hydrophilic compounds, PVA could act as a vector for transport up the food chain, similarly to more conventional plastics. During such phenomena, the contaminant concentrates, increasing its level of toxicity [61]. This area requires additional research in order to further elucidate the impact of intact PVA on the natural environment.Prior research has demonstrated that WWTPs are sources of microplastic pollution in natural and built environments. This is due to the fact that microplastics in treated sludge, termed “biosolids”, can have a variety of harmful effects on ecosystems beyond contaminant adsorption. A sizable fraction of biosolids is deposited on agricultural soils as they serve as a rich form of fertilizer, ultimately improving soil properties [64]. If biosolids are contaminated with microplastics, the particles destabilize the benefits of sludge by negatively affecting microbial activity, bulk density, and water holding capacity of the soils [57]. A portion of biosolids is also sent to landfills around the US. It is thought that this process sequesters materials in the long term, but landfill leachate has been found to carry microplastics, prompting the consideration of landfills as a source of microplastics into the environment [65]. Lastly, a portion of US biosolids is incinerated. Research has shown that incineration does not terminate plastic waste completely. On the contrary, residual ash can be considered a potential source of microplastic release into the atmosphere or environment [66]. If the plastics are completely incinerated, this process can produce airborne contaminants or pollutants [66]. As demonstrated, incomplete PVA breakdown within conventional water treatment results in a fraction of the material being sequestered within biosolids. The effects and behavior of residual PVA particles within biosolids are not well understood. More research is required to determine their impact on the environment relative to other, more conventional, plastics, whose physical presence in biosolids and ability to adsorb dangerous contaminants creates a threat to ecosystems.In summary, this research aimed to isolate trends within the current industrial output of PVA used for laundry or dish detergent pods in US wastewater; investigate the components; assess the biodegradability, solubility, and bacterial effect on its structure; and, lastly, outline the potential risks PVA poses as an environmental pollutant. We observed that PVA has low degradation rates within WWTPs; thus, its hydrophilicity and massive production numbers make it a cause for concern as a pollutant in the natural environment. Very little research exists that aims to monitor the biodegradability of PVA in the natural environment. This presents a challenge in determining its role or impact as a pollutant. Research into truly eco-friendly substitutes for PVA is warranted and should be further explored. Improving upon this research is essential for better understanding the link between PVA usage, and public and environmental health.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116027/s1, Literature review, survey questions, discussion of PVA degrading bacteria, Figure S1: Flow chart flow diagram, Table S1: Bacteria types and percent PVA degradation, Table S2: Soil PVA degradation rates, Table S3: Hydraulic retention times and solid retention times for WWTP segments, PVA Toxicity analysis.Conceptualization, C.R. and V.K.; methodology, C.R. and V.K.; software, V.K.; validation, C.R. and V.K.; formal analysis, C.R. and V.K.; investigation, C.R. and V.K.; writing—original draft preparation, C.R. and V.K.; writing—review and editing, C.R. and V.K.; visualization, C.R. and V.K.; funding acquisition, C.R. All authors have read and agreed to the published version of the manuscript.This research was funded in part by Blueland.Not applicable.Not applicable.The data presented in this study are available in the supplementary material.This work was supported in part by Blueland. Their sponsorship and critical input greatly assisted with the study. We thank Shireen Dooling for her assistance with graphic design.The funders had no role in the design of the study; collection, analyses, or interpretation of data; writing of the manuscript; or decision to publish the results.(A) Increasing number of publications per year focusing on polyvinyl alcohol (PVA) as well as a pie chart depicting percentage distribution of PVA applications [3] and (B) the chemical structure for partially hydrolyzed polyvinyl alcohol-acetate.Mass balance of PVA in a conventional activated sludge treatment plant, considering clarifier efficiencies and biodegradation efficiencies. Numbers in red indicate the percentage of PVA in respective treatment streams, and numbers in green represent the amount (% absolute) of degraded PVA in respective sections. RAS and WAS represent return activated sludge and waste activated sludge, respectively. Numbers in parentheses represent the degradation efficiencies of respective sections.PVA emissions across the U.S. in mtu/yr. The left panel is the spatial distribution of untreated PVA via wastewater that does not reach the treatment plants. The right panel represents the PVA from WWTP effluent streams, including aqueous and sludge disposal routes.Modeled PVA usage and emissions in metric tons per year (mtu/yr) in the US.Removal efficiencies of hydrophilic pharmaceuticals and their log Kow factors in filtration systems.The treatment section, corresponding degradation percentages, SRT, HRT, and other process conditions in a conventional sewage treatment plant. Other processes that do not contribute to degradation are excluded from this table.NA: Not applicable.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06028.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: elementary schoolteachers play a central role in shaping their students’ beliefs, attitudes, and behaviours related to health and oral hygiene. This study was designed to evaluate Turkish schoolteachers’ levels of oral health knowledge, attitudes, and behaviours. Methods: A cross-sectional survey-based study was conducted among elementary schoolteachers in Istanbul using a validated self-administered questionnaire. The questionnaire was composed of 36 multiple-choice items categorised into six sections, and the participants were recruited using convenience sampling. (3) Results: A total of 385 elementary schoolteachers participated in this study. The majority were female (62.2%), qualified with a licensure degree (81.3%) and working in public schools (86.5%). Female gender and greater work experience were found to be promoters for oral health knowledge and positive attitudes. The correlation between their perceived knowledge and actual knowledge was very weak, thus suggesting that the teachers are inclined to overestimate their knowledge. Conclusions: The Turkish elementary schoolteachers showed satisfactory oral health knowledge and attitudes toward oral health education. The teachers’ knowledge about dental trauma management was inadequate, necessitating urgent educational interventions, especially for physical education teachers, who are at the greatest risk of encountering such events during their work. The oral hygiene behaviours were not associated with teachers’ oral health knowledge, attitudes, or practice, thus requiring further investigation.Oral hygiene measures in childhood lead to healthy teeth and oral mucosa, providing optimal general health conditions [1]. Oral diseases are depicted as a major public health challenge, especially in school children. A total of 90% of them are suffering from dental caries, with increasing incidences in Asian and Latin American countries as reported by the Global Burden of Disease (GBD) in its 2005 report [2]. The prevalence of tooth decay in school children ranged between 60 and 90%. The incidence of caries is rising rapidly in developing countries [2].It had been reported by the United Nations Educational, Scientific and Cultural Organization (UNESCO) that the number of teachers at the primary school level worldwide by the year 1993 was around 23.9 million [3]. Schools continue to be an essential environment, offering an ideal and effective method to manage over 1 billion children worldwide [4]. Preschools and primary schools have great potential to influence child’s health behaviour [5,6]. Children spend a significant amount of time at school, especially at the age when their habits are shaped. Therefore, the role of teachers is critical in these developmental stages of the child [7].The education of school-age children in oral health is crucial because healthy oral habits occur at a young age. The importance of teaching children (infants, preschoolers, or schoolchildren) about oral hygiene was recognised as early as 1878 [8,9]. Schools are an optimal location for providing oral health education, as these services can be given similarly and widely to all children, especially those who do not have access to other health resources and cannot receive professional dental care [10,11].One of the critical issues in oral health is the treatment of dental injuries [12]. Children participate in sports activities at school, and in cases of close contact or physical activity, injuries may occur due to reasons such as falls or accidents [13]. In these trauma cases, successful management of the process from the moment of the event to the visit to the dentist significantly increases the chances of successful post-trauma treatments [14]. The good management of this process depends on the teacher’s level of knowledge. Correct guidance of the child and their parents can give the dentist a chance for early intervention. A teacher needs to know what to do in an emergency regarding primary and permanent teeth [15]. Dental trauma in industrialised countries ranges from 16 to 40% for six-year-olds and 4 to 33% for 12- to 14-year-olds; in some Latin American countries, it is about 15% of schoolchildren; in the Middle East, it is about 5 to 12% among 6- to 12-year-olds [16]. Many studies have shown that getting support from teachers is successful in improving the oral health of school children [17].However, according to some reports, teachers were hesitant to participate in oral health programs that require supervision [18]. The reason for this is thought to be due to the teachers’ lack of knowledge on oral health [19]. Since schoolteachers are models for school students, their oral health knowledge must be good, and their oral health behaviours and attitudes must follow professional advice [1]. For this reason, their knowledge and attitudes about oral health are important both for their oral health and for the children they influence and teach as a model [6]. Oral health is an integral part of overall health and well-being. Individuals with a healthy mouth live without pain, discomfort, and embarrassment while talking, eating, and socialising [20,21]. One study showed that school-age adolescents suffering from poor oral health were 12 times more often deprived of activities compared to their peers [22]. More than 50 million school classes are lost worldwide due to poor oral health. This can affect the student’s school performance and subsequent success in his life [23]. Ideally, primary schoolteachers can give information about good oral health, dental and gum care, proper oral hygiene, use of fluoride, proper dietary advice, and the benefit of routine dental visits [24]. The goal of oral health education is to improve knowledge within the target population, leading them to be individuals with better oral health and to adopt positive oral health behaviours [16].Recent studies revealed that oral-health promotion programs have a significant effect in improving dental health status and reducing the cost of healthcare systems. Therefore, oral-health promotion programs revealed their effects on children’s oral health and on parental dental treatment expenses [25]. Education in and through the media had an increasing tendency that can be seen in most health-promoting media literacy classes being taught in schools [26]. In order to prevent the spread of infectious diseases, health personnel are advised to use audiovisual media to educate the population about health [27]. Mass media campaigns tend to be helpful in influencing beliefs of the general public and healthcare practitioners by increasing the alignment between beliefs and current evidence and promoting self-management concepts [28].This study was conducted in Istanbul, Turkey, between July and September 2020 as an analytical cross-sectional survey-based study using an online self-administered questionnaire (SAQ) of multiple-choice items in order to evaluate the oral health self-perceived knowledge, actual knowledge, attitudes, behaviours, and practice of elementary school teachers. The questionnaire was electronically developed using SurveyMonkey (SVMK Inc. San Mateo, CA, USA 2020), and the target subjects received a uniform resource locator (URL) and a quick response (QR) code from their school administrator/principal to take part in the study.According to the official data of the Istanbul governorship, as of May 2019, there were 3,175,285 students enrolled in 7437 schools (3790 private and 3647 publicly funded) with 168,527 teachers. The adequate sample size for this study was calculated using Epi InfoTM version 7.2.4 (CDC. Atlanta, GA, USA 2020), which indicated that a total of 384 teachers was required in order to achieve conclusive results with a confidence level of 95% and an error margin of 5%. Participation in this study was entirely voluntary and not financially compensated or incentivised by other means to minimise self-selection bias. The participants were able to withdraw from the study at any time without the need to explain the reason.The items of the SAQ used in this study were adopted from previous studies with relevant purposes, including the assessment of schoolteachers’ oral health knowledge and attitudes, assessment of schoolteachers’ knowledge and practice regarding dental injuries, and evaluation of oral hygiene behaviours [1,2,5,6,7]. The instrument’s content validity was evaluated by a panel of experts in dental public health who reviewed the proposed items in terms of clarity and relevance.The SAQ had 36 multiple-choice items categorised into six sections. Section I (five items) was about demographic data such as gender, length of work experience, education status and branch. Section II (four items) was about perceived oral health knowledge assessed by a five-point Likert scale ranging from totally disagree to totally agree. Section III (10 items) was about actual oral health knowledge, which was designed to correlate with the items of Section II. Section IV (four items) was about the attitudes towards oral health education assessed by a five-point Likert scale ranging from totally disagree to totally agree. Section V (five items) was adapted from Hiroshima University Dental Behavioral Inventory (HU-DBI) aiming to evaluate oral health behaviours [29]. Section VI (eight items) was about the teachers’ experience and practice of oral health education (Supplementary Materials).This study was conducted following the Declaration of Helsinki for the ethical principles of medical research involving human subjects. It was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for cross-sectional studies [30,31]. Ethical approval was granted by the institutional review board (IRB) of Istanbul Medipol University under the code of 10840098-772.02-E.34168 on 8 July 2020. Written permission was obtained from each participating school’s principal/administrator to carry out the survey; electronic informed consent was required from each invited teacher before filling in the questionnaire.All statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 27.0 (SPSS Inc., Chicago, IL, USA, 2020). Descriptive statistics were carried out primarily for the normality of data, followed by the frequencies, percentages and central tendencies of demographic variables, oral health-related knowledge, attitudes, behaviours, and practice.For the total score of actual knowledge, each question had one point for the correct answer and zero points for wrong answers. The total score of perceived knowledge and the total score of attitudes were calculated by pooling the scores of each question in these sections that were originally rated by a five-point Likert scale. For the total score of oral hygiene behaviors, one point was given for the favorable behavior and zero points were given to the unfavorable behavior.Consequently, inferential statistics were carried out using the Chi-squared test (χ2), Mann–Whitney test (U), and Kruskal–Wallis (H) test to evaluate the impact of demographic variables on teachers’ oral health knowledge, attitudes, and behaviours. Spearman’s correlation test (ρ) was performed to evaluate the association between actual knowledge domains and perceived knowledge items. The strengths of correlation are described by the value of (ρ): 0.01 to 0.39 “weak”; 0.40 to 0.69 “moderate”; 0.70 to 0.99 “strong”; 1.0 “perfect” [32]. All inferential tests were carried out with a confidence level of 95% and significance value p ≤ 0.05.A total of 385 elementary schoolteachers from Istanbul, Turkey, participated in this study. The majority were females (62.2%), qualified with a licensure degree (81.3%) and working in public schools (86.5%). While 87 (22.6%) participants had one to five years of work experience, 111 (28.8%) have worked for 6 to 10 years, 89 (23.1%) had worked for 11 to 20 years, and 98 (25.5%) had worked for over 20 years. Eleven (2.9%) participants were teachers of physical education (Table 1).Oral health-related knowledge of teachers was evaluated in two main constructs—self-perceived knowledge and actual knowledge. Perceived knowledge is one’s self-assessment or feeling of knowing the information, and it should be distinguished from and assessed concurrently with actual knowledge in order to find out any knowledge discrepancies that indicate “knowledge illusion” [33,34,35]. The four domains of knowledge tested in this study were (i) primary dentition, (ii) oral diseases risk factors, (iii) oral hygiene methods, and (iv) dental trauma management.Regarding their perceived knowledge, 76.6%, 64.2%, 82.3%, and 36.9% of the participants agreed that they have sufficient knowledge about primary dentition, oral diseases risk factors, oral hygiene methods, and dental trauma management, respectively. The teachers with >10 years of work experience had higher levels of perceived knowledge in all four knowledge domains (U = 15,919, 14,819.5, 17,994.5, and 14,663.5; p = 0.01, <0.001, 0.589, and <0.001, respectively) than their colleagues with ≤10 years of experience. Female teachers had higher levels of all four perceived knowledge domains (U 13,690, 15,446, 14,602, and 16,976; p < 0.001, 0.054, 0.003, and 0.706, respectively) than their male colleagues. In general, the teachers at private schools had higher perceived knowledge levels than the teachers at public schools.Regarding their actual knowledge, the participants had to answer 10 multiple choice questions with one correct answer (two items: primary dentition; four items: oral diseases risk factors; two items: oral hygiene methods; two items: dental trauma management). All items of the second (oral diseases risk factors) and third domain (oral hygiene methods) were correctly answered by more than 50% of the participants. The two items of the fourth domain (dental trauma management) received the least number of correct answers 26% and 32.5%, respectively. The teachers with ≤10 years of experience had a significantly higher level of actual knowledge in the first domain (U = 16,446; p = 0.043) and a lower level in the fourth domain (U = 14,924; p = 0.007) than their colleagues with >10 years of experience. Female teachers had a significantly higher level of actual knowledge in the second domain (U = 13,648.5; p < 0.001) than their male colleagues. The difference between private schoolteachers and public schoolteachers in actual knowledge domains was not statistically significant (Table 2).Oral hygiene behaviours were evaluated using five items adapted from the Turkish version of HU-DBI (items no. 1, 4, 9, 12, and 16) [29,36]. While 243 (63.1%) participants reported that they do not worry about visiting the dentist, only 29 (7.5%) and 50 (13%) of them had used disclosing dye and a child-sized toothbrush. The teachers with ≤10 years of experience had a lower level of calculus detection (χ2 = 9.87; p = 0.002) compared to their colleagues with >10 years of experience. For the total behaviours score (0–5), no significant difference was found between participants in terms of gender, work experience, or school type (Table 3).The schoolteachers�� attitudes towards providing oral health education to their students were evaluated by four items assessed by a five-point Likert scale ranging from “strongly agree = 5 “to “strongly disagree = 1”. A total of 311 (80.8%) participants agreed that teachers should have a role in the oral health education of schoolchildren. The teachers with ≤10 years of experience (74.3%) had a significantly (χ2 = 11.22; p < 0.001) lower level of acceptance for their proposed role in providing oral health education compared to their colleagues with >10 years of experience (87.7%). A total of 294 (76.4%) participants agreed that teachers should receive oral health training as a part of general health training. The teachers with ≤10 years of experience (70.7%) had a significantly (χ2 = 7.23; p = 0.007) lower level of willingness for receiving oral health training compared to their colleagues with >10 years of experience (82.4%).A total of 334 (86.8%) participants believed that oral health could affect the psychology of schoolchildren. The teachers with ≤10 years of experience (89.4%) had an insignificantly (χ2 = 3.01; p = 0.083) lower level of belief of oral health impact on schoolchildren’s psychology compared to their colleagues with >10 years of experience (89.8%). A total of 351 (91.2%) participants agreed that oral health education could benefit their schoolchildren. The teachers with ≤10 years of experience (83.8%) had an insignificantly (χ2 = 1.60; p = 0.207) lower level of belief of oral health benefits for their schoolchildren compared to their colleagues with >10 years of experience (93%). There was no significant difference between participants in any of the attitude items according to gender (female vs. male) or type of school (public vs. private) (Table 4).Only 67 (17.4%) participants reported receiving training on how to provide education in general hygiene-related topics, with a significant (χ2 = 5.48; p = 0.019) difference between private schoolteachers (28.8%) and public schoolteachers (15.6%), and a significant (χ2 = 7.91; p = 0.005) difference between teachers with >10 years of experience (23%) and ≤10 years of experience (12.1%). An even smaller fraction of participants (9.4%) reported receiving training on how to provide education in oral hygiene-related topics, without a significant difference according to the length of work experience (χ2 = 2.5; p = 0.114), gender (χ2 = 0.31; p = 0.579), or school type (χ2 = 0.34; p = 0.56).While 51.7% of participants reported supervising the brushing habit of their schoolchildren, only 7.5% reported elevating the upper lip of their schoolchildren to check their teeth. Regarding their experience with teaching their own schoolchildren about oral health, 110 (28.6%) reported attempting to provide education on topics related to teeth and oral health, without a significant difference in terms of work experience (χ2 = 1.58; p = 0.209), gender (χ2 = 0.93; p = 0.334), and school type (χ2 = 0.5; p = 0.479) (Table 5).The most common topic for provided oral health education was oral hygiene (83.3%), followed by dental anatomy (9.8%) and oral disease (6.9%). While oral health videos (66%) were the most frequently used method in providing oral health education to schoolchildren, printed posters (8.7%) were the least common method. Regarding students’ feedback on the oral health education provided by their teachers, 75.9% had very favourable or favourable feedback, while only 5.8% had unfavourable or very unfavourable feedback (Figure 1).Female teachers had significantly higher levels of perceived oral health knowledge (U = 14,266.5; p = 0.003) and actual knowledge (U = 14,713.5; p = 0.011) than their male colleagues. Similarly, female teachers had insignificantly higher oral hygiene behaviours (U = 15,983.5; p = 0.175) than their male colleagues. Contrarily, male teachers had a higher level of attitudes (U = 17,099.5; p = 0.861) towards oral health education than female colleagues slightly.The teachers with >10 years of work experience had higher levels of perceived oral health knowledge (U = 14,864.5; p = 0.001), actual knowledge (U = 18,027.5; p = 0.651), oral hygiene behaviours (U = 18,166.5; p = 0.739), and oral health education attitudes (U = 15,917; p = 0.019) than their colleagues with ≤10 years of experience.The teachers with a master’s degree had slightly higher levels of perceived oral health knowledge (U = 10,101; p = 0.776), actual knowledge (U = 9213; p = 0.161), and oral health education attitudes (U = 9190; p = 0.162) than their colleagues with a licensure qualification. The only significant difference between teachers with a master’s degree and teachers with a licensure qualification was only found in terms of oral hygiene behaviours (U = 8024; p = 0.003).The teachers at public schools had insignificantly higher levels of actual oral health knowledge (U = 8212; p = 0.544) and oral hygiene behaviours (U = 7787.5; p = 0.222), while the teachers at private schools had insignificantly higher levels of perceived oral health knowledge (U = 7910; p = 0.312) and oral health education attitudes (U = 8513.5; p = 0.871).The physical education teachers had a significantly lower level of perceived oral health knowledge (U = 1260; p = 0.027) and actual knowledge (U = 1059; p = 0.005), and they had an insignificantly lower level oral health education attitudes (U = 2038; p = 0.956) and oral hygiene behaviours (U = 1321; p = 0.107). The physical education teachers were also the least knowledgeable in all actual knowledge domains, including dental trauma management (Table 6).The overall score of perceived oral health knowledge was weakly correlated with the overall score of actual oral health knowledge (ρ = 0.265; p < 0.001). The perceived knowledge and the actual knowledge were positively correlated in all domains, even though the relationships between both knowledge constructs were either negligible or weak (0.008 to 0.274).Both knowledge constructs were insignificantly (ρ = 0.080 and 0.071; p = 0.117 and 0.164) correlated in the first (primary dentition) and third domain (oral hygiene methods), respectively, while they were significantly (ρ = 0.208 and 0.274; p < 0.001 and < 0.001) correlated in the second (oral diseases risk factors) and fourth domain (dental trauma management), respectively.While actual knowledge of the third domain (oral hygiene methods) was not significantly correlated with any perceived knowledge domains, the actual knowledge of the fourth domain (dental trauma management) was significantly correlated with all perceived knowledge domains (Table 7).In all oral hygiene behaviour items of HU-DBI, the level of actual knowledge was higher in the teachers with favourable hygiene outcomes, except for the item of using disclosing agents, where the actual knowledge level was equal between the teachers who used disclosing agents and those who do not use them. Similarly, the level of perceived knowledge was higher in the teachers with favourable hygiene outcomes. On the other hand, the teachers who noticed white stick deposits had a significantly (U = 15,517.5; p = 0.019) higher perceived knowledge level. The overall behavior score (0–5) was weakly correlated with the overall score of perceived knowledge (ρ = 0.138; p = 0.007) and actual knowledge (ρ = 0.146; p = 0.004) (Table 8).The teachers with higher levels of perceived knowledge were significantly more in favour of the teachers’ role in oral health education (U = 2983.5; p = 0.002), in agreement that teachers should receive oral health training as a part of general health training (U = 3553.5; p = 0.001), in agreement that oral health can affect the psychology of schoolchildren (U = 2154; p = 0.041), and in agreement that oral health education can benefit their schoolchildren (U = 699.5; p = 0.001).While the teachers with higher levels of actual knowledge were significantly more in agreement that teachers should receive oral health training as a part of general health training (U = 4016.5; p = 0.016), they were insignificantly more in favour of the teachers’ role in oral health education (U = 3565; p = 0.058), and in agreement that oral health education can benefit their schoolchildren (U = 1399.5; p = 0.267) (Table 9).The teachers who received education about general hygiene had significantly higher levels of perceived oral health knowledge (U = 5862; p < 0.001), actual knowledge (U = 8376.5; p = 0.005), and attitudes towards oral health education (U = 7591; p < 0.001). Similarly, the teachers who received education about oral hygiene had significantly higher levels of perceived oral health knowledge (U = 4161.5; p = 0.001), actual knowledge (U = 4628; p = 0.008), and attitudes towards oral health education (U = 3938; p < 0.001).The teachers with higher levels of perceived oral health knowledge were more used to checking their students’ teeth (U = 4006; p = 0.043), supervising children’s brushing (U = 17751.5; p = 0.485), and teaching their students about oral health (U = 12,153; p = 0.002). Similarly, the teachers with higher levels of actual oral health knowledge were more used to checking their students’ teeth (U = 4513; p = 0.252), supervising children’s brushing (U = 13246.5; p < 0.001), and teaching their students about oral health (U = 12,129; p = 0.002).The more positive attitudes towards oral health education, the higher the frequency of teachers had checking their students’ teeth (U = 3574; p = 0.005), supervising children’s brushing (U = 15,479; p = 0.006) and teaching their students about oral health (U = 11,720; p = 0.001) (Table 10).The primary objective of this study was to evaluate the elementary schoolteachers’ level of oral health knowledge, oral hygiene behaviours, and attitudes towards oral health education. The secondary objectives were to estimate the correlation between teachers’ perceived knowledge and actual knowledge, explore the discrepancies of teachers’ oral health knowledge, and evaluate the impact of teachers’ knowledge on their behaviours, attitudes, and practice.Within the limits of our study, schoolteachers’ overall oral health knowledge was fair, except for dental trauma management. The correlation between perceived knowledge and actual knowledge was mainly weak in all the investigated domains. Our teachers’ perceived knowledge was significantly associated with their attitudes towards oral health education. The higher knowledge levels and the more positive attitudes were significant predictors for teachers’ actual engagement with providing oral health education to their children.Across gender, the female teachers in our sample had significantly higher levels of perceived oral health knowledge (U = 14,266.5; p = 0.003) and actual oral health knowledge (U = 14,713.5; p = 0.011) than their male colleagues. The previous studies of elementary schoolteachers’ oral health knowledge in Saudi Arabia, India, Tanzania, and Uganda found that female teachers were more knowledgeable than their male colleagues [6,37,38,39,40,41,42,43]. While the difference between female and male teachers in terms of attitudes towards oral health education was statistically insignificant (U = 17,099.5; p = 0.861), the male teachers had slightly higher levels of positive attitudes, including the willingness to receive oral health education. Similarly, in Madinah (Saudi Arabia) and Rungwe (Tanzania), male teachers had more positive attitudes towards oral health [6,42,43]. The lower levels of males’ perceived knowledge can be utilised to explain their higher levels of positive attitudes towards oral health education. Contrarily, in Upper Galilee (Israel) and Al-Kharj (Saudi Arabia), female teachers had significantly higher levels of both oral health knowledge and attitudes [7,39]. Al-Jundi et al., 2005 suggested that the female teachers’ positive attitudes could be related to their caring nature and closeness to children [44].The private schoolteachers had slightly higher levels of perceived oral health knowledge (U = 7910; p = 0.312) and lower levels of actual knowledge (U = 8212; p = 0.544) compared to the public schoolteachers. Vanka et al., 2012 found that the private schoolteachers in Bhopal (India) had a significantly lower level of oral health knowledge (χ2 = 15.421; p = 0.05) than the public schoolteachers [45]. However, in our sample, private schoolteachers (28.8%) reported more frequently that they received training in general hygiene than the public schoolteachers (15.6%); there was no significant difference in terms of receiving training about oral hygiene. Haloi et al., 2014 found that 82.2% of private schoolteachers had postgraduate degrees (PGs), while 65.2% of public schoolteachers held PGs in Mathura (India); this significant difference was found in our sample as 30.8% of private schoolteachers hold PGs while only 15% of the public schoolteachers do [46]. Therefore, the improved oral health outcomes of private school children in some developing countries should be fairly attributed to the fact that these children come from families with higher socioeconomic capacities that enable them to cover the tuition fees of private schools [47].The length of work experience was a significant promoter for the teachers’ perceived oral health knowledge (U = 14864.5; p = 0.001) and their positive attitudes towards oral health education (U = 15,917; p = 0.019). The teachers with >10 years of work experience had significantly higher levels of both perceived knowledge (U = 14,663.5; p < 0.001) and actual knowledge (U = 15,924; p = 0.007) about what to do in case of dental trauma than their colleagues with ≤10 years of experience. This result is consistent with the findings of Junges et al., 2015 where the elementary schoolteachers with more than 15 years of work experience were more likely to know the appropriate conduct for handling teeth avulsion and crown fracture situations, thus suggesting that greater work experience could provide opportunities for schoolteachers to witness dental trauma events at schools or home [48]. The same conclusion had been previously drawn from cross-sectional studies in Singapore and Brazil [49,50].On analysing the relationship between knowledge constructs, the schoolteachers’ perceived knowledge and actual knowledge were positively correlated in all the investigated domains, thus ruling out any critical knowledge discrepancies. Given the low correlation coefficients found in our sample, the schoolteachers’ perceived oral health knowledge can weakly predict their actual knowledge; therefore, it should not be exclusively used to evaluate the teachers’ knowledge in future studies. While the mean score of perceived knowledge was 14.91 ± 2.68 (total 20 points), the mean score of actual knowledge was 5.83 ± 1.73 (total 10 points), thus suggesting that the perceived knowledge score (74.6%) was considerably higher than the actual knowledge score (58.3%). Therefore, the schoolteachers, probably due to their social role as mentors for their students, are inclined to overestimate their oral health knowledge.The knowledge domain of “dental trauma management” received the lowest score in both perceived knowledge and actual knowledge compared to other knowledge domains, thus suggesting that this is the most critical knowledge area that requires urgent educational interventions for schoolteachers in Turkey. A recent Turkish study found that pre-schoolteachers had inadequate knowledge about dental trauma management that might be attributed to lack of trauma experience of the teachers and suggesting the same call for educational interventions for both schoolteachers and parents on how to handle dental trauma, which is common in these age groups [51].In Brazil, physical education undergraduates showed extremely low levels of knowledge about dental trauma [52]. The same was found in Hong Kong and India among physical education teachers [53,54]. In our sample, physical education teachers had the lowest frequency of correct answers to the actual knowledge questions of dental trauma management compared to teachers of other subjects. This finding supports the abovementioned call for educational interventions, especially for physical education teachers who are at the greatest risk of encountering dental fractures and avulsion events as part of their daily work with schoolchildren.In our study, the schoolteachers with higher levels of perceived oral health knowledge had significantly higher levels of positive attitude towards oral health education. Ramroop et al., 2011 found that primary schoolteachers in Trinidad had satisfactory knowledge about aetiology and the prevention of dental caries, while they had inadequate knowledge about dental trauma management. In spite of that, Trinidadian teachers had positive attitudes towards oral health education, suggesting that actual knowledge may not necessarily predict teachers’ attitudes [55]. In Greece, India, and Nigeria, the schoolteachers’ inadequate knowledge about oral health was significantly associated with their willingness to learn about oral health and involved themselves in its teaching [56,57,58].However, the overall oral hygiene behaviour score was weakly correlated with the overall score of perceived knowledge and actual knowledge; teachers’ hygiene behaviours were not found to be associated with positive attitudes towards oral health education, nor were teachers’ experience with teaching about oral health. Similarly, in Saudi Arabia, the high levels of oral hygiene were not associated with positive attitudes towards oral health education among schoolteachers [39]. These findings contrast with what was reported earlier by Maganur et al., 2017 in India and Dawani et al., 2013 in Pakistan, where schoolteachers with good oral hygiene had a significantly higher level of knowledge and more positive attitudes towards oral health education [59,60]. Our findings can be either explained as a methodological limitation because the oral hygiene items adapted from HU-DBI may not have been supposed to be taken out of their context, or as a true phenomenon where personal oral hygiene of the teachers’ may not affect their attitudes and practice of oral health education. This point warrants further investigation to understand better the role of schoolteachers’ oral hygiene in shaping their students’ oral health behaviours and attitudes.Our schoolteachers’ perceived oral health knowledge, actual knowledge and attitudes were significantly associated with their experience of providing oral health education to their students. This finding highlights the importance of increasing teachers’ awareness through educational interventions as a strategy to increase their likelihood of providing actual guidance and education to their students.The first limitation of this study is its sample, which was not equally stratified across gender, but this can be justified as the sample was aimed to be as representative as possible for the actual demographics of schoolteachers in Turkey who are predominantly female. The second limitation is that the vast majority of respondents were from public schools; however, the number of public schools is very close to the number of private schools in Istanbul. The third limitation was the selection bias, which was unavoidable in the convenience (non-random) sampling approach that we followed in this study; the teachers who voluntarily joined the study may have better awareness regarding oral health and oral hygiene issues.To the best of the authors’ knowledge, this is the first study to evaluate schoolteachers’ oral health knowledge, attitudes, and behaviours in the largest Turkish metropolis (Istanbul). It is the first study, to date, comparing teachers’ perceived knowledge and actual knowledge aiming to explore knowledge discrepancies of the target population. The study provides support to the prevailing evidence on the role of oral health knowledge in shaping teachers’ attitudes towards oral health education, their oral hygiene behaviours, and their experience with teaching their students about oral health.The findings of this study suggest that future studies of schoolteachers’ oral health knowledge should discriminate clearly between perceived knowledge and actual knowledge, and they should rely entirely on perceived knowledge outcomes as they can be overestimated. The schoolteachers, especially physical education teachers, exhibited low levels of knowledge about dental trauma management, thus calling for urgent educational intervention to increase their awareness and skills. A conceptual model for oral health knowledge, attitudes and practice of elementary schoolteachers should be proposed and tested in the future based on our study data and other similar studies which employed various oral health constructs, e.g., knowledge, attitudes, behaviours, etc. This suggested model will serve as a basis for promotion interventions targeting schoolchildren and their teachers.The elementary schoolteachers in Istanbul, Turkey, showed satisfactory oral health knowledge and attitudes toward oral health education. The correlation between their perceived knowledge and actual knowledge was very weak, suggesting that the teachers are inclined to overestimate their knowledge. The teachers’ knowledge about dental trauma management was inadequate, necessitating urgent educational interventions, especially for physical education teachers, who are at the greatest risk of encountering such events during their work. Female gender and greater work experience were found to be promoters for oral health knowledge and positive attitudes. The oral hygiene behaviours were not clearly associated with teachers’ oral health knowledge, attitudes or practice, thus requiring further investigation.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116028/s1, Questionnaire of Schoolteachers’ Oral Health KAB (in Turkish).Conceptualisation, G.Y. and A.R.; methodology, A.R.; validation, G.Y. and H.K.; formal analysis, A.R.; investigation, G.Y. and H.K.; resources, S.A.; data curation, H.K.; writing—original draft preparation, G.Y. and A.R.; writing—review and editing, M.K. and S.A.; supervision, H.K.; project administration, G.Y.; funding acquisition, S.A. All authors have read and agreed to the published version of the manuscript.The work of A.R. was funded by Masaryk University, grants number MUNI/IGA/1543/2020 and MUNI/A/1608/2020, in addition to the INTER-EXCELLENCE grant number LTC20031.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board (IRB) of Istanbul Medipol University Ref. 10840098-772.02-E.34168 on 8 January 2020.Informed consent was obtained from all participants involved in the study.The data that support the findings of this study are available from the corresponding author upon reasonable request.The authors thank all the participating teachers for their time and their eminent support for the study.The authors declare no conflict of interest.(a) Topics of oral health education provided to schoolchildren; (b) methods used in providing oral health education to schoolchildren; (c) schoolchildren’s feedback to oral health education provided by their teachers.Demographic characteristics of elementary schoolteachers, Istanbul, Turkey, 2020.Oral health-related knowledge of elementary schoolteachers stratified by work experience, Istanbul, Turkey, 2020.Chi-squared test (χ2) and Mann–Whitney (U) test were used with a significance level of ≤0.05.Oral health-related behaviours of elementary schoolteachers stratified by work experience.Chi-squared test (χ2) and Mann–Whitney (U) test were used with a significance level of ≤0.05.Oral health-related attitudes of elementary schoolteachers stratified by work experience.Mann-Whitney (U) test was used with a significance level of ≤0.05.Oral health education experience of elementary schoolteachers stratified by work experience.Chi-squared test (χ2) and Mann–Whitney (U) test were used with a significance level of ≤0.05.Social determinants of elementary schoolteachers’ oral health knowledge, behaviours, and attitudes.Mann–Whitney (U) and Kruskal–Wallis (H) test were used with a significance level of ≤0.05.Correlation of perceived oral health knowledge and actual knowledge of elementary schoolteachers.Prc. = perceived knowledge; Act. = actual knowledge; ρ. = Spearman’s correlation coefficient; ** correlation is significant at the level 0.01 level (two-tailed); * correlation is significant at the 0.05 level (two-tailed). Sig. = SignificanceImpact of elementary schoolteachers’ oral health knowledge on their behaviours.Mann–Whitney (U) test and Spearman’s correlation (ρ) test were used with a significance level of ≤0.05.Impact of elementary schoolteachers’ oral health knowledge on their attitudes.Mann–Whitney (U) test was used with a significance level of ≤0.05; Disagreement = Totally Disagree + Disagree; Agreement = Totally Agree + Agree.Impact of elementary schoolteachers’ oral health knowledge and attitudes on their practice.Mann–Whitney test was used with a significance level of ≤0.05.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06029.txt ADDED Viewed

	@@ -0,0 +1,7 @@

+The Yellow River Basin (YLRB) and Yangtze River Basin (YZRB) are heavily populated, important grain-producing areas in China, and they are sensitive to climate change. In order to study the temporal and spatial distribution of extreme climate events in the two river basins, seven extreme temperature indices and seven extreme precipitation indices were projected for the periods of 2010–2039, 2040–2069, and 2070–2099 using data from 16 Coupled Model Intercomparison Project Phase 5 (CMIP5) models, and the delta change and reliability ensemble averaging (REA) methods were applied to obtain more robust ensemble values. First, the present evaluation indicated that the simulations satisfactorily reproduced the spatial distribution of temperature extremes, and the spatial distribution of precipitation extremes was generally suitably captured. Next, the REA values were adopted to conduct projections under different representative concentration pathway (RCP) scenarios (i.e., RCP4.5, and RCP8.5) in the 21st century. Warming extremes were projected to increase while cold events were projected to decrease, particularly on the eastern Tibetan Plateau, the Loess Plateau, and the lower reaches of the YZRB. In addition, the number of wet days (CWD) was projected to decrease in most regions of the two basins, but the highest five-day precipitation (Rx5day) and precipitation intensity (SDII) index values were projected to increase in the YZRB. The number of consecutive dry days (CDD) was projected to decrease in the northern and western regions of the two basins. Specifically, the warming trends in the two basins were correlated with altitude and atmospheric circulation patterns, and the wetting trends were related to the atmospheric water vapor content increases in summer and the strength of external radiative forcing. Notably, the magnitude of the changes in the extreme climate events was projected to increase with increasing warming targets, especially under the RCP8.5 scenario.Global warming affects the frequency, intensity, and duration of extreme climate events such as droughts, heat waves, floods, hurricanes, and extreme cold and hot days [1,2,3]. Compared to the changes in conventional climate parameters such as the mean temperature and precipitation, extreme climate events can impose more significant stresses on human society and natural systems and can also exert severe socioeconomic and ecological impacts [4,5,6]. For example, concurrent drought and heat extremes can cause substantial decreases in barley yields worldwide, while extreme rainfall can cause floods and damage to urban infrastructure [7,8,9]. Vegetation sensitive to temperature and precipitation changes can be destroyed by extreme climate events, causing land desertification, soil erosion, and crop reduction [10,11,12,13]. Hence, it is necessary to study the spatial and temporal distribution characteristics, future development trends, and influencing factors of extreme climate events.There are many ways to define extreme climate events, and one of them involves the indices developed by the Expert Team on Climate Change Detection and Indices (ETCCDI) [14]. This set of indices contains 16 extreme temperature indices and 11 extreme precipitation indices which are widely applied in studies of extreme climate changes and in the establishment of different climate models [15,16]. In addition, the Coupled Model Intercomparison Project (CMIP) provides a set of coordinated global climate model experiments to simulate present and future climate changes [17]. CMIP data can also be adopted in the research of extreme climate events, including the calculation of ETCCDI indices [18,19]. Compared to CMIP Phase 3 (CMIP3) data, the performance of CMIP Phase 5 (CMIP5) data in the simulation of extreme climate indices exhibits certain improvements on the global and regional scales [20,21,22].On the global scale, increased warm events and decreased cold events have been reported over the past few decades, which indicates that changes are generally expected in a warming world [23,24,25]; extreme precipitation events will also increase in a warming climate due to the increased atmospheric humidity [26]. Extreme climate events in different regions worldwide always reveal different trends because of the contributions of atmospheric dynamics and thermodynamics [26,27]. For example, extreme precipitation exhibits a notable positive relation with a warming climate in midlatitude locations but a weak or negative relation in the tropics [26]. This result can also be observed in the northeastern and southeastern U.S. [28,29] and Southeast Asia [30,31].The extreme climate events in China are greatly affected by monsoons and geographical environments, such as territory and terrain conditions [32]. In addition, the urbanization in China has rapidly increased with economic development since the 1970s, and this increase has led to local climate changes [33]. Large-scale atmospheric circulation parameters also cause notable impacts on the extreme climate events in China, especially the El Niño-Southern Oscillation (ENSO), the Atlantic Multidecadal Oscillation (AMO), and the Pacific Decadal Oscillation (PDO) [34,35,36].Many studies have noted that significant increases have occurred in the frequency of warm extremes and decreases in cold extremes in China over the past few decades. For example, Shi et al. [37] mentioned the cold spell duration indicator (CSDI) decreased and the warm spell duration indicator (WSDI) increased in almost all parts of China during 1961–2005; Wu et al. 2019 [38] found that there was a significant increase in the frequency of compound dry/warm and wet/warm extremes while there was a decrease in compound dry/cold and wet/cold extremes for the period of 1988–2014 relative to 1961–1987 in China, which are consistent with global warming [39,40]. However, increase trends of extreme rainfall events coupled with decreased dry spells have been observed in many regions in China over the past century, such as the monsoon regions in China during 1964–2014 [41], northwestern China (1961–2010), and southeastern China (1961–2016) [42,43]. In northeastern and southwestern China, a notable drying tendency has been confirmed [34,42]. In the future, model calculations suggest that additional warming might increase extreme events in some areas, although there are considerable uncertainties in predicting future climates in specific localized areas. For example, an additional half a degree caused by global warming may increase the magnitude of extreme precipitation events [44]. However, Wang et al. [45] suggested a negative scaling of extreme precipitation with very high temperatures, thus raising doubts about future increases in precipitation extremes.These prior studies have mainly focused on China as a whole and revealed large-scale climate changes, but regional-scale research may help us to better understand the characteristics of climate change in different areas of China due to its different locations and terrains [16,46]. Furthermore, the impacts of extreme climate events on regions with different land use types are also different [41,47,48,49]. The Yangtze River Basin (YZRB) and Yellow River Basin (YLRB) are important population settlements and water supply sources, and they host several important economic belts [50]. In addition, the ecological environment in these river basins is fragile and more susceptible to extreme climate events [51,52]. Studies have focused on extreme climate events in the YZRB and YLRB, mainly employing station-observed data to study the spatiotemporal distribution characteristics of short-term extreme climate indices [53,54,55,56,57,58]. However, there is a lack of research on continuous spatial scales, as well as on the trends of extreme climate events in different future periods. Moreover, the impacts of atmospheric circulation patterns on extreme climate events also need to be further analyzed.In this study, we focused on the spatiotemporal distributions of extreme temperature and precipitation events in the YLRB and YZRB during the historical periods and different scenarios in the 21st century. We simulated several extreme climate indices based on data from multiple CMIP5 models in the YZRB and YLRB and compared the performance with observed data. Finally, we examined the influencing factors on the changes in extreme climate events in these two basins.The Yangtze River and Yellow River are the first and second longest rivers, respectively, in China, and their basin areas total approximately 2.55 × 106 km2, which accounts for nearly 26.6% of the landmass of China (Figure 1). The YZRB and YLRB consist of multiple economies, and more than 500 million people live in these basins, accounting for approximately 40% of the Chinese population [54,56]. The Yangtze River and Yellow River both originate on the Qinghai-Tibet Plateau at elevations exceeding 5000 m above sea level but ultimately flow into the East China Sea and Bohai Sea, respectively. Because of the sufficient water supply, the YZRB and YLRB are important wheat and maize production areas in China [52,59], but the complex terrain and climatic conditions threaten the ecological environment of these two river basins.The observed daily mean air temperature (Tm), daily maximum temperature (Tx), daily minimum temperature (Tn), and daily precipitation (Pre) over the period of 1961–2005 were obtained from 2472 national meteorological stations (excluding the two offshore island stations of Xisha and Coral Island) of the National Meteorological Information Center (NMIC), China Meteorological Administration (CMA; http://data.cma.cn/, accessed on 24 September /2019). The station construction and observation methods are consistent with the standards issued by the World Meteorological Organization (WMO), and the data have undergone strict quality control measures to ensure that the accuracy of the daily weather datasets approaches is 100% [60,61,62]. Hence, the data were converted to grid data at a 0.5° × 0.5° spatial resolution using the thin plate spline (TPS) method. There were a few missing values in the observational data, such as the precipitation data in 2002 and, therefore, the reference period was set to 1971–2000.The daily Tm, Tx, and Tn of 14 CMIP5 models and Pre of 10 CMIP5 models were adopted in this study and are available from the Earth System Grid Federation (ESGF, https://esgf-node.llnl.gov/projects/cmip5/, accessed on 2 October 2020) repositories. Table 1 provides the primary information of the various CMIP5 models with different horizontal and atmospheric resolutions. The data included historical simulations (from the 19th century to 2005) and future projections (2006–2300). The spatial distribution of future climate changes over the 30-year periods of the 2020s (2010–2039, or beginning-of-century), 2050s (2040–2069, or mid-century), and 2080s (2070–2099, or end-of-century) were analyzed relative to the reference period. Two different representative concentration pathways (RCP4.5 and RCP8.5) of future emissions were chosen, and they were named based on the radiative forcing in 2100, i.e., at 4.5 and 8.5 W/m2, respectively [63,64]. All models were bilinearly interpolated to a common 0.5° × 0.5° grid, consistent with the observations.Consequently, we generated Taylor diagrams to visualize the model simulation performance of the 30-annual mean values in the two basins relative to the observations of Tm, Tx, Tn, and Pre in the historical period (Figure 2) [65]. We selected 10 models for the estimation of Pre (CCSM4, CMCC-CMS, CSIRO-Mk3.6.0, CanESM2, HadGEM2-AO, HadGEM2-CC, IPSL-CM5B-LR, MPI-ESM-LR, MPI-ESM-MR, and INMCM4), as well as 14 models for the estimation of Tm, Tx, and Tn (ACCESS1.0, ACCESS1.3, CCSM4, CMCC-CM, CMCC-CMS, CSIRO-Mk3.6.0, HadGEM2-AO, HadGEM2-ES, IPSL-CM5A-MR, IPSL-CM5B-LR, MPI-ESM-LR, MPI-ESM-MR, NorESM1-M, and INMCM4).Here, we provide mean and trends of the Tm, Tn, Tx, and Pre values during 1971–2000 for the YLRB and YZRB in Table 2 and Figure 3 to compare the difference between the original CMIP5 data and observed data. The multi-year mean Tm, Tn, Tx, and Pre were 8.49, 2.22, 14.75 °C, and 499.18 mm in the YLRB, and the intervals of model values were 2.76–7.71 °C, −3.04–5.56 °C, 7.65–13.31 °C, and 515.1–1124.85 mm for Tm, Tn, Tx, and Pre, respectively. In the YZRB, the observed Tm, Tn, Tx, and Pre were 5.47, 11.06, and 19.88 °C and 1192.65 mm, and the intervals of the model values were 11.15–14.18, 6.94–11.39, 12.83–20.22, and 954.07–1916.59 mm for Tm, Tn, Tx, and Pre, respectively. In addition, the observed Tm, Tn, and Tx increased significantly at the rate of 0.34, 0.33, and 0.36 °C/decade in the YLRB during 1971–2000, while the Pre did not decrease significantly at −12.38 mm/decade. In the YZRB, the observed Tm and Tn increased significantly at 0.19 and 0.23 °C/decade during 1971–2000, while the increase trends of Tx and Pre were not significant (0.14 °C/decade and 26.14 mm/decade, respectively). At the same time, we also noticed that there was no single CMIP5 model which could better capture the annual mean value and multi-year trends of the observed temperature and precipitation data from 1971 to 2000. For example, in the YLRB and YZRB, the models usually underestimated the Tm, Tn, and Tx but overestimated the Pre, while the trends of MPI-ESM-MR model were close to the observed values in YLRB. In the YZRB, the INMCM4 model could better capture the multi-year mean Pre, but no model could better capture the trends of the observed Tm, Tn, and Tx.To reflect the extreme temperature and precipitation in multiple aspects, 14 extreme climate indices recommended by the ETCCDI were employed in this study (Table 3), including seven extreme temperature indices (the diurnal temperature range (DTR), the numbers of summer days (SU), the number of ice days (ID), the highest daily maximum temperature (TXx), the lowest daily minimum temperature (TNn), the warm spell duration index (WSDI), and the cold spell duration index (CSDI)) and seven extreme precipitation indices (the highest 5-day precipitation (Rx5day), the extremely wet-day precipitation (R99pTOT), the heavy precipitation days (R20mm), the total wet-day precipitation (PRCPTOT), the precipitation intensity (SDII), the consecutive dry days (CDD), and the consecutive wet days (CWD)) [20,21,40,66,67,68,69,70].The TXx, TNn, and DTR indices indicated the intensity of the extreme temperature, while the SU and ID indices represented the frequency and intensity, respectively, of extreme-temperature events, and the WSDI and CSDI indices represented the duration of extreme-temperature events. The Rx5day, R99pTOT, R20mm, PRCPTOT, and SDII indices represented different ways to assess extreme precipitation, and the CDD and CWD indices helped to distinguish between dry and humid areas in the two basins. The Rx5day and CDD indices can also be applied to evaluate potential floods because persistent heavy rainfall promotes the occurrence of floods and subsequent landslides [71]. These indices can also be classified into four categories: absolute indices, threshold indices, percentile indices, and duration indices [69,72]. The nonparametric Mann-Kendall trend test is applied to establish whether the trends of these indices are significant [73].The coarse resolution of the above models cannot provide reliable information at the local and regional scales, and a single CMIP5 model data cannot capture the multi-year average value and trends of the observed data. Thus, it is necessary to compensate for this deficiency by the application of downscaling methods, such as dynamical downscaling and statistical downscaling [69,74,75,76]. Dynamical downscaling produces finer-scale global climate models (GCMs) by nesting fine-resolution regional climate models (RCMs) [77,78], while statistical downscaling establishes and applies the historical statistical relationships between large-scale atmospheric variables and local climate variables [74,79]. In this study, we applied statistical downscaling because of its higher computational efficiency.The delta change method is a simple statistical downscaling method and is applied to correct the bias of the simulated temperature and precipitation data. The equation is given as follows:(1)xcor,i,j,k=xsim,i,j,k+(x¯obs,i,j,k−x¯sim,i,j,k)
+          where xsim,i,j,k and xcor,i,j,k are the simulated and bias-corrected i-th meteorological variables, respectively, at the j-th grid point on the k-th day, and x¯sim,i,j,k and x¯obs,i,j,k are the 30-year (1971–2000) averages of the simulated and observed i-th meteorological variables, respectively, at the j-th grid point on the k-th day [70,79,80].Previous studies noted that individual models perform differently in the simulation of different extreme indices [20,40,67,80,81,82,83]. Moreover, the performance of a multi-model ensemble is superior to that of most individual models [21,67,84]. In this study, we adopted the reliability ensemble averaging (REA) method to estimate the simulation extreme indices, and the multi-model weighted average change is defined as:(2)ΔT˜=A˜ΔT=∑iRiΔTi∑iRi
+          where the operator A˜ represents the REA operation and ∆Ti is the simulated change in the individual model output [85]. Variable Ri is a weight formulated as:(3)Ri=RB,im×RD,in=εTabs(BT,i)mεTabs(DT,i)n
+          where RB,i is a measure of the model performance criterion as a function of a bias factor (BT,i); RD,i is a measure of the model convergence criterion as a function of a distance factor (DT,i); BT,i is the bias between the simulated and observed output values over the baseline period (1971–2000); and DT,i is calculated by an iterative procedure. The initial hypothesis of DT,i was given by the difference between each model change and the simple ensemble averaging (defined as the mean of equally weighted models) change. Thereafter, the first guess of ΔT˜ was computed with Equations (2) and (3) and then subtracted from each model change to recalculate DT,i. The iteration was repeated until the procedure converged. The m and n parameters were employed to weigh each criterion, while εT is the difference between the maximum and minimum 10-year moving average values of the series after linear detrending [85].The performance profiles of the relative root mean square errors (RMSEs) of the extreme indices simulated by the CMIP5 models with respect to the observations in the 1971–2000 climatology are shown in Figure 4. To depict the RMSEs of the multiple variables at the same scale, the relative RMSE is defined as:(4)RMSEjR=100RMSEj-RMSEmedianRMSEmedian
+          where RMSEj is the RMSE of the j-th model and RMSEmedian is the median RMSE of the 14 and 10 CMIP5 models for extreme temperature and precipitation indices, respectively [20,86].Here, we applied three multi-model ensemble methods: the multi-model median (Median), simple model averaging (SMA, where each model is weighted equally), and REA methods. As shown in Figure 4, the performance of the individual models differed in terms of the simulation of the different extreme indices, especially the multi-model ensemble results. The RMSEs of the median, SMA, and REA were obviously lower than the RMSEs of single model; especially the REA results outperformed those of the individual models, and the values in the Yellow River basin were totally lower than those in the Yangtze River basin. Thus, we chose the REA results for further analysis and discussion in the following section.To evaluate the robustness of the model’s estimation in the future period, we calculated the uncertainty credibility of multi-model collective signals (SN):(5)SN=DN/DS
+          where the DN is the absolute differences of REA values between 2006–2100 and 1971–2000. The DS can be calculated as:(6)DS=1N∑i=0NEi−E¯2
+          where N is the number of models and the Ei is the annual mean value of i-th model in the future period, and E is the annual mean REA value in the future. When the SN is larger than 1, the model output result is robust, and the result is uncertain if SN is less than 1 [87].The spatial distribution of the annual mean observed and simulated extreme temperature indices from 1971–2000 is shown in Figure 5. In general, the ensemble TXx, TNn, DTR, ID, and SU index values were in good agreement with the observed data in the two basins, but the modeled WSDI and CSDI indices did not agree well with the observed indices (Table 4). The regional mean observed and simulated TXx ranged from 11 to 40 °C in the YLRB and YZRB, and the TNn ranged from −23.68 to −11.62 °C. The TXx and TNn index values were generally higher in the southeastern YZRB and the Sichuan Basin, and both were lower in the western region of the two basins. The observed and simulated WSDI and CSDI values were in the range of 4–26 and 8–25 days in the two basins, and the SU and ID ranged from 0 to 205 days and 0 to 242 days in the two basins. In addition, the DTR in the YLRB and YZRB ranged from 6 to 18 °C.The regional mean differences between the ensembled and observed TXx and TNn index values were 0.46 and 0.69 °C, respectively, in the YLRB and 0.15 and −0.15°C, respectively, in the YZRB. The TXx and TNn index values mainly revealed positive errors on the Loess Plateau and the central-western region of the two basins, and the maximum values were 1.5 and 2 °C, respectively, while negative errors mainly occurred in the western and central-eastern regions of the YZRB at −1.7 and −1.9 °C, respectively. The SU index value in the western region of the two basins was almost zero days, and the maximum value appeared in the southeastern YZRB at approximately 241 days. The average differences in the SU index values in the YLRB and YZRB were 1.15 and 2.56 days, respectively, and the maximum difference was observed in the Sichuan Basin at 10 days. The ID index value in most regions of the YZRB was smaller than 10 days, while the value was smaller than 100 days in most regions of the YLRB. A high ID index value mainly occurred in the western region (generally larger than 100 days), and the maximum value was approximately 202 days. Finally, the maximum and minimum differences of ID primarily occurred in the western and central regions of the two basins (10 and −7 days, respectively).The WSDI index value in the YLRB and western YZRB was generally lower than that in the eastern YZRB, with maximum and minimum values of approximately 6 and 26 days, respectively. However, the spatial distribution of the difference was the opposite to that of the observed values in general: the maximum and minimum differences occurred in the YLRB and eastern YZRB, and the difference ranged from −13 to 8 days. In addition, low CSDI index values were observed in the western regions of the two basins, with a minimum value of 8 days, and the highest value occurred in the Sichuan Basin at 22 days. The maximum and minimum differences in the CSDI index were observed in the western region of the two basins and the Sichuan Basin (−7 and 13 days, respectively), with absolute REs of 20.75 and 16.51% in the YLRB and YZRB, respectively. Finally, the DTR index values in the western regions of the two basins and on the Loess Plateau were much higher than those in the Sichuan Basin and the middle and lower reaches of the YZRB, and the value ranged from 6–17 °C. The ensemble DTR index values were higher than the observed values in the Sichuan Basin, with a maximum value of 0.11 °C, but were lower in the central region of the two basins, with a minimum value of −0.03 °C.Figure 6 shows the spatial patterns of the observed and simulated extreme precipitation indices in the YLRB and YZRB. In general, the ensemble values agreed very well with the observed values, especially the PRCPTOT index, but the ensemble Rx5day, CDD, and CWD index values were always larger than the observed values. Except for the CDD and CWD indices, all the other observed extreme precipitation indices were high in the central-eastern YZRB and low in the northwestern YLRB. The highest values of the PRCPTOT, Rx5day, R20mm, SDII, and R99pTOT indices were 1842 mm, 200 mm, 28 days, 12 mm/day, and 162 mm, respectively, and the minimum values were 106 mm, 19 mm, 0 days, 2 mm/day, and 6 mm, respectively. The maximum values of the CDD and CWD indices were observed in the northwestern YLRB and western YZRB (109 and 26 days, respectively), and the minimum values occurred in the Sichuan Basin and northern Loess Plateau (15 and 3 days, respectively).The regional mean observed and ensemble PRCPTOT were 411.24 and 438.41 mm in the YLRB, and 963.14 and 959.81 mm in the YZRB; the observed and ensemble SDII were 5.85 and 5.90 in the YLRB and 7.78 and 7.85 in the YZRB (Table 4). The spatial distributions of the differences in the PRCPTOT and SDII index values were scattered, and their extreme value distributions were not distinct, which was also true for those of the R99pTOT index. The differences in the R20mm and Rx5day index values were generally large in the central YZRB, with maxima of 33 mm and 1.7 days, respectively, but minimum values were observed in the southern YLRB and eastern YZRB (−0.6 days and −21 mm, respectively). Finally, the ensemble CDD and CWD index values were overestimated in the two basins, and the regional mean differences were 3.39 and 0.63 days, respectively, in the YLRB and 3 and 0.83 days, respectively, in the YZRB. The maximum difference in the CDD index value occurred in the western regions of the two basins at 20 days, and the CWD index exhibited the minimum difference at −3 days (Table 4 and Figure 6).Figure 7 shows the spatial distribution of the changes in the extreme temperature indices over the three periods of the 21st century based on the RCP4.5 and RCP8.5 scenarios. Under the RCP4.5 scenario, the TXx index showed an increase trend in the different periods in most regions of the YZRB and YLRB, and the largest increase occurred in the central-eastern part of the two basins in the 2050s period at approximately 0.8 °C/decade. However, in the 2080s period, a slight decline in the TXx index was found in the western YZRB, with a minimum value of approximately −0.1 °C/decade. Under the RCP8.5 scenario, the TXx index revealed the highest increase trend in the central YZRB in the 2050s period (1.5 °C/decade) and the lowest increase trend in most regions of the two basins (lower than 0.1 °C/decade). The TNn index in the 2080s period exhibited the largest decline in the central YZRB (−0.5 °C/decade) under the RCP4.5 scenario, but under the RCP8.5 scenario it showed increase trends in the three periods; the highest trend was found in the west region of the two basins (1.5 °C/decade) in the 2080s period.The SU index showed the highest increase trend in the YZRB in the 2050s period under the RCP4.5 scenario (approximately 10 days/decade), while in the 2080s period, it decreased at a rate of −3 days/decade in the central YZRB. Under the RCP8.5 scenario, the largest increase in the SU index also occurred in the western YZRB in the 2050s period, and the regional trends of the SU index in the 2020s and 2080s periods were 4.1 and 6.1 days/decade, respectively. The ID index mainly exhibited decrease trends in the YLRB and the western regions of the two basins under the RCP4.5 and RCP8.5 scenarios. The largest decline in the ID index under the RCP4.5 scenario occurred in the western YZRB and YLRB in the 2050s period (−4.5 days/decade), while in the 2080s period, slight increases in the ID index occurred in the YLRB, at a maximum value of 2 days/decade. Under the RCP8.5 scenario, the ID index decreased in the three periods, and the largest decline was found in the western YZRB at −18 days/decade in the 2050s period.The WSDI index under the RCP4.5 scenario revealed the largest increase in the 2050s period, but its increase trend decreased in the 2080s period, and there was a downward trend in some places (the largest decline was −4 days/decade). However, under the RCP8.5 scenario, the increase rate of the WSDI index increased over time, and the maximum rate occurred in the western YZRB in the 2080s period (23 days/decade). In addition, the CSDI index under both the RCP4.5 and RCP8.5 scenarios exhibited the largest declines in the 2020s period in the two basins, with a minimum value of approximately −6 days/decade, but slight increases occurred in the two basins in the 2050s and 2080s periods under the RCP4.5 scenario, while the maximum value was 1 day/decade. Under the RCP8.5 scenario, the regional decrease rates of the CSDI index in the two basins were −1.2 and −0.3 days/decade.Finally, the DTR index in the YZRB generally revealed an increase trend under the RCP4.5 scenario, especially in the middle and lower reaches, and the maximum value was 0.2 °C/decade in the 2050s period. However, under the RCP8.5 scenario, the largest increase was found in the central YZRB in the 2080s period (approximately 0.15 °C/decade). The DTR index generally exhibited a decrease trend in the northwestern region of the two basins, and the largest decline occurred in the 2020s period under the RCP8.5 scenario at −0.11 °C/decade.Figure 8 shows the spatial distribution of the changes in the precipitation extremes in the three periods based on the RCP4.5 and RCP8.5 scenarios. The Rx5day index in the YZRB exhibited the largest decline in the 2080s period under the RCP4.5 scenario at −28 mm/decade; under the RCP8.5 scenario, it revealed the highest increase trend in the 2080s period in the YZRB at 47 mm/decade. A regional decrease trend only occurred in the YLRB in the 2020s period at −0.37 mm/decade under the RCP4.5 scenario.The R99pTOT index under the RCP4.5 scenario generally exhibited increase trends in the YZRB and slight increases in the YLRB in the 21st century, and the regional mean trends were 5.8, 5.5, and 3.5 mm/decade, but the largest decline also occurred in the YZRB in the 2080s period at −43 mm/decade. Under the RCP8.5 scenario, the R99pTOT index in the YLRB and YZRB revealed the highest increase trend in the two basins, especially in the 2080s period in the YZRB, with maximum and average mean values of 106 and 21 mm/decade, respectively. The PRCPTOT index exhibited the highest increase trend in the 2020s period in the YZRB at a rate of 134 mm/decade and revealed the largest decline in the 2050s period at approximately −80 mm/decade under the RCP4.5 scenario. Under the RCP8.5 scenario, the PRCPTOT index had a decrease trend in the central region of the two basins in the 2020s period, with the largest decline equaling approximately −97 mm/decade, and exhibited the highest increase trend in the 2050s period at approximately 120 mm/decade.Under both the RCP4.5 and RCP8.5 scenarios, the SDII index values in the two basins revealed slight increases in the three periods, and the regional mean trends were all lower than 0.3 mm/day/decade. The highest increase trends (approximately 0.8 mm/day/decade) were primarily found in the eastern region of the two basins in the 2050s period under the RCP4.5 scenario. In the 2020s and 2080s periods, there were a few decreases in the central YLRB and southeastern YZRB at approximately −0.3 mm/day/decade. Under the RCP8.5 scenario, there was a slight decrease in the central YZRB in the 2020s period at approximately −0.15 mm/days/decade, while in the 2050s and 2080s periods, the increase trends in the YZRB were 0.18 and 0.27 mm/day/decade, respectively. The R20mm index also increased slightly in the three periods under the RCP4.5 scenario, and increases generally occurred in the YZRB. In the 2080s period, the R20mm index in the YZRB exhibited a decrease trend, with highest and regional mean values of −1.8 and −0.02 days/decade, respectively. Under the RCP8.5 scenario, slight decrease trends of the R20mm index were found in the central YZRB in the 2020s period at approximately −0.4 days/decade, but the regional mean trend in the two basins was 0.15 days/decade, and the largest increases appeared in the YZRB in the 2050s at 2.9 days/decade.Finally, under the RCP4.5 scenario, the CDD index mainly revealed a decrease trend in the YLRB and YZRB in the 2020s period at rates of −1.22 and −0.42 days/decade, while in the 2050s period, the CDD index increased at a rate of 0.18 days/decade in the YLRB but decreased at a rate of −0.25 days/decade in the YZRB. In the 2080s, the CDD index values in the two basins both increased at a rate of 0.1 days/decade. Under the RCP8.5 scenario, except in the YZRB in the 2080s period (0.55 days/decade), the CDD index values all revealed decrease trends, and the minimum rate was approximately −10 days/decade. The CWD index in the two basins exhibited decrease trends in the 2020s and 2080s periods (at −0.04 and −0.02 days/decade, respectively) but increases in the 2050s period at a rate of 0.15 days/decade. Under the RCP8.5 scenario, the highest increase trend of the CWD index was observed in the southeastern YZRB in the 2050s period at 2 days/decade, and the highest decrease trend occurred in the western YZRB with −2.2 days/decade.Figure 9 and Figure 10 show the projected changes in the regional mean extreme temperature and precipitation indices under the RCP4.5 and RCP8.5 scenarios. The variations in the extreme temperature indices in the YLRB and YZRB exhibited a strong consistency between the historical and future periods. There were general increases in most extreme temperature indices but notable decreases in the CSDI and ID indices in both basins. The trends of extreme temperature indices in the early 21st century were relatively similar under the RCP4.5 and RCP8.5 scenarios, but the values rapidly increased after 2040 under the RCP8.5 scenario in the two basins, except the DTR index. The values changed less after 2060, and the ID index even revealed an increase trend after 2080 under the RCP4.5 scenario. The extreme precipitation indices revealed trends were similar to those of the extreme temperature indices in the two basins, but the values in the YZRB were of a greater magnitude of change than those in the YLRB, except for the CDD index, indicating a wetter climate in the YZRB in future decades. However, after 2080, most of the indices indicated more notable trends than those in the early period under the RCP8.5 scenario but revealed the opposite trends under the RCP4.5 scenario.Table 5 shows the trends of extreme climate indices in the YLRB and YZRB for the historical period and under the RCP4.5 and 8.5 scenarios in the future period. Almost all the extreme climate indices showed significant trends in the two basins, except the CWD under the RCP8.5 scenario in the YLRB and YZRB. The extreme cold events in the two basins all showed down trends, like ID, CSDI, and CDD, and the decline in the YLRB was higher than that in the YZRB. For example, the trends of ID in the YLRB ranged from −2.32 to −4.95 days/decade (p < 0.01), while in the YZRB, the decreasing trends were in the range of −1.29—3.02 days/decade (p < 0.01). The increasing trends of TXx and SU were generally higher in the YLRB than those in the YZRB during the historical period, but lower in the future period. The DTR decreased significantly during the historical period but increased significantly under the RCP4.5 and 8.5 scenarios.The extreme precipitation indices generally increased faster in the YZRB than those in the YLRB, except the CDD, which showed significant decrease trends in the two basins, especially in the YLRB. At last, CWD generally showed significant increase trends in the two basins in the historical period and under the RCP4.5 scenario; under the RCP8.5 scenario, however, CWD decreased significantly in both the YLRB and YZRB.Figure 11 shows the proportion of the area with SN value greater than one in the study areas under the RCP4.5 and 8.5 scenarios. The reliability of the TXx, TNn, SU, and WSDI was higher than the ID, CSDI, and DTR in the YLRB and YZRB, and the reliability under the RCP8.5 scenario was higher than that of 4.5 scenario. The extreme precipitation indices generally showed lower reliability compared with the extreme temperature in the study areas and under different scenarios [88,89].The spatial distributions of the extreme temperature indices during the baseline period revealed notable relationships with the terrain or altitude [41]. For example, the TXx, TNn, SU, WSDI, and CSDI indices were generally low in the western regions of the two basins, which are the eastern parts of the Tibetan Plateau, where the altitude is generally higher than 3000 m, as well as on the Loess Plateau (northwestern YLRB). However, these indices were high in the Sichuan Basin and the lower reaches of the YZRB and YLRB, where the altitude is generally lower than 500 m. Most of the extreme precipitation indices also exhibited clear geographical differences. For example, extreme precipitation events generally occurred in the lower reaches of the YZRB but were rarely observed in the western regions of the two basins, which may be mainly influenced by monsoons.We also calculated the average value of water vapor flux and water vapor flux divergence at 850 hPa in summer and winter from 1971–2000 (Figure 12). Along the western margin of the Sichuan Basin, water vapor transport is blocked by the Tibetan Plateau, and a humid climate was thus observed in the southwest region of the Tibetan Plateau [35]. Overall, the risk of extreme cold events is relatively high on the eastern Tibetan Plateau and the Loess Plateau regions, where the extremely high precipitation is relatively low and the area is prone to drought events. The risk of extreme hot temperatures were relatively high in the middle and lower reaches of the YZRB, where the extreme precipitation events are relatively high.In general, the warm-temperature indices in the future decade in the two basins generally exhibited increase trends in the 2020s and 2050s periods in the two basins under the RCP4.5 and RCP8.5 scenarios, but the extreme high-temperature indices in the YZRB increased faster than those in the YLRB, such as the TXx, SU, and WSDI indices. In the YLRB, the extreme low-temperature index increased faster than that in the YZRB (the TNn index), and the ID index also decreased faster than that in the YZRB, which agrees with the study of Li et al. [90]. The warm regions were generally located on the southeastern Tibetan Plateau and northern YLRB. This area may be the focus because these relatively high-altitude regions receive more positive albedo-temperature feedback, and this capacity creates a higher temperature increase than that in the relatively low-altitude regions in the YLRB and YZRB [46]. However, in the 2080s period, the warming trends in the two basins slowed, and the temperature even became colder in certain regions. Under the RCP8.5 scenario, the two basins would continue to warm in future decades, and in the Tibetan Plateau region, the TNn, SU and WSDI indices would increase faster than those in the other regions, and the ID index would decrease considerably. Figure 13 shows the summer and winter mean downward solar radiation flux during 1971–2000. The radiation value was higher in northwest of China in summer and southwestern China in winter, while the spatial distribution of surface solar radiation was not the same as that of extreme high temperatures, indicating that surface solar radiation may not be the main cause of the spatial distributions of extreme high temperature, but the study of Hu et al. [91] noted that the net surface radiation flux and 500-hPa geopotential height indicated an enhancement of the net radiation in northern China and a weakening of the East Asian trough in winter under the RCP4.5 and RCP8.5 scenarios, which may explain the increase in the winter temperature in the YLRB.In the future, the regional mean extreme precipitation events exhibited limited increase trends in the two basins under the RCP4.5 scenario, and the increases were mainly found in the YZRB from 2020–2060. The flood peaks in the YZRB also increased 0.3–13.1% in this period [92]. Over the last decade of the 21st century, the frequency of extreme precipitation events was reduced in the middle and lower reaches of the YZRB. Under the RCP8.5 scenario, the rate of extreme precipitation events would continue to increase, especially on the eastern Tibetan Plateau and lower reaches of the YZRB. However, the number of extreme wet days would decrease, indicating that extreme precipitation would occur frequently.The air temperature is an important factor influencing the mean precipitation and extreme precipitation. With increasing air temperature, the water vapor in the atmosphere increases nonlinearly with the temperature, which is responsible for the increase in precipitation. However, the warming climate increases the water-holding capacity of the atmosphere and thus increases the atmospheric precipitable water. A more stable atmospheric structure makes it more difficult for water vapor to condense and precipitate, but the precipitation intensity can increase once an event occurs [46]. In addition, Wu et al. [93] noted that the water vapor flux convergence under the RCP8.5 scenario exhibited large increases in the eastern and western YZRB but decreases in the central YZRB and northeastern YLRB, which would contribute to the changes in extreme precipitation events in these two basins. Rai et al. [94] reported that the mean precipitation exhibited strong and significant relationships with extreme precipitation events in the mid-late 21st century. Finally, Zhou et al. [95] also distinguished changes in the precipitation characteristics due more to external radiative forcing than to the internal climate variability.In summary, the spatiotemporal distributions of seven extreme temperature indices and seven extreme precipitation indices based on 16 CMIP5 models in the YLRB and YZRB in China from 1961 to 2099 were analyzed. The statistical analysis indicated that 10 and 14 CMIP5 models met the requirements for the calculation of these extreme temperature and precipitation indices, respectively, and the REA method provided the best simulation results of the extreme climate indices for the baseline period (1971–2000); the future trends (2010–2099) of the extreme climate indices were studied based on the REA values.The spatial distributions of the warm-temperature indices generally revealed high values in the Sichuan Basin and the middle and lower reaches of the YZRB, followed by the Loess Plateau, and low values in the Tibetan Plateau areas. The REA method usually overestimated the temperature indices in the YZRB and YLRB; for example, the ensemble TXx, TNn, SU, ID, and DTR index values usually agreed well with the observed values, except WSDI and CSDI indices, and only the ensemble TNn and WSDI index values in the YZRB were underestimated. In addition, most of the observed extreme precipitation indices over the baseline period were generally high in the middle and lower reaches of the YZRB and low on the Tibetan Plateau and Loess Plateau. The ensemble extreme precipitation indices were also overestimated by the REA method, especially in the central-eastern YZRB, but the PRCPTOT index was underestimated in the YLRB and YZRB.Warming trends were observed in the YLRB and YZRB in the 21st century under the RCP4.5 and RCP8.5 scenarios. The warm-temperature indices (the TXx, SU and WSDI indices) in the YZRB increased faster, but the cold-temperature indices such as the TNn and ID indices changed faster than those in the YZRB. Under the RCP4.5 scenario, maximum rates of the warming trends generally occurred in the 2050s period in the YZRB, while in the 2080s period, the warming trends would slow down and even become negative in certain regions, such as in the central area of the YZRB. Under the RCP8.5 scenario, the warming trends continued, especially on the Tibetan Plateau and central area of the YZRB. The extreme precipitation events in the 21st century continued to increase in the YLRB and YZRB. Similar to the changes in the extreme temperature indices, the extreme precipitation indices increased faster in the 2020s and 2050s periods under the RCP4.5 scenario, especially in the central YZRB, the Tibetan Plateau, and the southeastern YLRB. In the 2080s period, the increase trends decelerated. Under the RCP8.5 scenario, extreme precipitation events continued to increase, especially in the last decades of the 21st century. In addition, the number of extreme wet days increased, indicating that extreme precipitation would appear more frequently.The long-term anomalies of the extreme temperature and precipitation indices in the YZRB and YLRB indicated that the warming and wetting trends in the two basins were the same under the RCP4.5 and RCP8.5 scenarios before 2040. The trends would continue to increase after 2050 under the RCP8.5 scenario, but the warming trends would decrease after 2060 under the RCP4.5 scenario, while the wetting trends would even decrease after 2090.The spatial and temporal distributions of the extreme temperature were controlled by the atmospheric circulation and the download solar radiation in this study. The future frequency and intensity changes of extreme temperature events are also affected by the background of global surface temperature warming, which is related to the increase in greenhouse gas emissions caused by human activities (such as the process of urbanization). The increase in the mean air temperature will cause increases in the mean precipitation and extreme precipitation, especially in the YZRB. The YZRB will experience stronger extreme precipitation processes in future decades than those experienced over the past half century, and an increased risk of floods will also occur. Extreme precipitation will appear more concentrated, along with a shorter duration. This outcome may occur because of the increased water vapor transport in summer and weakened East Asian winter monsoon (EAWM).There are still certain problems that have not been resolved, such as the accuracy of the different individual models, a comparison to the estimation results obtained with CMIP6 model data, and the impacts of climate change on surface processes. These aspects will be examined in depth in future work.Z.N. and L.F. designed the research; Z.N. and X.Y. performed the experiments and analyzed the data; Z.N. wrote the manuscript; X.C. and L.F. revised the manuscript. All authors have read and agreed to the published version of the manuscript.This work was financially supported by National Natural Science Foundation of China (41975044) and the Fundamental Research Funds for National Universities, China University of Geosciences (Wuhan).Not applicable.Not applicable.The daily meteorological data is provided by the National Meteorological Information Center (NMIC), China Meteorological Administration (CMA) (http://data.cma.cn, accessed on 24/9/2019). The CMIP5 model data is provided by the World Climate Research Programme (CWRP), Earth System Grid Federation (ESGF, https://esgf-node.llnl.gov/projects/cmip5/, accessed on 10/2/2020).We thank the National Meteorological Information Center (NMIC), China Meteorological Administration (CMA) for providing the temperature and precipitation data. Thanks to the three anonymous reviewers for their suggestions and comments for this paper, which provided many new ideas for our study. We also thank Carol Yang/Assistant Editor for helping us in the process of publishing this paper.The authors declare no conflict of interest.Geographical information of the Yellow River basin and Yangtze River basin.Taylor diagrams for the simulations of 30-annual mean Tm, Tx, Tn, and Pre of the two basins during the 30-year period (1971–2000).Annual mean Tm, Tn, Tx, and Pre from observed data and original CMIP5 model data during 1971–2000 in the Yellow River Basin and Yangtze River Basin.The relative RMSEs of extreme weather indices between the observed and the bias-corrected GCM simulations over the Yellow River Basin and Yangtze River Basin.Spatial distribution of seven extreme temperature indices for the annual mean observations (OBS, the first and third columns) and simulated (REA, the second and forth columns) values in the 1971–2000 period over the Yellow River Basin and Yangtze River Basin.Same as in Figure 5 but for the extreme precipitation indices.Spatial distributions of trends in seven extreme temperature indices for the 2020s (2010–2039), 2050s (2040–2069), and 2080s (2070–2099) periods in the Yellow River Basin and Yangtze River Basin under the RCP4.5 and 8.5 scenarios. The black solid dots indicate the trends are significant at the 95% significance level.Same as in Figure 7 but for the extreme precipitation indices.Annual regional mean of seven extreme temperature indices for the history and future periods.Annual regional mean of seven extreme precipitation indices for the historical and future periods.The proportion of the area with the SN value greater than one in the study areas under the RCP 4.5 and 8.5 scenarios.The summer and winter mean water vapor flux (vector arrow, g/cm/hPa/s) and water vapor flux divergence (shadow, g/cm2/hPa/s) at 850 hPa during 1971–2000.The summer and winter mean downward solar radiation flux (W/m2) during 1971–2000. It is not the net downward shortwave radiation flux.List of 22 CMIP5 models used in this study.The average and trends of Tm, Tn, Tx, and Pre from observed and downscaled model data during 1971–2000.* Significant at the 95% confidence level, ** significant at the 99% confidence level (MK test).Information of 14 ETCCDI extreme climate indices.Regional mean values of observed extreme climate indices (OSB) and differences of REA and OBS values (REA-OBS) in the Yellow River Basin and Yangtze River Basin.The trends of extreme climate indices (unit/decade) in the Yellow River basin and Yangtze River basin for the historical period and under the RCP4.5 and 8.5 scenarios in the future period.* Significant at the 95% confidence level, ** significant at the 99% confidence level (MK test).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06030.txt ADDED Viewed

	@@ -0,0 +1,4 @@

+While the COVID-19 has dramatically altered our lifestyle and sleep practices, the links between sleep, individual characteristics, personal experiences and mental health during the pandemic require further examination. This cross-sectional, multi-methods study examined differences in language used to describe personal experiences, and mental health, based on sleep quality during the early stages of the pandemic. N = 1745 participants (mean age 42.97 ± 14.46 years) from 63 countries responded to the survey. Sleep quality was assessed using the Pittsburgh Sleep Quality Index and mental health was examined using the Patient Health Questionnaire-9, the State Trait Anxiety Inventory, the Perceived Stress Scale and the UCLA-Loneliness Scale. Quantitative analysis of qualitative, language content of personal experiences was conducted using free-text responses and comments to a question on the survey. Almost 50% of the participants reported poor sleep quality, which was linked to a more negative emotional tone and greater mentions of money or finance related words. Good sleepers reported more positive emotional tone in their experiences. Greater reports of clinical state anxiety, moderate depression and moderate stress were observed in poor sleepers, even after accounting for demographics and pandemic-related factors such as loneliness, financial concerns and risk of contracting COVID-19 disease. Results from this study highlight an urgent need for sleep-related public health interventions. Practitioner education, sleep screening for those with mental health conditions, and encouraging people to adopt digital tools may help to reduce the burden of poor sleep on mental health. While the pandemic itself is a stressful and uncertain time, improving sleep can support positive emotion regulation, improving mood and consequential action.Good sleep quality plays a crucial role in maintaining positive well-being and mental health [1,2]. By contrast, sleep disturbances and related daytime dysfunction are a risk factor for reporting poor mental health [3,4,5,6,7]. As an example, sleep disturbances, such as insomnia, predict later onset of depression, anxiety, alcohol abuse and other psychotic illnesses [5,8,9]. The use of evidence-based treatments for sleep in individuals with psychiatric illness is beneficial for both sleep and mental health, even when the latter is not a primary target of the intervention [8,10,11]. Examining the interactions between sleep and mental health is now more important than ever, as the COVID-19 pandemic continues to disrupt daily life. While the negative effects of this pandemic on mental health are evident [12,13,14,15,16,17], their links with sleep disturbances require further investigation. Research investigating sleep disturbance may have translational benefits of improving sleep and reducing global mental health burden.A number of studies have emerged since the pandemic began, indicating an overall decline in sleep quality and quantity [18,19,20,21,22]. These findings mirror previous pandemic studies (e.g., Ebola), where higher rates of poor sleep and greater presence of sleep disorders was observed in regions with higher infection rates [23,24]. Some studies have also shown improvements in sleep characteristics during the pandemic [25,26,27]. For example, Wright and colleagues, [27] reported an increase in total time in bed amongst university students during stay-at-home orders while Blume and colleagues [26] observed a reduction in social jetlag. Together, this demonstrates that the stay-at-home orders and other government restrictions may allow for more flexibility in sleep-wake routines, which in turn can increase sleep opportunity. However, in both studies, authors noted that despite improvements in sleep-wake flexibility and reduced social jet-lag, sleep quality during the pandemic still decreased. Other studies have also observed this decline in sleep quality and increase in insomnia symptoms during different stages of COVID-19 pandemic across different countries and populations [28,29,30,31,32]. Changes in lifestyle, pandemic stress, risk of contracting COVID-19 disease and financial distress may reduce sleep quality despite the availability of greater sleep opportunity. Similarly, studies have also documented the impact of the pandemic on mental health, including increased rates of depression, anxiety, stress and loneliness [15,18,33,34,35,36].Based on pre-pandemic literature, poor sleep quality may be related to negative personal experiences and poor mental health during the pandemic. Poor sleep quality may explain why mental health continues to decline despite the increase in sleep opportunity. Significant changes to individual lifestyles and pandemic-related concerns (such as fears for personal or family health and anxieties over job stability) may increase stress and autonomic arousal, thus reducing sleep quality. In turn, these changes in sleep can exacerbate negative perception of events via an amplified amygdala response to stimuli [37], impacting emotion regulation and increasing stress. Together, this might explain the relationship between poor sleep and increased risk of depression and anxiety [7]. A positive feedback loop may exist, wherein increased presence of sleep problems and stressors progressively worsen each other through cyclic reinforcement, ultimately increasing the risk of experiencing anxiety and depression symptoms. As a result, it is important to understand stressors in people’s lived experiences, and how poor sleep quality may be associated with mental health during the pandemic.While it is unclear what the ongoing impacts of the pandemic on mental health will be, our treatment responses need to adapt. Knowing the nuanced experiences felt during the pandemic can enrich our understanding of sleep and mental health, helping improve intervention strategies. Computerized text analysis systems like the Linguistic Inquiry and Word Count (LIWC) can be used to capture common qualitative narratives, quantitatively, in large datasets by decreasing labor and increasing objectivity [38]. The LIWC uses word processing and dictionaries to detect the frequency of word types across a variety of psychological categories, including emotional tone, social processes and thinking styles. In the context of the current study, a combination of trends in the language used to describe personal experiences and a summary of quantitative sleep and mental health characteristics can help provide a nuanced and detailed understanding of peoples experience during the pandemic, which may help inform future treatments and mitigation strategies.Accordingly, this study aimed to; (a) determine the differences in language used to describe personal experiences of good sleepers compared to poor sleepers, and (b) explore associations between mental health and poor sleep. Specific demographic characteristics (such as age, gender and previously diagnosed mental health condition), and pandemic-related factors, such as loneliness, negative changes in financial situation (such as loss of job or income) and the risk of contracting the COVID-19 disease were also examined as covariates of associations between mental health and poor sleep.This study presents data from the first wave of surveys disseminated globally between 9 April and 25 May 2020. The survey is a part of a longitudinal study examining changes in sleep and mental health across the pandemic. Some of the data from the baseline survey has been published previously [17]. The study was approved by Monash University Human Research Ethics Committee and conducted in accordance with Declaration of Helsinki. Participants were recruited via social media channels, including Facebook, Twitter and LinkedIn. Participants were given an opportunity to respond to the survey anonymously or provide their email at the start of the survey if they were interested in participating in the longitudinal waves of the study.The survey included demographic items (age, sex, education, employment/student status, country of residence, ethnicity, marital status, and number and age of dependents), self-reported history of mental health diagnoses [“Have you been diagnosed with a mental health condition (either in the past or currently)?”], alcohol use (“Has the amount of alcohol, or the frequency with which you consume alcohol, changed since the COVID-19 crisis began? If yes, in what way?”), and questions related to stay at home orders [“Are you currently under lockdown or self-isolating due to the COVID-19 pandemic? (e.g., government orders for staying at home, working from home, limiting contact with others)?”], risk of contracting COVID-19 disease [“Are you at a high risk of COVID-19? (High risk refers to being over 70 years old; over 50 years old and of Aboriginal descent; pregnant; parent of a child under 12 months; under treatment for chronic health conditions, immune compromised.”)], financial situation (“Has the COVID-19 pandemic impacted your employment status?” “Has the COVID-19 pandemic impacted your financial situation?”), distress due to a change in financial situation (“How much distress has this change in financial status caused you?”), and type of change in their sleep patterns [“Have your sleep patterns changed since the start of the pandemic?”, “If yes, how have your sleep patterns changed?”, and “Do you think your sleep-wake routine (e.g., bed-time and get-up time) is more consistent with your personal preference or “body clock”, than before the lock down/self-isolation/quarantine period? (e.g., night owl able to go to bed later as you don’t have to get up early for work)]. Participants were also asked to self-report their sleep quality just before the pandemic in comparison to their sleep during the data collection period. An optional, free-text question on comments regarding sleep and experiences during the pandemic was also included (“Are there any comments you would like to make about your sleep or changes to your life during COVID-19 pandemic?”).The Pittsburgh Sleep Quality Index (PSQI) was used as a validated measure of self-reported sleep quality. The questionnaire has seven domains: subjective sleep quality; sleep latency; habitual sleep efficiency; sleep disturbances; use of sleep medication; and daytime dysfunction (e.g., sleepiness, lack of enthusiasm to carry on daily activities). The seven domains are scored from 0–3, with 3 indicating sleep disturbance in a particular domain, with a total score of 21. PSQI has high reliability (Cronbach’s alpha of 0.85) and validity [39] and is one of the primary measure used in assessment of sleep health. To classify participants as “good” or “poor” sleepers, PSQI ≥ 8 was used as a cut-off [39,40,41], which is a validated cut-off point used in other studies. PSQI score of eight was also the median score for this cohort.The Patient Health Questionnaire (PHQ-9) was used to examine depressive symptoms over the last 2 weeks. PHQ-9 scores range from 0 to 27, with scores ≥10 recommended as a cut-off for moderate depression. The scale has excellent reliability (Cronbach’s alpha of 0.89) and high construct validity, making it one of the primary instruments used for depression screening [42,43]. Anxiety was assessed using the well validated and reliable 6-item State Trait Anxiety Inventory (STAI) [44]. Six items related to anxiety were rated on a 4-point scale, with higher scores indicating more anxiety. A score of 40 or more was used as the cut-off for clinical anxiety. The Perceived Stress Scale (PSS) was used to assess current stress state and has excellent psychometric properties, with scores above 14 suggesting moderate to severe levels of stress [45,46]. Loneliness was examined using the UCLA-Loneliness scale short form (UCLA-LS), which has three questions that measure loneliness and social isolation.All quantitative data except for the LIWC group differences were analysed using IBM SPSS (version 26.0, IBM, Armonk, NY, USA). PSQI cut-off of eight or more was used an indicator of poor sleep and participants were accordingly divided into “good” or “poor” sleepers based on their PSQI scores. Chi-square analysis were used to examine differences in demographic characteristics among good and poor sleepers.To examine differences in the language participants used to describe their personal experiences during the pandemic, content analysis of language was performed using the Linguistic Inquiry and Word Count (LIWC) software [38] and further analyzed using R (version 3.6.3) [47]. The LIWC dictionary includes psychologically validated categories of words (e.g., anxiety words could include “nervous”, “afraid”, or “tense”). Frequency of a word category is measured by the percentage of occurrence of a particular word category within a piece of text. In doing so, the LIWC allows the gathering of quantitative summary data from qualitative writing. Prior to the LIWC analysis, data were cleaned for missing text (nmissing = 431) in comments (nremaining = 1103). Non-parametric Mann-Whitney U tests were then used to examine content differences on LIWC word categories between poor sleepers and good sleepers. The LIWC word categories of interest included the 33 sub-categories of psychological processes, including affective processes, biological processes, drives, perceptual processes, personal concerns, and social processes [48]. To control for alpha inflation, all significance values were compared to their critical value after a Benjamini-Hochberg correction [49]. Prior to conducting LIWC analysis, a logistic regression (in R version 3.6.3) was also performed on the demographics to identify what factors influenced the likelihood of providing a free-text comment (used in the LIWC analysis). The most frequently observed category was used as the dummy category (i.e., female, bad sleepers, Australia-New Zealand, in lockdown, bachelor’s degree or higher, and employed full-time).To examine the incidence of poor sleep, clinical levels of state anxiety, moderate depressive symptoms and moderate stress among participants in this sample, cut-offs for each scale (PSQI, STAI, PHQ-9 and PSS) were used to create dichotomous variables. Associations between these dichotomous variables and poor sleep were analyzed using binary logistic regression. Two models were examined. The first model included age and sex as covariates along with poor sleep. The second model examined prior diagnosis of a mental health condition, and pandemic-related factors such as, isolation, negative change in financial situation and risk of contracting COVID-19 disease as covariates along with poor sleep. Odds ratios (with 95% CI) were calculated to examine prevalence of anxiety, depression and anxiety in individuals with poor sleep.A total of 2555 participants responded to the survey. Data from 1745 respondents (who provided complete response to PSQI) were included in the study. Based on PSQI scores, 47% of the participants reported poor sleep quality. Within the PSQI subscales, 55% percent of respondents reported having difficulty falling asleep at least 2 nights per week. General sleep disturbances, such as feeling hot or cold at night, having pain or experiencing nightmares (at least 2 times per week) were reported by 38% of the participants. When asked to compare their sleep quality during the pandemic with their sleep pre-pandemic, 20% of the participants reported improved sleep, whereas 57% reported poorer sleep compared to their sleep pre-pandemic. Overall, 74% of participants reported that their sleep patterns had changed since the start of COVID-19 pandemic. Of these, the majority of respondents reported that they were sleeping and waking up later than their usual bedtimes/waketimes before the pandemic began (n = 361, 22% of the sub-sample). Forty-three percent of the sample (n = 685) reported that their sleep patterns were now more in sync with their body clock (i.e., their personal preferences for wake and sleep) compared to before the pandemic.Participant characteristics across good and poor sleepers are reported in Table 1. Differences in sleep quality and mental health based on participants’ country of residence have been reported previously [17]. Broadly, small, non-significant differences were observed between countries with the highest response rate (Australia, India, United Kingdom, South Africa and United States of America).More than 90% of the participants were in a government-imposed lockdown at the time of completing the survey, broadly defined as stay-at-home orders, or going out only for shopping essential items or for work when such work could not be performed from home. Poor sleepers reported higher levels of unemployment and disability than good sleepers. Poor sleepers also reported significant increase in their alcohol consumption since the start of the pandemic as opposed to good sleepers.Since the question related to personal experiences was optional, a logistic regression was conducted to determine which demographic groups were more likely to respond (Supplementary Table S1). Specifically, age, poor sleep quality, country of residence in Asia or Africa was associated with greater odds of responding to the question. Gender, lockdown status and education were not related to the likelihood of providing a free-text answer.Next, differences in the language used by participants to describe their personal experiences based on whether they were good or poor sleepers were analyzed using LIWC. Of the 33 LIWC categories relating to psychological processes, eight word-categories demonstrated significant differences between poor and good sleepers after Benjamini-Hochberg correction of critical values. Between poor sleepers and good sleepers’ significant differences were demonstrated in their emotionality (i.e., emotional tone), social processes, and money-related words (see Table 2). Notably, poorer sleepers used more negative emotion and anxiety related words in their comments. Examples of these comments from poor sleepers:
+        “Feeling anxious, either can’t sleep or sleep 12 h...” and,“I cannot find a job because of the situation, that has caused me a lot of distress, anxiety, and therefore I’m taking pills for sleep”.“Feeling anxious, either can’t sleep or sleep 12 h...” and,“I cannot find a job because of the situation, that has caused me a lot of distress, anxiety, and therefore I’m taking pills for sleep”.Additionally, poorer sleepers had a higher frequency of words relating to money, which may result from financial concerns as reflected in the following comment:
+        “The insecurity of income, oppressive and extended lockdown of undetermined duration and potential of violent uprisings due to hunger as well as worrying about unpaid rental and utility bills is causing extreme anxiety.”“The insecurity of income, oppressive and extended lockdown of undetermined duration and potential of violent uprisings due to hunger as well as worrying about unpaid rental and utility bills is causing extreme anxiety.”Conversely, good sleepers used more words relating to positive affect and social processes, including family and friends. One good sleeper commented:
+        “Able to sleep properly may be due to spending more time with family.”“Able to sleep properly may be due to spending more time with family.”Both regression models revealed greater odds of clinical state anxiety, moderate depression and moderate stress among participants with poor sleep (Table 3). In particular, Model 1 showed that poor sleep quality was associated with up to four times higher odds of reporting anxiety, depression and stress.Age was a significant covariate, with lower odds of poor mental health observed with increasing age. Pandemic-related factors, such as loneliness and negative change in financial status were associated with higher odds of reporting depression and anxiety. However, even after accounting for pandemic-related factors and prior diagnosis of a mental health condition, the risk of reporting depression and or anxiety was three times higher in poor sleepers [OR = 3.0 (CI = 2.37−4.27), OR = 3.01 (CI = 2.37−3.84), respectively] as opposed to good sleepers.Like depression and anxiety, greater loneliness and a negative change in financial situation were associated with higher reports of stress. Interestingly, individuals who self-reported as not at risk of contracting COVID-19 were more stressed that those who were at-risk.This study aimed to examine the links between sleep, differences in language used to describe personal experiences and mental health during the early stages of the COVID-19 pandemic. Overall, poor sleep quality was observed in the current sample, with almost half of the participants scoring above the cut-off of 8 on the PSQI. Similar to past research [12,13,50,51,52], elevated levels of stress, anxiety and depression were reported in this study.Additionally, the LIWC language analysis revealed that individuals with good sleep had a more positive emotional tone when reporting on their personal experiences, consistent with our findings on validated measures of mental health. Conversely, poor sleepers used more negative emotional tone words, displayed more anxiety, frequently mentioned finance related words as opposed to words related to social interactions or processes. This was also observed in validated questionnaires, where poor sleepers were more likely to indicate changes in their financial status, increased alcohol consumption, and were at least three times more likely to report depression, anxiety and stress. In particular, poor sleep was associated with the highest odds of reporting clinical state anxiety, moderate depression and moderate levels of stress, even after controlling for demographic and pandemic-related factors. Together, this suggests that independent of pandemic-related factors or demographic differences, sleep is important for overall wellbeing.While previous studies have reported an increase in sleep-wake flexibility and greater sleep opportunity [20,45,53,54], the current study observed poorer sleep quality and an increase in insomnia-related symptoms. This may be explained by a number of factors. Firstly, the increased mental health burden due to isolation, lockdowns and change in lifestyles may have heightened arousal, increasing both stress and poor sleep. Secondly, time gained from not needing to commute due to stay-at-home orders, may be used for work and not for sleep [55]. As a result, overall sleep quality may not have changed. Further, average bed-and wake-times have shifted during the pandemic [56], potentially leading to less sleep-wake regularity and decreased sleep quality. Finally, lockdowns, restriction in outdoor activities and excessive use of light-emitting screens may increase circadian misalignment due to lack of exercise and reduced natural light exposure. Pre-bed light hygiene like frequent phone use before bedtime, may have worsened sleep quality [57]. Interestingly, in the current study, only 25% of the participants were receiving more than 30 min of daily sunlight exposure, and individuals with poor sleep quality reported increased phone usage at night (in comparison to before the pandemic). This may involve ‘doomscrolling’, or increased consumption of negative content online, which can heighten arousal before bed and potentially affect both sleep and mental health. As an additional consequence, increased pre-bed phone usage can also disrupt melatonin secretion (i.e., major sleep promoting hormone in humans), resulting in delayed sleep timing or increased difficulties with sleep initiation [58].A novel aspect of this study was the qualitative synthesis of personal experiences of people experiencing good and poor sleep quality. Results revealed that more poor sleepers reported increased alcohol use, greater phone use and more loneliness. This suggests the precarious impact of a pandemic on lifestyles and overall wellbeing. Several themes also emerged from the language content analysis of comments that participants made about their sleep and mental health. For instance, the content from good sleepers had a more positive emotional tone. When describing their personal experiences, good sleepers also made more frequent use of words around social processes (such as “ally” or “friends”), which may indicate the protective role that social interactions can play in improving mental health and therefore sleep. Good sleepers used more family-related words, such as “husband” or “child” as opposed to poor sleepers. Because a full qualitative analysis was unable to be conducted (due to the large sample size), the context in which family-related words were used is not entirely clear. This could be positive (e.g., enjoying time with family) or negative (e.g., worrying about family). However, the current questionnaire findings of fewer good sleepers reporting feelings of isolation suggests that this group likely has a stronger support network, which may be protective against sleep disturbances. Additionally, the results are similar to another COVID-19 study [59], which suggested that increased sleep quality is associated with social capital, a concept that includes social trust, belongingness and participation. Potentially, individuals with more social capital may experience less loneliness and isolation, which could act as a buffer against poor sleep.By contrast, poor sleepers used more negative emotional tone words such as “hate,” “strange” and “isolation” when compared to good sleepers. Interestingly, the LIWC negative emotional tone category has been previously used to index for stress and depression [60]. As a result, the increased use of negative emotional tone may reflect greater levels of stress and depression experienced by poor sleepers. Poor sleepers also made more mentions of finance-related words such as “cash”, “money”, and “owe”. The accompaniment of higher finance related words and negative emotional tone might indicate that poor sleepers are experiencing more financial distress. This was also reflected in the quantitative analysis, wherein a greater proportion of poor sleepers indicated negative financial changes. Regression models revealed that these financial changes, related to loss of job or income accounted for increased risk of reporting depression, anxiety and stress. While we do not know whether any individuals were receiving some form of government financial support, it is possible that any support could have positive implications for their sleep and mental health.Results from the study demonstrate that poor sleepers had greater incidence of reporting anxiety, depression and anxiety. These associations were observed even when major demographic and pandemic-related concerns were accounted for results revealed that poor sleepers had more than three times higher odds of reporting clinical anxiety and moderate depression, and almost three times higher odds of reporting moderate levels of stress. This not only highlights the well-established links between sleep and mental health, but also shows that while changes to lifestyle during the pandemic may be related to mental health, addressing sleep could potentially help mitigate some of the negative effects.Given the cross-sectional nature of this analysis, it is difficult to determine the direction of this association. It is plausible that heightened anxiety and stress experienced as a result of the pandemic preceded disturbances in sleep. Conversely, changes in sleep may have occurred first, exacerbating daytime mood and stress symptoms [61]. Alternatively, both sleep and psychological distress may have emerged concurrently. Ongoing longitudinal assessments in the cohort will help us expand on these associations further, giving us an opportunity to examine causal effects between sleep and mental health.As shown in this study, poor sleepers had more negative personal experiences, greater stress and higher reports of depression and anxiety, independent of personal and pandemic-related factors. Considering that mood and anxiety disorders share bidirectional associations with sleep abnormalities [5], treating sleep can be a cost-effective, efficacious way of improving overall wellbeing. This can reduce the perpetuating effects of poor sleep on mental health. Previous studies have shown that treatments for sleep can also improve anxiety and depression symptoms [11,62,63], including both face-to-face and digital Cognitive Behavior Therapy for Insomnia (CBT-I and dCBT-I, respectively). There is also growing evidence of the efficacy of self-help tools, such as mindfulness apps, to treat subclinical and clinical symptoms of sleep disturbances [64]. While treating sleep may not directly address the negative experiences of a stressful, uncertain pandemic, it can help reduce their propensity and improve individual reaction to the events. For instance, brain regions responsible for rapid eye movement (REM) sleep (such as the amygdala) mediate the stress response [65]. These brain regions also show convergence towards emotional reactivity and consequential action [66], which can prepare individuals for responding to stressful events such as the COVID-19 pandemic. For example, a recent study reported that individuals who received evidence based digital treatment for insomnia had greater resilience and better mental health during the pandemic than those who did not [62]. Given that there is a dearth of consistent sleep training in healthcare professionals such as psychologists, general practitioners and pharmacists [67], upskilling the workforce in sleep treatment delivery, or increasing awareness and education about bidirectional associations between sleep and mental health, may improve overall mental health within communities.Lastly, providing sustained and consistent public health messages on sleep can also be the key to improving mental health outcomes. If research during this pandemic is any indication, people are experiencing a cluster of poor sleep and mental health issues, which can become chronic if left unaddressed. Public health message should be aimed at increasing the uptake of good sleep practices, which include modifiable behaviours. Given that poor sleepers had greater phone and alcohol usage, helping people create boundaries around phone usage at bedtime and alcohol consumption may be helpful. Additionally, the task force of the European CBT-I Academy suggests practices such as keeping a regular bed-time and wake-time (i.e., bringing more structure to sleep routines), having time to de-stress and reflect, and greater exposure to sunlight [26] for sleep retraining. Encouraging individuals to adopt these broader sleep practices can help improve sleep within the community.The study derives its strength from recruiting participants globally, representing different countries, ethnicities, age-groups and communities. It also uses a clear, well-validated measure of sleep quality and provides novel documentation of personal experiences during the pandemic. Furthermore, the study accounts for links between sleep characteristics and mental health after controlling for personal and pandemic-related factors, which provides greater understanding of how the pandemic itself may be related to poor mental health and what role sleep may play when it comes to anxiety, depression and stress.However, results from this study should not be over-stated. While there is significant convergence between sleep disturbances and negative mental health experiences, the data presented here is cross-sectional and cause-effect associations cannot be determined. Individuals with sleep or mental health issues may have been more inclined to respond to the survey, which may have led to bias. However, it must be noted that the study did not specifically set-out to recruit participants with sleep or mental health concerns, instead it broadly framed advertisements as questions about “sleep and mental health” during the pandemic. Further, the study compared participants who slept poorly versus those who slept well, which may help reduce any survey response bias.There were certain demographic groups that were more likely to respond to the question on personal experiences. While education itself was not a factor associated with likelihood of response, cross-cultural differences in the use of language may have an impact on the results. Cross-cultural studies comparing interview transcripts from Czech Republic, Poland, Turkey and Germany (translated to English) in differences of language about traumatic events suggests some differences in the use of LIWC categories such as cognition, but not affective processes that were used in this study [68]. Across cultures, differences in pronoun-use, use of perceptual and social language exist [69]. However, similar to Freitag et al., our study did not find any significant differences in use of affective processes across countries. Together this suggests that affective processes may be more robust to differences in culture and may be more reliable as a cross-cultural linguistic feature compared to other word types. Regardless, differences in language between good sleepers and bad sleepers should be interpreted cautiously.Building on the evidence from this growing body of evidence of the impact of the pandemic on both sleep and mental health is crucial. Since longitudinal analysis have shown that mental health may evolve during this period [70], more multi-wave studies are required to examine how both sleep disturbances and mental health symptoms may change, especially as different countries go in and out of lockdowns. Whether or not these conditions disproportionately affect certain ethnicities, nations, workers or individuals with pre-existing mental health conditions needs to be understood further.Results from the first-wave of our global, longitudinal survey indicate that poor sleep quality is common during the pandemic, and is associated with a 2–3 times increase risk of reporting state anxiety, moderate depression and stress in comparison to good sleepers. Specific factors such as prior diagnosis of mental health condition, financial changes and loneliness, along with sleep characteristics were linked to poor mental health, suggesting that while sleep problems increase the odds of experiencing poor mental health, not all individuals are impacted by the pandemic in the same way. People experiencing distress due to changes to their financial situation and employment, and individuals with pre-existing mental health conditions are some of the vulnerable groups in our community who reported poorer sleep and mental health outcomes, and whom may need additional and ongoing support. Sustained public health messaging on improved sleep practices and increased dissemination and accessibility of self-help tools to aid sleep and mental health are crucial to improve sleep and reduce psychological distress during these uncertain times.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116030/s1, Table S1: Logistic regression of demographics predicting answering of free-text question used in LIWC analysis.Conceptualization, P.V., H.M., M.J. and M.L.J.; Formal analysis, P.V., M.B. and M.L.J.; Investigation, P.V., M.J. and M.L.J.; Methodology, P.V. and M.B.; Resources, M.L.J.; Supervision, M.L.J.; Writing—original draft, P.V. and M.L.J.; Writing—review & editing, P.V., M.B., H.M., M.J. and M.L.J. All authors have read and agreed to the published version of the manuscript.This research received no external funding.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Monash University (protocol number 24018, approved April, 2021.Informed consent was obtained prior to the start of the survey from all subjects involved in the study.Data availability is currently restricted due to ongoing data collection and to maintain participant privacy.The authors would like to express gratitude towards all the participants for their time and commitment towards the study. Authors would also like to thank interns Shelley Webb, Stephen Ghosh and Comfort Kennedy, who supported us with data cleaning and data entry.The authors have no conflict of interest to declare.Participant characteristics (N = 1745).PHQ-9—Patient Health Questionnaire—9, PSQI—Pittsburgh Sleep Quality Index, PSS—Perceived Stress Scale, STAI—State-trait Anxiety Inventory (state subscale). Significance values obtained from independent samples t-test (for age and sleep duration), and chi-square (for all variables except age and sleep duration). a represents range. b total sample size was 1631. c cut-offs for poor sleep, moderate-to-severe levels of stress, clinical state anxiety, moderate-to-severe depression symptoms for PSQI, PSS, STAI and PHQ respectively.Differences between poor sleepers (n = 601) and good sleepers (n = 501) for LIWC categories based on the language used by participants to describe their experiences during the pandemic.1 Poor-sleepers and good sleepers based on cut-offs on Pittsburgh Sleep Quality Index (PSQI), where individuals with sleep quality score of PSQI < 8 were determined as ‘good’ sleepers and individuals with PSQI ≥ 8 were determined as ‘poor’ sleepers. 2 Emotional tone, as identified by language analysis of free-text comments by participants. For instance, positive emotional tone refers to using more upbeat language. Overall emotional tone includes both negative and positive emotions. Social processes include words that indicate social relationships like family, friends etc. 3 Benjamin-Hochberg critical value for false positive rate. * significant at respective critical value after applying Benjamini-Hochberg correction.Binary logistic regression examining the links of personal characteristics, pandemic and sleep-related factors with the severity of anxiety, depression and stress symptoms (based on cut-offs of each scale).1 UCLA-Loneliness Scale-short form (UCLA-LS). 2 Change in financial status due to the pandemic. 3 Not at risk of developing severe symptoms of COVID-19, self-reported by the participants. 4 Poor sleep defined as a score >8 on Pittsburgh Sleep Quality Index (PSQI) 5 Anxiety symptoms refers to state anxiety, measured by State Trait Anxiety Inventory (STAI). Depression symptoms was measured using Patient Health Questionnaire-9 (PHQ-9) and stress symptoms was measured using Perceived Stress Scale (PSS). Each variable was re-coded dichotomously, representing clinical state anxiety (STAI score >40), moderate-to-severe depression (PHQ-9 >10), and moderate-to-severe stress (PSS >14). ὰ Participants with incomplete/missed responses to any of the measures listed above were not included. § represents adjusted Odds Ratio. Model 1 (n = 1438) adjusted for age and sex. Model 2 (n = 1433) adjusted for age, sex, loneliness, prior mental health diagnosis, change in financial status due to COVID-19 pandemic, and at-risk of contracting COVID-19 disease. * p < 0.05, ** p < 0.01, *** p < 0.001.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06031.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Multiple gestations have become an increasing phenomenon that has impacted public health globally, largely due to the application of assisted reproductive technologies. The objective of this work was to find out the discourse that the health professionals involved in its follow-up have in our context. For this, a qualitative methodology was chosen, with semi-structured interviews recorded in audio, prior authorisation, and transcribed verbatim. It was based on a script designed for this purpose, with the following analysis categories: the current trend of multiple gestations, impact, and follow-up. The content analysis was based on the experiences, knowledge, and perceptions of the professionals interviewed. Professionals stated that the current socioeconomic and legal context hinders a single embryo transfer policy that decreases multiple gestation rates. They emphasised the importance of the psychic impact of such gestations on the couple, on the mother in particular, as well as the economic effect on families, health, and society in general. They expressed the need to create specific protocols to assist these gestations. Midwives, in particular, demanded that the health administration recognise and support the differentiated care they perform with this type of gestation. Work on specific models is needed to adequately size the impact of multiple gestations, as well as to generate social health policies that lead to co-responsible reconciliation measures that favour women having one pregnancy at a time.In the last fifty years, the incidence of multiple gestations has increased, acquiring epidemic dimensions, mainly due to delayed procreation and assisted reproductive technologies (ART). They have gone from representing 2% of births to 30–35% after the use of ART [1,2].In this regard, it should be noted that much effort has been made to identify a correct algorithm that considers women’s age and ovarian reserve markers as a tool to optimise the initial dose of recombinant follicle stimulating hormone (rFSH) in intrauterine insemination cycles. However, according to the current available evidence, applying this algorithm with respect to women with polycystic ovary syndrome, especially those with elevated anti-Mullerian hormone, does not seem appropriate [3]. Likewise, the current trend is to perform assisted reproduction treatments associated with ovarian stimulation, but in the case of obese women, these require significantly higher amounts of gonadotropins to achieve the success rates of in vitro fertilisation, similar to those of women with normal weight [4].Multiple gestations deserve special attention for their disproportionate contribution to maternal and perinatal morbidity, which has a special impact on public health [5]. They imply increased risks of maternal adverse outcomes such as hypertensive pregnancy states, gestational diabetes, bleeding, and postpartum depression (40% more likely), among others. They also involve foetal and neonatal risks such as higher rates of prematurity (50% of births occur before 37 weeks) or perinatal mortality and longer-term neurological developmental disorders (the risk of cerebral palsy is four times higher than in a single gestation) [6,7]. In addition, there are complications typical of twin pregnancy, i.e., growth discordance, intra-uterus foetal death of one of the twins, and twin-to-twin transfusion syndrome [8].In gestations based on ART, the risk of adverse maternal and neonatal outcomes is further accentuated not only by the higher incidence of multiple pregnancy, the most common undesirable side effect, but also by the manipulation involved in these processes [9]. This has led to the introduction of elective policies for the transfer of individual embryos and declarations of consensus at the international level [10].The impact of multiple gestations is not only on maternal (physical and emotional) [11] and neonatal health, but also on the health system itself (greater clinical and economic burden), families, and society in general [12]. In Europe, this impact differs among countries, and the number of foetuses in multiple gestation is one of the indicators of the European Perinatal Health Report (Euro-Peristat) programme to measure perinatal health in Europe [2].This study was raised with the aim of understanding the experience and perception that the involved specialists have through their discourse, hoping to contribute to improve the integral attention to multiple gestations.This was a qualitative study within the hermeneutic phenomenological perspective, with the aim of knowing and understanding the discourses of Andalusian health professionals involved in the monitoring of multiple gestations.The sampling was intentional, starting from two key informant professionals who facilitated the selection of other participants, resulting in a snowball sampling technique.In total, eight professionals directly involved in the monitoring of multiple gestations in the city of Seville participated: 4 gynaecologists working in Assisted Reproductive Units (responsible for 30% of multiple gestations), and 4 primary care (PC) midwives, for their linkage to women throughout the pregnancy and postpartum process. The eight planned interviews were carried out until the saturation level was reached.The targeting criteria were sex and professional category, which originated four profiles, resulting in two interviews per profile. In addition, service time, age, and scope of work were included as attributes that would make it easier to obtain greater discursive power (Table 1).The fieldwork was carried out in the capital of the province of Seville, between January and February 2018. Semi-structured interviews were conducted.An ad hoc script (Supplementary Materials S1 and S2) was used from the main study dimensions:Current trend of multiple gestations (embryo transfer policies).Impact and consequences on the family, society, and the health system.Monitoring of multiple gestations.Demands on health administration (emerging category).Current trend of multiple gestations (embryo transfer policies).Impact and consequences on the family, society, and the health system.Monitoring of multiple gestations.Demands on health administration (emerging category).It was the same script in the two groups, but some different questions were introduced to suit the idiosyncrasy of the assistance given by each group. The affected dimensions were Current Trend of Multiple Gestations with the question: “How do you address the issue of selective reduction?” in the script of gynaecologists and Follow-up with the question: “What strengths and weaknesses do you think the follow-up that is performed for pregnant women from primary care presents?” in that of midwives.Two scripts designed for the purpose (see Supplementary Materials S1 and S2) were used, adapted to each professional category and based on the main dimensions of the study: current trend of multiple gestations (its relationship with ART and embryo transfer policies), monitoring of multiple gestations, and their impact, while also incorporating another emerging category of relevance and interest (demands to the health administration).The interviews were recorded, lasting between 50 and 70 min. The exact date and place were set by the participants, with special consideration of the place, being free of interruptions, ensuring privacy and allowing them to be relaxed. Before starting each interview, the interviewer described the purpose of the study, explained the basic rules, and offered confidentiality guarantees.To minimise the variability of the interview process, all of them were performed by the same researcher who took field notes from each of them with the aim of detecting and recording nonverbal language. The researcher applied summary and paraphrase techniques to cross-validate the collected information from the interviewees and minimise the possibility of data distortion. The transcripts were carried out by an expert in the field from outside the research team.Content analysis was used as an analysis technique to look closer at the experiences and knowledge of professionals, and to try to grasp the subjective meaning of their discourse regarding the categories of analysis raised [13]. This was carried out in terms of the main categories raised (current trend of multiple gestations, monitoring, and their impact) and incorporating another emerging category of relevance and interest (demands to the health administration). As an indicator of the validity of the results, a triple triangulation was performed that involved performing data triangulation, methodological (intra-method) triangulation, and triangulation of the research team.Atlas.ti textual analysis software (Windows v8.0, Microsoft, Redmond, WA, USA) was used as an analysis tool. All transcripts of interviews were encoded, grouped by nodes or topics (categories), and text fragments were assigned to previously established categories.Analysis of the interviews showed a series of arguments linked to the different categories of analyses created:Opinion on multiple gestations.Attribution of multiple gestations.Approach to embryo transfer policies.Circumstances, motivations, or needs to undergo an ART. Impact of ART on the couple, at the work and social level.Motivations for the delay of maternityImpact of multiple gestation.Follow-up of pregnant women from primary care.Emotional monitoring for women and couples undergoing ART.Negative situation in emotional follow-up.Follow-up in the postpartum.Maternal mental health during follow-up.Treatment of health administration of multiple gestations.Role of the professionals of the administration in the treatment.Suggestions.Opinion on multiple gestations.Attribution of multiple gestations.Approach to embryo transfer policies.Circumstances, motivations, or needs to undergo an ART. Impact of ART on the couple, at the work and social level.Motivations for the delay of maternityImpact of multiple gestation.Follow-up of pregnant women from primary care.Emotional monitoring for women and couples undergoing ART.Negative situation in emotional follow-up.Follow-up in the postpartum.Maternal mental health during follow-up.Treatment of health administration of multiple gestations.Role of the professionals of the administration in the treatment.Suggestions.This qualitative analysis has been carried out through a rooting criterion, hierarchically categorising the generated codes. In this way, the codes with the highest number of appointments have been prioritised. For its part, the marked cut-off point is located at the lower limit of the second quartile of the distribution, using the codes found in the first two quartiles of the appointment distribution for the analysis.This study adhered to the principles articulated in the Declaration of Helsinki, updated in 2013 in Brazil. To ensure anonymity, personal identifiable data were replaced by numbers. All the participants signed an informed consent. Authorisation was obtained from the Research Ethics Committee of the University Hospital Virgen Macarena in Seville, Spain, with code #09876, on 18 July 2014.The data collected in the study will be treated with absolute confidentiality in accordance with the provisions of Spanish laws, specifically Organic Law 3/2018, of December 5th, on the Protection of Personal Data and Guarantee of Digital Rights, and Law 41/2002, of November 14th, basic regulation of the autonomy of the patient and rights and obligations in matters of information and clinical documentation.Data organisation made according to the categories of analysis facilitated the exposure of the results presented through the most relevant verbatims. Similarly, each category of analysis shows, through figures, the resulting cognitive maps from having raised issues to the participants to analyse these categories, complementing the information provided in the results.There is a coincidence in the increase in multiple gestations, mostly due to the use of ART. Although the current trend is to try and decrease them, it does not seem so easily applicable due to hindering social and legal connotations.It is considered that this may be due to variables such as the socioeconomic context of couples where multiple gestation is an advantage over only one (infertility treatments are very expensive and allowing more than one is highly unlikely), the employment situation of women (there is no real reconciliation that will protect them and shield their reproductive desires), and due to legal protection so as not to implement single transfer policies (Table 2).Figure 1 and Figure 2 show the different codes that make up the opinion section on the number of multiple gestations in recent years and the relationships between them.In Figure 1, in the face of the possible situation of cycle reduction or multiple gestations, the importance of information meetings on reduction risks and multiple gestations is evident. Despite being much-desired pregnancies, both the mother and babies are exposed to risks that future parents should be aware of.As for ART, reference is made to the need for a specific unit of psychology, as traumatic situations occur during the procedure that require psychological support. The main views of participants regarding the number of multiple gestations in recent years are illustrated in a cognitive map below.With regard to Figure 2, it must be highlighted that this includes a possible disadvantage, since, for assisted reproductive treatments, there is an important waiting list due to its high demand, thus slowing down the gestation process. With regard to the attribution of multiple gestations, the discourses of the participants can be summarised as follows:Figure 3 shows a “cognitive map” based on the circumstances, motivations, and needs for assisted reproductive treatment. For this issue, the interviewees allude to a series of clear ideas to be discussed below:The fundamental reason for couples to decide to undergo ART is because they have been trying to get pregnant for over a year with no success. However, there is also a circumstance in which there is no real infertility and the recommended waiting times have not been respected. The people interviewed allude to several reasons in this regard: assisted reproductive treatments have been standardised, they can be relatively easily financed, and above all, the constant programming of peoples’ lives, where motherhood is not left to chance and is highly planned according to women’s work wishes, in particular. This leads them to unfailingly delay pregnancy because there is no family reconciliation that actually safeguards their reproductive and career desires together.The collective of gynaecologists agrees on the serious emotional wear and tear of couples facing multiple gestations. Everyone, in the context of parenting, refers to the important psychic impact on mothers, because of the lack of time to internalise these emotional changes since all their energy and time are focused on this double upbringing. Stress is greatest in cases of prematurity, especially when there are differentiated needs between the two twins, and they feel they are providing unequal care.There is a sense of critical emptiness and frustration in mothers once husbands regain their former life, while they remain totally immersed in parenting.Fathers are often very involved in parenting and household chores, either because they have no choice when the mother is fully immersed in parenting or because they have had a hard time achieving that pregnancy and now they are enjoying a one-month parental leave that puts them at the forefront of the reality of parenting (although it is said that some fathers use these leaves for their personal time, often older men who are not involved).There is agreement regarding the family support network being critical for coping positively.The economic impact for couples is great, not only due to the investment made in reproductive treatments through private health, but also for what double upbringing implies. For the health system, there are economic demands as well, with the increasing maternal and perinatal morbidity associated with these gestations (Table 3).Through Figure 4, an analysis is made on the impact of multiple gestations at the various levels contemplated in the research according to the testimonies collected through these interviews.This figure corroborates the above results, highlighting the relevance, as the study states, of the need for couples to have co-responsibility.It also clearly shows the idea that older males are less involved than younger males, and eventually mentions the important role of family in supporting multiple gestation.All interviewees express the need for a specifically emotional, multidisciplinary and protocolised follow-up.Although there is no differentiated protocol, midwives agree that their differentiated follow-up is much greater and that the primary-specialised inter-level relationship for proper gestation monitoring is satisfactory.Infertility specialists refer to the fact that infertility is poorly sized, starting in primary care by family physicians, which delays and negatively impacts the follow-up of these couples, and emphasises the role of primary care midwives to be relevant and necessary for the successful monitoring of these gestations (Table 4).Midwives demand recognition of the differentiated care that they provide. Many post-night visits are not counted, so the problem of multiple gestations cannot really be sized and their impact on the system remains unnoticed. Public health gynaecologists understand that reproductive medicine is a second-rate specialty and that no investment is made in the resources needed for its proper development, which translates, for example, into unmanageable waiting lists for many women who require assisted reproductive treatment (Table 5).Figure 5 shows the perception that women with multiple gestations have of the follow-up made from primary care. It also reflects the differentiated care that midwives refer to in their follow-up, demanding to be recognised in order to make this work visible.Among the suggestions provided by professionals to address the problem of multiple gestations are: raising awareness among women and their partners about the risks of multiple gestations for both them and their offspring; educating about the importance of the age of the first pregnancy and how that age will have a definite impact on women’s fertility; and generating multidisciplinary protocols that specifically deal with the emotional sphere and strengthen the economic support for these families (Table 6).The tag cloud (Figure 6) also shows critical points and improvement proposals after category analysis: health management demands and suggestions.Following the analysis of the discourses and as a summary of the most important results, we emphasise that professionals stated that the current socioeconomic and legal context hinders a single embryonic transfer policy that decreases multiple gestation rates. They emphasised the importance of the psychic impact of such gestations on the couple, on the mother in particular, as well as the economic effect on families, health, and society in general. They expressed the need to create specific protocols to assist these gestations. The midwives, in particular, demanded for the health administration to recognise and support the differentiated care they perform with this type of gestation.Significant similarities have been found in the expressed opinions and professional demands, although they are a barely explored subject, since most offer a clinical vision of multiple gestation and, in particular, its relationship with ART [14].Spain tops the ranking of European countries where the most cycles of assisted reproduction take place, and also ranks third globally below the United States and Japan [15].The recommended multiple pregnancy rates after ART (in vitro fertilization) are below 10% [16]; therefore, over the past decade, the international scientific community has implemented transfer policies aimed at reducing multiple gestation rates by reducing the number of transferred embryos [17]. Currently, the recommendation is to carry out a single transfer in young patients of good prognosis as their pregnancy rate is not significantly affected, being, therefore, comparable with spontaneous pregnancies [18]. However, fertility societies currently raise the ethical dilemma between the right of patient autonomy and the recommendations of performing single transfer to avoid the risks associated with multiple gestation, which is a strong moral, socioeconomic, and scientific debate still unresolved. These data coincide with our research, where the opinion of professionals in this regard was: “Legal protection along with the socio-economic circumstances of couples make elective transfer difficult even if recommended”.In Spain, since 2006 and under Law 14/2006 [19], the transfer/cycles of embryos were limited to a maximum of three. This law remains in force today and, under its protection, the percentage of double transfers continues to account for more than half of the total transfers made in the country. This is reflected in the statistical report of 2017 of the Spanish Society of Fertility on assisted reproductive technologies [20]; 60.3% of the transfers were of two embryos that, in 23.5%, resulted in multiple gestation, although since 2015, the embryo transfer policy is to increase the number of unique transfers. This is largely explained by the pressure that health professionals receive on behalf of patients requesting double transfers as they are considered more profitable in terms of their socioeconomic and work context [21,22].The Guide to Assisted Human Reproduction in the Public Health System of Andalusia [23] contains a specific section on the human and material resources necessary for the operation of these centres. However, this section does not cite the psychological profile as part of the equipment necessary to serve the applicants for these services. However, on the other hand, the testimonies contained in our results show the importance of this group for the correct care of users of ART, demanding its presence in public health as it is in private assistance.Treatments of ART in public health are free of charge, resulting in such high demand that waiting lists are sometimes impossible to cope with by the applicants. This causes a leak to private assistance where such waiting lists do not exist, even if treatments come at a very high cost. This is a common complaint that is reflected in the testimonies of participating gynaecology professionals who demand more economic investment for public centres.Currently, in Spain, the mean age for motherhood is 32.25 years and the mean number of children per woman is 1.31 [24]. In the lives of women, motherhood becomes a complement to their profession, which is now a life project, causing changes in motherhood patterns with a delayed and reduced number of sons or daughters [25]. Late motherhood beyond the age of 35 is the most decisive factor in infertility, where multiple gestation after ART is more than probable. More than 50% of patients come from fertility treatments and older maternal age is more common [26]. In this same line, the literature highlights the important implications of the age of 35 for fertility and for the added risks of multiple gestation [27]. This idea is reflected in our results as the people interviewed express the delay of motherhood as one of the main causes for the use of ART, a delay related, among other causes, to the professional aspirations of women.Multiple gestations account for 10% of overall perinatal mortality and overall twin mortality is 5–10 times higher than in simple gestations, with higher prematurity rates, resulting in a large impact on the morbidity of descendants. Prematurity is reflected in our research as a source of added stress in the context of twin pregnancy. In addition, the overall maternal mortality associated with multiple births is 2.5 times higher than that of single births [28]. In parallel with the drastic increase in multiple gestations in recent decades, there has also been an increase in caesarean rates, largely due to the belief that these improve perinatal results. In fact, elective caesarean is the most common form of birth in twin gestations [29,30].The importance of support networks and the psychic impact of multiple gestations is highlighted in the testimonies of the interviewees and is, thus, reflected in the literature, where we find that, regardless of the family situation or structure, support networks are fundamental to the maternity process and other daily tasks [31]. However, in the case of multiple gestations, there is almost twice the risk of postpartum depression and depressive symptoms, compared to single gestations [32]. Likewise, the psychological impact on the relationship is significant and the mental health of these parents is worse [33].This impact on the emotional sphere is exacerbated by the economic difficulties that already arise at the start of such gestations in order to resort to ART, and then, upbringing. For families, socioeconomic costs are approximately 4 to 11 times higher in twin gestations than in simple gestations [34]. Therefore, economic burdens, as well as the potential to reduce the quality of life of twins, need careful evaluation.In addition, due to the increased risk of complications, women with multiple pregnancies need more health monitoring and greater contact with health care and their professionals, which will have an impact on health system resources [35]. On this line, the people interviewed allude to the economic impact of multiple gestations on the family, but also on the health system and society in general.The Anglo-Saxon context reflects the need for differentiated follow-up for multiple gestations [36], which is an opportunity to ensure that guidelines reflect best practices and women receive the best possible care. In Andalusia, the follow-up of all gestations in the Autonomous Community is carried out through a care programme which is defined and delimited by the Integrated Care Process (PAI, for its acronym in Spanish) of Pregnancy, Childbirth, and Postpartum (EPP, for its acronym in Spanish). However, this process does not consider the special needs for emotional support that arise in cases of multiple gestations.Additionally, although there are concrete guidelines for the management of such gestations by some scientific societies such as the Spanish Society of Obstetrics and Gynaecology (SEGO) [37], these do not contain general recommendations adapted to this type of gestation or the need for greater emotional support. In contrast, our results do illustrate the need for such circumstances when the great psychic impact of twin gestations manifests in most testimonies. However, these particular needs, in the case of midwives, who are regarded as developing a key role in the preparation of mothers and fathers of twins [38], are being met even though they do not have any official protocol in this regard. Such would be the case of advancing vaginal BGS screening. This is reflected in the “Protocol of Assistance to Pregnancy and Childbirth of Multiple Gestations” of the Clinic Hospital in Barcelona, though this is not a scientific society, but a specific hospital [39].In the region of Andalusia (Spain), the electronic medical history support system (Diraya) integrates all the health information of the people assisted in the health centres, and also serves as the management of the health system. In the case of the postpartum visits, only those made the first ten days after delivery are recorded, even though in the case of multiple gestations, midwives continue to assist and offer care beyond this date. In contrast, the current literature reflects the enormous importance that registered midwives can have on maternal satisfaction, as well as the need to support their professional role and positively impact teamwork, organisational processes, and research [40,41]. The testimonies of the interviewed midwives point out the current configuration of this programme as a great disadvantage, because it does not allow the differentiated attention they are giving to these pregnant women to be faithfully viewed.Although Organic Law 3/2007 [42] for the effective equality of men and women, in article 44, determines the rights to reconcile family and work life, even if they have some legal protection, many women encounter serious problems related to work leaves during fertility treatments or in the case of pregnancy. In short, as Gracia-Maroto et al. already stated, the increase in fertility rates implies the transformation of different social aspects related to restructuring productive and reproductive work, the development of social services, enhancement of institutional aid, and the revision and improvement of health care policies [43].There are a number of limitations in the present study. First, it is worth considering the exploratory nature of this research and the participation of only eight subjects. This is why our results have limited generalisation. For future research, collecting other significant opinions, such as those of couples undergoing ART and mothers with multiple gestations, is being considered.On the other hand, in the current literature, no article has been found that has studied the testimony of the participants selected in the research at hand on multiple gestations and their relationship with ART. This limitation is understood as an opportunity to identify new gaps in the literature and, consequently, new research that allows a greater body of knowledge on the subject of research and improved practice.Another limitation is the time available to investigate the issue; since the interviews were conducted in the workplaces of the participants, we adjusted to agenda gaps and moments of rest which, in some cases, did not allow the interviews to expand as much as desired.The study only reflects the discourse of the collectives of gynaecology and midwifery, while this impact could have included other health actors such as primary care physicians, who are directly referred to by gynaecologists, thus providing complementary results.For the collective of gynaecology, it is important to apply single embryo elective transfer policies in cases of ART. This would allow the rates of multiple gestations after the use of these techniques to be maintained at the recommended percentages and control the clinical and economic impact of multiple gestations.Midwives stress the need to improve clinical documentation registration systems to be a true reflection of the follow-up needs posed by multiple gestations.For both professional groups, the socioeconomic implications and impact on maternal–foetal and neonatal health must be considered by political/health authorities and, also in this regard, educate future parents.As for the practical implications of this study, progress should be made in protocols that contribute to better care and monitoring of women’s emotional health in particular, and couples in general, optimising clinical decisions and promoting safe, quality, and lower cost care. In addition, developing sociohealth policies that lead to co-responsible reconciliation measures that favour women having one pregnancy at a time is necessary (although, moral, socioeconomic, and scientific debates are still unresolved).More research is needed to investigate the opinions and experiences of other health professionals and couples to delve into the factors that prevent the complex problem of multiple pregnancy from actually being determined today.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116031/s1. Supplementary Material S1, Planned Script for the Interview: Midwives; Supplementary Material S2, Planned Script for the Interview: Assisted Reproduction Unit Professionals.Conceptualization, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M.; Data curation, E.J.-G., A.B.-H., N.N.-R., and R.C.-M.; Formal analysis, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M.; Investigation, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M.; Methodology, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M.; Project administration, E.J.-G., A.B.-H. and R.C.-M.; Resources, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M.; Software, E.J.-G., A.B.-H., J.G.-S., N.N.-R., and R.C.-M.; Supervision, E.J.-G., A.B.-H., J.G.-S., N.N.-R., and R.C.-M.; Validation, E.J.-G., A.B.-H., J.G.-S., N.N.-R., and R.C.-M.; Visualization, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M.; Writing original draft, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M.; Writing—review and editing, E.J.-G., A.B.-H., F.J.F.-C., J.G.-S., N.N.-R., and R.C.-M. All authors have read and agreed to the published version of the manuscript.This research received no external funding.This study adhered to the principles articulated in the Declaration of Helsinki, updated in 2013 in Brazil. To ensure anonymity, personal identifiable data were replaced with numbers. All the participants signed an informed consent. Authorisation was obtained from the Research Ethics Committee of the University Hospital Virgen Macarena in Seville, Spain, with code #09876, on 18 July 2014. The data collected in the study will be treated with absolute confidentiality in accordance with the provisions of Spanish laws, specifically Organic Law 3/2018, of December 5, on the Protection of Personal Data and Guarantee of Digital Rights, and Law 41/2002, of November 14, basic regulation of the autonomy of the patient and rights and obligations in matters of information and clinical documentation.Written informed consent was obtained from all subjects involved in the study.All data are available within this article and its supplementary file.The authors declare no conflict of interest.Opinion on the number of multiple gestations in recent years. Source: Own elaboration.Multiple Gestation Attribution (ART). Source: Own elaboration; ART: Assisted Reproductive Technologies.Circumstances, motivations or needs to undergo an ART. Source: Own elaboration; ART: Assisted Reproductive Technologies.Impact of multiple gestation. Source: Own elaboration.Follow-up to pregnant women from primary care. Source: Own elaboration; ART: Assisted Reproductive Technologies.Tag cloud with critical points and improvement proposals. (a) Critical points. (b) Proposals for improvement.Profiles of interviewees.ART, assisted reproductive technologies; PC, primary care.Current trend of multiple gestations (embryo transfer policies).Impact of multiple gestations on women, couples, society, and the health system.Follow-up of multiple gestations.Demands to the health administration.Suggestions.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06032.txt ADDED Viewed

	@@ -0,0 +1 @@

+ These authors contributed equally to this work.Soybean koji refers to steamed soybeans inoculated with microbial species. Soybean fermentation improves the health benefits of soybeans. Obesity is a serious health concern owing to its increasing incidence rate and high association with other metabolic diseases. Therefore, we investigated the effects of soybean and soybean koji on high-fat diet-induced obesity in rats. Five-week-old male Sprague-Dawley rats were randomly divided into four groups (n = 8/group) as follows: (1) regular diet (RD), (2) high-fat diet (HFD), (3) HFD + steamed soybean (HFD+SS), and (4) HFD + soybean koji (HFD+SK). SK contained more free amino acids and unsaturated fatty acids than SS. In a rat model of obesity, SK consumption significantly alleviated the increase in weight of white adipose tissue and mRNA expression of lipogenic genes, whereas SS consumption did not. Both SS and SK reduced serum triglyceride, total cholesterol, and low-density lipoprotein cholesterol levels, and increased high-density lipoprotein cholesterol levels. SS and SK also inhibited lipid accumulation in the liver and white adipose tissue and reduced adipocyte size. Although both SS and SK could alleviate HFD-induced dyslipidemia, SK has better anti-obesity effects than SS by regulating lipogenesis. Overall, SK is an excellent functional food that may prevent obesity.Soybeans are an important dietary source because they contain high-quality proteins and are rich in bioactive compounds such as isoflavones and saponins [1]. Nowadays, diet supplementation has been widely accepted as a useful strategy to modulate biochemical and molecular pathways to improve physiological and pathological conditions [2,3]. It has been reported that soybeans have many therapeutic properties, including hypocholesterolemic [4], anti-hypertensive [5], and anti-cancer effects [6]. In Asia, soybeans are primarily fermented dietary sources to give them a unique flavor and improve their shelf life. Fermented soybeans are used to produce soybean sauces, such as doenjang, gochujang, mirin, sake, and miso. In recent years, fermented foods have gained attention because fermentation can boost the bioactive components responsible for health benefits [7]. Commercially available fermented soybean products use soybean koji as a starter in soybean fermentation to speed up the process and ensure food safety [8]. Soybean koji is made by steaming soybeans and then inoculating molds or bacteria that are recognized as safe for consumption. Soybean koji has an abundance of hydrolytic enzymes that break down soybeans and synthesize new components during fermentation [9]. Previous studies have reported that soybean koji is a potent bioactive food material, as it exerts better antioxidant activity than soybeans [6].Obesity refers to the accumulation of excess body fat due to an energy imbalance. When energy intake exceeds energy expenditure, the surplus energy is stored in the body, leading to adipose tissue enlargement, dyslipidemia, and fatty liver, and is linked to other metabolic diseases such as diabetes mellitus, cardiovascular disease (CVD), and certain types of cancer [10,11]. In modern society, the rate of obesity is increasing due to energy-dense foods and a sedentary lifestyle [12]. According to the National Health and Nutrition Examination Survey (NHANES), the rate of severe obesity in the US increased from 4.7% to 9.2% during 1999–2000 and 2015–2016, respectively [13]. More recently, it has been reported that people are more likely to become obese due to restrictions on physical activity caused by the COVID-19 pandemic [14]. Zhu et al. reported that during the COVID-19 outbreak people spent most of their time at home, resulting in decreased physical activity and increased food intake, which was closely related to weight gain [15]. Obesity is a modifiable factor by regulating food consumption as well as physical activity; therefore, dietary intervention holds more weight for preventing/overcoming obesity when people reduce their physical activity, such as during the COVID-19 lockdown. We may also consider therapeutic regulations in obesity. To date, five drugs have been approved by the US Food and Drug Administration (FDA) for the treatment of obesity, which act as appetite suppressants, metabolic stimulants, or nutrient absorption inhibitors primarily through hormonal action [16]. The aforementioned therapies have significant limitations owing to side effects and weight-loss efficacy. Therefore, dietary and lifestyle interventions are fundamental to countering obesity. As a result, natural materials, including red pepper [17], ginger [18], and green tea [19], which have been reported to possess anti-obesity effects, are preferred for preventing obesity [20].In previous studies, soy protein has been reported to affect obesity by lowering body weight, fasting glucose levels, and hepatic fat accumulation in animal models [21]. Soybean polysaccharides and genistein prevent high fat-induced body weight gain, dyslipidemia, oxidative stress, and inflammation in mice [22]. However, the effects of fermented soybeans on obesity are not fully understood yet. Recently, Kim et al. developed soybean koji inoculated with Bacillus amyloliquefaciens CJ 14-6, which has excellent protease and amylase activities [23]. Gochujang, a typical traditional Korean sauce produced using the soybean koji, showed an anti-obesity effect [24]; however, the effect of soybean koji on metabolic alterations induced by a high-fat diet has not yet been elucidated. Therefore, this study aimed to investigate and compare the effects of soybean and soybean koji inoculated with Bacillus amyloliquefaciens CJ 14-6 on high-fat diet-induced obesity and obesity-induced metabolic changes in rats, and to evaluate whether soybean koji could be used as a better alternative to prevent obesity.The steamed soybean (SS) and soybean koji (SK) used in this study were supplied by the CJ CheilJedang Corporation (Suwon, South Korea). For the preparation of SS, soybeans were immersed in water at 15 °C for 15 h and then steamed at 40 °C for 30 min. After cooling and drying at room temperature (RT; 20–25 °C) for 24 h, the soybeans were ground into powder. Bacillus amyloliquefaciens CJ 14-6 is a patented strain of the CJ CheilJedang Corporation (KCCM 11718P). To obtain SK, Bacillus amyloliquefaciens CJ 14-6 was cultured at 37 °C for 24 h on a rotary shaker at 200 rpm. Then, 2% (v/w) of the culture solution, based on the weight of the soybeans, was inoculated into the steamed soybeans and incubated at 37 °C for 36 h. The fermented soybeans were then dried at 40 °C for 24 h and pulverized into a powder.The general composition analysis of SS and SK was carried out in accordance with the Association of Official Analytical Chemists method [18]. Moisture content was determined using the oven-drying method at 105 °C. Crude protein content was determined using the micro-Kjeldahl method. The crude fat content was analyzed using a Soxhlet apparatus. The ash content was analyzed by ignition in an electric furnace. Carbohydrates were calculated by subtracting the moisture, crude protein, crude fat, and ash contents from 100. The fiber content was determined using an enzymatic-gravimetric method. All analyses were performed in triplicate.Free amino acid analysis was conducted by hydrolyzing samples (0.5 g) with 3 mL of 6 N HCl at 121 °C for 24 h. Excess acid was removed using a rotary vacuum evaporator, and the sample was then dissolved in 10 mL of sodium phosphate buffer (pH 7.0) for further analysis [25]. The solution (1 mL) was filtered through a membrane filter (0.2 μm), and amino acid analysis was carried out in a Biochrome 20 Amino Acid Analyzer (Pharmacia Biotech, Cambridge, UK). All analyses were performed in triplicate.Fatty acid analysis was conducted following the Wijngaarden method [26]. The lipids in the samples (2 g) were extracted with ether, filtered, and concentrated under reduced pressure. Lipid samples (100 mg) were transferred to Erlenmeyer flasks and stirred with 4 mL of 1 N KOH ethanol until the lipid droplets disappeared. Following the addition of 5 mL of 14% BF3-methanol, the samples were heated at 80 °C for 5 min to synthesize methyl ester. The samples were cooled again; 3 mL of saturated NaCl solution and 1 mL of hexane were added and allowed to stand to separate the solution. The supernatant was transferred to a new tube and mixed with anhydrous Na2SO to remove moisture. The fatty acids in the solution were analyzed using gas chromatography (GC-10A, Shimadzu, Kyoto, Japan). All analyses were performed in triplicate.All animal studies were approved by the Chosun University Institutional Animal Care and Use Committee (C IACUC No. 2015-A0028). After 1 week of acclimatization, 5-week-old male Sprague-Dawley rats (Orient Bio, Inc., Seongnam, Korea) were housed in cages for 8 weeks until sacrifice. The animals had access to food and water ad libitum throughout the experimental period. The experimental rats in each group (n = 8) were fed one of the following dietary compositions: (1) a regular diet (RD) based on the AIN-93G formulation (15.8% of energy from dietary fat), (2) a high-fat diet (HFD, 39.5% of energy from dietary fat), (3) an HFD diet mixed with 3% SS (HFD+SS, 39.2% of energy from dietary fat), or (4) an HFD diet mixed with 3% SK (HFD+SK, 39.2% of energy from dietary fat). Individual dietary compositions are listed in Table 1. Body weight (BW) and food intake were measured weekly. The food efficiency ratio (FER) was calculated by dividing the total BW gain by total food intake. At the end of the study, final body weights were measured after overnight fasting, and whole blood samples were collected from the heart and the serum was isolated after clotting procedures. The organs were harvested, weighed, and snap-frozen in liquid nitrogen after being sacrificed by thoracotomy after CO2 narcosis. Isolated serum and organ samples were stored at −80 °C until further analysis.The enzymatic activities of alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), lactate dehydrogenase (LDH), triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), and fasting glucose (GLU) levels were measured using a chemistry analyzer (Fujifilm Dri-C hem 3500i, Fujifilm, Tokyo, Japan) as previously described [27]. The values of low-density lipoprotein cholesterol (LDL-C), atherogenic index (AI), and cardiac risk factor (CRF) were also calculated as previously described [28,29].Lipids were extracted from ~0.1 g of liver and adipose tissues as previously described [30]. The TG and TC levels were measured from the lower (lipid-abundant) layer following previously described methods [31,32].Total RNA was extracted from the samples using NucleoSpin RNA Plus (Macherey-Nagel GmbH & Co., Düren, Germany) according to the manufacturer’s protocol. RT-PCR was performed as described previously [33]. The expression of each experimental gene was normalized to the expression of β-actin, which did not significantly vary between the different dietary settings. The gene-specific oligonucleotide primers used in this study are listed in Table 2.Liver and epididymal adipose tissues were fixed, sectioned, and stained as previously described [27].The experimental data were analyzed using one-way analysis of variance (ANOVA; GraphPad PRISM 8, San Diego, CA, USA). Subsequently, Tukey’s post-hoc test was applied to distinguish groups that varied significantly (p < 0.05).To investigate the changes in the composition of soybeans following fermentation, the proximate composition and free amino acid and fatty acid contents in SS and SK were analyzed. As shown in Table 3, SK contained more moisture and crude fat than SS; however, the carbohydrate content in SS was higher than that in SK. On the other hand, the crude protein, ash, and dietary fiber contents did not significantly differ between SS and SK.Free amino acid analysis of SS and SK revealed 23 amino acids (Table 4). The total free amino acid content in SK was 4.3-fold higher than that in SS (p < 0.05). Specifically, the contents of urea, threonine, glutamic acid, glycine, valine, isoleucine, leucine, tyrosine, phenylalanine, histidine, and tryptophan were higher in SK than in SS (p < 0.05), and the difference in the contents of leucine and phenylalanine was the largest.The fatty acid compositions of SS and SK are listed in Table 5. SS contained more saturated fatty acids than SK, whereas SK had more monounsaturated fatty acids (MUFA) and polyunsaturated fatty acids (PUFA) (p < 0.05). The predominant components of SS were linoleic acid (C18:2n6c), heneicosanoic acid (C21:0), oleic acid (C18:1n9c), and palmitic acid (C16:0). Similarly, linoleic acid (C18:2n6c), oleic acid (C18:1n9c), palmitic acid (C16:0), and heneicosanoic acid (C21:0) were the most abundant fatty acids in SK. Compared to those in SS, the amounts of linolenic acid (C18:3n3), stearic acid (C18:0), linoleic acid (C18:2n6c), and oleic acid (C18:1n9c) were significantly higher in SK (p < 0.05).At the end of the experiment, the body weight and daily weight gain of the HFD group were significantly higher than those of the RD group (p < 0.05) (Figure 1A,B). The body weight and daily weight gain of rats supplemented with SS or SK tended to be lower than those of the HFD group; however, the differences were not statistically significant (Figure 1A,B).To determine whether the amount of food intake was associated with changes in the body weight, daily food consumption and FER were measured. Daily food consumption of the HFD group was slightly lower than that of the RD group, but there was no statistical difference; nevertheless, rats fed HFD with SK showed significantly lower food consumption compared to the RD group (p < 0.05) (Figure 1C). The FER, determined by dividing the body weight gain by the amount of food consumed, was significantly higher in the HFD group than in the RD group (p < 0.05). SS and SK did not affect the FER increased by HFD, indicating that weight changes between HFD and SS- and SK-supplemented groups were related to the alterations in food intake (Figure 1D).As the liver and adipose tissue have central roles in regulating whole-body metabolism, we measured the weights of the liver and white adipose tissue fat pads. The liver weights of the rats were not statistically different among the four groups (Figure 2A). White adipose tissue (WAT) weight was significantly higher in the HFD group than in the RD group. In contrast, SK significantly inhibited HFD-induced WAT accumulation (p < 0.05) (Figure 2B). The weight of epididymal adipose tissue (EAT) did not significantly differ among the groups; however, mesenteric adipose tissue (MAT), retroperitoneal adipose tissue (RAT), and perirenal adipose tissue (PAT) weights were significantly higher in the HFD group than in the RD group (p < 0.05) (Figure 2C–F). Specifically, the weight of MAT was lower in HFD+SK group than in the HFD group (p < 0.05) (Figure 2D).Serum ALT, AST, ALP, and LDH activities are typical indicators of liver function [34]. Leptin, an adipokine that exerts endocrine function, regulates metabolic homeostasis by regulating appetite, energy expenditure, and glucose utilization [35]. In this study, HFD increased the serum ALT, AST, ALP, and LDH activities compared to the RD (p < 0.05) (Figure 3A–D). However, rats that consumed the HFD+SS and HFD+SK diets had lower serum ALT activities than those in the HFD group (p < 0.05) (Figure 3A). Additionally, ALP activity was significantly lower in the HFD+SK group than in the HFD group (p < 0.05) (Figure 3C).Serum leptin levels in the HFD group were significantly higher than those in the RD group; however, SS and SK inhibited HFD-induced increase in serum leptin levels (p < 0.05) (Figure 3E). The HFD group had a higher fasting serum glucose level than the RD group (p < 0.05), and the levels in the HFD+SS and HFD+SK groups did not significantly differ from those in the HFD group (Figure 3F).Serum lipid levels are one of the most important predictors of metabolic disease. High LDL-C and low HDL-C levels are highly correlated with the incidence of CVD. In this study, the HFD group had higher serum TG, TC, and LDL-C and lower HDL-C levels than the RD group (p < 0.05) (Figure 4A–D). On the other hand, SS and SK significantly inhibited the HFD-induced increase in serum TG, TC, and LDL-C and the reduction in HDL-C (p < 0.05) (Figure 4A–D). Consequently, SS and SK significantly inhibited the increase in the HFD-induced atherogenic index and cardiac risk factor (p < 0.05) (Figure 4E,F).The liver metabolizes most nutrients, and chronic HFD consumption can lead to fatty liver due to lipid overload. Therefore, we measured the levels of total lipid, TG, and TC in the liver to investigate whether SS and SK had a beneficial effect on hepatic lipid accumulation induced by HFD. As shown in Figure 5A, the HFD group had a higher hepatic lipid level than the RD group (p < 0.05). The hepatic levels of TG and TC were also significantly higher in the HFD group than in the RD group (p < 0.05) (Figure 5B,C). Rats that consumed SS and SK had lower total lipid, TG, and TC levels in the liver compared to those in the HFD group (p < 0.05) (Figure 5A–C). Moreover, hepatic TC levels were significantly lower in rats fed with SK than in rats fed with SS (p < 0.05) (Figure 5C).Figure 5D shows the representative liver images of each group. The liver of the HFD group appeared brighter than that of the RD group; however, SS and SK ameliorated the morphological changes caused by HFD (Figure 5D). Similarly, Oil Red O-stained images of the liver sections showed that HFD increased lipid accumulation in the liver. In contrast, SS and SK improved hepatic lipid accumulation induced by HFD (Figure 5E).Adipose tissue is a central organ that accumulates excessive energy in the form of triglycerides (TG). Therefore, we analyzed the levels of total lipids, TG, and TC in WAT. We also measured adipocyte size from EAT, because lipid accumulation contributes to adipocyte hypertrophy. In EAT, the HFD group had higher total lipid, TG, and TC levels than the RD group (p < 0.05). In contrast, SS and SK significantly inhibited TG and TC accumulation (p < 0.05) (Figure 6A–C). In MAT, total lipid, TG, and TC levels were also significantly higher in the HFD group than in the RD group (p < 0.05); however, the TG level in the HFD+SS group and the TC levels in the HFD+SS and HFD+SK groups were significantly lower than those in the HFD group (p < 0.05) (Figure 6D–F).Figure 7A shows the H&E-stained images of the EAT sections. The adipocyte size of the HFD group was greater than that of the RD group; however, SS and SK appeared to ameliorate adipocyte hypertrophy induced by HFD. Consistently, the adipocyte size in the HFD group was significantly larger than that in the RD group; however, the adipocyte size was reduced considerably by SS and SK consumption (p < 0.05) (Figure 7B).To examine the possible mechanism by which SS and SK ameliorate obesity, we measured the mRNA expression levels in epididymal adipose tissue. Acetyl-CoA carboxylase (ACC), fatty acid synthase (FAS), and glucose-6-phosphate dehydrogenase (G6PDH) are vital enzymes that regulate lipid synthesis. In this study, the mRNA expression of Acc, Fas, and G6pdh was higher in the HFD group than in the RD group (p < 0.05) (Figure 8A–C). On the other hand, the mRNA levels of Acc and G6pdh were significantly lower in the HFD+SK group than in the HFD group (p < 0.05) (Figure 8A,C). The Fas mRNA levels were also lower in the HFD+SS and HFD+SK groups than in the HFD group, but a significant difference was observed only in the HFD+SS group (p < 0.05) (Figure 8B).Soybean fermentation improves the health benefits of soybeans, as fermentation increases the bioactive peptides and free isoflavones [36]. Whereas traditional soybean fermentation relies on spontaneously colonized inoculum, industrial fermentation uses koji to maintain consistent quality. Koji refers to steamed grains inoculated with well-characterized microbial species under controlled conditions. The anti-obesity effects of soybean protein and soybean polysaccharides have been well documented [22,37]. Moreover, Shin et al. reported that gochujang, made using soybean koji, alleviates obesity [24], but there is a lack of information on whether soybean koji directly influences obesity. Therefore, we investigated whether soybean koji prevents high-fat diet-induced obesity and metabolic disorders better than soybean in rats. The present study, for the first time, demonstrated that eight weeks of SK consumption ameliorated HFD-induced obesity more effectively than SS, and both SS and SK have beneficial effects on HFD-induced dyslipidemia in the experimental rats.The microorganism used in the present study was Bacillus amyloliquefaciens CJ 14-6. Bacillus species, including B. amyloliquefaciens and B. subtilis, have high growth rates and abundant hydrolytic enzymes [38]. According to proximate composition analysis, SK had a higher moisture and crude fat ratio than SS, but the carbohydrate ratio was lower in SK than in SS. Microorganisms produce moisture as a byproduct through respiratory oxidation, using carbohydrates as an energy source. Therefore, the increase in moisture and fat content after fermentation may be due to a decrease in the carbohydrate ratio in the total mass, resulting in a redistribution of nutrient percentages. Nevertheless, the percentage of protein did not increase significantly, which may be due to the partial degradation of proteins during fermentation. Consistently, the free amino acid analysis showed that SK contained more free amino acids, especially leucine and phenylalanine, than SS. Both leucine and phenylalanine are essential amino acids with hydrophobic side chains. Leucine has been reported to inhibit high-fat diet-induced weight gain, hyperglycemia, and hypercholesterolemia in mice [39]. In addition, phenylalanine has been reported to promote cholecystokinin, a satiety hormone that reduces energy intake [40]. It is not clear whether the increased leucine and phenylalanine in this study contributed to metabolic alterations through increased bioavailability in rats. However, previous studies have reported that the requirement and digestibility of leucine and phenylalanine are relatively high compared to other amino acids, suggesting that the beneficial effects of SK may be due in part to the leucine and phenylalanine [41]. Regarding fatty acid composition, SK comprised more MUFA and PUFA than SS; however, the differences were not as significant as those in amino acids. Therefore, we logically postulated that SK might more effectively improve metabolic changes in obesity than SS, which may be related to the enhanced composition of free amino acids.Excessive energy intake from a high-fat diet is one of the leading causes of obesity in modern society. In a previous in vivo study, chronic high-fat diet consumption induced body weight gain, hepatic steatosis, and insulin resistance [42]. In this study, eight weeks of high-fat diet consumption resulted in a 12.1% increase in body weight compared to RD consumption, despite the indifferent dietary intake. In addition, HFD increased the white adipose tissue weight, the sum of EAT, MAT, RAT, and PAT mass, by 37.9% compared to the RD group, indicating that the majority of body weight gain in the HFD group may be due to the weight increase of WAT. However, SK suppressed the WAT increase caused by a high-fat diet, and most of the reduction was found in MAT. Similar tendencies were observed in the SS group, but the differences were not as significant as those in SK, indicating that SK inhibits obesity more effectively than SS. Although the SK group had lower food consumption than the RD group, the anti-obesity effect of SK may not have been only due to an anorexic effect, because the difference in FER between the HFD and SK groups was not statistically significant. In previous studies, the anti-obesity effect of soy protein was linked to its ability to modulate adipogenic and oxidative capacity and adipokine production [43]. Therefore, we suggest that SK may have protective effects against obesity by partially regulating adipose tissue metabolism.The liver is the main organ that metabolizes nutrients and detoxifies harmful substances in the body. In obesity, nutrient overload and increased inflammatory mediators hinder hepatic function, thereby increasing the enzymatic activities of AST, ALT, ALP, and LDH [44]. AST and ALT are enzymes that participate in gluconeogenesis by transferring an α-amino group of aspartic acid and alanine to ketoglutaric acid to generate oxalic acid and pyruvic acid [34]. Both AST and ALT are concentrated in the liver, but liver damage eventually increases serum aminotransferase levels [34]. In general, mild aminotransferase elevation is found in non-alcoholic fatty liver disease [45]. ALP is an enzyme involved in the transport of metabolites through the cell membrane. Although elevated ALP levels do not always imply liver damage, it has been reported that the most common cause of elevated ALP levels is liver disease [45]. LDH is present in most living cells, but at higher levels in the muscle, liver, and kidney. When cells are damaged, LDH leaks from the cells into the blood, so elevated levels of LDH indicate some form of organ damage. In the present study, HFD increased AST, ALT, ALP, and LDH activities, similar to a previous study [46], whereas SK alleviated the increase in ALP levels. As ALT is mostly present in the liver, an increase in ALT activity is a more specific indicator of liver damage than other such indicators [34]. Therefore, from these results, we assumed that both SS and SK were not hepatotoxic, but SK was thought to better alleviate the decline in liver function due to obesity.Leptin is an adipokine secreted from adipose tissue that acts as a hormone through systemic circulation. Leptin was first revealed to regulate appetite, and recently, its regulatory effects on inflammation and blood glucose levels have been reported [47]. In this study, both SS and SK effectively inhibited the HFD-induced increase in serum leptin levels. Since leptin production is related to the size of the adipose tissue [48], the decrease in leptin may be associated with body fat reduction. Furthermore, in this study, we have shown that the HFD+SK group had a lower dietary intake than the RD group, whereas the HFD and HFD+SS groups did not. Therefore, further research on the effect of SK on appetite control is plausible. Previous studies have shown that obesity leads to systemic inflammation through dysregulated adipokine production, resulting in insulin resistance [49]. However, despite the anti-obesity action, the serum glucose increase during HFD feeding was not ameliorated by SK. Therefore, it is assumed that HFD interferes with glucose utilization through other metabolic changes apart from obesity, and SS and SK did not effectively alleviate these pathways.Obesity is a risk factor for CVD, partly due to impaired lipid homeostasis. The hallmark of dyslipidemia in obesity is increased TG, high LDL-C, and low HDL-C levels [50]. Normally, adipocytes regulate energy homeostasis by storing energy in the form of TG and releasing free fatty acids. However, in an obese state, hypertrophic adipocytes lose their ability to store energy, thus increasing the release and the hepatic influx of fatty acids. Hepatic accumulation of TG increases very-low-density lipoprotein synthesis, which delays the lipolysis of chylomicrons [50]. These TG-rich lipoproteins stimulate TG transfer to LDL and HDL particles, leading to the formation of atherosclerotic LDL particles and decreased HDL-C concentration [51]. In the present study, SS and SK alleviated HFD-induced dyslipidemia by lowering TG, TC, LDL-C, and increasing HDL-C levels. Moreover, SS and SK inhibited hepatic lipid accumulation caused by HFD by reducing TG and TC levels, indicating that both SS and SK may have functions against dyslipidemia and fatty liver in obesity. These results concur with previous studies reporting that soybeans fermented with Enterococcus faecium and Lactobacillus jugurti reduce cholesterolemia in Wistar rats fed a hypercholesterolemic diet [52]. Anthony et al. reported that isoflavones are responsible for the effects of soy protein on serum lipid profiles during a mild hypercholesterolemic diet [53]. In this study, SK and SS alleviated dyslipidemia and fatty liver, despite increased body fat. Therefore, we suggest that SS and SK may directly regulate lipid metabolism, and further research is required to elucidate the underlying mechanism.Chronic high-fat consumption increases adipocyte size due to increased fat storage. Therefore, we also measured the adipose tissue lipid concentration and adipocyte size. SS and SK effectively alleviated HFD-induced TG and TC accumulation in adipose tissue. Histological analysis also showed that SS and SK alleviated adipocyte hypertrophy. To maintain lipid homeostasis, adipocytes carry out two reciprocal processes, lipogenesis and lipolysis. ACC, FAS, and G6PDH are critical enzymes involved in lipid synthesis in adipocytes. It has been reported that mRNA expression levels of Acc, Fas, and G6pdh are significantly increased in mouse models of obesity [54,55]. ACC plays a role in the synthesis of malonyl-CoA, a major substrate for long-chain fatty acids [54]. FAS catalyzes the last step in the fatty acid biosynthetic pathway, synthesizing palmitate in the presence of malonyl-CoA and NADPH. G6PDH produces the cellular NADPH required for the biosynthesis of fatty acids [54]. Therefore, we analyzed the mRNA expression of Acc, Fas, and G6pdh to verify the possible mechanism underlying the anti-obesity effect of SK. Consistent with the results on adipose tissue weight, the mRNA expression of Acc, Fas, and G6pdh was increased by HFD, but SK alleviated the mRNA expression more effectively than SS. Therefore, it is postulated that the better anti-obesity effects of SK than SS could be related to the inhibition of lipogenesis via regulation of Acc, Fas, and G6pdh mRNA expression.In conclusion, SS and SK improved the serum lipid profile, hepatic steatosis, and lipid accumulation in the adipose tissue of rats fed a high-fat diet. Notably, SK showed a better anti-obesity effect by reducing white adipose tissue mass compared to SS. Therefore, it is assumed that SK could be used as an effective dietary source to ameliorate obesity and dyslipidemia.Conceptualization, S.P., J.-J.L., and J.-H.H.; methodology, S.P.; software, S.P., J.-J.L., H.-W.S., S.J., and J.-H.H.; validation, S.P., J.-J.L., S.J., and J.-H.H.; formal analysis, S.P. and J.-H.H.; investigation, J.-J.L., S.J., and J.-H.H.; resources, J.-J.L. and H.-W.S.; data curation, S.P., J.-J.L., S.J., and J.-H.H.; writing—original draft, S.P., J.-J.L., H.-W.S., S.J., and J.-H.H.; writing—review and editing, S.P., J.-J.L., H.-W.S., S.J., and J.-H.H.; visualization, J.-H.H.; supervision, J.-J.L., S.J., and J.-H.H.; project administration, J.-J.L.; funding acquisition, J.-J.L. All authors have read and agreed to the published version of the manuscript.This work was supported by a grant funded by CJ Cheiljadang Corp. (No. 2015-00000533).The study was approved by the Chosun University Institutional Animal Care and Use Committee (C IACUC No. 2015-A0028).Not applicable.The data presented in this study are available from the corresponding author upon request.The authors declare no conflict of interest.Effects of steamed soybean (SS) and soybean koji (SK) on final body weight, daily body weight gain, daily average food consumption, and food efficiency ratio in HFD-fed Sprague-Dawley rats. Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. (A) The final body weight, (B) daily body weight gain ((the final BW after dietary feeding—the initial BW)/day), (C) daily average food consumption, and (D) food efficiency ratio were measured or calculated. RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are displayed as a box-and-whisker plot with means (expressed as ‘+’), n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison. Means labeled without a common letter differ significantly, * p < 0.05. *** p < 0.001.Effects of steamed soybean (SS) and soybean koji (SK) on the relative liver and white adipose tissue weight in HFD-fed Sprague-Dawley rats. Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. Relative weights of (A) liver, (B) white adipose tissue (WAT), (C) epididymal adipose tissue (EAT), (D) mesenteric adipose tissue (MAT), (E) retroperitoneal adipose tissue (RAT), and (F) perirenal adipose tissue (PAT). Relative tissue weights (%) were calculated as organ weight (g)/final BW × 100. RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are displayed as a box-and-whisker plot with means (expressed as ‘+’), n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison. a,b Means labeled without a common letter differ significantly, p < 0.05.Effect of steamed soybean (SS) and soybean koji (SK) on biochemical markers of hepatic function, leptin, and glucose in HFD fed Sprague-Dawley rats. Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. (A) Alanine aminotransferase (ALT), (B) aspartate aminotransferase (AST), (C) alkaline phosphatase (ALP), and (D) lactate dehydrogenase (LDH) activities and (E) leptin and (F) fasting glucose levels were analyzed enzymatically or biochemically. RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are displayed as a box-and-whisker plot with means (expressed as ‘+’), n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison. Means labeled without a common letter differ significantly, p < 0.05. *** p < 0.001.Effect of steamed soybean (SS) and soybean koji (SK) on lipid panels, atherogenic index, and cardiovascular disease risk factors. Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. Serum levels of (A) triglyceride (TG), (B) total cholesterol (TC), and (C) high-density lipoprotein-cholesterol (HDL-C) were measured in the experimental rats. (D) Low-density lipoprotein-cholesterol (LDL-C), (E) atherogenic index (AI), and (F) cardiac risk factor (CRF) were processed. RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are displayed as a box-and-whisker plot with means (expressed as ‘+’), n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison. Means labeled without a common letter differ significantly, p < 0.05. *** p < 0.001.Effect of steamed soybean (SS) and soybean koji (SK) on fat accumulation in the liver. Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. (A) Hepatic lipid, (B) hepatic triglyceride (TG), and (C) hepatic total cholesterol (TC) were measured and normalized by the measured tissue weight (g). (D) Harvested whole liver and (E) liver samples fixed and stained with Oil red O staining from each experimental group. RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are displayed as a box-and-whisker plot with means (expressed as ‘+’), n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison. Means labeled without a common letter differ significantly, p < 0.05.Effect of steamed soybean (SS) and soybean koji (SK) on lipid, triglyceride (TG), and total cholesterol (TC) deposition in the epididymal adipose tissue (EAT) and mesenteric adipose tissue (MAT). Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. (A,D) Lipid, (B,E) TG, and (C,F) TC were measured from the epididymal adipose tissue (EAT) and mesenteric adipose tissue (MAT) and normalized by the measured tissue weight (g). RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are displayed as a box-and-whisker plot with means (expressed as ‘+’), n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison. Means labeled without a common letter differ significantly, p < 0.05.Effect of steamed soybean (SS) and soybean koji (SK) on epididymal adipocyte size. Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. The epididymal adipose tissue (EAT) was stained with hematoxylin and eosin (HE). Magnification, 100×. The EAT surface area was quantified using the Image J program. (A) Representative EAT images. (B) Quantification of the surface area of EAT. RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are displayed as a box-and-whisker plot with means (expressed as ‘+’), n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc comparison. Means labeled without a common letter differ significantly, p < 0.05.Effect of steamed soybean (SS) and soybean koji (SK) on epididymal adipocyte mRNA expression. Experimental rats were fed a regular diet (RD) or a high-fat diet (HFD) with 3% SS (HFD+SS) or with 3% SK (HFD+SK) for 8 weeks. (A) Acetyl-CoA carboxylase (Acc), (B) fatty acid synthase (Fas), and (C) glucose-6-phosphate dehydrogenase (G6pdh) mRNA expression was analyzed using qRT-PCR and normalized by β-actin. RD, regular diet; HFD, high-fat diet; HFD+SS, high-fat diet + 3% steamed soybean; HFD+SK, high-fat diet + 3% steamed soybean. Values are means ± standard deviation, n = 8. Data were analyzed using one-way ANOVA followed by Tukey’s post-hoc test. Means labeled without a common letter differ significantly, p < 0.05.Composition of experimental diet.(1) RD: regular diet, (2) HFC: high-fat diet, (3) AIN-93-GX mineral mixture, and (4) AIN-93-VX vitamin mixture.RT-PCR primer sequences (5′ to 3′).Proximate composition of steamed soybean and soybean koji.* p < 0.05, *** p < 0.001.Free amino acid composition in steamed soybean and soybean koji.** p < 0.01, *** p < 0.001. (1) <LLOQ; lower limit of quantification.Fatty acid composition in steamed soybean and soybean koji.* p < 0.05, ** p < 0.01, *** p < 0.001. (1) <LLOQ; lower limit of quantification.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06033.txt ADDED Viewed

	@@ -0,0 +1,11 @@

+Corporate environmental responsibility (CER) is an important component of the corporate social responsibility (CSR) report, and an important carrier for enterprises to disclose environmental protection information. Based on the corporate micro data, this paper evaluates the effect of a mandatory CSR disclosure policy on the fulfillment of corporate environmental responsibility by adopting the difference-in-differences model (DID) with the release of a mandatory disclosure policy of China in 2008 as a quasi-natural experiment. The study draws the following conclusions: First, a mandatory CSR disclosure policy can promote the fulfillment of CER. Second, after the implementation of a mandatory CSR disclosure policy, enterprises can improve their CER level through two channels: improving the quality of environmental management disclosure and increasing the number of patents. Third, the heterogeneity of the impacts of mandatory CSR disclosure on CER is reflected in three aspects: different CER levels, different corporate scales and a different property rights structure. In terms of the CER level, there is an inverted U-shaped relationship between the CER level and mandatory CSR disclosure effect. In terms of the corporate scale, mandatory disclosure of CSR plays a greater role in large-scale enterprises. In terms of the structure of property rights, mandatory CSR disclosure has a greater effect on non-state-owned enterprises.Corporate environmental responsibility (CER) is an important component of a corporate social responsibility (CSR) report and an important carrier for enterprises to disclose environment related behavior information. With the development of global economy, the environment is deteriorating day by day, especially by way of the pollution of the air, water and ocean becoming more and more serious, which brings great threat to the survival and development of human beings [1,2]. Environmental problems have become the bottleneck of economic development. In the 1980s, CSR movements began to rise in developed countries in Europe and the United States, including environmental protection issues. Some non-governmental organizations (NGOs) and public opinions involving Greenpeace, environmental protection, social responsibility, human rights, and other issues also constantly call for the connection between social responsibility and trade. CER requires enterprises to take precautions against environmental problems, take the initiative to assume the responsibility of environmental protection, and promote the development and popularization of environmental technology [3,4,5]. Under increasing external pressure and their own development needs, many European and American multinational companies formulate responsibility codes to make necessary commitments to society, or to meet the needs of different interest groups through environment, occupational health, and social responsibility certification [6,7,8].Although governments attach great importance to CSR, there is no consistent policy on CSR disclosure, especially in terms of mandatory disclosure and voluntary disclosure. In the second half of 2007, ASEAN countries adopted a semi-voluntary disclosure policy [9], i.e., they allowed enterprises to disclose CSR under some restrictive conditions. Before 2013, Malaysia required the disclosure of corporate environmental information in its annual reports so as to disclose CER and promote “Green Malaysia” [10]. India’s company law of 2013 stipulates that in CSR disclosure, enterprises need to disclose environment and other relevant information in the audited annual report [11]. The EU Directive 2014/95/EU on the disclosure of non-financial information by public interest organizations came into force in 2017 [12,13], explicitly stating that non-financial information disclosure should include CER contents. Italy made it mandatory for companies to publish social and environmental reports for the first time in 2018 [14]. As a developing country, China has implemented a series of environmental policies to encourage enterprises to disclose their CSR, especially emphasizing the popularization of environmental awareness and the implementation of environmental protection measures [15,16,17,18]. As in the primary stage of development, China’s stock exchanges did not make mandatory requirements for CSR disclosure until 2007, i.e., they implemented voluntary disclosure. Accordingly, before 2007, less than 3% of China’s listed companies disclosed CSR reports. In order to reduce the opportunities for enterprises to hide environmental pollution, as well as reduce the cost of government supervision, Shanghai Stock Exchange issued the “Guidelines on Environmental Disclosure of Listed Companies on the Shanghai Stock Exchange” (Guidelines for short) in May 2008, which clearly stipulate the disclosure of CSR reports of related companies listed on the Shanghai Stock Exchange. The Guidelines stipulate that relevant companies must disclose environmental information in the form of a temporary announcement, as well as specify the scope of information that must be disclosed by enterprises identified as seriously polluted by the environmental protection department. At the same time, the procedural requirements of environmental information disclosures are clarified. On December 31, 2008, in the “Notice on Doing a Good Job in 2008 Annual Reports of Listed Companies”, three kinds of companies are required to disclose the CSR report, namely, the sample companies of “Shanghai Stock Exchange Corporate Governance Sector”, the companies issuing overseas listed foreign shares, and financial companies. Additionally, other qualified companies are encouraged to disclose CSR voluntarily [19,20,21]. In both mandatory disclosure and voluntary disclosure of CSR, environmental information is regarded as one of the key aspects.Global environmental challenges have been transformed into practical environmental responsibility management mechanism [22,23,24]. CER is an important component of CSR, which plays an important role in improving environmentally efficient conservation and protection behaviors. In Hexun’s CSR professional assessment system, which is one of the most widely used systems in China, the weight of environmental responsibility is 20% by default; while for the manufacturing industry, the environmental responsibility weight is as high as 30%. The important role of CER is embodied in the following two aspects: on the one hand, CER can significantly promote innovation performance and improve environmental performance [25,26]. On the other hand, CER can promote the green governance of enterprises, with corporate reputation as a mediator [27,28]. Managers report information on environmental management, policies, and impacts to the public, and publicize the positive aspects of such governance to achieve better results [29]. From the perspective of management mechanism, it is not enough for enterprises to only disclose environmental information, and it is necessary to study the quality of CER [30,31]. Peng et al. [25] put forward an operationalization of the corporate environmental reporting credibility concept, identifying possible determinants, relevant measures, and indicators. Referring to the CER measurement system proposed by Li et al. [32], this paper takes the CSR mandatory disclosure stipulated by Shanghai Stock Exchange of China in 2008 as a quasi-natural experiment, using the DID model to study the impact and effect of the CSR mandatory disclosure policy on CER.The main work and marginal contribution of this paper is to study the impact of mandatory disclosure policies on CER, as well as the mechanism and heterogeneity of the impact. First, the impact of mandatory disclosure of CSR on CER is evaluated. This empirical study shows that mandatory disclosure of CSR can significantly improve CER as a whole, which has a short-term effect. From the analysis of corporate behavior characteristics, mandatory CSR disclosure improves the relevant management level and technical level of enterprises, then promotes enterprises to fulfill CER. The short-term feature of the effect indicates that the policy is the bellwether of the corporate future strategy. After significantly improving the level of corporate environmental responsibility, the whole market threshold rises, and the improving degree weakens. Second, the impact mechanism of compulsory CSR disclosure on CER is realized by enhancing green management and increasing the moderating effect of patent application. By improving the management level and innovating technology, enterprises can achieve a sustainable development strategy, which not only improves the corporate reputation, but also realizes the long-term economic effect. Third, the heterogeneity of the impact of mandatory CSR disclosure on CER is reflected in the CER quantile level, corporate scale, and the difference of the property rights structure, and this heterogeneity is dependent on the discretionary decision-making behavior of enterprises. Regarding the CER level, this study found an inverted U-shaped relationship between the CER level and the effect of mandatory CSR disclosure. Since the environmental protection foundation of enterprises is different, the promotion ability of mandatory CSR disclosure on CER is also different. The promotion speed is first accelerated and then slowed down. For Chinese enterprises, the goal and purpose of non-state-owned enterprises are not consistent with environmental responsibility to a certain extent. The mandatory disclosure policy forces the shareholders and managers to pay more attention to the performance of environmental responsibility. Therefore, the mandatory CSR disclosure can enhance the CER of non-state-owned enterprises more obviously.The rest of this paper is arranged as follows: the Section 2 is the research design, describing the basic hypotheses, model setting, data, and core variable measurement. The Section 3 is the empirical test of the effect of compulsory CSR disclosure on CER. The effect is obtained by parameter estimation, and the robustness of the parameter results is tested. The Section 4 analyzes the impact mechanism of the mandatory CSR disclosure on CER. The Section 5 is the heterogeneity analysis of the impact of mandatory CSR disclosure on CER. The Section 6 draws the conclusion.Compulsory disclosure of CSR reports can promote the full disclosure of corporate information, thus encouraging enterprises to increase their environmental protection practices and improve the quality of their environmental responsibility information disclosure. When China was still in the stage of voluntary disclosure, the government did not stipulate the contents of disclosure, so less than 3% of China’s listed companies disclosed CSR reports, and the quality of CSR reports was uneven, lacking effective information that could reflect corporate behavior. With the increasingly serious environmental problems, both the government and the market require enterprises to strengthen environmental responsibility. However, with the pursuit of profit maximization, enterprises largely ignore the environmental responsibilities, therefore the effect of mandatory CSR disclosure is different from that of voluntary CSR disclosure to the listed companies. Compulsory and voluntary disclosure of CSR have different impacts on corporate behavior and strategy [33]. The will of the government is a bellwether of the market to some extent [34]. Compulsory disclosure policies force more enterprises to fulfill their environmental responsibilities, over time guiding them to adjust their decision-making behaviors, accelerate the promotion of environmental protection work, and improve the quality of environmental responsibility information disclosure. Based on the above analysis, this paper puts forward the following hypothesis:
+            Mandatory CSR disclosure policy can promote enterprises to fulfill CER.
+          In the context of mandatory CSR disclosure, in order to meet the disclosure requirements, enterprises will strengthen the performance of environmental responsibility from two aspects, namely management and technology [35,36]. In management, enterprises will increase the emphasis on their own environmental protection concepts and environmental protection goals, and establish an environmental management system, environmental knowledge education and training, as well as an environmental emergency response mechanism to regulate their own behavior. In the technical level, through the investment of green technology innovation, enterprises can produce more green patents and realize their sustainable development strategy [37]. Thus, they tend to strengthen environmental management and develop more patents to fulfill their environmental responsibilities so as to meet the mandatory CSR disclosure requirements. Based on this, this paper puts forward the following hypothesis:After the implementation of mandatory CSR disclosure, enterprises fulfill their environmental responsibilities by improving the disclosure quality of environmental management and patents, so as to improve the level of CER.In the face of a mandatory CSR disclosure policy, enterprises will make decisions based on profit optimization as well as their own attributes and external characteristics, so as to achieve a balance between economic benefits and environmental benefits. Therefore, there is heterogeneity in the impact of mandatory CSR disclosure on CER. Generally speaking, different levels of CER represent different levels of legal awareness, social evaluation, low-carbon technology and output, and green management. When the CER of an enterprise is at a low level, its environmental problems are usually serious and there is more room for improvement. Therefore, the mandatory CSR disclosure can prompt the enterprise to carry out corresponding innovation and improve its environmental responsibility level at a faster speed. When the enterprise’s environmental protection level is improved to a certain extent, its development encounters bottlenecks, such as limited space for improvement of management and technology, so its CER increase speed slows down. Therefore, the promotion of mandatory CSR disclosure to CER may have an inverted U-shaped feature. At the same time, the corporate behavior will also be affected by the nature of property rights. Different natures of property rights lead to different stakeholders, as well as different interest demands of all parties. For example, shareholders of state-owned enterprises will pay more attention to sustainable strategic development [38], while those of non-state-owned enterprises pay more attention to economic benefit maximization, therefore paying less attention to environmental issues than shareholders of state-owned enterprises. The implementation of the mandatory CSR disclosure policy makes shareholders of non-state-owned enterprises pay more attention to the performance of environmental responsibility. Therefore, the CER improvement level of non-state-owned enterprises may be higher than that of state-owned enterprises. Based on this, this paper puts forward the following hypothesis:The impact of compulsory CSR disclosure on CER is heterogeneous.Compulsory disclosure of CSR will have an impact on the behavior of enterprises, forcing them to pay attention to the disclosure of environmental protection information, therefore under the pressure of market competition and supervision there will be differences in the fulfillment of environmental responsibility between enterprises with mandatory disclosure and those without mandatory disclosure. The mandatory CSR disclosure policy implemented by China’s Shanghai Stock Exchange in 2008 can be regarded as a quasi-natural experiment. There are many methods to evaluate the effect of such policies. The difference-in-differences (DID) model can examine the implementation effect of policies comprehensively. Therefore, this paper chooses the DID model to study the effect of mandatory CSR disclosure on CER.In recent years, the DID model has been mostly used for quantitative assessment of the implementation effect of public policies or projects in econometrics. In general, the large-scale public policy research is different from ordinary scientific research, and it is difficult to guarantee the complete randomness of sample allocation between the policy implementation group and the control group. The experiment of non-randomly assigned policy implementation group and control group is called natural experiment, which has significant features, i.e., there may be ex-ante differences between different sample groups before the implementation of the policy, which will be ignored through simple before-and-after comparison or lateral comparison, leading to biased estimates of the effect of policy implementation. The DID model is based on the data obtained from natural experiments, and it can effectively control the ex-ante differences between research objects and effectively separate the real results of the policy impact [39].This paper analyzes the impact of mandatory CSR disclosure on CER, meaning relevant variables need to be set based on the policy implementation time. Therefore, samples selected in this paper are the enterprises that disclosed their CSR reports from 2004 to 2012. The enterprises that are subject to mandatory disclosure belong to the treated group, while the enterprises that are not subject to mandatory disclosure belong to the control group. It is expected that the CER scores of the two groups will change significantly after the implementation of mandatory disclosure of CSR. The basic form of the DID model is as follows:(1)CERit=∝+βTreatt ×PLi+∑j=1nγjXjit+λi+vt+εikt
+        where CERit is the environmental responsibility rating of company i in year t; Treatt is the dummy variable of policy implementation, indicating whether the year t is the year of policy implementation or later, if yes it is 1, otherwise it is 0; PLi is a grouping dummy variable, indicating whether the enterprise i is the object of a mandatory disclosure policy, if yes it is 1, otherwise it is 0; Treatt × PLi is the interaction term between the policy implementation dummy variable and the grouping dummy variable; β is the coefficient that this paper focuses on. If β > 0, it indicates that the mandatory disclosure policy has a positive impact on corporate environmental responsibility. Xjit is the j-th control variable in the t-th year of company i; λi is the individual fixed effect; vt is the time fixed effect. This paper studies H1 through Model (1).The explained variable in this paper is the CER score, which will be used as referring to Li et al. [32] to assess the environmental responsibility of enterprises from five dimensions: legal awareness, evaluation, environment-friendly output, low-carbon technology, and green management. The CER score is calculated as follows: First, as data related to enterprise environmental information are largely unavailable, qualitative indicators are usually obtained by content analysis, and quantitative indicators are obtained by weighted aggregation. Second, the qualitative indicators of 2006–2012 come from Chinese Research Data Services Platform (CNRDS) (www.cnrds.com, accessed on 10 April 2021) and CSMAR, including the business research reports, corporate social responsibility reports of listed companies, and so on. Further, the qualitative indicators of 2004–2005 come from the annual reports of the company (www.cninfo.com.cn, accessed on 10 April 2021) and its CSR reports (if any) by the web crawler technology. Third, the Hexun (www.hexun.com, accessed on 10 April 2021) social responsibility scoring system is used to assign weights to the five dimensions respectively to calculate the final CER scores.Specifically, the dimension of legal awareness accounts for 10% of the weight, mainly indicating whether the enterprise has the awareness to abide by the laws and regulations related to environmental protection, which is reflected by three indicators. The first indicator is whether the enterprise follows the GRI Sustainability Reporting Guidelines, which provides the direction of the code of conduct that enterprises need to abide by. The second indicator is environmental and sustainable development disclosure. The third indicator is whether environmental penalties are imposed. The environmental reputation of an enterprise in the society is reflected by two indicators: the first one is whether it has received environmental recognition; the second is whether it has an environmental advantage.The dimension of environment-friendly output accounts for 25% of the weight. It primarily observes whether the corporate production and operation activities are environmentally friendly, which is reflected by three indicators: the first indicator is whether there is a circular economy; the second indicator is the availability of environmentally beneficial products; and the third indicator is whether there are pollution emissions.Low-carbon technology dimension accounts for 25% of the weight, primarily reflecting whether the enterprise saves energy and whether there are measures to reduce three processing wastes, namely, waste gas, waste water, and waste residue.Green management dimension accounts for 25% of the weight, reflecting the impact of corporate management on the environment. The detection index is whether there is third-party inspection and whether green office is adopted.In order to ensure the consistency of the scoring, for the two indicators of whether there is environmental penalty and pollution emission, the enterprise is scored as 0 if there is, and 1 if there is not. For all other indicators, the enterprise is scored as 1 if it is affirmative, and 0 if it is not. The weighted score of all the indicators is the CER score.Based on the above analysis, the environmental responsibility scoring system established in this paper is shown in Table 1.In this paper, the implementation date of the mandatory CSR disclosure policy in 2008 is taken as the policy implementation node, and the enterprises that are forced to disclose are taken as the treated group, while those that voluntarily disclose CSR are taken as the control group. In the treated group, the Treat × PL value before 2008 is 0 and after 2008 is 1. In the control group, the Treat × PL values before and after 2008 are 0. For the control variables, this paper refers to Feng et al. [40], mainly considering the internal characteristics of the enterprise: (1) Corporate scale, which is expressed by the logarithm of total assets. Small enterprises have less incentive to reform and be green, thus have less advantage in technological innovation, while large ones have more advantage in technological innovation. (2) The corporate age, which is represented in this paper by the difference between the study year and the year the enterprise was founded. (3) Capital labor density, which is expressed by the ratio of net fixed assets to total number of employees. Generally speaking, the degree of environmental pollution caused by capital-intensive industries and non-capital-intensive industries are not the same. (4) Ratio of fixed assets, i.e., the ratio of net fixed assets to total assets. The smaller the ratio is, the stronger the enterprise’s liquidity, and the stronger its innovation initiative. (5) Capital structure, namely asset-liability ratio. The higher the debt ratio is, the greater the debt risk the enterprise assumes, which may limit the cost of environmental protection investment and the effect of technological innovation to a certain extent [41]. (6) Return on assets. This indicator reflects the business environment of the enterprise, and to some extent, it will affect its ability to implement environmental protection measures.This paper selects two variables as the mediating effect indicators: (1) The quality of environmental management disclosure. The index reflects the corporate environmental management ability, which consists of eight contents, reflecting whether the company is disclosing: the environmental protection concept; the environmental protection objectives; the environmental management system; the environmental protection education and training; the environmental protection special action the environmental protection emergency mechanism; the environmental protection honor or reward; and the “three simultaneity” system. If the answer to one item is “yes”, 1 point will be scored; otherwise, 0 points will be scored. The final score is the quality score of the environmental management disclosure. (2) Number of patent applications: the total number of patent applications filed by a listed company and a subsidiary joint venture company for the year in which the company applied for the patent. The index reflects the ability of companies to fulfill their environmental responsibilities from the technical level.To implement the DID model, considering the availability of data, this paper only takes into account the data of 4 years before and after the implementation of the policy, i.e., this paper selects all Chinese A-share listed enterprises that disclose their CSR from 2004 to 2012 as the initial research samples. Due to the loss of corporate performance, its operation and environmental protection practice will change significantly, which makes it fail to fully reflect the decision-making behavior of normal operations in the face of policy implementation. Therefore, this paper follows the standard practice and excludes the companies with ST/ST* from January 2004 to December 2012. In addition, the financial industry differs greatly from entity enterprises in operation and environmental protection operations, which cannot be studied uniformly, therefore, enterprises in the financial industry are excluded from this paper. The annual data in this paper start from 2004, thus excluding the data of listed companies after January 1, 2004. After the above sample processing, the sample size of the empirical research in this paper is 460. The financial and social responsibility data in this paper are from the CSMAR database and the China Research Data Service Platform (CNRDS).After collecting relevant data, this part also makes descriptive statistical analysis on the variables. Considering that the variables are different among enterprises with different property rights, the samples are also divided into the state-owned-enterprise (SOE) group and the non-state-owned-enterprise (NSOE) group. The differences of CER mean value changes between the treated group and the control group, before and after the policy implementation are reported.As shown in Table 2, the CER of both the treated group and the control group increased significantly after the policy implementation in 2008, and the CER of the treated group is significantly higher than that of the control group. The change in CER in state-owned enterprises is almost the same as that of the full sample, but the change in CER in non-state-owned enterprises is more significant. The mean value change in the treated group of non-state-owned enterprises is higher than that of the treated group of state-owned enterprises, while the mean value change in the control group of non-state-owned enterprises is lower than that of the control group of state-owned enterprises, and the difference between the treated group and the control group of non-state-owned enterprises is twice that of the state-owned enterprises. The results of descriptive statistics confirm the previous hypothesis.One of the preconditions for the validity of the DID estimation is that the treated group and the control group meet the common trend assumption before being processed [42]. Therefore, in order to verify the appropriateness of the DID model in this paper, we conducted a parallel trend test on the CER scores of the treated group and the control group according to the parallel test procedure. The growth trends of CER scores in the treated group and the control group are shown in Figure 1.As can be seen from Figure 1, before and after the mandatory disclosure of CSR, the CER scores, as well as the growth trends of the treated group and the control group, are significantly different. Before the release of the mandatory CSR disclosure policy, the CER scores of the control group and the treated group maintained the same growth trend, but after the release of the policy, the CER scores and growth trends of the treated group and the control group changed significantly. Therefore, the result of the DID model is in line with the common trend hypothesis.According to Model (1), OLS is used to conduct benchmark regression on the impact of a mandatory CSR disclosure policy on CER, and the benchmark DID regression results are shown in Table 3.In Table 3, column (1) shows the traditional DID regression. Column (2) displays the results of the fixed effect of time and individual. Column (3) displays the results with control variables. The results show that the coefficients of Treat × PL are 25.33, 9.726 and 9.522, respectively, all of which have significant positive effects under 1%, indicating that the mandatory CSR disclosure policy can significantly improve the CER score of enterprises, and promote enterprises to fulfill their environmental responsibility.In order to reach a more reliable conclusion, we conducted two robustness tests. Firstly, a more accurate parallel trend test was conducted for the treated group and the control group. Secondly, we set up a virtual sample by randomly setting the treated group and the control group to conduct the placebo test.In the previous part, we made the CER growth trend curve of the treated group and the control group before the benchmark DID regression, and roughly verified that the two groups had the same trend before the policy implementation. In order to verify the parallel trend hypothesis more accurately, the following dynamic effect model is established:(2)CERit=∝+∑t=20042012βtTreatt ×yeart+∑j=1nγjXjit+λi+vt+εikt
+          where the base year is 2008; yeart is a time dummy variable; βt represents a series of estimated values from 2004 to 2012; and the definitions of other variables are the same as those of the benchmark regression model (1). Taking the year before the policy, i.e., 2007, as the base period, if the regression coefficients from β2004 to β2006 are not significant, then the treated group and the control group meet the parallel trend test. The specific parameter estimation results are shown in Table 4.The dynamic effects of the estimated values in Table 4 illustrate the parallel trend hypothesis, and the DID model can be used in this quasi-natural experiment. In Table 4, the year before the policy, i.e., 2007, is taken as the base period. Column (1) displays the dynamic effect results of the fixed effects of individual, time and industry, and column (2) displays the results with control variables. βt are not significant from 2004 to 2006, indicating that there is no significant difference in environmental responsibility scores between the treated group and the control group before the policy was implemented, satisfying the parallel trend hypothesis. βt in 2008 is significantly positive, indicating that the CER score of enterprises that are forced to disclose their CSR in this year is significantly higher than that of enterprises that are not forced to disclose their CSR. The policy can promote enterprises to fulfill their environmental responsibility. Since then, the coefficient has decreased but remained significant, indicating that, compared with 2007, there is still a significant gap between the CER scores of the two groups of enterprises after 2008, as the growth trend shown in Figure 1. The possible reasons are as follows: With the advance of the policy, the number of enterprises subject to mandatory disclosure increased in 2009 and beyond. Meanwhile, voluntary disclosure enterprises in the policy environment strengthen their own environmental responsibility. To better demonstrate the results of the parallel trend test, the point estimates of coefficient βt and the corresponding 95% confidence interval estimates are plotted in Figure 2. The figure also reflects the parallel trend of the treated group and the control group, as well as the future trend changes.In order to further examine the differences in CER scores between enterprises subject to mandatory CSR disclosure and enterprises with non-mandatory CSR disclosure, this study conducts a placebo test by randomly assigning a treated group and a control group. This method ensures that the independent variable TREL × PL has no influence on the CER score. The benchmark model includes 460 sample enterprises, among which 160 are subject to mandatory CSR disclosure (the treated group) and 300 are voluntary disclosure enterprises (the control group). The corresponding numbers of treated groups and control groups are randomly established in this paper, and then regression is performed according to Equation (1). This step was repeated 500 times thereafter. The t-value density distribution of these 500 estimates is shown in Figure 3. On the one hand, the t-value presents an inverted U-shaped distribution, which proves that the regression results are robust. On the other hand, the straight dotted line in the figure is the t value of the fundamental regression, which is at the far right of the 500 estimates, also verifying that the fundamental regression is robust.Mandatory CSR disclosure has a significant positive effect on CER, but the influence mechanism may have multi-path characteristics. In practice, enterprises tend to strengthen the performance of environmental responsibility from two aspects: management and technology. On the one hand, enterprises should strengthen management and establish scientific working procedures to ensure sufficient execution to achieve strategic goals quickly [43]. On the other hand, enterprises can improve the quality of products and save energy consumption by constantly innovating technology to ensure their core competitiveness and realize sustainable development [44]. Through many experiments and theoretical analysis in the hypotheses, this paper found that mandatory CSR disclosure enhances CER through strengthening green management and Research and Development patent applications. Based on this, this paper uses the mediating effect test to verify H2 by examining the impact of the environmental management disclosure quality and patent application number on the performance of CER.The basic model forms to investigate the mediating effect of environmental management disclosure quality are as follows:(3)CERit=α1+θ1Treatt ×PLi+∑j=1nγjXjit+λi+vt+εikt
+        (4)MNGit=α2+θ2Treatt ×PLi+∑j=1nγjXjit+λi+vt+εikt
+        (5)CERit=α3+θ3 · Treatt ×PLi+θ4 · MNGit+∑j=1nγjXjit+λi+vt+εikt
+        where MNG represents the quality of an enterprise’s environmental management disclosure. In this part, the stepwise test of regression coefficients is adopted, and the meanings and calculation methods of other variables are consistent with the fundamental DID regression model. θ1 measures the impact of the mandatory CSR disclosure policy on CER. θ2 measures the impact of the mandatory CSR disclosure policy on MNG. θ3 measures the direct impact of the mandatory CSR disclosure policy on CER. θ4 measures the impact of MNG on CER. Parameters of models (3)~(5) are estimated, and the results are shown in Table 5.It can be seen from Table 5 that improving the quality of environmental management disclosure plays a mediating role in the influence mechanism of mandatory CSR disclosure on CER. Specifically, column (1) in Table 5 verifies the direct impact of mandatory CSR disclosure on CER; column (2) verifies the impact of mandatory CSR disclosure on the quality of corporate environmental management disclosure. The regression coefficient is 0.281 and significantly positive, indicating that mandatory CSR disclosure does lead to an increase in the quality of environmental management disclosure. Column (3) shows that the impact coefficients of mandatory CSR disclosure on CER and environmental management disclosure quality are significantly positive, indicating that the environmental management disclosure quality plays a partial mediating role in the effect of mandatory CSR disclosure on CER. Meanwhile, the Sobel test is used to verify that the mediating effect is robust [45,46].The basic model forms to investigate the mediating effect of the number of patent applications are as follows:(6)CERit=α4+β1Treatt ×PLi+∑j=1nγjXjit+λi+vt+εikt
+        (7)Patentit=α5+β2Treatt ×PLi+∑j=1nγjXjit+λi+vt+εikt
+        (8)CERit=α6+β3 · Treatt ×PLi+β4 · Patentit+∑j=1nγjXjit+λi+vt+εikt
+        where Patent represents the number of an enterprise’s patent applications. The meanings and calculation methods of other variables are consistent with the fundamental DID regression model. β1 measures the impact of a mandatory CSR disclosure policy on CER. β2 measures the impact of a mandatory CSR disclosure policy on Patent. β3 measures the direct impact of a mandatory CSR disclosure policy on CER. β4 measures the impact of the number of an enterprise’s patent applications on CER. Parameters of models (6)–(8) are estimated, and the results are shown in Table 6.It can be seen from Table 6 that the number of corporate patent applications plays a mediating role in the influence mechanism of mandatory CSR disclosure on CER. Specifically, column (1) in Table 6 verifies the direct impact of mandatory CSR disclosure on CER, while column (2) verifies the impact of mandatory CSR disclosure on the number of corporate patents application. The regression coefficient is 79.98 and significantly positive, which indicates that mandatory CSR disclosure does lead to an increase in the number of corporate patents application. Column (3) shows that the impact coefficients of mandatory CSR disclosure on CER and the number of corporate patents application are significantly positive, indicating that the number of corporate patent applications plays a partial mediating role in the effect of mandatory CSR disclosure on CER. Meanwhile, the Sobel test is used to verify that the mediating effect is robust.Combining the above two parts, we can draw a conclusion that the empirical results indicate that the impact of mandatory CSR disclosure on CER is to some extent realized by improving the quality of enterprise environmental management and increasing the number of patent applications.Different CER levels reflect different levels of legal awareness, social evaluation, low-carbon technology and output, and green management [47]. Due to the differences in enterprises’ characteristics, the dispersion degree of CER is great, and the estimated results may be affected by extreme values. This part refers to the practice of Li et al. [32], adopts the benchmark regression Equation (1), as well as quantile regression to study the impact of a mandatory CSR disclosure policy on CER scores at different levels to verify Hypothesis 3. The quantile regression model is used to conduct the parameter estimation, and the results are shown in Table 7.It can be concluded from Table 7 that there is an inverted U-shaped relationship between the CER level and mandatory CSR disclosure effect. Table 7 shows the regression results of CER scores at the 10% to 90% quantile levels. It can be seen that the mandatory disclosure policy has a significant impact on CER scores from 10% to 90%. The coefficient first increases, reaches the maximum at 30%, and then decreases, which is consistent with Hypothesis 3. On the whole, mandatory disclosure policies can promote the CER, but when the level of CER is very low, the environmental problems of enterprises tend to be more serious, with much room for improvement, so the mandatory disclosure of CSR forces enterprises to carry out corresponding innovation in order to meet the disclosure requirements, and improve their level of environmental responsibility at a faster speed. When an enterprise’s CER increases to the level of 0.3 or above, its development will encounter a bottleneck, and the growth rate of its CER will slow down due to management and technical limitations.Past studies have shown that companies of different sizes usually perform differently [48,49,50]. Fortunati et al., analyzed that small and medium sized enterprises in the Italian agri-food industry could provide an adequate level of circular strategies and social responsibility practices [51]. This part adopts the benchmark regression Equation (1), as well as the quantile regression of corporate scale, to study the impact of a mandatory CSR disclosure policy on CER scores to verify Hypothesis 3. The quantile regression model is used to conduct parameter estimation, and the results are shown in Table 8.It can be seen from Table 8 that the promotion effect of a compulsory CSR disclosure policy on CER levels is heterogeneous at different corporate scales. In short, the promotion effect of a compulsory CSR disclosure policy on CER increases with the increase in the corporate scale. In the sample group of about 20% of the largest enterprises, the promotion effect of the policy is not significant. This is due to the number of large-scale enterprises in the treated group and in the control group being unbalanced, which leads to the bias of results. When the policy is promulgated, large-scale enterprises have to make stronger responses to meet their regulatory requirements and social reputation, due to their social influence.Different property rights reflect different stakeholders and different ways of making decisions [52]. Therefore, in view of the difference in property rights structures, this paper continues to discuss the impact of mandatory disclosure of CSR on CER. In 2008, when the policy was issued, there were 342 state-owned sample enterprises (SOEs) and about 118 non-state-owned sample enterprises (NSOEs), which were respectively tested according to the benchmark regression Equation (1), and the results are shown in Table 9.It can be concluded from Table 9 that the impact of mandatory CSR disclosure on CER is heterogeneous in terms of different property rights structures. Table 9 shows the impact of mandatory disclosure policies on the CER scores of SOEs and NSOEs. Columns (1) and (3) show the regression results of SOEs and NSOEs, respectively. Columns (2) and (4) show results with control variables based on the above results, respectively. We can see that the policy has a significant positive impact on the CER performance of enterprises with different property rights structures. However, the regression coefficient of non-state-owned enterprises is about twice that of state-owned enterprises, indicating that the promotion effect of the policy on non-state-owned enterprises is greater than that on state-owned enterprises. One of the stakeholders of state-owned enterprises is the state (investor), while the investor of non-state-owned enterprises is the private sector, which is more focused on economic interests, and, due to the mandatory CSR disclosure policies, non-SOEs are forced to take environmental objectives into account, so the CER changes are more dramatic than those in SOEs that reflect the will of the state to some extent.Faced with environmental deterioration, China pays special attention to the popularization of environmental awareness and the implementation of environmental measures, especially in enterprises. On the basis of voluntary CSR disclosure, mandatory disclosure policies for three types of enterprises are implemented, namely, the sample companies of “Shanghai Stock Exchange Corporate Governance Sector”, the companies issuing overseas listed foreign shares, and financial companies, in order to reduce the opportunities for enterprises to hide environmental pollution and reduce the cost of government supervision. In this context, this paper uses the DID model to test the impact of mandatory CSR disclosure policies on corporate environmental responsibility (CER). The main conclusions are as follows:First, mandatory CSR disclosure policies can promote the fulfillment of CER by enterprises. The CER scores of enterprises subject to mandatory disclosure are higher than those of voluntary disclosure enterprises. However, this policy has a short-term effect. In the long run, the CER gap between mandatory disclosure enterprises and voluntary disclosure enterprises gradually narrows. With the promotion of the policy, the number of enterprises subject to mandatory disclosure in 2009 and later increases; at the same time, the voluntary disclosure enterprises strengthen their own environmental responsibility in the policy environment, and the threshold of the whole society has also been raised.Second, after the implementation of mandatory CSR disclosure policies, enterprises can improve their CER level through two channels: improving the quality of enterprise environmental management and increasing the number of patent applications. In order to meet the disclosure requirements, enterprises will strengthen the implementation of environmental responsibility from two aspects, namely management and technology. From the management level, enterprises will increase the emphasis on their own environmental protection concept and environmental protection goals, and establish an environmental management system, environmental knowledge education and training, and an environmental emergency response mechanism to regulate their own behavior. From the technical level, through the investment of green technology innovation, enterprises can produce more green patents and realize their sustainable development strategy.Third, the heterogeneity of the effect of a mandatory CSR disclosure policy on CER is reflected in three aspects: CER levels, corporate scales, and property rights structure. In terms of the CER level, there is an inverted U-shaped relationship between the CER level and the mandatory CSR disclosure effect. In terms of the corporate scale, mandatory disclosure of CSR plays a greater role in large-scale enterprises. In terms of property rights, mandatory disclosure of CSR plays a greater role in non-state-owned enterprises. Specifically, a mandatory CSR disclosure policy has different impacts on enterprises with different levels of CER. When CER is at or below the 30% quantile level, the promoting effect of the mandatory CSR disclosure policy increases; when CER is above the 30% quantile level, the promoting effect decreases, presenting an inverted U-shaped change. A mandatory CSR disclosure policy promotes change in non-state-owned enterprises better than in state-owned enterprises. Shareholders of state-owned enterprises will pay more attention to sustainable strategic development, while non-state-owned enterprises will pay more attention to economic benefit maximization, so their shareholders pay less attention to environmental issues than shareholders of state-owned enterprises. The implementation of the mandatory CSR disclosure policy makes shareholders of non-state-owned enterprises pay more attention to the performance of environmental responsibility. Therefore, the CER improvement level of non-state-owned enterprises is higher than that of state-owned enterprises.The conclusions provide following implications: First, policy makers can use mandatory policies in the field of corporate environmental responsibility to achieve green production and green management. Second, enterprises should concentrate on the quality of enterprise environmental management and number of patent applications, which are key to improving their CER levels. In the absence of mandatory policies, enterprises can also improve the above two parts through independent efforts to gradually realize environmental responsibility. Finally, governments should formulate appropriate policy strategies in the face of different types of corporate entities and pay more attention to those with extreme CER levels, as well as small and medium size and state-owned enterprises.Furthermore, as corporate social responsibility is an important and complex issue, it is not enough to consider only the environmental responsibility. Future research should focus on multi-dimensional research of corporate social responsibility [53]. The existing research on mandatory disclosure policies also lacks multi-dimensional consideration, which will be the main direction of our future research.Conceptualization, Y.L. and P.F.; methodology, L.C.; software, L.C.; validation, Y.L., P.F., and L.C.; resources, Y.L.; data curation, L.C.; writing—original draft preparation, Y.L.; writing—review and editing, P.F.; visualization, L.C.; supervision, P.F. All authors have read and agreed to the published version of the manuscript.This research was funded by Humanities and Social Sciences Foundation of Ministry of Education of China, grant number: 19YJCZH108; Key project of Education Department of Hunan Province, grant number: 20A139; General Project of Philosophy and Social Science Foundation of Hunan Province, grant number: 19YBA069.Not applicable.Not applicable.Not applicable.The authors declare no conflict of interest.Growth trends of the CER score.Dynamic effect of the DID model.Results of the placebo test. Note: Parameter estimates (dots) and corresponding 95% confidence intervals (lines) are based on Model (2).Corporate environmental responsibility scoring system.Descriptive statistics of the variables.DID regression results.** p < 0.05, *** p < 0.01, t values are in brackets.Dynamic effects of mandatory CSR disclosure policy on CER.* p < 0.10, *** p < 0.01, t values are in brackets.Mediating effect results of environmental management disclosure quality.*** p < 0.01, t values are in brackets.Results of mediating effect of the number of patents application.*** p < 0.01, t values are in brackets.CER estimation results at different quantile levels.** p < 0.05, *** p < 0.01, t values are in brackets.CER estimation results of different corporate scales.* p < 0.10, ** p < 0.05, *** p < 0.01, t values are in brackets.Estimation results of different property rights structures.* p < 0.10, *** p < 0.01, t values are in brackets.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06034.txt ADDED Viewed

	@@ -0,0 +1,3 @@

+Objectives: China is the country most afflicted by hepatocellular carcinoma in the world. However, little is known about the epidemiology of hepatocellular carcinoma in China. This study aimed to examine the trends of the prevalence, incidence, and mortality of hepatocellular carcinoma in China, and to investigate the effects of age, period, and birth cohort on the epidemiological trend. Methods: The data were obtained from the Urban Employee Basic Medical Insurance claims database (2003–2017) in Tianjin, China, which covers 5.95 million individuals. The average annual percentage change of the prevalence, incidence, and mortality were accessed using joinpoint regression. Age-period-cohort models were produced to quantify the effects of age, period, and cohort. Results: The hepatocellular carcinoma prevalence rate increased by 5.13% annually from 20.12/100,000 in 2008 to 30.49/100,000 in 2017, and the incidence rate was almost unchanged, from 13.91/100,000 in 2008 to 14.09/100,000 in 2017, but mortality decreased by 1.80% annually from 8.18/100,000 in 2008 to 7.34/100,000 in 2017. The age-period-cohort analysis revealed that the prevalence rate was remarkably increased from age 25, peaked in age 60, and decreased at age 70 and over. In the period index, the prevalence rate increased gradually from 2008 to 2016, and decreased a little in 2017. In the cohort index, the prevalence rate decreased approximately linearly from the 1925 cohort to the 1990 cohort. The result for the incidence was similar to the prevalence. The mortality rate increased approximately linearly from age 45 to 85, decreased from the 1925 cohort to the 1990 cohort, but it changed a little with the change of period. Conclusions: The findings of this study could inform the necessity of conducting earlier screening for high-risk individuals and improving the treatment of hepatocellular carcinoma, which may also help to predict future changes in hepatocellular carcinoma epidemiology.Primary liver cancer is one of the most prevalent and deadly cancers worldwide, with about 905,667 new cases and 830,180 deaths occurring in 2020 [1]. China is the country most afflicted by liver cancer in the world, and about half of global newly diagnosed cases and deaths occur in China (410,038 new cases and 391,152 deaths in 2020) [1,2]. In patients with primary liver cancer, the major histological type is hepatocellular carcinoma (HCC), comprising about 90%, followed by intrahepatic cholangiocarcinoma and other rare types [3]. As HCC is responsible for a significant incidence and mortality around the world, resulting in a substantial economic burden, the description of the changing HCC epidemiological data is critical for the healthcare system.Previous studies have reported worldwide or national trends on HCC epidemiology. Rich et al. suggested that the highest incidence of HCC in the world was in Asia and Africa, and the HCC incidence may have plateaued or begun to decrease in some Asian countries. Goh et al. also stated that Asia had the highest incidence of HCC worldwide (the age-standardized incidence rate of HCC in males in Eastern Asia was 31.9/100,000), and some Asian countries or regions had success in reducing HCC by introducing hepatitis B virus (HBV) vaccinations, such as Taiwan and Singapore [4]. Yeesoonsang et al. reported that the HCC incidence was expected to decrease among males and stabilize among females in the future [5]. Unlike in Asia, recent studies have reported that the HCC incidence has increased in low and medium incidence areas such as Western Europe and North America [6,7,8,9]. Wallace et al. reported that the age-adjusted incidence rate of HCC among Australian increased from 1.38/100,000 in 1982 to 4.96/100,000 in 2014, and White et al. revealed that the HCC incidence increased from 4.4/100,000 in 2000 to 6.7/100,000 in 2012 in the United States [7,9]. In addition, Bertuccio et al. reported that HCC mortality was observed to reduce in East Asia (e.g., Japan, Hong Kong and Korea) due to the control of HBV and hepatitis C virus (HCV) infections, but they still remained around 10–24/100,000 men and 2–8/100,000 women in Asia, which was 2- to 5-fold higher than those in most European and American countries [10].However, there is no data on the epidemiology of HCC in mainland China, while these have been widely reported in the United States, Australia, Denmark, Thailand and many other counties and areas [4,5,6,7,8,9,10]. Only a few studies have reported the incidence and mortality rates of primary liver cancer in some regions of China, including Shanghai, Guangzhou, Shenzhen, Chongqing, Fuzhou, Nantong, Sihui and other cities [11,12,13,14,15,16,17,18,19,20,21]. However, almost all of the cities mentioned above are located in the south of China; the changes in prevalence, incidence, and mortality of primary liver cancer have not yet been examined in Northern China, and much less for HCC. In addition, the diet and lifestyle for people living in different countries or areas are different, and people in Northern China prefer to drink alcohol, which is an identified risk factor for HCC [22]. Therefore, it is essential to examine the epidemiology of HCC in mainland China, especially in Northern China.Tianjin is one of the four municipalities in China, which is also the largest coastal opening city located in Northern China, ranking the seventh among the total 31 provinces/municipalities in terms of the gross domestic product per capita in mainland China. With the development of society, the living conditions, diet, and lifestyle, as well as the level of disease diagnosis and treatment have changed significantly, which may all affect the epidemiology of HCC. These factors in Tianjin also reflect the changes in many other cities in Northern China. This study aimed to describe the trends in HCC prevalence, incidence, and mortality rates over the past decade in Tianjin, Northern China, and to investigate the effects of age, period, and birth cohort on HCC prevalence, incidence, and mortality.This population-based study was conducted using data obtained from the Urban Employee Basic Medical Insurance (UEBMI) claims database (2003–2017) in Tianjin, China. China has almost achieved the universal coverage of medical insurance through two systems: UEBMI and Urban and Rural Resident Basic Medical Insurance (URRBMI), which covers 1354.36 million inhabitants, accounting for 96.7% of the total Chinese population in 2019 [23]. The URRBMI program covers children, students and other unemployed adult residents living in urban and rural areas, and UEBMI enrollees represent all adult (at least 18 years) employees and retirees of the public and private sectors. Due to the different reimbursement benefits between the two types of medical insurance systems (the URRBMI program covers fewer healthcare items and pays less than UEBMI), inhabitants who are enrolled in the URRBMI program always have a lower utilization rate of healthcare resources than those enrolled in the UEBMI plan [24]. The evidence from the UEBMI claims database could more truly reflect the level of diagnosis and treatment, as well as the epidemiology data of a city. In Tianjin, there were about 11.37 million enrollees (UEBMI: 5.95 million; URRBMI: 5.42 million) in 2019, of which UEBMI enrollees accounted for over half [25]. The analytical sample in this study was thirty percent of the UEBMI enrollees randomly sampled based on their unique identification number. The UEBMI database consisted of inpatient, outpatient, and pharmacy service claims (the database from 2003 to 2007 included only the inpatient claims). The enrollment history, patient demographics (age, sex, working status), dates of service, diagnoses, medical prescription, and procedure information, as well as the related costs, were included in this database. Both the International Statistical Classification of Diseases and Related Health Problems 10th Revision (ICD-10) codes and medical records were used to identify the disease diagnoses. In addition, the all-cause death information was included in a separate dataset, which could be linked by patients’ unique identification number. This study was exempted from the application for ethical approval by the Safety and Ethics Committee of the School of Pharmaceutical Science and Technology in Tianjin University.The target population were the enrollees of UEBMI in the analytical sample during each calendar year from 2008 to 2017. The definition and identification of prevalent cases, incident cases and death cases are introduced as follows. The prevalent cases were patients with a diagnosis of HCC (ICD-10: C22.0 supplemented with Chinese descriptions) through the inpatient and outpatient claims, who were identified by the calendar year during 2008–2017 and included newly diagnosed patients and previously diagnosed cases. The incident cases (i.e., newly diagnosed patients) were identified from the prevalent cases. The year of the initial HCC diagnosis for each prevalent case was identified. The patients who had any diagnosis of HCC between 1 January 2003 and 31 December 2007 were excluded. Death cases (i.e., patients with HCC who died in each year) were also identified from the prevalent cases. The individual identification numbers were used to identify the death cases from a separate dataset mentioned above.The crude rate and age-standardized rate (ASR) were calculated. The crude prevalence, incidence, and mortality rate were defined as the number of prevalent cases, incident cases and death cases divided by the target population size, respectively. The crude rates were also calculated by sex (male/female), age groups (divided according to Segi’s World Standard Population, i.e., 20–24, 25–29, and 30–80 by 5 years, ≥85) and birth cohorts (1915–1919, 1920–1990 by 5 years). The ASR was calculated by summing up the products of the age-specific rates (ai, where i denotes the ith age class) and the number of persons (or weight, wi) in the same age subgroup i of the chosen standard population, then dividing the sum of the standard population [2] (or weights):(1)ASR=∑i=1Aaiwi∑i=1Awi×100,000Segi’s World Standard Population was used to standardize the prevalence, incidence, and mortality in this study [26].The joinpoint regression analysis was established to estimate the annual percent change (APC) for each segment and the average annual percent change (AAPC) over the entire period to quantify the trends of the age-standardized HCC prevalence, incidence, and mortality. This analysis was composed of a few continuous linear segments, which are always used to describe the trends in the outcome, and are also named as piecewise regression, segmented regression, broken line regression and multi-phase regression [27]. A general form of this model for observations (x1, y1), (x2, y2),……,(xn−1, yn−1), and xn, yn  could be written as follows:(2)Ey|x = β0+β1x+δ1x−τ1++……+δkx−τk+
+          where  τ1, τ2,……,τk−1, τk are the unknown joinpoints and x−τk+ = x−τk  for x−τk > 0,  and 0 otherwise [27]. The analysis starts with the zero joinpoint, representing a straight line, and tests whether more joinpoints are statistically significant and should be added to the model. The significant joinpoints were identified by a Monte Carlo Permutation method [27]. In the final model, each joinpoint denotes a significant change in the trend of the line segment separated by this joinpoint. The annual percent change for each line segment and the corresponding 95% confidence intervals (CI) are reported in the final model.The age-period-cohort analysis was performed to evaluate the net effects of age, period, and cohort on HCC prevalence, incidence, and mortality simultaneously. Both age (from 20 to 85 years old) and period (from 2008 to 2017) were subdivided by 1-year intervals, and the birth cohort was calculated by subtracting the age from the period. The age-period-cohort models in this study were based on a Poisson log-linear model with an intrinsic estimator (IE), which was widely used to avoid linear dependency (i.e., period = age + cohort) and to disentangle the three effects of age, period, and cohort [28]. The IE approaches the estimator of the age-period-cohort model by applying the estimable functions and the singular value decomposition of the matrices, and it generates the coefficients of the effects, which are exponentially expressed as rate ratios [28]. The model could be generally expressed as:(3)logYj = μ+α agej+β periodj+γ cohortj+ε
+          where the Y denotes the HCC prevalence, incidence, and mortality of the corresponding age group j; the α, β and γ denote the corresponding age, period, and cohort effects; μ is the intercept item; and ε   is the random error. A full age-period-cohort model fitted better than any combination of age, period, and cohort factors (Supplementary Table S1).The statistical analyses were performed using Joinpoint Regression Program V.4.7.0.0 and Stata V.13.0. The significant level was defined as two-sided alpha = 0.05.Over the study period from 2008 to 2017, a total of 3811 men and 1834 women with HCC were identified, of which 3563 men and 1742 women were new cases, and 2163 men and 722 women died during the follow-up. The prevalence, incidence, and mortality of males were about twice those of females (Figure 1). The number of patients with HCC increased from 447 (crude prevalence rate: 38.07/100,000; ASR: 20.12/100,000) in 2008 to 1141 (crude prevalence rate: 71.37/100,000; ASR: 30.49/100,000) in 2017. There were 334 new cases (crude incidence rate: 26.65/100,000; ASR: 13.91/100,000) in 2008 and 524 new cases (crude incidence rate: 32.78/100,000; ASR: 14.09/100,000) in 2017. The number of deaths increased from 196 (crude mortality rate: 15.64/100,000; ASR: 8.18/100,000) in 2008 to 281 (crude mortality rate: 17.58/100,000; ASR: 7.34/100,000) in 2017 (Table 1). Figure 2 shows the age-specific prevalence, incidence, and mortality rates of HCC from 2008 to 2017. The prevalence rate of HCC was lower among individuals under the age of 30 (<5.0/100,000; average value between 2008 and 2017, and the same below); it increased for the 30–34 age group (6.6/100,000) and peaked in 80–84 age group (217.1/100,000). The incidence rate of HCC also increased from 30–34 age group (5.6/100,000) and peaked in 80–84 age group (156.2/100,000). The mortality rate of HCC was lower under the age of 45 (<5.0/100,000), and continuously increased from the 45–49 age group (6.6/100,000) to the ≥85 age group (114.6/100,000). Figure 3 showed the temporal trends of the prevalence, incidence, and mortality rates of HCC by age group from 2008 to 2017. The prevalence rate of HCC among individuals aged 50–54, 55–59, 60–64, 65–69 and 70–74 was found to increase from 2008 to 2017, but there was no apparent trend in any of the other age groups. The incidence and mortality rates of HCC among the different age groups had no obvious temporal trends from 2008 to 2017. The cohort-specific prevalence, incidence, and mortality rates of HCC by age group are shown in Figure 4. While the prevalence rate among those aged 55–59 and 60–64 was found to increase steadily with each successive birth cohort, there was no apparent trend in any of the other age groups. The incidence rates in the different age groups all decreased with the birth cohort even though there were some fluctuations in several groups. The mortality rate in the different age groups almost all decreased with the birth cohort, except for 40–44 and 45–49.Table 2 shows the results of the joinpoint regression analysis. The prevalence rates among the overall patients, men, and women all increased from 2008 to 2017, but the changes for the women were not statistically significant (overall: AAPC: 5.13%; 95% CI: 1.56%–8.83%; p = 0.005. Men: AAPC: 3.62%; 95% CI: 1.54%–5.76%; p = 0.001. Women: AAPC: 7.01%; 95% CI: −1.52–16.27%; p = 0.110). There was no remarkable trend in the HCC incidence rates among the overall patients (AAPC: −0.82%; 95% CI: −3.62%–2.07%; p = 0.576), men (AAPC: 3.98%; 95% CI: −5.6%–14.53%; p = 0.429) or women (AAPC: −0.1%; 95% CI: −5.85%–5.99%; p = 0.969) from 2008 to 2015, but a significant increasing trend was found in the incidence rate among men during 2008–2015 (APC: 2.71 %; 95% CI: 0.48%–4.98%; p = 0.026). The mortality rates among the overall patients and women significantly decreased from 2008 to 2017 with the AAPC of −1.8% (95% CI: −3.37%–−0.2%; p = 0.032) and −12.19% (95% CI: −22.32%–−0.75%; p = 0.040), respectively, and the mortality rate among men remained stable (AAPC: −0.1%; 95% CI: −5.85%–5.99%; p = 0.969).Figure 5 and  Supplementary Tables S2–S4 reflect the effects of age, period, and birth cohort on the prevalence, incidence, and mortality of HCC. The prevalence rate of HCC was remarkably increased from age 25, peaked in age 60, and decreased at age 70 and over. The prevalence rate for people aged 70 years was 24.2 times higher than that for people at 25 years old. In the period index, the prevalence rate increased slightly from 2008 to 2017, which in 2017 was twice as many as 2008. In the cohort index, the prevalence rate decreased approximately linearly from 1925 cohort to 1990 cohort, and that for people born in 1920 was 12.7 times as many as in 1990. The tendencies of the age, period, and birth cohort effects on the incidence rate were similar to the prevalence rate. The incidence rate for people aged 70 years was 21.8 times higher than that for people at 25 years old, and it was 14.2 times as many as those born in 1920 for people born in 1990. The incidence rate in 2017 was 1.3 times as many as in 2008. There were smaller age, period, and birth cohort effects on mortality than the prevalence and incidence. The mortality rate of HCC increased from age 40 to 85, which was 12.2 times higher for people at 85 years old than for people aged 40 years. However, the period had no significant effect on the mortality rate. The tendency of the birth cohort’s effect on the mortality rate was similar to incidence and prevalence rate, and the mortality rate for people born in 1920 was 6.3 times as many as that for people born in 1970.This is the first study to reveal the prevalence, incidence, and mortality rates of HCC and the effects of age, period, and birth cohort on them in Northern China, taking Tianjin as an example. In addition, this is also the first study to report the prevalence, incidence, and mortality of HCC in mainland China. This population-based study revealed that the age-standardized incidence rate of HCC in Tianjin ranged between 13.91 and 19.98/100,000 during 2008–2017, which was substantially higher than that in other countries [7,8,9,29]. The age-standardized mortality rate of HCC based on the world standard population among men in Tianjin was much higher than that in Europe and America, such as France, Germany, United Kingdom, United States, and Australia, but lower than Japan, Hong Kong, and Republic of Korea [10]. The age-standardized mortality rate among women in Tianjin was much higher than that in France, Germany, United Kingdom, United States, Australia, Japan, Hong Kong and many of the other countries or areas, but lower than Republic of Korea [10]. The population in this study were employees and retirees derived from the UEBMI database in Tianjin who had stronger awareness of health and had easier access to better medical care than inhabitants who were enrolled in another basic medical insurance (i.e., URRBMI), which may partly explain why the mortality rate in this study was lower than that in Japan, Hong Kong, and the Republic of Korea. In addition, the incidence, and mortality of HCC among men are always much higher than those among women, which could be explained by some environmental exposures with differential risks between the sexes, e.g., heavy alcohol use.Because most of the published studies reported the prevalence, incidence, and mortality rate for liver cancer in China but did not focus on HCC, we compared the time trends on the prevalence, incidence, and mortality of HCC in Tianjin with those of liver cancer in other areas in mainland China over the past decade. The prevalence rate in this study showed a significant increasing trend from 2008 to 2017, which was similar to the trend for liver cancer in China [30]. The incidence rate of HCC in Tianjin was almost unchanged over the past decade, and it was similar to Chongqing [21]. However, there were other studies that reported that the incidence of liver cancer was decreasing in China, Shenzhen city and Taizhou city [17,19,31]. The mortality rate of HCC in this study showed a decreased trend during 2008–2017, and this was similar to the trend of liver cancer in China as well as and Tieling city [12,14,31].The age-period-cohort analysis results suggested that age exhibited a strong association with HCC prevalence, incidence, and mortality in Tianjin. Age is always considered as the main factor in age-period-cohort analysis, because it could represent consistent extrinsic factors, such as the accretion of mutations and accumulative exposures to carcinogens over time, which would increase the risk of developing cancers [32]. As is known to us, the elderly also have a higher risk of death than the young. To be mentioned in this study, the incidence rate increased substantially with age before 50 years old, and then flattened, which meant that the incidence rate of the population aged 50–80 was the highest in all age groups. The range of ages with the highest incidence in this study is wider than that which was reported in the United States [33]. Liu et al. reported that the incidence of HCC increased from ages 25–29 years before peaking at ages 50–54 years and declined thereafter in the United States [33]. The high-incidence age in this study was also much earlier than that in other countries [5,7]. The earlier typical onset age and the larger proportion of elderly HCC patients result in the huge economic burden of HCC in China. In addition, it is conceivable that the number of high-risk individuals in the future will increase with China’s rapidly aging population. Therefore, earlier screening for high-risk individuals could be conducted to prevent HCC.The period effect is usually affected by socioeconomic development and historical events, including public health emergencies (e.g., economic crises, epidemics of infectious diseases), public health interventions (e.g., disease screening, detection levels), lifestyle behaviors in different periods, wars, and other factors. With the accelerating pace of life and work, lifestyle has changed over the past decade. More and more people are suffering from increasing mental stress, staying up late, drinking, obesity, and long-term fatigue, which may do damage to the liver. The impact of unhealthy lifestyles and environmental degradation may be the reason why the prevalence and incidence rate of HCC in 2017 were 1.5–2 times as many as those in 2008. In addition, there was no period effect on the mortality of HCC, which meant that the treatment of HCC was not improved from 2008 to 2017. Because this study only examined the ten-year trend, an analysis based on longer-term data may reflect a more remarkable period effect on the epidemiology of HCC.Cohort effects represent variations of some characteristics over time among groups of individuals who were born in the same year or decade or are defined by other shared experience. The cohort effect obtained by the age-period-cohort model was the net effect related to birth year after deducting the impact of age and period effects. The HCC prevalence, incidence, and mortality for individuals born in recent years were remarkably lower than the earlier cohorts in this study. The probable reasons were the effective control of HBV and HCV infection attributed to the increasingly wide availability of HBV vaccines and the use of new antiviral drugs for HBV and HCV, and the lower infection risk of aflatoxins, as well as the stronger awareness of health and disease prevention among the later birth cohorts than the earlier birth cohorts. There were two fluctuations around 1935 and 1960 which reflected the influence of the Japanese invasion of China and the social and economic system changes on inhabitants’ health. This is similar to the previous study [34].This study was subject to several limitations. First, the cases identified from the claim database are “prevalent-treatment cases”, and the number may be lower than the true number of prevalent cases. However, this underestimation would not affect the analyses of the time trends on the prevalence rates. Second, because the database used in this study was the claims for employees and retirees, the population under the age of 18 was not included, which may underestimate the prevalence and incidence as well as overestimating the mortality. However, a recent study reported that the prevalence was under 1.00/100,000 for individuals under 20 years old [30], which may have a very small effect on the prevalence of the total population. Third, this study was conducted only on the UEBMI database in Tianjin; the study population may not represent patients with different types of insurance or no insurance coverage. However, the diagnosis of HCC is unlikely to differ by insurance status; thus, the results in this study may apply to other cities with similar levels of economy to Tianjin in Northern China. Fourth, this study only reported the trend in the prevalence, incidence, and mortality of HCC from 2008 to 2017, as it was limited by the current database. As Sorafenib was covered by the National Reimbursement Drug List in 2017 and Lenvatinib was approved by the National Medical Products Administration in 2018 in China, a future study focused on longer-term trends could be conducted to examine the impact of new interventions on the mortality of HCC patients. Lastly, as there is no epidemiological study focused on HCC in China, we compared the time trend in the prevalence, incidence, and mortality of HCC with those of liver cancer.In the past decade, HCC incidence has changed a little and mortality has decreased, so the prevalence has shown an increasing trend in Tianjin, China. The strong age and cohort effects and a small period effect on the prevalence, incidence, and mortality were found. The findings of this study could inform the necessity of conducting earlier screening for high-risk individuals and improving the treatment of HCC, which may also help to predict future changes in HCC epidemiology.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116034/s1. Table S1: Goodness of fit statistics for the combination of age, period and cohort factors. Table S2: Results of the age-period-cohort analysis on the prevalence of HCC. Table S3: Results of the age-period-cohort analysis on the incidence of HCC. Table S4: Results of the age-period-cohort analysis on the mortality of HCC.Conceptualization, J.W. and C.L.; Methodology, J.W. and C.L.; Software, J.W. and C.L.; Formal Analysis, J.W. and C.L.; Writing—Original Draft Preparation, J.W. and C.L.; Writing—Review and Editing, J.W., C.L. and Z.C.; Project Administration, J.W.; Final approval of the Manuscript, all authors. All authors have read and agreed to the published version of the manuscript.This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.Because the data used in this study was the de-identified data extracted from claims database which did not involve patient privacy and other ethical issues, the Safety and Ethics Committee of the School of Pharmaceutical Science and Technology in Tianjin University waived the requirement of ethics approval for this study.Not applicable. The data that support the findings of this study were available from the Tianjin Healthcare Security Administration. Due to the requirement from the data owner, these data could only be used for this study under the license, which could not be shared to others.The authors declare no competing interest.Trends of the age-standardized prevalence, incidence, and mortality rates of HCC from 2008 to 2017.Age-specific prevalence, incidence, and mortality rates of HCC from 2008 to 2017.Temporal trends of the prevalence, incidence, and mortality rates of HCC by age group from 2008 to 2017. Note: As there was no apparent trend among individuals aged 75−79, 80−84, and ≥85, these three age groups were not shown to enhance the clarity and readability of this figure.Cohort-specific prevalence, incidence, and mortality rates of HCC by age group.The age, period, and cohort effects on the prevalence, incidence, and mortality of HCC. Note: Some categories of the age and cohort factors are not shown in the figure due to the lack of space. Please refer to Supplementary Table S2 for the detailed statistics.Estimated prevalence, incidence, and mortality of HCC from 2008 to 2017.†ASR—age-standardized rate, and Segi’s world population was used for ASR. ‡ The death cases used to calculate the mortality rate in a particular year were defined as the patients who had claims related to HCC in that year, and they do not include the patients with a diagnosis of HCC in the early year but who did not see a doctor in that year.Results of the joinpoint regression on the age-adjusted prevalence, incidence, and mortality rates of the HCC from 2008 to 2017.* p < 0.05. APC: annual percentage change; AAPC: average annual percentage change; 95% CI: 95% confidence interval.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06035.txt ADDED Viewed

	@@ -0,0 +1,2 @@


1	+ The prevalence of nonalcoholic fatty liver disease (NAFLD) is increasing worldwide. Recent experimental studies suggested that phthalates might induce NAFLD. Therefore, this study aimed to investigate the relationship between phthalates metabolites and NAFLD in the human population. This cross-sectional analysis was performed using data from the Korean National Environmental Health Survey II (2012–2014) among Korean adults (n = 5800). NAFLD was diagnosed using the hepatic steatosis index (HSI) in the absence of other causes of chronic liver diseases. Among the participants (mean age 46 years, 47.5% male), the prevalence of NAFLD was associated with urinary levels of mono(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), mono-(2-ethyl-5-oxohexyl) phthalate, mono-(2-ethyl-5-carboxypentyl) phthalate, mono-benzyl phthalate (MBzP), and mono-n-butyl phthalate (MnBP) compared to the reference group. In the multivariate model, the odds ratios (ORs), 95% confidence interval (CI) for NAFLD were 1.33 (1.00–1.78) and 1.39 (1.00–1.92) in the 3rd and 4th quartile of MEHHP, respectively. Based on the study findings, high levels of urinary phthalates are associated with the prevalence of NAFLD in Korean adults. Further investigation is required to elucidate the causal relationship.Nonalcoholic fatty liver disease (NAFLD) is a condition whereby there is a significant accumulation of fat in the liver (>5% fat content), without excessive consumption of alcohol or other causes of liver disease [1]. The global prevalence of NAFLD is estimated to be approximately 30%, and it is likely to increase [1,2,3]. The presence of metabolic disorders, including obesity, type 2 diabetes, dyslipidemia, and hypertension, is considered to be a risk factor for NAFLD [4,5].While a Westernized diet and sedentary lifestyle may contribute the rapid increase of the NAFLD occurrence [2,6], recent studies suggests that exposure to endocrine disrupting chemicals (EDCs) may also associated with the pathogenesis of NAFLD through insulin resistance (IR) [7,8,9]. Although IR is one of the hallmarks of NAFLD progression, EDCs can directly affect hepatic lipid homeostasis [10,11] because most of EDCs are metabolized in the liver, which is a central organ in energy metabolism.Among the various EDCs, phthalates are the most commonly used plasticizers in industries and in daily life. Phthalates are a group of phthalic acid esters that are widely used to increase the flexibility in various consumer products including medical devices, children’s toys, personal care products, and food wrap made of polyvinyl chloride (PVC). This widespread usage of phthalate has also led to a marked increase in the exposure of the public to these agents [12]. Phthalates entering the body are rapidly hydrolyzed to the monoesters in the liver and gut [13,14,15], and most of them are excreted through the urine and feces [16,17]. Previous studies have focused on the development and reproductive toxicity after phthalate exposure [18,19]. Recent studies have suggested that phthalates can induce obesity, insulin resistance, and metabolic disorders [20,21,22,23,24,25].The occurrence of metabolic diseases after phthalate exposure might be linked to the activation of peroxisome proliferator-activated receptors (PPARs) [25,26], which are modulated by the sterol regulatory element binding proteins (SREBPs) [27]. In experimental studies, the disruption of gene expression related to fatty acid metabolism was suggested as a plausible mechanism in the development of NAFLD after phthalate exposure [28,29,30,31,32,33,34]. However, there is a lack of clinical evidence to support the phthalates-NAFLD relationships.Therefore, this study aimed to investigate whether urinary phthalate metabolites are associated with the development of NAFLD in the population.This study was based on cross-sectional data obtained from the Korean National Environmental Health Survey (KoNEHS) II (2012–2014). KoNEHS is conducted every 3 years, and this is to measure the human exposure level of environmental chemicals, examine influential factors, and continuously investigate the factors of spatiotemporal distribution and changes [31]. It is a data set that includes national, multistage, stratified, and clustered probability sampling designs, so as to develop representative samples of the South Korean population.The 6478 subjects in the KoNEHS II were aged 19 years and older (2274 men and 3704 women). They were interviewed with questionnaires on demographic characteristics, socioeconomic status, indoor/outdoor environment, lifestyle factors, and history of disease experience. Biological samples (urine and blood) were collected for clinical analysis and to measure the level of environmental chemicals. The published data of questionnaire, clinical tests, and urinary phthalate metabolite levels were used in this study.Based on the data, 678 participants were excluded due to a lack of data of phthalate metabolites concentrations (n = 247), no data of alanine aminotransferase (ALT) or aspartate aminotransferase (AST) levels (n = 21), significant alcohol consumption (n = 262, including males who consumed alcohol more than 3 times per week and 7–9 cups per time (n = 238) and females who consumed alcohol more than 3 times per week and 5–6 cups per time (n = 24)), pregnant women (n = 29), those with hepatitis or hepatic disease (n = 59), and those with an AST/ALT ratio exceeding 2 (n = 60). Finally, 5800 participants (2355 men and 3445 women) were included in this analysis (Figure 1).Data on the participants’ general characteristics, including age, sex, drinking and smoking status, physical activity, socioeconomic status, and education level, were obtained via interviews using questionnaires. These characteristics were categorized based on their responses as follows: education (less than high school, high school, and college and more), drinking and smoking (never, past, present), physical activity levels (no, moderate, yes), socioeconomic status (high, mid-high, mid-low, low), and marital status (single, married, and divorced/separated).Participants who were diagnosed with hepatitis or fatty liver disease and who were currently undergoing treatment or taking medication were considered to have hepatic disease. Hypertension was defined as a self-reported history of hypertension or the use of antihypertensive drugs. Diabetes mellitus (DM) was defined as a self-reported history of DM or the use of antidiabetic drugs. Hyperlipidemia was defined as a self-reported history of hyperlipidemia, use of anti-hyperlipidemia drugs, triglyceride (TG) ≥240 mg/dL, or high-density lipoprotein cholesterol ≤40 mg/dL. Data on serum TG, ALT, and AST levels were obtained. Body mass index (BMI) was calculated by dividing body weight (kg) by height squared (m2).The hepatic steatosis index (HSI) was used as a screening tool for NAFLD [35,36,37]. The variables in the HSI formula were levels of ALT, AST, BMI, sex, and presence of DM. Subjects were categorized into two groups based on the HSI score: ≤36 was defined as non-NAFLD and >36 was defined as NAFLD.
2	+ HSI=8×ALTASTratio+BMI +2, if diabetes;+2, if femaleThe weighted mean or frequency and standard errors were provided. Comparisons between groups were performed using the t-test (continuous variable) or χ2 test (categorical variables). The concentrations of phthalate metabolites were categorized into quartiles based on the weighted sample distribution. The lowest quartile group was considered as the reference. Because the levels of phthalate metabolites were skewed to the right, log transformation was required. Multivariate logistic regression analysis was performed to predict the relationship between phthalate levels and NAFLD after considering potential demographic and clinical variables. The included demographic covariates were age, sex, smoking, drinking, exercise level, marital status, education level, and socioeconomic status. Hypertension, DM, and hyperlipidemia were included as clinical variables. In logistic regression, BMI was not used as an independent variable. The reason for this decision was that BMI is used as part of the HSI calculation formula and to avoid collinearity between BMI and HSI. However, in the HSI formula, women’s +2 is intended to compensate for women’s lower BMI compared to men. For DM, +2 was used for the same reason, so in logistic regression, sex and DM were used as independent variables. Data analyses were performed using STATA (version 15.0 StataCorp LP College Station, TX, USA). p-values < 0.05 were considered significant.The participants were categorized according to their HSI score (Table 1), with 4405 and 1395 in the non-NAFLD and NAFLD groups, respectively. The proportion of gender was not statistically different between the non-NALFD and NAFLD groups (p = 0.061). The mean age and BMI were significantly higher in the NAFLD group than in the non-NAFLD group (p = 0.005 and p < 0.001, respectively). The statuses of alcohol consumption were not significantly different between the two groups (p = 0.054); however, the status of smoking and regular exercise were significantly different (both p = 0.013). Socioeconomic status, education level, and marital status also showed significant differences (p = 0.040, p < 0.001, and p < 0.001, respectively). The proportions of participants with hypertension, DM, and hyperlipidemia were significantly higher in the NAFLD group than in the non-NAFLD group (all p < 0.001).The distribution of urinary phthalate metabolites in the participants is shown in Table 2. Unadjusted urinary mono-(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP) and mono-(2-ethyl-5-carboxypentyl) phthalate (MECPP) levels in the NAFLD group was significantly higher than those in the non-NAFLD group (both p < 0.001). The urinary levels of mono-(2-ethyl-5-oxohexyl) phthalate (MEOHP) was significantly higher in the NAFLD group compared to the non-NAFLD group (p = 0.031). Mono-n-butyl phthalate (MnBP), and mono-benzyl phthalate (MBzP) in urine did not differ between the groups (p = 1.000 and p = 0.085, respectively).The multivariate odds ratios (ORs) and 95% confidence intervals (CIs) for the HSI score according to urinary phthalate levels are shown in Table 3. The lowest level of each phthalate (first quartile) was considered as the reference value.Urinary MEHHP and MECPP in 3rd and 4th quartiles showed higher ORs than those in the lowest quartile, in crude analysis. After adjusting for age, sex, and creatinine level (Model 1), adjusted ORs (95% CI) of MEHHP in 3rd and 4th quartiles were 1.40 (1.08–1.81) and 1.43 (1.04–1.95), compared with those in the lowest quartile.Additionally, adjusting for smoking status, drinking status, regular exercise, marital status, education level, and socioeconomic status was performed (Model 2). The adjusted ORs of MEHHP in the 3rd and 4th quartiles were significantly higher than those in the lowest quartile (1.34 [95% CI 1.02–1.76] and 1.40 [95% CI 1.02–1.93], respectively).When further adjusted for hypertension, DM, hyperlipidemia, and urinary bisphenol A levels (Model 3), adjusted ORs (95% CI) were 1.33 (1.00–1.77) and 1.39 (1.00–1.92) in the 3rd and 4th quartile of MEHHP, compared with those in the lowest quartile.This study was performed to investigate the association between urinary levels of phthalate metabolites and NAFLD in the population. The prevalence of NAFLD based on the HSI score was associated with higher levels of urinary MEHHP and MECPP in univariate analyses. After we adjusted the covariates, the 3rd and 4th quartiles of MEHHP showed significantly higher ORs compared to the lowest levels. The risk of developing NAFLD significantly increased as the quartiles of MEHHP increased in univariate and multivariate analyses.Humans can be exposed to phthalates through ingestion (e.g., phthalate-contaminated food and water), dermal absorption (e.g., cosmetics and other personal care products), and inhalation (e.g., nail polish, hair spray, and other phthalate-containing products) [12]. Phthalates entering the body are rapidly hydrolyzed to monoesters and then metabolized in the liver and the gut [13,14,15] and excreted through the urine, with half-lives of less than 24 hours [16,17]. Phthalate monoesters represent the major human metabolites; however, the monoester of di-(2-ethylhexyl) phthalate (DEHP) is further metabolized and produced the secondary oxidized metabolites [14,16,17]. Thus, urine samples are commonly used to measure phthalate exposure.Phthalates were classified as obesogen based on the previous animal studies on obesity and metabolic derangement [25,38,39]. Although NAFLD is closely associated with obesity [4], it also occurs in non-obese people [40,41]. That is, excess accumulation of visceral fats, which is related to IR, play a key role in NAFLD pathogenesis [41,42]. In cross-sectional studies, urinary phthalate metabolites showed an association with BMI, waist circumstances, and IR [20,21,23]. In addition, urinary MEHHP levels, which is the secondary oxidized metabolite of DEHP, showed a positive association with obesity and IR in prepubertal girls, but not with that of MEOHP and MECPP [43]. In human liver cell line, DEHP reduced insulin receptor levels and glucose oxidation, and subsequently increased IR [44]. Also, DEHP treated male rats experienced reduced insulin levels in serum and liver glycogen, and increased thyroid hormone levels in serum [22]. Because obesity and IR can contribute to the development of NAFLD [7,8,9], we hypothesized that phthalates could increase the risk of developing liver disease.In experimental studies, exposure to DEHP, the most widely used phthalate, along with high fat or oleic acid in diets increased lipid content and inflammation in rat liver [28,33] and in a hepatocellular carcinoma cell line (HepG2 cells) [29]. The increase in hepatic lipid accumulation after phthalate exposure might be due to the up-regulation of genes related to hepatic lipid metabolism such as SREBPs and PPARs [28,29,33]. In transcriptome analysis, supporting evidence was identified for the association between exposure to a low dose of DEHP and the disruption of genes related to hepatic fatty acid metabolism in the zebrafish [30]. In addition, mono-2-ethylhexyl phthalate (MEHP), which is the primary metabolite of DEHP, increased the TG content and modulated the gene related to fatty acid metabolism in HepG2 cells [32]. In particular, perinatal exposure to DEHP induced hepatic TG content in adult male pups through up-regulation of diacylglycerol acyltransferase 1, regardless of obesity [34]. This study indicated that DEHP alone could influence re-esterification and increase small lipid droplets in the liver instead of activation of de novo lipogenesis. Although the experimental studies suggested that phthalates could accelerate lipogenesis and thereby contribute to the development of NAFLD, further study is needed to clarify the relationship between DEHP and NAFLD pathogenesis.In addition, because thyroid hormones are important to regulate the glucose and lipid metabolism [45], the alteration of thyroid hormone levels due to phthalates may increase the risk of NAFLD. Previously, DEHP treated rats showed the decrease of thyroid hormones, sodium iodide symporter, and thyroid peroxidase levels [22,46]. It seems that DEHP can disrupt the thyroid hormone synthesis, transport, and metabolism. The negative association between phthalate metabolites levels and thyroid hormones also showed in adult men [47]. A previous meta-analysis study showed the association between hypothyroidism and NAFLD [48]. However, because the mechanism whereby phthalates induced NAFLD is associated with thyroid dysfunction is unclear, further study is needed. To our knowledge, this is the first study to report the association between urinary phthalate levels and prevalence of NAFLD in the general population. However, some limitations should be noted. First, we could not determine whether a noninvasive marker could truly indicate the prevalence of NAFLD. Liver biopsy is the most sensitive toll for clinical diagnosis of NAFLD. Nevertheless, this is not suitable for population-based studies. Among the various noninvasive tools for predicting NAFLD, we used the HSI score. The area under the receiver-operating curve of HSI was 0.812 (95% CI 0.801–0.824) in the Korean population [35]. Because HSI correlated with IR [49], it could estimate the phthalates induced IR, which is a primary factor of NAFLD. However, the problem of classifying the intermediate group (30 ≤ HIS ≤ 36) as a control group was an inevitable one. Further prospective studies are required to confirm the reproducibility and accuracy of our observations. Second, this study used cross-sectional survey data; thus, it was difficult to assess the causal relationship between phthalate exposure and NAFLD. Further study is warranted to investigate the influence of phthalates on NAFLD development. Third, KoNEHS represented the drinking times in the last month and the number of glass per times instead of the amount of alcohol (g/day). Heavy drinkers who consumed alcohol more than 3 times in a week and 7–9 glasses per time in men (5–6 glasses per time in women) were defined by referring to a previous study [50]. Last but not least, the inter- and intra-subject variations of urinary phthalate levels were a limitation because of the relatively short half-lives. Although KoNEHS collected single spot urine samples, a sufficient number of samples may adequately reflect the average exposure level of the population to phthalate.This study showed epidemiological evidence that exposure to phthalates is associated with the occurrence of NAFLD. Therefore, a reduction in phthalate exposure might help prevent NAFLD.Conceptualization, Y.-J.Y.; methodology, Y.-J.Y. and T.K.; formal analysis, Y.-J.Y. and T.K.; writing—original draft preparation, Y.-J.Y.; writing—review and editing, Y.-P.H. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Ethical review and approval were waived for this study, due to the use of existing information/data, documents, and records.Not applicable.This study used data from the Second Korean National Environmental Health Survey (KoNEHS) which was conducted by Ministry of Environment, National Institute of Environmental Research. The data presented in this study are available on request from the corresponding author. The data are not publicly available due to protect personal information.This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2021R1I1A3046386).The authors declare no conflict of interest.Study population in the present study obtained from the Korean National Environmental Health Survey II (2012–2014). AST: aspartate aminotransferase; ALT: alanine aminotransferase.General characteristics of study participants according to the hepatic steatosis index score.Data are shown as the weighted mean or frequency ± standard error as appropriate. NAFLD: non-alcoholic fatty liver disease; BMI: Body mass index.Urinary phthalates levels according to hepatic steatosis index score.GE: geometric mean; GSE: geometric standard error; NAFLD: non-alcoholic fatty liver disease; MEHHP: mono (2-ehtyl-5-hydroxyhexyl) phthalate; MEOHP: mono (2-ethyl-5-oxohexyl) phthalate; MECPP: mono (2-ethyl-5-carboxypentyl) phthalate; MnBP: mono-n-butyl phthalate; MBzP: mono-benzyl phthalate.The association between hepatic steatosis index score and urinary phthalates levels (ug/L).OR: odds ratio; CI: confidence interval; MEHHP: mono (2-ehtyl-5-hydroxyhexyl) phthalate; MEOHP: mono (2-ethyl-5-oxohexyl) phthalate; MECPP: mono (2-ethyl-5-carboxypentyl) phthalate; MnBP: mono-n-butyl phthalate; MBzP: mono-benzyl phthalate. *: p values were shown the test of trend of odds. Crude: hepatic steatosis index, each type of phthalate metabolites. Model 1: Crude + age, sex, creatinine. Model 2: Model 1 + smoking, drinking, exercise, marital status, education, socioeconomic status. Model 3: Model 2 + hypertension, diabetes mellitus, hyperlipidemia.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06036.txt ADDED Viewed

	@@ -0,0 +1 @@

+ (1) Background: Sexual violence (SV) has become common in universities for reasons related to unwanted social/peer pressures regarding alcohol/drug use and sexual activities. Objectives: To identify perceptions of SV and alcohol use and estimate prevalence among nursing students in Catalonia, Spain. (2) Methods: Observational descriptive cross-sectional study of a convenience sample of nursing students attending public universities. (3) Results: We recruited 686 students (86.11% women), who reported as follows: 68.7% had consumed alcohol, 65.6% had been drunk at least once in the previous year, 62.65% had experienced blackouts and 25.55% had felt pressured to consume alcohol. Drunkenness and blackouts were related (p < 0.000). Of the 15.6% of respondents who had experienced SV, 47.7% experienced SV while under the influence of alcohol and were insufficiently alert to stop what was happening, while 3.06% reported rape. SV was more likely to be experienced by women (OR: 2.770; CI 95%: 1.229–6.242; p = 0.014), individuals reporting a drunk episode in the previous year (OR: 2.839; 95% CI: 1.551–5.197; p = 0.001) and individuals pressured to consume alcohol (OR: 2.091; 95% CI: 1.332–3.281; p = 0.001). (4) Conclusions: Nursing instructors need to raise student awareness of both the effects of alcohol use and SV, so as to equip these future health professionals with the knowledge and skills necessary to deal with SV among young people.Sexual violence (SV) is one of the most frequent forms of gender violence and a major public health problem worldwide [1]. The World Health Organization (WHO) defines SV as “any sexual act, attempt to obtain a sexual act, or other act directed against a person’s sexuality using coercion, by any person regardless of their relationship to the victim, in any setting. It includes rape, defined as the physically forced or otherwise coerced penetration of the vulva or anus with a penis, other body part or object” [2]. This definition of SV includes sexual harassment, abuse and assault. While most of these acts are directed by men against women, men also experience SV [2,3].According to WHO estimates [2], 1 in 3 women aged 15–45 years worldwide has experienced some form of sexual or physical violence in their lives. More specifically, between 0.3% and 12% of women have reported an experience of non-partner SV from the age of 15 years [4]. SV-related incidents are difficult to quantify because many cases are not reported and no fully reliable record-keeping system exists [5,6]. A Latin American study estimates that only 5% of adult victims of SV notify the police [7]. SV incidence is also higher in low-income countries in Africa and Latin America, although several studies conducted in the USA, a high-income country, have reported widespread sexual harassment, abuse and aggression in universities [8]. A study of SV for the period 1995–2013 by the US Department of Justice [9] concluded that women in the 18–24 age bracket (the typical age of attendance at university) were at greater risk of experiencing some form of SV than women in other age brackets. Alarming rates of SV among university students were documented for the first time 30 years ago and trends have not improved; indeed, the US Department of Justice reported in 2015 that incidence had increased by 10% over the previous decade [10,11].Alcohol and drug use is frequent in university settings, where SV typically occurs for reasons related to social coercion, i.e., the feeling that to be socially accepted one has to give in to peer pressure to consume alcohol/drugs and engage in sexual activities, even if these make a person feel uncomfortable [12]. A study conducted at a university in the US northwest reported that 15% of first-year women were victims of rape when incapacitated by alcohol/drugs [13]; the explanation is that parties in universities take place in permissive environments where alcohol/drugs are freely available and affordable [14]. In relation to alcohol, several studies have linked risk behaviours and sexual relations in leisure settings with its consumption, as alcohol is often used to increase the chances of a sexual encounter or to modify a person’s behaviour or will, so they become more amenable to sex [15,16]. Alcohol intake is the most important risk factor for non-consensual sexual relations and forced touching [17,18], although its consumption affects men and women differently: in women, the capacity to react to alarm signals is reduced, i.e., protective behavioural strategies are undermined [19], whereas in men, impulses are disinhibited and aggressiveness increases [20]. Alcohol also heightens sexual arousal and desire, in turn increasing interest in sex and fostering other risky behaviours [21,22,23]. Drug-facilitated sexual assault (DFSA) is a term used to describe sexual intercourse with someone who is partially or fully incapacitated (and so unable to grant consent) due to the effects of the involuntary consumption of alcohol/drugs [24]. While different substances are associated with DFSA, ethyl alcohol is most commonly used [25]. Around three quarters (72%) of sexual assaults occur when the victim—usually a woman—is intoxicated and so unable to control or stop what is happening [26,27].While evidence reported in numerous US and Canadian studies correlates alcohol intake with an increase in SV in leisure settings, few studies on this topic have been conducted among university students in Spain or, more specifically, among nursing students.In Spain in 2016, there were 18.79 cases of SV per 100,000 inhabitants: 2.69 rapes and 16.10 cases of sexual abuse [28]. In the Spanish autonomous region of Catalonia in 2016, 20.9% of cases of SV against women by men who were not their partners took place in leisure contexts [29]. In Spain, alcohol is intimately associated with leisure among most young people. Alcohol is reported to be consumed regularly by young people aged 20–24 years [30] and, in 2017, 75.6% of students aged 14–18 years said they had consumed alcohol in the previous year [31].The objectives of this study were, for a sample of nursing students in Spain, to identify perceptions of SV and alcohol use, to estimate SV prevalence in leisure settings and to analyse possible associations between SV, drunk episodes, blackouts and being pressured to consume alcohol.This cross-sectional descriptive non-randomized study, based on a self-administered questionnaire, was carried out between September 2017 and June 2018.The study population was 2471 nursing students attending four public universities in Catalonia (Autonomous University of Barcelona, University of Barcelona, University of Girona and Rovira i Virgili University in Tarragona).The sample, taking into account the 20% prevalence rate for sexual abuse and assault reported in other studies conducted in Spain [32,33,34], was calculated on the basis of a 95% confidence level, a 3% precision level and a design effect of 1, resulting in an estimated sample size of n = 626. Inclusion criteria were to be present in the classroom on the day of data collection and to complete the questionnaire. A total of 686 participants were recruited.The survey was based on an ad hoc questionnaire in two parts. The first part collected (a) demographic data, namely, age, gender (respondents self-identified as a man or woman) and academic year (first to fourth year) and (b) data on alcohol use (“Do you consume alcohol in leisure settings?”), drunk episodes in the previous year (“In the last year, how often have you been drunk?”), blackouts in the previous year (“Have you had a blackout—failed to remember what happened during several hours—in the last year?”) and pressures to consume alcohol (“Have you ever felt under pressure to take alcohol?”). The second part reflected reported experiences of sexual abuse and sexual assault. The questionnaire was based on the Sexual Experiences Survey-Short Form Victimization (SES-SFV) instrument developed by Koss et al. [35,36], which addresses four different experiences of SV as follows: (1) “Someone fondled, kissed, or rubbed up against the private areas of my body (lips, breast/chest, crotch or butt) or removed some of my clothes without my consent (but did not attempt sexual penetration)”; (2) “Someone had oral sex with me or made me have oral sex with them without my consent”; (3) “A man put his penis into my vagina, or someone inserted fingers or objects without my consent”; (4) “A man put his penis into my butt, or someone inserted fingers or objects without my consent”.Students were also asked to indicate the frequency of SV experiences since starting university, whether their experiences had occurred in a leisure context, whether they had been under the influence of alcohol to the point that they were insufficiently aware to stop what was happening and whether, even though nothing happened, someone had attempted something. Also explored was if they or someone they knew had been raped.Nursing faculties were contacted to agree a day and time to administer the questionnaires to first-, second-, third- and fourth-year nursing students in situ in their classrooms.Researchers visited the classrooms to inform the students of the objectives and purposes of the study, to recruit volunteers and to administer the questionnaire. Female researchers informed the participants of the importance of the study, that the anonymity of their responses was guaranteed and that they could at any time notify their withdrawal. Given the sensitivity of the topic, participants were asked to be sincere in their responses. The questionnaire was administered in electronic format using a QR code, while a paper format was also available to participants in case of technical problems or for participants who felt this format better assured their confidentiality. The estimated time required to complete the questionnaire was 10 min.The study was approved by the ethics committee of the University of Barcelona. Permission to carry out the study was obtained from nursing faculty deans and teachers were duly informed. Students who participated in the study were informed about the study, were guaranteed data confidentiality and anonymization in accordance with Spanish legislation on personal data protection [37] and gave their signed informed consent.A univariate descriptive analysis was conducted of all variables; continuous variables were described in terms of mean and standard deviation (SD) and median and interquartile range (IQR), while categorical variables were reported as percentages. In a bivariate analysis, associations between categorical variables were tested using Pearson’s chi-square test or Fisher’s test, while logistic regression analysis was performed with variables that proved significant in the bivariate analysis. Results were considered statistically significant for p < 0.05 and a confidence interval (CI) of 95%. Statistical analyses were performed using SPSS for Windows, version 23 (SPSS, Chicago, IL, USA).The sample was composed of n = 686 nursing students, 86.11% (n = 591) of whom self-reported as women and 13.9% (n = 95) as men. Mean (SD) age was 21.36 (4.14) years and median (IQR) age was 20 (3). By academic year, first-, second-, third- and fourth-year students accounted for 31.9%, 28.75%, 23.89% and 17.36% of the sample, respectively.Most students (84.58%; n = 576) consumed alcohol always, almost always or sometimes when they went out at night, while 15.42% (n = 105) never or almost never consumed alcohol. No differences were found by gender or academic year (p = 0.861 and p = 0.102, respectively).Two thirds of the respondents (65.6%; n = 448) reported having experienced drunk episodes at least once in the previous year, while 12% (n = 82) reported having this experience more than 10 times. Differences by gender and academic year were not statistically significant (p = 0.114 and p = 0.192, respectively).Around a quarter of the respondents (25.55%; n = 174) reported feeling pressured to consume alcohol in leisure settings. There were no statistically significant differences by gender (p = 0.817) or academic year (p = 0.151).Of the respondents pressured to consume alcohol, 66.09% reported having experienced a drunk episode at least once in the previous year. No statistically significant differences were observed between these students and those who had not experienced pressures (Table 1).Almost two thirds of the respondents (62.65%; n = 425) declared having experienced a blackout, with no significant differences by gender (p = 0.589). By academic year there were statistically significant differences in blackout experiences, with increased percentages for second-, third- and fourth-year students (p < 0.000) (Table 2).Of the respondents who reported a drunk episode in the previous year, 78.03% (n = 348) also declared having experienced a blackout, compared to 21.97% (n = 98) who reported drunk episodes but no blackouts (p = 0.000).Of the 686 nursing students in our sample (Table 3), 15.6% (n = 107) stated that they had experienced some type of SV (women 93.5% (n = 100) versus men 6.5% (n = 7); p = 0.017). Of the students who had experienced some type of SV, 47.7% (n = 51) experienced SV while under the influence of alcohol and so were not sufficiently alert to stop what was happening (women 92.2% (n = 47) versus men 7.8% (n = 4); p = 0.603). Most of those experiences, therefore, primarily affected women and most (72.9%; n = 78) occurred in a leisure context.A total of 21 (3.06%) respondents—2.2% of the women (n = 19) and 3.3% of the men (n = 2)—confirmed that they had been raped, while 104 (15.16%) respondents—15.6% of the women (n = 90) and 14.9% of the men (n = 14 men)—reported that they knew someone who had been raped. Results by gender and academic year were not statistically significant (p = 0.556 and p = 0.640, respectively).Of the respondents who said they had been raped, equal proportions said they had experienced drunk episodes or blackouts in the previous year (87.71%) and over half (51.14%) had been pressured to consume alcohol; differences in relation to respondents who had not experienced rape were statistically significant (Table 4).A logistic regression analysis adjusted by gender and specific variables as identified in the bivariate analysis showed that the following were more likely to experience SV in SES-SFV terms: women who reported a drunk episode or a blackout in the previous year or who had been pressured to consume alcohol (Table 5).Krebs et al. [38] reported a relationship between leisure, alcohol use and SV. In our study of university nursing students aimed at investigating this possible relationship, we found that women were more likely to have experienced SV in leisure settings if they were pressured to consume alcohol and if they reported drunk and blackout episodes in the previous year.Corroborating a similar finding reported for the French university population [39], 68.57% of the nursing students in our sample said they consumed alcohol in leisure settings (with similar distributions by gender and academic year). Other European studies have reported lower levels of alcohol consumption; in the UK, high-risk alcohol consumption was observed in 55.6% of nursing students [40], while in Germany, approximately 60% of university students reported alcohol consumption at least once a week, with higher consumption in men than in women [41]. A study conducted in Chile reported higher alcohol consumption by fourth-year students [42]. The higher alcohol consumption reported in our study may be due to cultural differences; teenagers in Spain begin to drink alcohol between 14 and 18 years of age, for an incidence rate in 2016 of 47.7% [31].Detecting cases of DFSA is frequently difficult due to the lack of evidence and victim recall problems. One alert, however, is blackout [43,44]. Of the students in our sample, 62.65% reported having experienced a blackout—a percentage higher than reported by other authors—while 78.30% of those who reported a drunk episode in the previous year declared having experienced a blackout. One study found that although women consumed less alcohol, those who reported recent blackouts were more likely to have had to deal with unsolicited sexual advances [45].In terms of the situations covered by the SES-SFV instrument, 15.6% of our respondents said they had experienced an episode of SV. Higher percentages of SV have been reported by other studies. Untied et al. [46] reported that 37% of women reported a history of unwanted sexual contact, attempted rape or sexual coercion and that 8.4% reported rape. In our study, 3.06% of the students reported rape, while 3.4% and 1.46% reported that “A man put his penis into my vagina, or someone inserted fingers or objects without my consent” and that “A man put his penis into my butt, or someone inserted fingers or objects without my consent”, respectively. Other authors report rates of 22.7% [47] and 28% [48] for unwanted sexual experiences at university. Carey et al. [13] reported that 15.4% of the first-year university students in their study reported ‘incapacitated rape’ (their term for DFSA). In our study, 7.4% of the students reported having experienced some form of SV while under the influence of alcohol and being insufficiently alert to stop what was happening. As noted by other authors, there appears to be a link between drunk episodes/blackouts and the probability of experiencing SV [48]. Our study reflects that same pattern: women who were drunk, experienced blackouts and felt pressured to consume alcohols were more likely to report SV.In our study, the behavioural reasons for the consumption of alcohol, which happens to be easily accessed and affordable in Spain, have not been addressed. While several studies have found that pluralistic ignorance influences drinking behaviours [49], further research is needed into student opinions and motives regarding alcohol consumption, so as to investigate beliefs in relation to the effects of alcohol and sexual behaviours and develop and implement suitable interventions. Providing information on peer behaviours and attitudes has been reported to reduce alcohol consumption and foster more responsible behaviours [50], so such a strategy might, in the case of our students, be useful in fostering healthy and responsible sexual behaviours.The limitations of this study include the fact that the truthfulness of responses to the survey cannot be assured, that clear cut cause-effect relationships cannot be established given the cross-sectional nature of the study and that our use of a convenience sample means that results cannot be extrapolated to the general student population. Our study, furthermore, did not identify lesbian, gay, bisexual, transgender and intersex (LGBTI) respondents (students could only choose to self-report as a man or woman). This issue needs to be addressed in future studies for Spain, as has been done for the USA. One study of 27 university campuses by the Association of American Universities [18] reported that while 23.1% of women and 5.4% of men had experienced some form of SV, more LGBTI individuals (21%) reported having experienced SV than non-LGBTI individuals (18%).Although we report lower prevalence of SV compared to other studies, the number of SV cases reported gives cause for concern, especially in relation to pressures exerted regarding alcohol consumption. Nursing students—our health professionals of the future—need to be fully informed regarding SV and the impact of alcohol on SV. Indeed, academic curricula in general need to include training on the effects of alcohol and other drugs and actions aimed at preventing SV, while social and political leaders need to tackle this dual public health problem by implementing actions aimed at fostering health.The consumption of alcohol, the association with blackouts and the relationship between alcohol consumption, leisure settings and SV constitute sufficient reason to implement changes in nurse training, to ensure that future nurses are equipped with the necessary knowledge and skills to become agents of societal change regarding attitudes and beliefs about SV among young people.Conceptualization, C.F.-P., D.B.-F. and D.R.-M.; data curation, A.B.-S. and Z.R.-A.; funding acquisition, D.R.-M.; investigation, C.F.-P., D.B.-F., D.R.-M., A.B.-S., Z.R.-A., E.M.-G., A.R.-C., C.R.-H., P.G.-E. and M.D.B.-M.; methodology, C.F.-P., D.R.-M. and D.B.-F.; project administration, C.F.-P.; supervision, D.B.-F. and C.F.-P.; visualization, C.F.-P. and D.R.-M.; writing—original draft, C.F.-P. All authors have read and agreed to the published version of the manuscript.This research was funded by a research grant from the University of Barcelona Nursing School (reference PREI009-I).The study was approved by the Bioethics Committee of the University of Barcelona, Spain (reference IRB00003099). The research and authors complied with Good Clinical Practice guidelines, the Code of Good Research Practice and the Helsinki Declaration of the World Medical Association on Ethical Principles for Research Involving Human Subjects. Informed consent was obtained from participants and Spanish Organic Law 3/2018 guaranteeing personal data protection and digital rights was respected.Informed consent was obtained prior to the start of the survey from all subjects involved in the study.Data availability is currently restricted due to ongoing data collection and to maintain participant privacy. The authors wish to thank all the students who participated in this study, Josep Garre-Olmo for statistical assistance and Ailish Maher for translation and revision of the manuscript.The authors declare no conflict of interest. The funders had no role in the design of the study, in the collection, analyses, or interpretation of data, in the writing of the manuscript or in the decision to publish the results.Drunk episodes and pressures to consume alcohol (n = 678).* Chi-square test: significance p < 0.05.Blackouts by academic year (n = 668).Chi-square test: significance p < 0.05.Prevalence of sexual violence by gender (n = 686).* Respondents may have had the experience more than once. ** % with respect to the total of affirmative answers. *** Chi square test: significance p < 0.05.Rape associated with drunk episodes, blackouts and pressures to consume (n = 668).* Chi square test: significance p < 0.05.Regression analysis of sexual violence: gender, drunk episodes, blackout episodes and pressures to consume alcohol.* Gender: man 1, woman 2. B (beta coefficient); CI (confidence interval); df (degrees of freedom); Exp(B) (exponentiation of the B coefficient, odds ratio value); SE (standard error); sig (statistical significance).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06037.txt ADDED Viewed

	@@ -0,0 +1,47 @@

+Type 2 diabetes (T2DM) prevalence is three times higher among West African Immigrants compared to the general population in the UK. The challenges of managing T2DM among this group have resulted in complications. Reports have highlighted the impact of migration on the health of the immigrant population, and this has contributed to the need to understand the influence of living in West Africa, and getting diagnosed with T2DM, in the management of their condition in the UK. Using a qualitative constructivist grounded theory approach, thirty-four West African immigrants living in the UK were recruited for this study. All participants were interviewed using Semi-structured interviews. After coding transcripts, concepts emerged including noticing symptoms, delayed diagnosis, affordability of health services, beliefs about health, feelings at diagnosis, and emotions experienced at diagnosis all contribute to finding out about diagnosis T2DM. These factors were linked to living in West Africa, among participants, and played significant roles in managing T2DM in the UK. These concepts were discussed under finding out as the overarching concept. Findings from this study highlight important aspects of T2DM diagnosis and how lived experiences, of living in West Africa and the UK, contribute to managing T2DM among West African immigrants. The findings of this study can be valuable for healthcare services supporting West African immigrants living in the UK.Type 2 Diabetes Mellitus (T2DM) is a public health concern globally. T2DM is a major public health challenge in the UK due to the increasing prevalence of the condition in the country. Specifically, there are more than 4 million people living with T2DM, an increase from 2% in 1999 to over 7% in 2019 [1,2]. The National Health Services (NHS) spend more than 10% of its budget on managing T2DM annually [3,4]. However, immigrant population are disproportionately affected by this condition [5]. Immigrant populations have been reported to be disproportionately affected by T2DM prevalence with more than 3 times higher risk among African immigrants, particularly among the West African population, compared to the white population [6]. In addition, West African immigrants are reported to have poorer T2DM management when compared to the general population in the country [7]. The increasing challenges of managing T2DM are impacted by delays in the diagnosis of the condition. T2DM onset has been reported to occur 4–7 years before the condition is clinically diagnosed [8]. However, immigrant populations present even later for T2DM diagnosis [9]. Delayed diagnosis contributes to the complications reported in living with T2DM, which has resulted in poorer management of T2DM reported [10]. To begin, studies have reported higher rates of T2DM complications, including kidney failure, amputations, blindness, and retinopathy among Africans in Africa because of delayed diagnosis [11,12]. These challenges can be associated with cultural and environmental factors that impact accessing healthcare services [13]. This study is important as African immigrants, particularly West African immigrants, are among the fastest-growing immigrant populations in the UK with over 1.3 million (17%) living in the UK [14,15].To understand the challenges faced by West African immigrants, in managing T2DM in the UK, it is important that research explores the diagnosis process of the population. This study focused on first-generation West African immigrants who have experienced living in the Africa region.Studies have reported on the experiences of West African immigrants in the UK [16]. However, no study has been conducted among this population to understand their experiences before migration to the UK and how these experiences impact their current T2DM management in the UK. It is important to understand this aspect of their lives, as exploring their management in isolation of their prior experience may not give a holistic picture of their experiences. In addition, understanding these individuals’ experiences in receiving T2DM diagnosis can impact the support offered in managing their condition.Therefore, the study aims to explore West African immigrants’ knowledge and perception of being diagnosed with T2DM, and how these impact managing the T2DM condition, in the UK.This study is part of a research project on West African immigrants living with T2DM in the UK with the purpose of understanding their experiences of living with T2DM. The current challenges experienced in living with T2DM among this population have been published [17]. However, it was important to explore aspects beyond the current management of their condition to their prior experiences that influence the management of T2DM currently in the UK. This project was approved by Bournemouth University ethical committee, UK (Approval number 13441). All participants consented to participate in this study.The process of participant recruitment has been described in a previous publication [17]. Briefly, we recruited West African immigrants, living in the UK, who have been diagnosed with T2DM. Majority of individuals recruited were diagnosed with T2DM in West Africa (see Appendix A). Qualitative data was collected from participants to help understand the influence of their experiences of living in West Africa and how this might contribute to their current approach to managing T2DM in the UK. Participants were recruited through support groups and within communities (see Table 1). Of the 50 participants approached to participate in the study, 34 accepted while 16 declined. From those that declined, reasons offered include confidentiality concerns, workload, and preference of a quicker research approach, such as a questionnaire.The first author (FA) conducted data collection using semi-structured interviews. The questions asked focused on the process of diagnosis, these include: Where were you diagnosed, how was the experience of being diagnosed with T2DM, what are the changes that has occurred as of the diagnosis, factors impact on your diagnosis experience, do think your environment impacted on the diagnosis and why? In addition, discussions focused on the process of diagnosis with T2DM and the impact of living in West Africa on T2DM management in the UK. Prompt questions were used to expand on participants’ responses and clarification where required. Each interview session lasted an average of one hour. All interviews were conducted in English, because participants could communicate in the language. All interviews were audio-recorded, with participants’ consent, and transcribed for analysis. The other authors (AH and ATW) reviewed interview transcripts.The interview transcripts were analysed using Constructive Grounded Theory (CGT) broad concept abductive approach [18]. Computer software (Nvivo 11 by QSR international) was used to support the analysis of collected data. Initially, open coding was carried out according to Charmaz [19], using this approach; each transcript was labelled, line by line, with codes that support the narrations of participants. All transcripts were open coded and focus coded to derive substantive codes. Throughout the analysis, constant comparison was done to incidents within each transcript and between transcripts. Emerging concepts were related to each other as analysis progressed to give further insight into the experiences of participants. Using the bottom-up approach, codes were identified, and they formed concepts, while concepts were aggregated to form the category that explained the experiences of participants. Memos were written throughout the process of analysis to help expand on the conceptual understanding of the findings from this study [20]. Each transcript was read multiple times to understand the underlying meaning in participants narrations. This was followed by story written from the understanding of each narration.To ensure credibility in the analysis process, FA carried out coding and category development. AH and ATW reviewed the analysis separately, and all authors discussed and reached a consensus where disagreements occurred. In addition, the findings from this study were presented at the support groups in line with the triangulation method [21]. Fictitious names were used to ensure confidentiality of participants’ identity. To present the findings from participants, data excerpts were used to support the findings. The fictitious name, age, gender, and place of diagnosis of each participant follow supporting quotes to provide a context of each participant.Theoretical Category: Finding Out.This theoretical category was identified as ‘finding out’ which represents the process of being diagnosed with T2DM. These factors influence the management regime of each participant and future management that, in turn, influenced the ways blood glucose is controlled. Participants described how they found out about their T2DM, and the impact of finding out on recommended T2DM management regimes by healthcare practitioners, as a category that emerged from this study. Several concepts further explained factors that contribute to the management of T2DM among West African immigrants in the UK.One of the key concepts, found in this study, is noticing symptoms. This concept outlines the stage at which participants began experiencing health challenges and the realisation that there is a need to seek support with their health. Participants reported symptoms including urinating frequently, fatigue, feeling thirst, weight loss, as well as mouth and body ulcers. The noticing of symptoms gives specific reasons participants felt something was not right about their health and needed to seek medical attention. This was mainly reported among participants diagnosed in the UK. All but four participants expressed noticing at least one form of symptom, which made them seek medical attention from healthcare facilities prior to finding out.
+            “I found out, … the way it started I was feeling weak, dizzy and dry mouth, dry lip when I shouldn’t. I was filling my mouth with everything I could but it was not helping”
+          And:
+            “I noticed a symptom of diabetes which is the frequency of urination and on a consultation of my doctors, I asked that I was having to drink a lot and a lot urinate frequently and they agreed to start the process of diagnosis of diabetes”
+          However, there were reports of delay in diagnosis among participants. Responses such as “getting worse” or “family insistence” made participants seek medical explanations, which led to a diagnosis with T2DM. The participants discussed that what prompted them to notice something was not right about their health was the frequency at which the symptoms occurred.It became apparent that people in this study not only noticed symptoms but also experienced an increase in symptoms occurrence that propelled them need to seek medical attention. However, most tried local medicines and self-treatment before accessing medical healthcare services as a last resort. In taking this further, the frequency of symptoms was perceived as more severe when they interfered with daily life of participants. The analysis found that the severity of symptoms, which interfere with daily activities of participants, made them seek medical attention.As highlighted, finding out about T2DM among participants was influenced by the environment (location) of diagnosis. T2DM diagnosis in West Africa differs significantly from diagnosis in the UK. For participants that were diagnosed in West Africa, delayed diagnosis affected their management of T2DM.Most participants, particularly those diagnosed in West Africa, continuously tried home remedies to manage the symptoms before seeking medical help. The unawareness and the environment influenced the length of delay before seeking medical support and diagnosis.
+            “But you know our culture, it was thought to be something else my sexual life was affected, tried with my wife; it was the same thing so I concluded that there was something wrong, it began to sink in that I had diabetes. Because of that side effect it sank in that I had diabetes, I tried all African herbs that I could think of you know”
+          Being unaware of the importance of early diagnosis among most of the participants contributed to the delayed diagnosis reported in this study. Participants discussed how lack of awareness contributed to delay in seeking medical attention when symptoms were noticed in West Africa.
+            “So my mum had diabetes, I used to have palpitations, but I didn’t know what they were each time I get angry or upset, it’s palpitations. So I think it also worked with my blood sugar, yes but I didn’t know I had blood sugar, I didn’t know”
+          And:
+            “That is something that we never took as it would affect us because …. Ignorance is the biggest killer”
+          As the analysis progressed, the location (West Africa or the UK) of diagnosis influenced how participants found out about their T2DM status. This highlights the need to go beyond individual factor in the finding out concept. The environmental influences include social norms of using alternative medicines, limited access to healthcare services, lack of medical services availability, and health beliefs in West Africa were discussed as contributing factors to delay seeking medical attention for T2DM diagnosis. Participants living in West Africa, at the time of diagnosis, discussed how they delayed seeking medical attention. This was attributed to several factors in the environment.
+            “There is the issue of misdiagnosis or unknown condition in Africa, making it difficult to properly treat. Lack of knowledge about diabetes also contributes to the issue. I would say there is better management of diabetes in the UK than Africa”
+          And:
+            “I noticed my symptom as went to the hospital, they found nothing and sent me back home in 1987, it was when I came to the UK in 1999 that I was diagnosed as having diabetes”
+          Participants discussed how the limited availability of medical facilities in the environment (West Africa) affected accessing medical advice. It was mentioned that challenges of inaccessible medical facilities contributed to the difficulty of seeking medical attention after noticed symptoms.
+            “Well, I was in my country (Nigeria) when I was diagnosed with diabetes. I was feeling very tired and thirsty initially I thought it was the stress that I was going through so I did not do much about it. It later became apparent that something was not right with my health. This was when my wife insisted I go to the hospital. I went to a private hospital because it is easier and faster to see the doctor there. It was there that I was asked to run some test and was then diagnosed with diabetes”
+          Following further analysis, it emerged that the ease of accessing medical attention in the UK contributed to early diagnosis among participants diagnosed in the UK. The easy accessibility of medical facilities by the population allowed most participants, diagnosed in the UK, to seek medical attention almost immediately after symptoms were noticed. For example, four participants were diagnosed with T2DM from routine checks before even noticing symptoms.
+            “One thing about me … I am very concerned about my health, so I go on to check a lot … do you understand? So, I just go there to do the general test, my blood and all general. So, I was like okay do this for me … as my mum got it so let me know if am close or so. So after doing that for like … after some time they sent me a letter, that I have to go and do a further test or something like that and then that I had to go to a group and start going for lectures so I can prevent”
+          As analysis progressed, the affordability of healthcare services become important to the finding out concept. Among participants who were diagnosed with T2DM in West Africa, the cost of accessing health services in Africa was part of the reasons for the delay in presenting health conditions to medical professionals. They discussed how paying their own medical costs impacted their access to medical facilities as the last resort, which resulted in worse symptoms before diagnosis. However, participants diagnosed in the UK did not mention cost as a factor. This might be because of the non-payment at the point of accessing healthcare service in the UK, so participants presented earlier for medical attention, when symptoms were noticed, than those diagnosed in West Africa.
+            “I felt I might need to see a doctor for these symptoms but I had to get some money to register at the private hospital. So am not sure exactly how long but it took a while to go anyway, with my wife’s insisting I go when she got worried about my symptoms. Going to the general hospital is not a pleasant experience was why I waited to go to private”
+          And:
+            “Well … we don’t really have the opportunity to check our sugar level on routine checks … it cost money to do that so we only check when we notice concerns”
+          Prejudgement about health was influential in seeking medical attention for symptoms noticed in T2DM diagnosis. Environmental beliefs about health contribute to when and how participants seek medical support for symptoms noticed. In most cases, participants diagnosed in West Africa expressed beliefs that other factors, such as witchcraft or religious belief about the cause of symptoms, which differs from allopathic understanding.
+            “But you know our culture, it was thought to be something else, someone has bewitched me so I concluded that something was wrong”
+          And:
+            “For me, I have faith in God, how our health turns out does not depend on our actions, God has destined what will happen irrespective. This has been my focus to make sure I live life the best way God want me to”
+          And:
+            “I believe I need to seek medical attention when I noticed the symptoms for my condition, however, I wanted to try other readily available self-medication. Medical service back home is not the easiest especially in government hospitals, so I tried to help myself on my own first”
+          Going further into the finding out brought the analysis to participants’ emotions about getting diagnosed. Being diagnosed was described as a turning point in their life and marked the beginning of living with T2DM. It was explained that being diagnosed is a significant, and unforgettable, event in their life. This led to different feelings after diagnosis that is explained below.
+            “Well … being diagnosed with diabetes was one of the worst news I have received ever … it was disappointing and really worrisome. I was so shocked especially because of the limited knowledge we have in Africa concerning the disease”
+          And:
+            “It was just not something I suspected even when I was not feeling well, diabetes was beyond strange for me until I was diagnosed …. I was scared for my future, what will happen to me”.
+          Most participants in this study discussed the feelings that they experienced after being diagnosed with T2DM. The shock was discussed as the first emotion experienced at the discovery of having T2DM because of the unexpected event of being diagnosed. Although two participants talked about suspecting T2DM when some symptoms were noticed, it still came as an enormous shock to confirm a positive T2DM status.
+            “However, the sudden shock passed and I started to think of how I can logically live with the disease and still be as normal as possible. Anyway, it has been by God’s grace so far since then we have been able to give thanks to God on the journey”
+          After the shock of being diagnosed with T2DM, panic at the realisation that they will live with T2DM for the rest of their life is another emotion that was well discussed among most participants. The panic feeling affects how they manage T2DM because of the prognosis of the condition.One participant discussed how the initial panic of finding out about her T2DM made her buy all meals labelled as T2DM food. She also talked about not enjoying these foods, but her panic state made her consider so many other options in the effort to manage her T2DM. The feeling described was as though it was a death sentence when informed of T2DM diagnosis, and this made her feel vulnerable about her health.
+            “I have always been health conscious, … the very first time they said I am diabetic …. In any situation, you panic, in my panic, I was going around looking for diabetic meals and I actually picked a few things that said it’s a diabetic meal. I prepared it and it was horrible, I couldn’t take it”
+          Panic because of the extra burden of being diagnosed with T2DM was another feeling that participants had after diagnosis. Most participants felt that being diagnosed with T2DM has placed extra burden on their health. This is particularly because of the comorbidity of T2DM with other health conditions. Feeling panic is mainly towards the burden of managing T2DM.
+            “You know I have arthritis and so I cannot walk properly, having diabetes scared me so much. I just feel my issues have double now”
+          And:
+            “I have to tell you, I have heart issues that am already managing, to then be diagnosed with diabetes just makes me feel my heart will be overwhelmed with these issues, it will not be able to manage as long as it should”
+          Further analysis showed that uncertainty about the future mainly caused panic and fear among participants. This takes the analysis to a higher conceptualisation, which helps in understanding the emotions expressed at being diagnosed with T2DM. The decisions made in the early phase, after diagnosis, seem to affect the management of T2DM as time goes on. This highlights the importance of the initial phase in the management process.Being diagnosed contributed to the management regime that participants follow to control their T2DM. All participants diagnosed in West Africa were placed on medication to manage their T2DM immediately after diagnosis, which is in addition to lifestyle changes such as dietary and physical activity.
+            “They said yes I have got diabetes Type 2 and on that same day they put me on metformin. And I have been on metformin ever since from that day on”
+          And:
+            “I was diagnosed by chance as I was made to do a general test before retirement, I only noticed I had grown lean but nothing alarming before then. I was immediately placed on medications because of the late stage of my condition”
+          However, more participants diagnosed with T2DM in the UK were recommended lifestyle changes only to manage their sugar level. One participant explained he is still on lifestyle changes to manage T2DM without being placed on any medication. Similarly, another participant discussed how he lived with his T2DM for over 10 years on lifestyle management before being placed on first-line T2DM medication. All participants diagnosed in West Africa were immediately placed on T2DM controlling medication after diagnosis because of the delayed diagnosis in such an environment.
+            “I was able to manage without … taking any tablet (Metformin) for 10yrs … I started taking the tablet in 2015 less than 4 yrs. ago so and ummm …. it has been relatively good control of diabetes”
+          The analysis showed that findings emerged into individual and environmental (external) factors. The upper circle (Figure 1) shows codes relating to individual factors that are within control of individuals, in seeking medical attention to symptoms, while the lower one shows the environmental (external) influence that contributes to the finding out process. The influence of family and friends, that encouraged and supported the individuals noticing symptoms to seek medical attention, can be seen as within individual control. The lower environmental (external) circle shows factors that are beyond participants’ control in seeking medical attention. This is more pronounced in West Africa, where the availability of health, accessibility, and affordability can be challenging to achieve. Their environment was actively involved in finding out. The overlap between the two circles shows how both individual factors and environmental factors interact to influence the finding out of T2DM. This was apparent in the stories of participants that found out about their T2DM status in relation to the location of diagnosis.In conclusion, different factors affect the finding out concept and how these interact, to impact the T2DM management process, among West African immigrants in the UK is an important exploration. In terms of the interaction between these contributing factors, individual factors are less pronounced and influential among people diagnosed with T2DM in West Africa. Environmental barriers to finding out were more influential in the diagnosis of participants from this population. Individual factors were less impactful in being diagnosed in West Africa (see Figure 2). This means that individuals’ choice on finding out about their T2DM was restricted due to the unavailability and inaccessibility of health care facilities in this environment.On the other hand, people diagnosed in the UK emphasize the influence of individual factors on finding out about T2DM. In this environment, there is better availability of medical services with easy access to these services. In this study, participants diagnosed in the UK reported that getting diagnosed is more dependent on individual factors such as busy schedule or not placing importance on healthcare services. These barriers are factors that are mainly influenced by self-efforts to seek the cause of their symptoms.This study aimed to expand West African immigrants’ experiences, and perception, of being diagnosed with T2DM and how this impacts managing the T2DM condition in the UK. The impact of the finding out concept showed factors that directly, and indirectly, contribute to the management of T2DM. In understanding the management of T2DM among West African immigrants in the UK, it is apparent that the experiences of living in West Africa have impacted the management approach among West African immigrants living in the UK. As stated by Romano [22] experiences are part of life that shapes us, and we will never be the same after undergoing such events. They unconsciously contribute to our perceptions of the future, beliefs and actions, irrespective of our environment. It can be inferred that West African immigrant’s lifestyle practices in their management of T2DM are outcomes of their lived experiences in West Africa prior to migration to the UK. It is essential to improve the management outcome of West African immigrants because of poor management reported.The theoretical concept ‘finding out’ highlights different aspects of the journey of West African immigrants in the diagnosis of T2DM, which relates to lived experiences. This category explained the interactions between participants and the healthcare system prior to diagnosis and the influences on the management of T2DM in the UK. This further highlights how participants situate themselves as patients who are shaped by their lived experiences, such as accessing healthcare services before migration to the UK. In the narrations of participants in this study, all participants draw on their experiences of accessing healthcare services in West Africa, which influences their management of T2DM in the UK. These experiences were majorly unpleasant and significant in finding out about their T2DM condition, as it resulted in delayed diagnosis. This is particularly pronounced among people that were diagnosed in West Africa. Alzubaidi et al. [23] reported that negative experiences, in accessing healthcare services, can be barriers to accessing healthcare services in the management of T2DM. This may explain the delay in diagnosis that most participants discussed in their narration.All participants diagnosed in West Africa were diagnosed after noticing symptoms, highlighting the delay before getting tested for T2DM. Similar findings have been reported in the literature in terms of delayed diagnosis [24]. The delayed diagnosis has also been reported among immigrant populations living with T2DM [10,25]. In addition, the delay in diagnosis, among participants in this study, has also been reported in other research among African immigrants [4,26]. Although T2DM onset can begin 10 years before the diagnosis of the condition [8], this may be longer among West Africans where limited access to T2DM testing and diagnosis is noted.Delay in diagnosis has been implicated for higher risks of complication, which can cause mortality [8,10,27]. For example, because of the delayed diagnosis among Africans, survey research reported that 21–25% of patients already have developed retinopathy complications of untreated T2DM [28]. Nephropathy prevalence was reported to be between 32–57% with a mean duration of 5–10 years. In addition, lower extremity amputations can be up to 7% of T2DM patients and about 12% of T2DM patients hospitalised have foot ulceration, as reported in the study [28]. However, the study did not further investigate the cause of the delayed diagnosis as a common occurrence among people in these communities. The delay in diagnosis is an event that, in most cases, leads to difficulty in meeting dietary and lifestyle recommendations in managing T2DM among this population.The environment, in which individuals live, influences the finding out about T2DM and the management of the condition. In this aspect, the limited availability of healthcare services in the West African region has affected the finding out process. The impact of this is more significant in West Africa, where access to healthcare services is limited [29]. Some issues experienced by West African immigrants in accessing healthcare services in West Africa can influence their management of T2DM, due to the limited resources allocated to healthcare services.Negative experiences and beliefs about healthcare services made seeking medical attention for symptoms noticed the last resort. In the findings of this study, participants express how they try other sources of treating their symptoms before going to a healthcare facility as a later resort. Similarly, Alzubaidi et al. [30] reported the delay of healthcare services among Arabic speaking immigrants, by accessing as a last resort, after the use of alternative treatments. This shows beliefs in other, alternative options to treat T2DM prior to accessing healthcare services.In addition, there was a lack of trust, in terms of beliefs about efficiency of healthcare services, in West African environment. This may explain the lack of awareness about seeking medical explanation of noticed symptoms referenced by most of the participants in this study. Having information about healthcare availability can be influential in managing T2DM [31]. There is a lack of routine testing services in West African areas, which may be a barrier to early diagnosis of West African immigrants prior to noticing symptoms [32]. In addition, there is the issue of out-of-pocket payment, in West Africa, for patients to get diagnosed and treated for T2DM. This influenced the lack of belief that some participants express in healthcare services, which, in turn, contributes to the late diagnosis.Similarly, Social Determinants of Health (SDH) have been implicated in the poorer management of non-communicable diseases in Africa [33,34]. The findings from this study agree with the impact of SDH among West African immigrants. This is highlighted in the explanation of environmental influence in the delayed diagnosis and poorer management of T2DM among participants prior to migration to the UK. In terms of the influence of finding out the concept in the management of T2DM in the UK, participants continued to refer to their experiences of lifestyle habits during, and prior to, the diagnosis of T2DM. This highlights the need to improve the management of T2DM while understanding the impact of lived experiences among the West African population.The understanding of finding out, among participants, presents the impact of getting diagnosed with T2DM in West Africa and the UK. The influence of living in West Africa contributes to the finding out, getting diagnosed with T2DM, and managing the condition in the UK. Findings from this study, on the T2DM diagnosis process among West African immigrants, can be valuable to health professionals supporting this group with managing T2DM in the UK. This study highlights the importance of lived experiences of West African immigrants in the management of T2DM. Managing T2DM goes beyond current presentations in the UK, but should include understanding the lived experiences of immigrants before migration.Conceptualization, F.A.; Methodology, F.A., A.H. and A.T.-W.; Data analysis: F.A., A.H. and A.T.-W., Data curation; Writing-Draft and review; Supervision; A.H. and A.T.-W. All authors have read and agreed to the published version of the manuscript..This is part of findings from first author’s PhD project which was sponsored by Bournemouth University. The authors are grateful for the funding.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Bournemouth University (protocol code 13441 and 09/01/2017).Informed consent was obtained from all subjects involved in the study.Presented data from this study are available upon considerable request from corresponding author.The authors declare no conflict of interest. Contributing factors to Finding Out of T2DM.Controllable and Uncontrollable factors by individuals in finding out.Participants characteristics.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06038.txt ADDED Viewed

	@@ -0,0 +1,6 @@

+Cadmium (Cd) is one of the most important heavy metal toxicants, used throughout the world at the industrial level. It affects humans through environmental and occupational exposure and animals through the environment. The most severe effects of oral exposure to Cd on the male reproductive system, particularly spermatogenesis, have not been discussed. In this study, we observed the damage to the testes and heritable DNA caused by oral exposure to Cd. Adult male Sprague–Dawley rats were divided into four groups: a control group and three groups treated with 5, 10, and 15 mg Cd/kg/day for 17 days by oral gavage. Our results revealed that Cd significantly decreases weight gain in 10 and 15 mg/kg groups, whereas the 5 mg/kg groups showed no difference in weight gain. The histopathology showed adverse structural effects on the rat testis by significantly reducing the thickness of the tunica albuginea, the diameter of the tubular lumen, and the interstitial space among seminiferous tubules and increasing the height of the epithelium and the diameter of the seminiferous tubules in Cd treated groups. Comet assay in epididymal sperms demonstrated a significant difference in the lengths of the head and comet in all the 3 Cd treated groups, indicating damage in heritable DNA, although variations in daily sperm production were not significant. Only a slight decrease in sperm count was reported in Cd-treated groups as compared to the control group, whereas the tail length, percentage of DNA in head, and tail showed no significant difference in control and all the experimental groups. Overall, our findings indicate that Cd toxicity must be controlled using natural sources, such as herbal medicine or bioremediation, with non-edible plants, because it could considerably affect heritable DNA and induce damage to the reproductive system.Demand for different products has increased as a result of the massive growth of the human population, and different industries have been established to meet these demands; unfortunately, these industries have brought about increased mining and industrial processing activities, resulting in environmental pollution [1,2]. Several life-threatening pollutants are found in the soil, water, and air [3]. Among these pollutants, Cadmium (Cd) is considered one of the most lethal heavy metal toxicants because of its hazardous properties, for example, severe toxicity, global availability, transferability, and persistence [4]. According to the Agency for Toxic Substances and Disease Registry (ATSDR), Cd is the sixth most hazardous chemical for living organisms [5]. The World Health Organization reported that, along with natural activities (e.g., volcanic eruption, river transport, weathering and erosion), the concentration of Cd in the environment increases with human activities, such as combustion of fossil fuel, mining, smelting, and refining of non-ferrous metals, smoking of tobacco, production of phosphate fertilizers, incineration of municipal wastes containing Cd batteries and plastics, and recycling of Cd-plated scrap [6,7,8]. Jarup and Akesson (2009) reported that increased mining and industrial activities and the use of Cd-containing fertilizers result in the contamination of soil and absorption of large amounts of Cd by plants grown for human and animal consumption [9]. In most parts of the world, the diet has been reported as the chief source of environmental Cd exposure in non-smokers [9,10,11].The concentration of Cd in the diet varies significantly. Common contributors of Cd in humans are fiber-rich diets, such as vegetables, shellfish, and cereals; in some areas, rice is reported to be a common source of Cd [12,13]. Another major cause of Cd exposure is tobacco smoking. A cigarette usually contains approximately 1–2 µg of Cd, depending on the brand; approximately 10% of this Cd is inhaled, and an estimated 50% of inhaled Cd is absorbed by the lungs [14,15]. The average Cd consumption from the diet usually varies from 8 µg to 25 µg per day [9,16,17,18,19,20,21,22,23]. This consumption may be even higher in some parts of the world (e.g., Japan) [9].Studies have measured the damage caused by Cd to the body by using various routes of administration because the distribution and absorption of different elements in food are affected by the administration route [9,24]. The quantity of Cd absorbed by the body is higher when the metal is administered intraperitoneally (i.e., i.p. injection) than when it is administered orally, and through inhalation of cigarette smoke or occupational exposure to fumes containing high concentrations of Cd (ATSDR 2008) [23,25,26]. In daily life, exposure to Cd commonly occurs through food sources; therefore, knowledge of the damage caused by Cd, when it is absorbed in the intestine, is important. Moreover, the effect of Cd exposure on heritable DNA has rarely been discussed. The present study aimed to observe the microscopic damage brought about by Cd on the testis and heritable DNA of Sprague–Dawley (SD) rats.Adult male SD rats (age, 70–85 days) were obtained from the Animal Facility of College of Animal Sciences, Jilin University, Changchun, China, and kept in plastic cages with a stainless steel top at a controlled temperature of 24 ± 2 °C and 50–60% humidity for 1 week to acclimatize to the lab environment. All the rats were maintained at a 12 h/12 h light/dark cycle and fed with standard laboratory food. Tap water was made available ad libitum. The animal handling, treatment, and sacrifice protocols were approved by the College of Animal Sciences, Jilin University China (Permit Number SY201909012).Twenty-four adult male SD rats were divided into four groups of six animals each. The first group served as the control group and was given 1.5 mL of saline via a feeding tube. The remaining three groups were treated with 5, 10, and 15 mg Cd/kg/day in the form of cadmium Chloride (CdCl2) (Tianjin Guangfu Technology Development Co., Ltd; Tianjin, China) solution with a feeding tube. In brief, a stock solution of 0.1 M CdCl2 was prepared in double-distilled water and then a required quantity of Cd from this stock solution was mixed with saline in accordance with the weight of the animals to obtain a total solution volume of 1.5 mL for each animal.The three doses of Cd were selected based on the results of previous studies. According to ATSDR, the lethal dose of Cd with a 50% kill rate (LD50) is 100–300 mg Cd/kg [27,28,29,30,31]. The current doses (i.e., 5, 10, and 15 mg Cd/kg/day) were selected because the smallest possible lethal oral dose of Cd is 15.3 mg/kg/day. All our doses were under that range, and these doses did not kill the animals examined in previous studies [32,33].All of the treatments in our experiment were based on data from the ATSDR, which divides Cd exposure into three categories on the basis of health effects: acute (i.e., exposure for 2 weeks or less), intermediate (i.e., exposure for 15 days to 1 year), and chronic (i.e., exposure for 365 days or more) [34]. Cilenk (2016) and his team studied cadmium toxicity by using an i.p. injection for 17 days [34]. In daily life, Cd exposure usually occurs through the diet; thus, all the doses in the current experiment were administered orally ((i.e., oral gavage) for 17 days, including weekends. On day 18, roughly 24 h after administration of the last dosage, the animals were anesthetized with 750 mg/kg of i.p. injection of 2,2,2-tribromoethanol solution and then sacrificed according to the guidelines of Jilin University (https://sydw.jlu.edu.cn/info/1009/2196.htm (accessed on 4 May 2021); http://202.198.25.15/uhtbin/cgisirsi/x/0/0/5?searchdata1=548824{ckey}; (accessed on 4 May 2021); https://www.lac.pku.edu.cn/docs/20200227111544292237.pdf (accessed on 4 May 2021)). The testes and epididymis were dissected out. The left testis and epididymis were fixed in 10% formaldehyde for histological processing, while the right testis and epididymis were stored at −80 °C for evaluation of daily sperm production (DSP) and comet assay.During the animal trials, the weight of animals was checked after every 4 days to update the Cd doses. After dissection of animals, the relative weight of testis was calculated by dividing testis weight (mg) by animal weight (g).
+        Relative weight of testis (mg/g)) = testis weight (mg)/animal weight (g)
+        whereas the total weight gain/loss was calculated by subtracting the final weight from the initial weight.
+        Weight gain/loss (g) = final weight (g) − initial weight (g)The testes and epididymis were fixed in 10% formaldehyde and embedded in paraffin wax for slides preparation. In brief, 5 µm-thick sections were cut from the paraffin blocks using a microtome. These sections were affixed on glass sides, on a slide warmer, and deparaffinized prior to staining with hematoxylin and eosin (H&E) stain. The slides were examined under an Olympus microscope (Model IX2-ILL100) equipped with a micro-photographic system.During histological analysis of the testis, five parameters (i.e., interstitial space, thickness of tunica albuginea, diameter of the seminiferous tubules, diameter of the tubular lumen, and height of the epithelium) were measured. In the histological analysis of the epididymis, the area and diameter of the lumen and the height of the epithelium were measured. All the measurements were obtained using Image J software. The standard scale (Supplementary Figure S1) used for image J was taken with the same magnification as the remaining pictures.Testicular tissues stored at −80 °C were defrosted at room temperature for 2–5 min prior to homogenization. Spermatids that were resilient to homogenization were calculated according to the protocol of Robb et al. [35]. The testes were weighed, and tunica albuginea was removed. Approximately 100 mg of testis parenchyma was homogenized in 2 mL of saline and diluted, according to Jahan et al. [36]. A small portion (5.5 µL) of the sample was placed in Neubauer chambers (hemocytometer), and late spermatids were counted under a microscope at 40× magnification. DSP was calculated according to the following formula.
+        Y=(x/16)×100×5×5.5×1000
+        where Y is the total number of spermatids, x is the number of spermatids counted on hemocytometer, 16 is the total number of squares observed, 100 is the total number of squares, 5 is the dilution factor, 5.5 µL is the sample volume loaded into the hemocytometer, and 1000 is the conversion factor from microliters to milliliters.Damage to heritable DNA was determined using comet assay [37], with some modifications. Slides were prepared by placing 100 µL of 1% regular melting point agarose and covered by a large coverslip. The slides were then placed in a refrigerator for 30 min to solidify the agarose. After 30 min, the slides were placed at a slide warmer at 37 °C and the coverslips were carefully removed. Next, 20 µL of a suspension of sperm from the cauda epididymis and 65 µL of low-melting-point agarose were mixed in Eppendorf using a micropipette. The mixture was placed on the agarose slides, and a coverslip was used to spread it. The slides were placed in a wooden slide box to avoid exposure to direct light and the resultant DNA damage, and the agarose slides were solidified in the refrigerator. After solidification, the coverslips were removed and the slides were submerged in staining jars containing freshly prepared cold lysing solution (100 mM EDTA disodium salt, 10 mM Tris, 2.5 M NaCl, pH 10, with 1% Triton X-100 added just before use). The slides were soaked in this solution overnight, and the staining jars were covered with aluminum foil to avoid direct exposure to light and DNA damage. A gel electrophoresis tank (horizontal) was filled with electrophoresis solution (300 mM NaOH + 1 mM EDTA, pH 12.5), and 12 slides were placed in it side by side in two rows, with the agarose end facing the positive terminal. The slides were left in the tank for some time, and the DNA fragments were separated by electrophoresis for 10 min at 25 V and 300 mA. The alkaline detergent was washed after electrophoresis with 0.4 M Tris solution to avoid interactions with the stain.Exactly 100–200 mL of 20 mg/mL acridine orange solution was overlaid on the slides by using a coverslip for comet scoring. The slides were observed under a fluorescent microscope, and comets were analyzed using Casplab_1.2.3b2.One-way analysis of variance followed by Tukey’s test was applied to compare the experimental data of different groups by using Graphpad Prism 5 software. All results are presented as mean ± SEM, and the significance level was set to * p < 0.05, ** p < 0.01.In the current experiment, significant weight loss was observed in Cd-treated groups (Figure 1). The weight of animals in the control group was increased by 28 g, while the weight in the 15 mg Cd/kg treatment group decreased by the same amount (Table 1, Figure 1). The total weight gain in the 5 mg Cd/kg treatment group was similar to that in the control group. The weight of the testes and epididymis showed no variations among the control and Cd-treated groups (Table 1). However, the relative mass of the testes was significantly higher in the 15 mg Cd/kg treatment group compared to that in the control group (p < 0.05) and mg Cd/kg treatment groups (p < 0.01) (Table 1).Microscopic analysis of the testes showed that the control group had closely arranged seminiferous tubules with normal spermatogenesis (Figure 2A,B). Germ cells of all stages were observed in germinal epithelium, and the lumens of the tubules were narrow and filled with sperm. In brief, all the tubules observed in the control group could be divided into two categories based on the morphologic appearance, the first showing the early (initial) half of the spermatogenic cycle (stage 1–8) and the 2nd showing the other half (stage 9–14). In interstitial space, Leydig cells of different shapes (round, oval, and irregular) were present, along with blood vessels surrounded by seminiferous tubules (Figure 2A,B). The diameter of the seminiferous tubules in the Cd-treated groups showed variations from the control group (Figure 2C,D, Table 2), indicating decreased spermatogenesis. The high-dose groups (10 and 15 mg Cd/kg/day) showed the greatest deterioration in testicular tissues. The size of the interstitial space (Table 2), the height of the epithelium (Figure 2), and the thickness of the tunica albuginea (Figure 2 and Table 2) showed remarkable variations compared to the control group (Table 2).The thickness of the tunica albuginea (Figure 2) in the 10 mg Cd/kg treatment group decreased significantly (p < 0.001) compared with that in the control group. The effect of Cd on the tunica albuginea was dose-dependent, as evidenced by the lack of a significant difference between the control and 5 mg Cd/kg treatment groups. The exterior walls of the tunica albuginea in animals in the 15 mg Cd/kg treatment group were affected in a non-continuous manner, where the wall was thicker at some points and thinner at others (Figure 2H). Overall, the mean thickness was similar to the control. However, as can be seen in Figure 2H, the tunica albuginea was severely affected, and the size of the wall was significantly (p < 0.001) thicker than that in the 10 mg Cd/kg treatment group (Table 2, Figure 2).The average space between different seminiferous tubules (Figure 2) showed no significant (p > 0.05) difference between the 5 mg Cd/kg treatment and control groups. The interstitial space significantly decreased in the 10 and 15 mg Cd/kg treatment groups (p < 0.001) compared with that in the control and 5 mg Cd/kg treatment groups (Figure 2E,G). The number of Leydig cells in the interstitium was similar in all groups, but the interstitial space in Cd-treated groups was remarkably reduced compared with that in the control group (Figure 2, Table 2). The average diameter of seminiferous tubules increased with increasing Cd concentration. In the 10 mg Cd/kg treatment group, the diameter of tubules increased significantly (p < 0.001) compared with that in the control and 5 mg Cd/kg treatment groups. All other groups showed increases in the diameter of seminiferous tubules, but the differences noted were not significant (p < 0.05).The epithelial height and tubular lumen showed interesting results. The average height of the epithelium (both early and late phase of spermatogenesis) increased significantly (p < 0.001) in the 10 and 15 mg Cd/kg treatment groups compared with that in the control group (Figure 2). A significant increase in epithelium height (p < 0.05) was also noted compared with that in the 5 mg Cd/kg treatment group (Table 2), but the process of spermatogenesis appeared to be impaired in all Cd-treated groups (Figure 2). It was observed in the current study that the epithelium of the control group was much denser, having cells of all stages of spermatogenesis (Figure 2A), whereas, in Cd treated groups, the overall number of cells appear to be much lower compared to the control group (Figure 2C,E,G). The DSP showed non-significant variation (Table 1), which could be explained by the histological deformities observed here. The size of the tubular lumen diameter in the 15 mg Cd/kg treatment group showed a significant difference compared with that in the control and the two other Cd-treated groups. The spermatozoa in the lumen of all Cd-treated groups were premature, thus demonstrating the marked impact of Cd on spermatogenesis.Histological analysis of the epididymis showed that the 2D area of the tubular lumen significantly increased with increasing Cd dose. In the 5 mg Cd/kg treatment group, the 2D area of the lumen slightly increased, but this increase was not significant. However, a significant increase in the area of the tubular lumen of the 10 mg Cd/kg and 15 mg Cd/kg treatment groups in comparison with the control and 5 mg Cd/kg treatment groups was observed. The diameter of the lumen significantly increased in all Cd-treated groups, but the increase in the 10 mg Cd/kg treatment group was the highest. The epithelium showed no remarkable difference among all groups (Figure 3).The mean value of DSP in the Cd-treated groups slightly decreased compared with that in the control group (Table 1). Among the groups assessed, the 10 mg Cd/kg treatment group revealed the lowest amount of sperm produced. However, overall, no significant difference in the amount of sperm produced was found among the groups.Damage to heritable DNA was determined by comet assay. The results of the Cd-treated and control groups are presented in Table 3, and relevant microphotographs are shown in Figure 4. The comets produced in different groups varied. The numbers of comets observed in each group are not shown in this paper because some of the comets may have been washed away by prolonged soaking in the buffer. Overall, the number of comets observed in the control group was lower compared with that found in the Cd-treated groups. DNA damage was estimated by considering different parameters (i.e., comet length, head length, tail length, % DNA in head, % DNA in tail, and tail moment).The head and comet lengths showed significant increases. The length of the head in the 5 mg Cd/kg treatment group was significantly (p < 0.001) shorter than that in the control group, whereas, in 10 mg/kg, it was p < 0.05 and 15 mg/kg p < 0.01 (Table 3). The percentage of DNA in the tail was lower in the control group than in the treatment groups, but the difference observed was not significant (p < 0.05). The lengths of the comet and head were comparable among all Cd-treated groups (p < 0.01 for the 5 mg Cd/kg treatment group and p < 0.05 for the 10 and 15 mg Cd/kg treatment groups when compared with the control) (Table 3, Figure 4).A major concern related to the increase in the global population is heavy metal toxicity. Cd exposure occurs through water, food, and air. In previous literature, the average Cd absorbed from food in some parts of the world was approximately 15.5 µg/day, and the average Cd contents in the blood and urine are 0.74 µg/L and 0.34 µg/g, respectively [1,38,39]. The results of previous molecular cell biological experiments indicate that Cd has more than one complex effect on different cells and pathways, especially the pituitary-hypothalamus sex organ pathway [1,40,41]. Cd disturbs cell proliferation, differentiation, cell cycle progression, and DNA replication and repair; the apoptotic pathways were also impaired [1,42,43,44]. In the present study, we observed the microscopic damage caused by oral administration of Cd on the reproductive system and the process of spermatogenesis in male SD rats.A 28 g increase in the weight of the control group was noted (Figure 1, Table 2). Rats in the high-dose Cd treatment groups (10 and 15 mg Cd/kg) lost approximately 30 g of weight (Table 2, Figure 1). Rats in the low-dose Cd treatment group (5 mg Cd/kg) showed a weight gain similar to that in the control group. Hence, besides initiating reproductive toxicity, Cd affects the weight of animals. This finding contradicts some of the data reported in the literature [44,45]. According to Sagba et al. [46], Cd has a negative effect on weight gain. In our findings, the effect of Cd on weight gain was not significant, and these findings are in accordance with references having weight loss, for example. Leach et al. and Santose et al. [46,47,48]. However, when we compared the difference in weight gain, the result became interesting due to a significant variation in the control and Cd-treated groups.In the current study, major Cd-induced damage was investigated in the reproductive system of rats. No variations in the weight of the testes or epididymis were noted. Earlier studies reported contrasting results on the weight of reproductive organs in animals exposed to Cd [48,49,50]. According to some research groups, Cd ions exert marked effects on the weight of reproductive organs, as well as the bodyweight of the animals; specifically, the weight of the reproductive organs significantly decreased following Cd exposure. Some research groups reported no change in accessory glands [51,52], and our findings appear to agree with these groups, i.e., the weight of the animals decreased but no significant difference in the weight of the testes and epididymis was noted. One of the major factors in weight gain could be the route of administration. In experimental studies, the path of administration affects the delivery and absorption of Cd [9,24]. According to Ryan, Wilhelm, Ysart, and their respective teams, the absorption of Cd is much higher when the route of administration is i.p. injection, compared to intestinal absorption or occupation and cigarette smoke inhalation [23,25,26]. In the current experiment, the animals were exposed to Cd through oral gavage, and Cd absorbed by the intestine had minimal effect on the weight of accessory organs.Disparities in morphology are directly related to physiological problems [53] and decreased DSP; these disparities result in issues related to spermatogenesis, which is fulfilled in the epithelium of seminiferous tubules [54,55,56]. We performed histopathology of the testes to determine morphological damage in these organs. We observed several deformities in the testicular seminiferous tubules of Cd-treated groups, consistent with the findings of previous research groups [57,58,59,60]. According to one research group [61], pro-inflammatory cytokines stimulate the inflammatory process, such as vascular congestion, interstitial mononuclear cell infiltrations, tissue degeneration, and necrosis. Figure 2 reveals the presence of blood cells in a cluster, as well as damage to the epithelium, thus indicating the onset of necrosis. Increases in the diameter of the seminiferous tubules and length of the epithelium (Table 2) were also observed, thus indicating the beginning of inflammation. Despite the variations in histopathology noted in this work, the daily sperm produced was not in accordance with that reported in previous studies. In the current study, we observed a non-significant decrease in DSP. However, morphological changes are directly related to physiological changes [53]. Cd-induced oxidative stress has been reported to be the chief source of decreases in sperm count [62]. Some researchers speculate that genome instability and DNA damage may sometimes result from Cd toxicity, resulting in malignancy [10,40,63,64]. Correction of inaccurate base pairs, deletion of unrequired base pairs, and repair of damaged pairs are usually blocked by Cd via the recognition of damaged DNA and attachment of proteins to affected sites [10,40,65,66]. Our results are in accordance with those studies. In the present study, the comet and head length were remarkably extended, and a higher percentage of DNA in the tail of comets was found in the Cd-treated groups, which points out the breaks in heritable DNA.Observation of various parameters led us to conclude that higher concentrations of Cd in food and water could result in reproductive deformities, along with other damages to the reproductive and body physiology.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116038/s1, Figure S1: Scales for histological calculations.This study was designed by X.Z., C.L., and T.I. and was performed by T.I., M.C., T.C., Y.Z. and Z.Z. the data was analyzed by T.I., and X.Z. the initial draft of manuscript was written by T.I. which was reviewed by X.Z., C.L and L.C. All authors have read and agreed to the published version of the manuscript.This work was supported by the National Natural Science Foundation of China (31772596, and 31872983); Jilin Provincial Department of Education Project (grant number: JJKH20190 176KJg) and Science and Technology Development Project of Changchun (18FP005).Animal handling, treatment, and succeeding sacrifice were approved by the college of Animal Sciences, Jilin University China (Permit number SY201909012).Not applicable.No data set is linked to current study; the summary of experiments performed is presented in current manuscript.The authors alone are responsible for the writing and content of this article and declare no conflict of interest.Mean (with SEM) weight gain/loss during animal trials, showing about 30 g decrease in weight of 15 mg/kg animals, whereas the same weight was gained in the control group. (Probability: * = p < 0.05 and ** = p < 0.01)Microphotograph of testicular tissue (seminiferous tubules). (A,B) The control group showed normal seminiferous tubules with a lumen filled with sperm. The interstitial space revealed normal Leydig cells. (C,D) The 5 mg Cd/kg treatment group showed a decrease in epithelial height and a marked increase in interstitial space. The (E,F) 10 mg Cd/kg and (G,H) 15 mg Cd/kg treatment groups revealed prominent damage. In particular, the number of Leydig cells in interstitial space are decreased and spermatogenesis in epithelium appeared to be disrupted. (LC: Leydig cells, SM; smooth muscle, ST: spermatids, PS: primary spermatocytes, IS interstitial space, SC: Sertoli cells, LD: (lumen diameter) tubular lumen, EH: epithelial height, TD: diameter of seminiferous tubules TA: tunica albuginea.Microphotographs of the epididymis. (A) The control group showed normal morphology and lumen full of sperm. (B) The 5 mg Cd/kg treatment group showed damaged tubules with a partially empty lumen. (C) The 10 mg Cd/kg and (D) 15 mg Cd/kg treatment groups showed lumen with a large surface area. The (E) area and (F) diameter of the lumen. (G) Height of the epithelium (magnification 10×). (a = comparison to control, b = comparison to 5 mg/kg group. Probability: * = p < 0.05, ** = p < 0.01 and *** = p < 0.001).Fluorescent microphotograph of rat sperm after Cd treatment as determined by single-cell gel electrophoresis with acridine orange staining. (A) Control group presenting intact DNA, (B) 5 mg Cd/kg treatment, (C) 10 mg Cd/kg treatment, and (D) 15 mg Cd/kg treatment groups showing damaged DNA represented by the tail.Mean and SEM of body weights, testes weight and epididymis weights. Daily sperm production per 100 mg of the testis, and the relative weight of testis.All values are expressed as mean ± SEM (a = comparison to control, c = comparison to 10 mg/kg group. Probability: * = p < 0.05 and ** = p < 0.01).Mean ± SEM of different parameters of testicular histology.All values are expressed as mean ± SEM (a = comparison to control, b = comparison to 5 mg/kg group, c = comparison to 10 mg/kg group. Probability: ** = p < 0.01 and *** = p < 0.001) Mean ± SEM of DNA damage in control and Cd treated adult rats after 17 days of treatment.(Probability: * = p < 0.05, ** = p < 0.01, *** = p < 0.001, * = significant difference compared to the control).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06039.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: Intimate partner violence (IPV) has previously been recognized as a major public health issue. Oral healthcare providers, such as dentists, are crucial to the screening and identifying of individuals experiencing IPV, since most injuries occur in the head and neck region. A lack of knowledge and awareness regarding teaching and learning about IPV in dental school curricula has been identified. Based upon the overall lack of knowledge, the objective of this study was to conduct a longitudinal assessment of knowledge, awareness, and beliefs regarding IPV among dental students in their first year of education. Methods: All students (n = 245) from three consecutive, first-year dental student cohorts (n = 81, n = 82, n = 82) were provided a brief and voluntary in-class survey in conjunction with an instructional workshop. The survey included questions designed to ascertain knowledge, awareness, and beliefs regarding IPV as a healthcare and dental issue before and after the instructional session. Differences in responses to the questions before and after the IPV educational workshop were measured using paired t-tests. Results: A total of n = 232 completed pre- and post-responses were received from all three first-year dental student cohorts (n = 76, n = 80, n = 76), representing an overall 94.6% response rate. Analysis of these data showed that the student population was predominantly male and white (non-minority), aged in their mid- to late twenties, with most students reporting no previous IPV education. The few students reporting previous IPV education were mainly younger (<25 years), which may represent more recent endeavors to increase awareness of IPV among students currently attending colleges and universities. Conclusions: The results of this study may suggest that information-specific seminars within the curriculum might be adequate to provide dental students with awareness and knowledge of IPV and specific information regarding local resources and referrals for any patients experiencing IPV.Intimate partner violence (IPV) has previously been recognized as a major public health issue among dental and oral healthcare professionals [1,2]. This has led to calls for increased data regarding awareness and knowledge among dental professionals to recognize and address deficiencies in training and surveillance for signs of IPV in clinical practice [3,4]. Increased knowledge and awareness of IPV among dental professionals has been demonstrated to significantly improve referrals for IPV-specific support services and other effective intervention programs [5,6].Oral health care providers, such as dentists, are crucial to the screening and identifying of individuals experiencing IPV, since most injuries occur in the head and neck region [7,8]. The most commonly reported IPV traumas involve facial contusions and lacerations, dental concussion, and mandibular fractures [9,10]. However, only a few select reports have assessed awareness and knowledge among dental students regarding IPV recognition and the appropriate resources and referrals needed to implement IPV curricular integration [11,12].These efforts may contribute, in part, to a larger focus on the development and implementation of IPV prevention programs that are being created in different and varied settings to address these important issues [13,14]. For example, community-based programs have been developed that demonstrate significant progress may be possible with significant input and feedback from victims, abusers, as well as the healthcare professional providers and prevention teams [15,16]. In addition, these efforts have found models among refugee and humanitarian missions that clearly demonstrate these types of approaches may provide significant benefits and might be well placed for integration into medical and healthcare education and curricula to improve knowledge, awareness, and willingness to engage in these critical areas [17,18,19].As other medical and healthcare disciplines move towards comprehensive curricular reform to include domestic and interpersonal violence, the need for more information regarding awareness and knowledge among dental students becomes crucial for dental school administrators interested in these specific curricular reform efforts [20,21]. As evidence continues to emerge regarding misinformation and misconceptions about IPV among other healthcare students, the need to accurately assess and evaluate knowledge among dental students becomes critically important [22,23]. Based upon the overall lack of knowledge in this area, the objective of this study was to conduct a longitudinal assessment of knowledge, awareness, and beliefs regarding IPV among dental students in their first year of education.The protocol for this study was reviewed and deemed Exempt by the Office for the Protection of Research Subjects (OPRS) and the Institutional Review Board (IRB) at the University of Nevada, Las Vegas (UNLV), under protocol “Retrospective Investigation of Course Content Evaluation by Students: A Survey of Domestic Violence Education and Experience among UNLV-SDM Dental Students” (OPRS#1103-3752M). Informed Consent was waived pursuant to the exemption under the Basic HHS Policy for Protection of Human Research due to Subjects, (46.101) Subpart A (b) regarding IRB exemption for research involving the use of education tests (cognitive, diagnostic, aptitude, achievement), in which the subjects cannot be identified directly or through identifiers.In brief, students (n = 245) among three consecutive, first-year dental student cohorts (n = 81, n = 82, n = 82) were provided a brief and voluntary in-class survey, in conjunction with an instructional workshop, as part of the normal instructional curriculum in Treatment Planning and Diagnosis. The survey included questions designed to ascertain knowledge, awareness, and beliefs regarding IPV as a healthcare and dental issue before and after the instructional session. Inclusion criteria included students enrolled in the DS1 student curriculum and the exclusion criteria included any student that chose not to participate. The workshop was facilitated by the UNLV Student Wellness Center’s IPV/Domestic Violence prevention and outreach coordinator, a certified provider of policies, procedures, and information for all UNLV departments, faculty, staff, and students. The workshop was originally developed and organized through a collaborative effort between the UNLV Jean Nidetch Women’s Center, the Office of Student Conduct, Counseling and Psychological Services, the Office of Civic Engagement and Diversity, and the Multicultural Center, and was supported by a grant from the Office on Violence Against Women, US Department of Justice (Grant 1009-WA-AX-0022).The objective of the workshop was to provide an introduction and overview of interpersonal and domestic violence (DV), followed by step-by-step directions for what a student, faculty, or staff member can do when someone (patient, student, faculty, or staff) discloses IPV or DV. This included on- and off-campus resources for victims and survivors of IPV/DV and relevant UNLV regulations, as well as Nevada Revised Statutes (NRS) code law for reporting IPV/DV. The pre-and post-survey was a brief questionnaire that addressed IPV awareness, resources, professional beliefs, and responsibilities, and personal IPV education and intervention beliefs, which had been previously used and validated [11,24,25].The pre- and post-surveys were physically attached to one another and were distributed at the beginning of the session prior to the commencement of the workshop. Surveys were color coded to ensure the students that chose to participate were completing the “pre” survey at the appropriate time. Once the workshop was completed, students were asked to complete the “post” survey and return all items for analysis. Each survey (pre- and post-) was assigned a numerical, non-duplicated identifier to prevent disclosure (and ensure confidentiality) of survey participants. Basic demographic information, such as age, race, and sex were included in the survey at the end of the post survey questionnaire.All responses and demographic information were manually transcribed into an Excel spreadsheet (Microsoft Excel for Microsoft 365, Version 2104; Redmond, WA, USA) for subsequent analysis. Demographic analysis of the study participants was summarized and presented as descriptive statistics. Differences between the study participants and the overall cohort demographics were analyzed using Chi square statistics, which are appropriate for categorical (demographic) data analysis. Differences in responses to the questions before and after the IPV educational workshop were measured using paired t-tests, which are appropriate for measuring changes in parametric data analysis.All students enrolled in three consecutive dental student cohorts (n = 245) were asked to participate in an in-class voluntary questionnaire that was administered before and after an educational session specifically targeted towards IPV (Table 1). A total of n = 242 students were present in class on the days when the educational session in each of the cohorts was presented, with a total of n = 232/245 or 94.6% of the student participants completing the survey. The demographic analysis of the study participants was approximately one third females (36.6%) and two thirds males (62.5%), with no significant differences between the reported sex of the study participants and overall cohort demographics, p = 0.8368. Similarly, the majority of the study participants reported their race/ethnicity as non-minority or white (50.4%), which was similar to the overall demographics of the three cohorts (56.6%), p = 0.2268.The first survey question sought to evaluate knowledge and awareness among these dental students by asking whether they had previous educational experience with IPV or DV (Figure 1). More specifically, approximately two-thirds (64%) of respondents indicated that they had no formal educational experience with either IPV or DV in a curricular setting or educational environment. As many changes to educational systems and curricular interventions occurred over time, the data from each cohort were evaluated independently, which demonstrated that there were no significant changes over time during the time period evaluated by this study, p = 0.411. In addition, slight differences between males and females were noted, with a slightly higher percentages of males (71.1%) reporting no previous IPV or DV experience than their overall percentage in the school population (62.4%), which was not statistically significant, p = 0.1229.The majority of respondents (64%) indicated no previous IPV-specific education, with a minority of students (36%) reporting some previous educational experience. No temporal changes were noted over time, as each cohort was evaluated separately (C1, 36%; C2, 39%; C3 36%, p = 0.411). Note: The percentage of males and females reporting previous IPV education was different. Although males represented 62.4% of students, 71.1% of the No previous IPV education responses were from males, which was not statistically significant (χ2 = 2.380, d.f. = 1, p = 0.1229).The second question sought to assess whether dental students perceived IPV or DV as a healthcare or dental healthcare issue (Figure 2). These data demonstrated that slightly more than half of all respondents in the pre-survey (n = 119/232 or 51.3%) indicated that IPV or DV was a healthcare or dental issue. However, more in-depth analysis revealed that responses from females (average 69%) were significantly higher than responses from males (41%) in the pre-test, p = 0.041. Following the educational seminar, the overall percentages of students who felt DV/IPV was a healthcare or dental issue rose to 81%. The responses among males (77%) and females (86%) were more closely aligned and were not significantly different, p = 0.221.The last question sought to determine the percentage of students with awareness of resources and referral information for DV or IPV (Figure 3). The data analysis revealed that the vast majority of initial responses (pre-test) indicated students were unaware of specific resources, referral procedures, or other pertinent materials specific to Nevada or UNLV (Average 18.1%). A detailed analysis by sex revealed these deficiencies were not sex-specific, which may suggest that any previous educational experience may not have included resources, or that it was not provided in a local setting (e.g., UNLV), p = 0.812. However, following the educational seminar, the overall percentage of students who responded that they felt they knew the process for referral and could identify the resources and referral information rose to 83%, which did not vary significantly between females and males, p = 0.566.Based upon the limited number of previous studies available in this area, the objective of this study was to conduct a longitudinal assessment of knowledge, awareness, and beliefs regarding IPV among dental students in their first year of education to determine whether the results from the initial study from this group were representative of these responses [11]. This study was able to successfully analyze three additional years of dental student responses, which increases the strength of inferences that can be drawn from this type of longitudinal analysis [26,27]. In addition, this study also provides additional evidence that this type of educational intervention may be sufficient for implementation in healthcare settings to increase knowledge and awareness of graduate and professional students regarding resources and referrals for patients with DV or IPV needs—an important objective for most programs that strive to improve screening and support services [28,29].This study also demonstrated significant differences in the pre-survey between male and female perceptions regarding whether DV and IPV are healthcare issues, which may reflect the findings of other students among medical and nursing graduates and professional students [30,31,32]. It is critically important that healthcare curricula address the differences in perception and education of all students with regard to DV and IPV, particularly when it comes to the differences and gaps between awareness and recognition among males and females and how these differences may disproportionately affect decisions to provide referrals and support to patients in need [33,34,35].Despite the lack of awareness and knowledge among some females in the pre-survey group regarding DV and IPV, another significant finding from this study was that the overwhelming majority of students (including female students) did not appear to have adequate knowledge of resources or specific referral services [36,37]. More data regarding this overall lack of knowledge and whether this is commonly found among graduate and professional healthcare students may help to facilitate discussions regarding the curricular importance of including DV and IPV instruction in foundational healthcare educational instruction, particularly among dentists and other oral healthcare providers [38,39,40].Although these findings provide a significant advance in our longitudinal assessment of dental student knowledge and awareness of DV and IPV within the oral healthcare setting, there are some limitations which should also be considered. For example, this study was completed in only one dental-school specific setting and may therefore not be representative of many graduate and professional healthcare students; however, limited information is available to make these comparisons [41,42]. In addition, no long-term follow up was completed to assess whether the knowledge and awareness of resources and referrals might be retained for use in subsequent years when more intensive clinical training occurs; however, other studies have shown some preliminary data regarding long-term retention of closely related training topics among medical and other healthcare students [43].As more evidence accumulates that demonstrates many of the skills and competencies carried into professional practice are first developed in the educational setting, reviews, assessments, and feedback of preclinical dental education regarding important, but often overlooked, healthcare issues such as DV and IPV become more imperative [44,45]. The function of educational research in the assessment and evaluation of this type of training in the context of public health, particularly for healthcare students with areas of specialty that may be uniquely suited to finding, reporting, treating, and referring patients at risk for DV and IPV, has also become evident [46,47,48].Although recent data may be lacking due to the interruption of routine work practices and healthcare visits due to the COVID-19 pandemic, some studies have suggested that the incidence and prevalence of DV and IPV may be increasing in some areas, including Nevada [49,50]. These data support the official position statement of the American College of Preventive Medicine that IPV and DV are important sociomedical problems that deserve professional, academic, and curricular attention to help provide capable, competent, and engaged healthcare providers with sufficient knowledge and awareness, not only of this issue, but of the resources and referrals that may help provide care and support for patients experiencing DV or IPV [51].Although this study provides a longitudinal assessment of dental student knowledge and awareness of DV and IPV, this was institution-specific and was not performed in multiple sites to determine whether this is more broadly generalizable. In addition, the cross sectional (one-time sampling) nature of this study does not allow for an assessment of long-term information retention. It is hoped that future studies based upon these results will help to broaden and strengthen these findings and will incorporate these features to provide more in-depth and robust data to support these conclusions. However, the results of this study strongly suggest that targeted, information-specific educational seminars incorporated into a healthcare curriculum may be sufficient to provide dental students with an understanding of the key issues regarding IPV. Moreover, these trainings may be sufficient to increase student awareness of available resources and referrals, which may be of specific use when diagnosing and treating patients experiencing DV or IPV. With the acquisition of this type of specific knowledge and training, these students and future clinicians may be able to better provide specific information about resources and referrals for services to their patients who may be experiencing the adverse effects of DV or IPV.K.K. and R.J.E. were responsible for the overall project design. C.B. and K.K. were responsible for data generation. All authors have read and agreed to the published version of the manuscript.No external funding was obtained for this study. The authors would like to acknowledge the Office of Research and the Department of Biomedical Sciences at the University of Nevada, Las Vegas, School of Dental Medicine for support of this project. K.K. is a co-investigator on the National Institute of Health (NIH) grant R15DE028431.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Office for the Protection of Research Subjects (OPRS) and the Institutional Review Board (IRB) at the University of Nevada, Las Vegas (UNLV), as “Exempt” under protocol “Retrospective Investigation of Course Content Evaluation by Students: A Survey of Domestic Violence Education and Experience among UNLV-SDM Dental Students” (OPRS#1103-3752M) originally approved 7 April 2011.Informed Consent was waived pursuant to the exemption under the Basic HHS Policy for Protection of Human Research due to Subjects, (46.101) Subpart A (b) regarding IRB exemption for research involving the use of education tests (cognitive, diagnostic, aptitude, achievement), in which the subjects cannot be identified directly or through identifiers.The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the study protocol data protection parameters requested by the IRB and OPRS for the initial study approval.Preliminary data from this study were submitted for presentation to the American Association for Dental Research (AADR) 2019 conference.Authors have declared that no competing interests exist.Previous IPV education among respondents. Approximately two-thirds (64%) of respondents indicated that they had no formal educational experience with either IPV or DV in a curricular setting or educational environment, with no significant changes over time during the time period evaluated by this study, p = 0.411. In addition, slight differences between males and females were noted, with a slightly higher percentages of males (71.1%) reporting no previous IPV or DV experience than their overall percentage in the school population (62.4%), which was not statistically significant, p = 0.1229.Percent of student responses indicating that vomestic violence (DV) or interpersonal violence (IPV) is a healthcare or dental issue (pre- and post). More than half of respondents in the pre-survey (n = 119/232 or 51.3%) indicated that IPV or DV was a healthcare or dental issue, which was initially higher among females (average 69%) than males (41%), p = 0.041. Overall percentages of students who felt DV/IPV was a healthcare or dental issue rose to 81% in the post-test, with responses among males (77%) and females (86%) more closely aligned, p = 0.221.Percent of student responses indicating awareness of domestic violence (DV) or interpersonal violence (IPV) resources or referral information (pre- and post). Most initial responses (pre-test) indicated students were unaware of specific resources or referral (Average 18.1%), which were not sex-specific, p = 0.812. However, following the educational seminar, students who could identify the resources and referral information rose to 83%, which did not vary significantly between females and males, p = 0.566.Study participants.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06040.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Palliative care improves quality-of-life and extends survival, however, is underutilized among gynecological cancer patients in the United States (U.S.). Our objective was to evaluate associations between healthcare access (HCA) measures and palliative care utilization among U.S. gynecological cancer patients overall and by race/ethnicity. We used 2004–2016 data from the U.S. National Cancer Database and included patients with metastatic (stage III–IV at-diagnosis) ovarian, cervical, and uterine cancer (n = 176,899). Palliative care was defined as non-curative treatment and could include surgery, radiation, chemotherapy, and pain management, or any combination. HCA measures included insurance type, area-level socioeconomic measures, distance-to-care, and cancer treatment facility type. We evaluated associations of HCA measures with palliative care use overall and by race/ethnicity using multivariable logistic regression. Our population was mostly non-Hispanic White (72%), had ovarian cancer (72%), and 24% survived <6 months. Five percent of metastatic gynecological cancer patients utilized palliative care. Compared to those with private insurance, uninsured patients with ovarian (aOR: 1.80,95% CI: 1.53–2.12), and cervical (aOR: 1.45,95% CI: 1.26–1.67) cancer were more likely to use palliative care. Patients with ovarian (aOR: 0.58,95% CI: 0.48–0.70) or cervical cancer (aOR: 0.74,95% CI: 0.60–0.88) who reside >45 miles from their provider were less likely to utilize palliative care than those within <2 miles. Ovarian cancer patients treated at academic/research programs were less likely to utilize palliative care compared to those treated at community cancer programs (aOR: 0.70, 95%CI: 0.58–0.84). Associations between HCA measures and palliative care utilization were largely consistent across U.S. racial-ethnic groups. Insurance type, cancer treatment facility type, and distance-to-care may influence palliative care use among metastatic gynecological cancer patients in the U.S.Palliative care is an integral aspect of high-quality cancer treatment. The United States (U.S.) National Comprehensive Cancer Network (NCCN) recommends palliative care should begin at cancer diagnosis and be delivered concurrently with regular cancer life-prolonging treatment to relieve symptom burden [1]. Palliative care can improve health-related quality of life by addressing symptoms frequently experienced during cancer treatment, such as pain, nausea, fatigue, neuropathy, as well as psychosocial symptoms [2]. Due to its documented benefits in symptom relief and improvements in survival among patients with advanced cancer [3,4,5,6,7,8], palliative care is also recommended by the American Society of Clinical Oncology (ASCO) and the Society of Gynecologic Oncology (SGO) throughout the cancer care continuum [1,9,10].Palliative care is underutilized among U.S. patients with gynecological cancers despite the high symptom burden during treatment for metastatic disease (stage III-IV cancer at diagnosis) [11,12,13,14]. For example, in a retrospective study of deceased patients with ovarian cancer, only 28% were referred to palliative care, and the most common type of palliative care was a referral to hospice care rather than palliation of adverse symptoms [13]. Reasons for the underutilization of palliative care among U.S. gynecological cancer patients remain unclear [15]. Additionally, existing U.S. studies have focused on palliative care during the end-of-life setting rather than across the cancer care continuum.Several factors associated with the use of palliation in the context of end-of-life care among cancer patients have been identified, including differences by race/ethnicity. Racial minorities are less likely to utilize hospice services at end-of-life compared to their White counterparts, and these differences have been attributed to knowledge gaps and attitudes among patients and providers alike [16,17]. In contrast, however, Black gynecologic cancer patients have been found to be more likely to utilize inpatient palliative care services before death compared to non-Hispanic White gynecologic cancer patients [18]. Near end-of-life, Medicaid-insured and uninsured gynecological cancer have been found to be more likely to utilize palliative care compared to those with Medicare insurance [18]. Other measures related to health care access, such as distance from provider to patient, have not been evaluated among metastatic gynecological cancer patients in the palliative care context. In our prior work, we demonstrated that racial/ethnic disparities exist in palliative care use among metastatic gynecological patients, specifically that NH-Black and Hispanic patients are less likely to use palliative care compared to their NH-White counterparts in the U.S. [14,19]. A current knowledge gap exists regarding the role of measures of health care access with palliative intervention use among metastatic gynecological cancer patients, and in particular, whether healthcare access factors may explain any observed racial/ethnic differences in palliative care utilization in the U.S. The objective of this study was to evaluate the following health care access measures: patient’s insurance status, Medicaid expansion status of the state the patient was treated, distance from patient-to-provider, cancer treatment facility type, percent of adults without a high school degree in patient’s zip code (a surrogate measure for educational level), and median household income of adults in patient’s zip code (a surrogate measure for income level). We evaluated associations of palliative care utilization with these health care access measures by metastatic gynecological cancer site of origin overall and by race/ethnicity. We hypothesized that measures of high health care access (i.e., private insurance; receiving treatment at an academic medical center; residing in a highly educated area, etc.,), will be positively associated with using palliative care. Data for this study were obtained from the latest 2004–2016 Participant Use Files (PUF) of the U.S. National Cancer Data Base (NCDB), a United States hospital-based oncology database combining data on patients seen at any of the 1500 Commission on Cancer (CoC) accredited institutions in the United States [20,21]. The NCDB registry is a joint project of the American Cancer Society and the Commission on Cancer of the American College of Surgeons. The NCDB registry includes more than 29 million unique cases or 70% of all patients with newly diagnosed cancer in the United States [22]. Data reported to the NCDB are highly standardized similar to other state health departments and federal cancer registry data systems, including the U.S. Surveillance, Epidemiology, and End Results (SEER). Data included in the NCDB are from patient charts abstracted by Certified Tumor Registrars (CTR) who undergo training specific to cancer registry operations [23]. The data abstractors use standardized methods to collect sociodemographic, including race/ethnicity, and clinical data, including tumor type, stage, grade, and treatments. To ensure high-quality and accurate data, CoC-accredited sites undergo an external review of hospital charts and registry abstracts to verify the NCDB registry data correctly reflect the information documented in individual patient records using a sample of at least 10% of records [24]. The study was approved by Duke University Institutional Review Board (Durham, NC, USA) under a general study protocol (IRB#: Pro00102834). Study participants included patients with Stage-III and IV ovarian, cervical, and uterine cancers at diagnosis. We included patients diagnosed between 1 January 2004 to 31 December 2016 using the following International Classification of Diseases for Oncology, Third Edition topography codes: ovarian (C569), cervical (C530, C531, C538, C539), and uterine (C559). Patients with missing or unknown cancer stages were excluded (n = 35,346, 9.9%). We excluded patients with missing data on palliative care utilization (n = 1018, 0.6%). Overall, our study population included 176,899 patients. In sensitivity analyses, we excluded metastatic gynecological cancer patients who were known to be alive based on the vital status of the patient as of the last date of contact (n = 52,170, 32.2%). We evaluated deceased metastatic gynecological cancer patients (n = 124,729) as palliative care has historically been prioritized among those near the end-of-life. Results based on sensitivity analyses are summarized in Supplementary Material. The main outcome was palliative care as defined by the NCDB, as in previously published studies [13,21,25,26,27]. The NCDB includes information on any palliative care from patients’ clinical medical records during their treatment at the reporting facility. The NCDB codes treatments as palliative only if the patient’s medical records explicitly mentioned that the goal of treatment is palliation and not cure. Specifically, any procedure was categorized as palliative care if treatment was provided to “prolong a patient’s life by controlling symptoms, to alleviate pain, or to make the patient more comfortable [28].” Types of palliative care documented and abstracted from the patient’s medical record could include pain management therapy, surgery, radiation therapy, or systemic chemotherapy administered to alleviate symptoms. Patients utilizing palliative care in the NCDB may also concurrently be undergoing curative treatment. The NCDB does not document hospice services or referral to hospice and was therefore not included in the definition of palliative care. Palliative care utilization was compared to those who did not utilize palliative care.We evaluated several health care access variables, which we define here. Insurance type was identified as the patient’s primary insurance carrier at the time of initial diagnosis and/or treatment. Types of insurance include private insurance, which is traditionally provided by the patient’s employer or union, Medicare, Medicaid, and other Government insurance. Medicaid and Medicare are U.S. government-sponsored national health insurance programs; Medicaid is available to eligible low-income adults, children, pregnant women, elderly adults, and people with disabilities [29]. The Medicare program is a federal health insurance program for people who are 65 years or older, certain younger people with disabilities, and people with end-stage renal disease [30]. Other government insurance may include the Indian Health Services Insurance, which is offered to U.S. adults who identify as Native Americans; the Veterans Health Administration or the VA, which is offered to U.S. adults who have served the U.S. military. Also, we evaluated the potential association of Medicaid expansion status of the state the patient was diagnosed. In the U.S., states have an option to expand the eligibility requirements to enroll in Medicaid to cover more low-income Americans [31]. As of 2019, Medicaid has fully expanded in 33 states and the District of Columbia. It is important to evaluate the impact of Medicaid expansion on the use of health services in the U.S. to provide evidence of the benefits of the program and the potential long-term cost savings after investment into the program developed by the Affordable Care Act (ACA). We only included data starting from 1 January 2011 to 31 December 2016 as the Affordable Care Act was passed in 2010 and enacted in the following year [32]. We evaluated the greatest circle distance from provider to patient, which is a measure of distance in miles between the patient’s residence (residential latitude and longitude based on the patient’s zip code centroid or the city if the zip code was not available) and the provider’s hospital location (street address for the facility). Additionally, we evaluated the percentage of adults in the patient’s zip code without a high school degree and median household income in the patient’s zip code. These zip code level variables were derived from the 2012 American Community Survey (ACS) data, spanning years 2008–2012, and adjusted for 2012 inflation [33]. The ACS is an ongoing survey conducted by the U.S. Census Bureau that provides annual demographic data on U.S. communities. Further details regarding the ACS can be found here: https://www.census.gov/programs-surveys/acs/about.html (accessed on 25 May 2021). The percentage of adults with a high school degree and median household income was categorized as quartiles based on equally proportioned income ranges among all United States zip codes. Race/ethnicity, as defined by the U.S. Census Bureau [34,35], were captured in the NCDB based on self-report or as reported by the patient’s providers. In the U.S., ethnicity is defined as Hispanic or Latino, and Not Hispanic or Latino. The U.S. Census Bureau defines “Hispanic or Latino” as a person of Cuban, Mexican, Puerto Rican, South or Central American, or other Spanish culture or origin regardless of race [35]. U.S.-specific definitions of racial categories can be found here: https://www.census.gov/topics/population/race/about.html (accessed on 25 May 2021). We evaluated the health care access variables overall and stratified by race/ethnicity. We combined reported race/ethnicity to create the following categories: Non-Hispanic (NH) White (NH-White), NH-Black, Hispanic, Asian, American Indian/Alaskan Native, Native Hawaiian/Pacific Islander, and other Race. For the main analysis, we focused on comparisons of NH-White, NH-Black, Hispanic, and Asian metastatic gynecological cancer patients to ensure adequate patient size across racial groups for statistical modeling. We summarized patient characteristics as percentages by palliative care utilization among metastatic (Stage III/IV at diagnosis) ovarian, cervical, or uterine cancer patients at the time of presentation. We evaluated racial/ethnic differences in health care access measures using bivariate statistical analyses (χ2 tests). Next, we evaluated multivariable associations of health care access measures stratified by cancer site and next by race/ethnicity using multivariable logistic regression. We estimated adjusted odds ratios and 95% confidence intervals and adjusted for study covariates identified using directed acyclic graphs based on prior literature for health care access and palliative care. Adjustment sets for each health care access measure are summarized in Tables accordingly. We accounted for non-independence within clusters at the facility level to account for correlated patient characteristics within hospitals and calculated cluster-robust standard errors. We assessed each covariate for collinearity and used a complete case approach. All analyses were performed with Stata statistical software, version 15.0 (StataCorp, College Station, TX, USA).Overall, the median age of patients was 62 years, and most were non-Hispanic White (72%) and either insured with Medicare (41%) or privately insured (36%). Twenty-one percent of patients had a Charlson-Deyo comorbidity score of one or above. Most lived in urban areas (96%) and were treated at either a comprehensive community cancer program (36%) or an academic/research program (38%). Five percent of patients with metastatic gynecological cancer at the time of presentation utilized palliative care at any time during their disease course. Overall, 4% of patients with ovarian, 9% with cervical, and 11% with uterine metastatic cancer utilized palliative care. Among patients who did receive palliative, the most common types included either surgery, radiation, or chemotherapy alone (62%) and 12% received pain management only (Table 1). Table 2 summarizes each health care access measure stratified by race/ethnicity. NH-White (86%) and NH-Black (69%) patients were mostly either privately or Medicare insured. Asian patients commonly lived in the Western census region (43%), whereas NH-Black patients commonly resided in the South (53%) (p < 0.001). We observed the highest proportion of Hispanic patients (58%) residing in zip codes with ≥17.6% of adults without a high school degree (i.e., less educated), and the lowest among NH-White patients (16%) (p < 0.001). Patients were commonly treated at academic/research programs, particularly NH-Black (46%), Hispanic (43%), and Asian (45%) patients. Compared to privately insured patients, uninsured patients with ovarian (aOR: 1.80, 95% CI: 1.53–2.12) and cervical cancer (aOR: 1.45, 95% CI: 1.26–1.67) were more likely to utilize palliative care after adjustment for age, race/ethnicity, Charlson-Deyo comorbidity score, and median household income (Table 3). Medicaid-insured patients with ovarian cancer (aOR: 1.89, 95% CI: 1.64–2.19) and cervical cancer (aOR: 1.41, 95% CI: 1.26–1.57) were more likely to utilize palliative care. When evaluating the greatest circle distance from provider to patient, we observed that compared to patients <2 miles away from their provider, the odds of utilizing palliative care decreased with increasing distance for all gynecological cancer sites after adjustment for age, urban or rural area of residence, and census region. Compared to ovarian cancer patients who were treated at comprehensive community cancer programs, academic/research program patients were less likely to utilize palliative care (aOR: 0.70, 95% CI: 0.58–0.84). Sensitivity analyses revealed similar patterns of palliative care utilization across cancer sites for each health care access measure among deceased metastatic gynecologic cancer patients (Table S1). Figure 1 provides a summary of health care access measures evaluated by race/ethnicity by gynecological cancer sites, and the point estimates are available in Table S2. Uninsured NH-White (aOR: 1.98, 95% CI: 1.74–2.24), NH-Black (aOR: 1.47, 95% CI: 1.18–1.83), and Hispanic (aOR: 1.86, 95% CI: 1.38–2.51) patients were more likely to utilize palliative care when compared to patients with private insurance. NH-White and NH-Black patients with Medicaid or Medicare were more likely to use palliative care compared to their privately insured counterparts. Asian patients residing in more educated areas were more likely to use palliative; For example, compared to Asian patients residing in areas with ≥17.6% of adults without a high school degree (i.e., least educated), Asian patients in the most educated areas (<6.3% of adults without a high school degree) had over two times the odds of using palliative care (aOR: 2.41, 95% CI: 1.43–4.07). Compared to NH-White, NH-Black, and Asian patients residing <2-mile circle distance away from their provider, those who reside more than 45 miles away had 37% (aOR: 0.63, 95% CI: 0.53–0.74), 54% (aOR: 0.46, 95% CI: 0.35–0.61), and 66% (aOR: 0.34, 95% CI: 0.14–0.83) lower odds of palliative care utilization, respectively. NH-White (aOR: 0.73, 95% CI: 0.62–0.86) patients treated at academic/research programs were less likely to utilize palliative care compared to those treated at comprehensive community cancer programs. Conversely, Asian (aOR: 2.01, 95% CI: 1.15–3.51) patients treated at Integrated Network Cancer Programs were more likely to utilize palliative care. Increasing distance from patient to provider also led to lower odds of palliative care, utilization specifically among NH-White, NH-Black, and Asian patients. Sensitivity analyses revealed similar findings among deceased metastatic gynecologic cancer patients across racial/ethnic groups (Table S3). In our study of metastatic gynecological cancer patients treated at Commission on Cancer (CoC) accredited institutions in the United States, we observed that several measures of health care access were important predictors of palliative care use. Distance-to-care plays an important role in palliative care use among patients with metastatic gynecological cancers: Patients living farther from their providers were less likely to utilize palliative care than those living closer to their provider, and this trend was consistent across racial groups. Uninsured patients and patients with Medicaid or Medicare insurance were more likely to utilize palliative care compared to the privately insured, particularly patients with ovarian or cervical cancer. Understanding health care access measures that influence palliative care use may reveal areas for intervention to improve access to equitable high-quality cancer care among gynecologic cancer patients in the U.S. and globally, where patients may experience similar barriers to palliative care use.Our study suggests that patients living farther from their providers were less likely to utilize palliative care, and this finding was consistent across racial groups. Our finding is similar to prior studies demonstrating distance-to-care impacts high-quality cancer care, including studies evaluating cancer treatment outcomes such as receipt of guideline adherent care [36]. For example, a prior study found that urban women receiving curative treatment for cervical cancer who lived farther than 15 miles away from their provider were less likely to initiate timely treatment compared with those <5 miles from their provider [37]. Prior work evaluating distance-to-care using the NCDB has also reported similar findings: for example, stage III colon cancer patients who traveled 50 to 249 miles for treatment were less likely to receive adjuvant chemotherapy than patients with a travel distance less than 12.5 miles [38]. Limited studies have evaluated distance-to-care in the palliative care context. The disparity we observed may be attributable to geographic disparities that exist in the distribution of gynecologic oncologists across the United States [39]. A survey of gynecologic oncologists showed that almost three-quarters practiced in an urban setting and only 13% practiced in an area with a population <50,000, i.e., rural areas [39]. Patients living farther from their cancer provider may be more likely to live in a rural community and therefore receive cancer treatment at a smaller community hospital. Smaller community hospitals may be less likely to have palliative care programs in place, leading to the underutilization of palliative care services [40]. Future research evaluating rural and urban differences in palliative care use among cancer patients should be prioritized to optimize rural cancer care. The NCDB provides a unique opportunity to evaluate the role of health care access from the perspective of several types of insurance providers. Our work demonstrates patient’s insurance type plays an important role in palliative care use. We observed that patients with ovarian and cervical cancer insured through Medicaid and patients without health insurance were more likely to utilize palliative care compared to those on private insurance. Our finding is consistent with a prior NCDB study of patients with colon, lung, melanoma, and prostate cancer patients, which also demonstrated that Medicaid insurance was a determinant of increased palliative care use when compared to privately insured patients [41]. In the U.S, patients insured through Medicaid or without health insurance are more likely to be low-income or without employment, which in turn leads to poor access to U.S. health care due to the prohibitive costs associated with cancer treatment in the U.S. As such, uninsured or Medicaid-insured patients are less likely to access care at early stages of cancer development leading to metastatic cancer at diagnosis. Prior research shows that palliative care is prioritized near the end-of-life or when cancer has progressed, which may explain this finding. Further, previous research has shown that the delivery of palliative care services to individuals with Medicaid insurance results in lower healthcare costs to the hospital and providers, especially when treatment was delivered to patients at the end of life and died due to their cancer [42,43,44]. These cost savings associated with palliative care services provided to Medicaid patients may also explain our results, however, further research is needed to evaluate the impact of insurance type. Metastatic gynecologic cancer patients receiving care at academic, or research hospitals were less likely to receive palliative care compared to comprehensive community cancer programs. Also, our data suggest ovarian cancer patients receiving care through community cancer programs may be more likely to use palliative care, particularly those who may have passed due to their cancer. Our finding is in contrast to prior work that found public hospitals, sole community provider hospitals, and for-profit hospitals are less likely to have palliative programs and services compared to hospitals affiliated with medical schools or large hospitals [45]. However, this prior work explored the broader question of the existence of palliative care programs generally while we evaluated patient-level utilization of the palliative care programs specifically among gynecologic cancer patients. Our findings are also in contrast to a prior NCDB study which showed that colorectal cancer patients receiving care at academic or research programs were more likely to utilize palliative care when compared to non-academic programs, which is in contrast to our comparison group and potentially contributing to the different findings [27]. Further research is needed to delineate differences in palliative care use by cancer care facility types in the United States.The use of palliative care is influenced by several factors, including patient characteristics, disease characteristics, and provider characteristics. We are limited to the data available in the NCDB and are not able to evaluate unmeasured factors that may influence the choice of physician or patient to opt for palliative care. For example, patients may not receive palliative care due to personal choice or beliefs regarding end-of-life care [46]. Additionally, further investigation of the role of area-level societal factors that may play a role in access to care at the patient level will be an important area of research to deliver equitable cancer care in the U.S. This analysis was limited to zip-code level proxy measures for educational attainment and income level. However, prior research has demonstrated the limitations of leveraging zip-code level measures as a proxy for individual level socioeconomic status [47]. Future research should prioritize leveraging a more precise measure of area-level socioeconomic status, such as census-tract level measures. It is also important to acknowledge the NCDB data on palliative care services are of uncertain accuracy. Palliative intent must be inferred from clinical records, and therefore, there is an opportunity for misclassification of palliative care use and the type of palliative care treatment. However, the NCDB has established protocols to ensure the data are captured accurately and as noted in prior work, record abstraction methods used to develop the NCDB is the approach leveraged by all hospital-based studies evaluating palliative care and is likely to be more accurate than health insurance claims. We explored health care access measures to inform the under-utilization of palliative care services among patients with metastatic gynecological cancers in the United States. We observed that patients who had Medicaid or who were uninsured were more likely to use palliative care, that individuals living far away from their provider were less likely to receive palliative care, and that individuals receiving care at academic, or research hospitals were less likely to receive palliative care compared to the referent group. Our results suggest racial and ethnic identities may play an important factor in palliative care utilization among women with metastatic gynecological cancers, potentially due to structural barriers racial minorities experience to obtain high-quality cancer care: Limitations may exist in the race/ethnicity data captured in the NCDB due to differences in race recording practices across participating institutions. For example, some institutions may rely on an individual’s last name to assign Hispanic ethnicity, which is an unreliable measure of racial/ethnic identity. Future research conducted to evaluate racial/ethnic disparities in palliative care use should prioritize efforts to optimize the capture of self-defined racial/ethnic identity. Also, given the differences in the demographic composition of patient populations by cancer type, future research should investigate the role of health care access factors in palliative care use in the context of other cancer sites to improve uptake and accessibility of palliative care for all cancer patients. Equitable access to palliative care is an important metric of high-quality cancer care in the United States, and efforts to improve the delivery of palliative care services using insights from our analysis should be prioritized. The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116040/s1, Table S1: Associations of Health Care Access Factors with Palliative Care Use Among Deceased Metastatic Gynecological Cancer Patients by Cancer Site; Table S2: Associations of Health Care Access Measures with Palliative Care Receipt Among All Metastatic Gynecological Cancer Patients by Race/Ethnicity; Table S3: Associations of Health Care Access Indicators with Palliative Care Receipt Among Deceased Metastatic Gynecological Cancer Patients by Race/Ethnicity.J.Y.I. and T.A. conceptualized the project. J.Y.I. carried out statistical analyses and led manuscript development. V.S., R.A.P., T.A. provided critical and expert interpretation of results. All authors contributed to writing and finalizing the manuscript.All authors have read and agreed to the published version of the manuscript.There was no funding obtained for this project. Islam is supported by UNC’s Cancer Care Quality Training (2T32CA116339-11). Previs is supported by grants from the AAOGF-GOG Foundation and the Emerson Collective. Akinyemiju is supported by an R37 funded by the National Cancer Institute (7R37CA233777-02).The study was approved by Duke University Institutional Review Board under a general study protocol (IRB#: Pro00102834).Patient consent was not applicable as this manuscript summarizes a secondary data analysis of an existing large database. The data are publicly available by application through the American College of Surgeons at the following link: https://www.facs.org/quality-programs/cancer/ncdb/publicaccess. (accessed on 1 April 2021).The authors acknowledge and thank the American College of Surgeons for making the National Cancer Database available to researchers to conduct this work.Authors report no conflict of interest.Associations of Health Care Access Measures with Palliative Care Use Among Metastatic Gynecological Cancer Patients by Race/Ethnicity.Health Care Access Measures Among Metastatic Gynecologic Cancer Patients by Palliative Care Utilization (n = 176,899).Abbreviations: No.: Number; Row %: Row percentages; Col %: Column percentages; SD: Standard Deviation. * Data were restricted to 2011–2016 as the Affordable Care Act was passed in 2010 and enacted in the following year (n = 88,637).Health Care Access Measures Among Metastatic Gynecological Cancer Patients by Race/Ethnicity (n = 165,211) *.* Excludes metastatic gynecological cancer patients of other races (n = 11688). † χ-squared test p-value to test differences across racial/ethnic categories. ‡ Data were restricted to 2011–2016 as the Affordable Care Act was passed in 2010 and enacted in the following year (n = 88,637).Associations of Health Care Access Factors with Palliative Care Use Among All Metastatic Gynecological Cancer Patients by Cancer Site (n = 176,899).Abbreviations: aOR: Adjusted odds ratio; CI: Confidence Interval; Ref.: Reference. * Adjusted for age, race/ethnicity, Charlson-Deyo comorbidity score, and median household income quartile of patient’s zip code. † Adjusted for age, race/ethnicity, census region, and median household income quartile of patient’s zip code. ‡ Adjusted for age, race/ethnicity, census region, and % of high school degree in quartile of patient’s zip code. ▲ Data were restricted to 2011–2016 as the Affordable Care Act was passed in 2010 and enacted in the following year (n = 88,637); Adjusted for age, Charlson-Deyo comorbidity score, and race/ethnicity. §Adjusted for age, race/ethnicity, area of residence and census region. ¶ Adjusted insurance type, area of residence, census region, and greatest circle distance to care.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06041.txt ADDED Viewed

	@@ -0,0 +1 @@

+ During the first year of the COVID-19 pandemic, dental faculties had to rethink their way of teaching and interacting with students and of delivering solid theoretical knowledge and practical skills to students. Background: The purpose of the study was to assess dentistry students’ opinions about the online activity, together with a self-evaluation of their mental and physical health, during the first wave of the pandemic. Methods: A cross-sectional study was conducted using an online survey. Three hundred and three students, enrolled across all six years of study, were included in the research. Socio-demographic and academic data were collected, along with a self-evaluation of physical and mental status. Some items investigated students’ opinions about distance learning and the impact of that online activity on their achievement. The answers were rated using a five-item Likert-like scale. Data were analyzed using SPSS (v.24). Results: statistical analyses showed that more than 20% of the students strongly agreed with the statement that they felt more anxious and depressed during the first months of the pandemic, and more than 30% were totally satisfied with their relationships with their family members. One-fifth of the respondents declared that they were totally dissatisfied with the relationships with their colleagues and friends. Overall, 50.60% of the students attended the courses/labs in their entirety when they were connected online. Two-thirds of the respondents considered that their practical training was affected due to the online activity, and that not all of the subjects could be taught online. More than half of the respondents agreed that the most objective evaluation method is that of the multiple-choice exams administered at school, and considered that exclusively utilizing online assessments of students encourages unethical behaviors. Age, involvement in online activity, and active participation using video cameras were strongly correlated with satisfaction with academic results. Conclusions: The results of the present study showed that online activity was a good alternative for dentistry students during the pandemic restrictions. The positive aspects, together with the negative consequences, of distance learning should also be taken into consideration by university teachers and academic institutions to improve teaching experiences and to ensure a solid professional formation for dentistry students.After the outbreak of the pandemic, universities were forced to align with the restrictions imposed by the governments of each country. The rate of spread of the virus in each country has led to various measures, from the total closure of courses to partial closures, depending on the epidemiological data identified at the regional level.In the academic environment, the preventive measures taken were, in general, to completely stop all activity in the classroom, to move the didactic activity to an online environment, and to avoid coexistence on university campuses.According to the American Dental Education Association (ADEA), on 15 May 2020, some key changes were implemented in United States (USA) dental training institutions specifically due to the COVID-19 crisis. Students were sent home for a period, and institutions planned to evaluate their return strategies [1]. On 4 December 2020, the American Academy of General Dentistry (AGD) assembled a list of AGD-sponsored and other online education programs available to members to help them through the pandemic period [2]. On 14 October 2020, the ADEA updated the recommendations for higher education programs in dentistry that intended to initiate the return of dentistry students to campus. The new guidelines were proposed in order to avoid risk and the need for crisis management in higher education, to diminish financial losses, to ensure good learning environments, and to support students’ mental health and the well-being of dentistry faculty members [3].Many studies conducted in the USA one year after the outbreak of the COVID-19 pandemic also pointed out several positive changes among dental faculties, including curricular changes; students’ involvement in frontline healthcare services; and the adoption of new strategies and methods of learning, teaching, and evaluating students. In order to continue the educational process, policy makers, institutions, teachers, and students worked together to cope with the pandemic crisis. A swift implementation of creative methods was adopted to avoid disruptions in practical training relating to patients and clinical competencies [4]. In many educational fields, especially among medical specialties, technology-based learning (TB learning) was used as an efficient educational tool. The virtual educational system was found to help dental students improve their clinical skills [5].Dental treatments were considered to be of very high risk due to aerosol generation, so the greatest challenge for the teaching staff was to decrease the risk of COVID-19 infection while ensuring the continuity and quality of the dental education system. In light of national COVID-19 infection-control recommendations, dental faculties had to suspend in-person instruction, and teaching activity was conducted online. Some researchers who focused on online dental education pointed out the rapid change adopted by dental faculties in many countries, such as the USA [6,7,8,9], Germany [10], Italy [11], China [12,13], Chile [14], Brazil [8,15,16], Pakistan [17], Nepal [18,19], Indonesia [20], India [21,22], Jordan [23], Romania [24,25,26], Portugal [27,28], Cyprus [29], Australia [8,30], and Spain [31]. During the first year of the pandemic, the academic research teams additionally wanted to identify the differences between countries in what aspects of dental education were disrupted. For example, a multi-country survey conducted by Ammar et al. [32] questioned academic teachers working in dental academic institutions across 28 countries about multiple challenges they experienced during the COVID-19 outbreak. The findings showed the need for human, financial, and technical resources to ensure high quality in online schooling, sharing teaching resources and best practices, and training academics.The nature of teacher–student interactions changed, and the new type of virtual communication had to use more rich visual content, interactive video tools, graphs, or web-based interaction to meet students’ need for interaction but also to help students create their own learning styles. Individual tasks or online working group sessions were implemented. The need for developing practical skills proved that online learning was an efficient tool in the medical sciences, including dental education, during these special pandemic times.Distance learning has facilitated the ability for students to continue their studies in dentistry, thus creating the opportunity to access information and simulate practical skills. The online development of the didactic activities also required students’ access to audio—video connection means and to a stable connection to the internet. Various studies [33] showed that in addition to the lack of direct teacher–student and student–patient interaction, the online activity should be maintained due to the highlighted benefits, such as increased satisfaction with learning; individual working styles; synchronous or asynchronous support activities, which give students more freedom to organize their learning and free time; and the use of more interactive methods that provoke students to become more involved in activities. In medical specialties, various studies conducted among dentistry students showed less satisfaction with the teacher–student relationship and a strong disagreement with the effectiveness of online classes [16].In the field of dentistry education, the evaluation of TB teaching and learning in non-prepared situations, such as those imposed by the COVID-19 pandemic, has not been fully investigated in Romania.In Romania, between March and October 2020, the faculties of dentistry decided to suspend face-to-face courses. Regarding the next academic year, various methods of carrying out teaching activities were proposed (online or on-site), and practical stages with patients were considered a high-risk activity. This study aimed to evaluate dentistry students’ opinions about academic activity after the first wave of the pandemic and to identify different aspects related to online activity among undergraduate dental students.The present study was conducted for two weeks, between 18 November and 2 December 2020. The study period represents the middle of the first academic semester that usually starts in Romania on 1 October and lasts until 10 February. During the period that this study was developed, due to the epidemiological conditions and restrictions imposed among all universities in Romania and Europe, the main parts of the academic activities in the medical sciences were conducted online. The dentistry faculty organized only online activity during this first semester, so students were located at home (i.e., foreign students were in their countries of origin and national students were in their native regions).The questionnaire was created and distributed using the Google Forms application (Alphabet, Mountain View, CA, USA) to all students of the dentistry faculty across all academic years. The participants were informed about the purpose of the study. Filling in the questionnaire was assumed as informed consent to participate in the study. No incentives were offered to students, and they had the ability to withdraw from the research with no consequences. The inclusion criteria were that participants had to be students enrolled in dentistry studies (in years 1 to 6 of study), and the questionnaire had to be submitted before the deadline. Exclusion criteria were students being enrolled in master’s degree or doctoral research programs and/or questionnaires being submitted after the deadline.The survey form was sent to 527 students who were registered across all years of study. A total of 314 participants answered the questionnaire before the deadline; after eliminating 11 respondents who declared that they were postgraduate or Ph.D. students, a total of 303 participants were finally included in the research (See Figure 1).After researching the literature about e-learning and online activity during the COVID-19 pandemic, a survey was constructed in order to assess students’ opinions about the educational process during restrictions. The questionnaire contained four parts:-The first part collected socio-demographic-, academic-, and health status-related data (i.e., age; gender; year of study; being suspected/confirmed as infected with the SARS-Cov2 virus; chronic disease; consumption of cigarettes, alcohol, energy drinks, and drugs; living environment or sharing room during the academic year).-The second part collected information about students’ physical and mental health status and their fear of infection with COVID-19. Self-rated items were constructed, and answers were rated using a five-item Likert-like scale.-The third part of the survey inquired about students’ satisfaction with their academic results and relationships with friends, colleagues, and family members, as related to the restrictions imposed by the COVID-19 pandemic.-The fourth part collected data regarding students’ opinions about their online activities; their practices and habits during lectures; practical stages that were developed online; the teaching and evaluation process; and the impact of online classes on their academic achievement and results, together with ethical concerns about evaluation.-Self-rated items were constructed, and answers were rated using a five-item Likert-like scale.The first part collected socio-demographic-, academic-, and health status-related data (i.e., age; gender; year of study; being suspected/confirmed as infected with the SARS-Cov2 virus; chronic disease; consumption of cigarettes, alcohol, energy drinks, and drugs; living environment or sharing room during the academic year).The second part collected information about students’ physical and mental health status and their fear of infection with COVID-19. Self-rated items were constructed, and answers were rated using a five-item Likert-like scale.The third part of the survey inquired about students’ satisfaction with their academic results and relationships with friends, colleagues, and family members, as related to the restrictions imposed by the COVID-19 pandemic.The fourth part collected data regarding students’ opinions about their online activities; their practices and habits during lectures; practical stages that were developed online; the teaching and evaluation process; and the impact of online classes on their academic achievement and results, together with ethical concerns about evaluation.Self-rated items were constructed, and answers were rated using a five-item Likert-like scale.All analyses for this research were performed using the IBM Statistical Package for Social Sciences (SPSS) Statistics for Windows, version 24 (SPSS Inc., Chicago, IL, USA). The descriptive statistics of the socio-demographic- and academic-related data were expressed as means (M) and standard deviations (SD), frequencies, and percentages (%). In order to assess comparative results considering gender, marital status, and academic activity, a Mann–Whitney test was performed. The relationship between variables was assessed using Spearman correlations. The level of statistical significance was set at p < 0.05.The present study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethical Committee of Faculty of Medicine at the University of Oradea, Romania, with the registration No. 12/20.11.2020.Socio-demographic-, academic-, and health-related data were collected from all of the participants enrolled in the undergraduate dental program within the University of Oradea at the time of the research. The majority of the respondents were females, which is in congruency with the worldwide gender distribution in medical studies. For descriptive analyses, subjects were presented by year of study; however, this variable was restructured for comparative results as preclinical (years 1, 2, and 3 of study) and clinical (years 4, 5, and 6 of study) academic years. The descriptive analysis of data is presented in Table 1.Social distancing requirements, restrictions against visiting family members and friends, and the interruption of academic activity at the university or in dental clinics were all reported to have had a great impact on the psychological health status of students. Students were also asked about the activities that they engaged in during the first months of the pandemic in order to cope with stress related to the various restrictions and to keep them healthy. More than one-third of the respondents declared that they had watched documentaries, films, and show series (n = 102, 33.66%); more than one-fifth indicated that they had read books or journals (n = 66, 21.78%); a minority of the group declared that they had cooked (n = 37, 12.21%); and some of the respondents said that they had spent their time doing physical activities (n = 36, 11.88%).Students were asked to self-rate their mental health status regarding the following different problems: sleep, eating, depression, anxiety, and getting bored. The respondents had to indicate their level of agreement with the statements using a five-item Likert-type scale. Answer options included 1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree, and 5—strongly agree. The students’ scores are detailed in Table 2.Six items investigated students’ satisfaction regarding their physical and mental health; their relationships with friends, family members, and colleagues; and their academic results during the first months of COVID-19 restrictions. The answers were rated using a five-item Likert-type scale with response options ranging from 1—totally unsatisfied to 5—totally satisfied.We identified that more than half of the students surveyed were satisfied with their physical health, but that almost 30% were unsatisfied with their mental health. Furthermore, more than 50% were unsatisfied with their relationships with friends and colleagues, while 57% were satisfied with their relationships with family members. The responses are presented in Table 3.During the present research, the academic activities of dentistry students were developed online. The majority of the students connected to the online platforms from their homes (n = 295, 97%), and a low percentage of students indicated that they participated in online activities from other places, where the internet connection was better than that at their homes (n = 8, 3%).The need to move teaching activities to an online environment resulted in determination on the part of the universities to find platforms quickly in order to carry out courses and practical activities in the best possible conditions. Students were asked what platforms they used most often to conduct various school activities, such as those used for online connections for theoretical or practical meetings, the completion of assignments and their submission to teachers, and online courses.A total of 63 students (20.70%) declared that they used the Microsoft Teams University Platform for their academic activity; more than half of the respondents indicated that they used Zoom, Google Meetings, and Skype (n = 202, 66.40%) more often for their online activity, and a small number of them (n = 39, 12.80%) reported that they usually preferred using other ways of communicating with colleagues and teachers (e.g., via text messages, emails, etc.). Moreover, when communicating with their colleagues regarding their academic tasks, dentistry students preferred to use social platforms (n = 299, 98.40%) or Microsoft Teams (n = 5, 1.6%).Where communication with their teachers was concerned, students declared that they predominantly used Microsoft Teams (n = 140, 46.10%), social networks (n = 77, 25.3%), email (n = 50, 16.40%), or communicated through faculty administration staff (n = 37, 12.2%).Most students preferred to use laptops for online activities (n = 191, 62.80%). A lower percentage said they used their smartphones to connect (n = 102, 33.60%), and an even smaller percentage said they participated in online classes using a desktop computer (n = 11, 3.60%).A series of sixteen items were constructed in order to identify dental students’ opinions about online education. The results showed that over two-thirds of the respondents (79.90%) believed that their practical training had been affected due to the use of online activity and that not all subjects could be taught online (73.90%). We also identified that more than half of the students disagreed with the following statements: that the didactic activity was more interactive online than within the institution (44.60% totally disagreed and 18.80% disagreed with the item); that they had preference for online activity over the type carried out within the institution (51.20% totally disagreed and 14.20% disagreed with the item), and that lectures and practical stages were more interesting online.Regarding the involvement of teachers, students appreciated that the academic staff made the courses available to students in the digital format, but they considered that the teachers were not well-trained in using the online platforms and audio–video tools.The analysis of the responses proved that online participation did not encourage students to be active; we found that only 8.80% of students connected with a video-camera. However, 22.40% of students agreed that they could focus better on new information than when they attended the course/laboratory in person.Two items were used to evaluate students’ opinions about the amount of time available for study and hobbies during the restrictions imposed by the COVID-19 pandemic. The results showed that more students indicated that they invested time in hobbies and recreational activities than in studying or academic tasks. The detailed results are presented in Table 4.The e-learning processes and online evaluations were two of the most challenging aspects during the first months of the pandemic, because the teaching staff had to develop and use different types of instruments. Therefore, the students had to adjust to the new academic methods and tools. A number of items aimed to investigate students’ opinions regarding online evaluation as compared to the classic one, as well as the ethical aspects of online evaluation methods.The results proved that 46.20% of students totally disagreed with the statement that online oral evaluation is the most objective method, and that 40% of the respondents totally agreed with the statement that an online multiple-choice exam is the most objective method available for the evaluation of the students.More than 60% of the respondents agreed that the most objective evaluation method was that of the multiple-choice question exam (MCQ) administered at school, with 37.30% of the total students surveyed totally agreeing with that. It is also noteworthy that almost half of the questioned students indicated that they agreed with the opinion that exclusive online assessment of students encouraged unethical behaviors. The detailed results are presented in Table 5.Three items investigated students’ fear regarding the risk of COVID-19 infection. This survey was conducted before their return to school. The respondents had to indicate their level of agreement with the statements using a five-item Likert-type scale rated as follows: 1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree, and 5—strongly agree. The responses showed that students were more afraid of being infected by patients during their internship than of contracting the virus during classes or from classmates. The frequency of answers is presented in Table 6.There were significant differences between women and men in terms of e-learning teaching activity. Men (Mdn = 4.00) had more positive opinions than women (Mdn = 3.00) about the internet connection quality that influenced the quality of the teaching activity (z = −4.643, p < 0.001).The results of the present research showed that, unlike women (Mdn = 2.00), men (Mdn = 3.00) considered that the teaching activity was more interesting in the online environment (z = −2.906, p < 0.001). At the same time, men (Mdn = 2.00) were more involved in the online activity, keeping their video camera connected during the courses (z = −2.906, p = 0.004), as opposed to women.Comparative analyses showed that dental students suffering from chronic diseases (Mdn = 3.00) were less satisfied with the relationships with their family members during the restrictions related to the COVID-19 pandemic (z = −3.261, p = 0.001) than healthy students (Mdn = 4.00). Moreover, the results of the Mann-Whitney test (z = −2.262, p = 0.024) showed that students suffering from chronic diseases noticed that during the pandemic they had more sleep disorders (Mdn = 4.00), unlike the rest of the students (Mdn = 3.00). At the same time, chronically ill students considered that the amount of information taught in the online activity was higher (Mdn = 4.00), unlike healthy students (Mdn = 3.00), (z = −2.390, p = 0.017).Significant differences were identified between the years of study in terms of online teaching. Students in their clinical education years preferred online activities because they were more interactive, more interesting, helped them organize their time better, and they could better focus on the new information received. Additionally, older students had a higher degree of fear regarding the possibility of COVID-19 infection during practical activities within the faculty. The detailed results for these items are presented in Table 7.The comparative analyses also showed significant differences between students from urban areas (Mdn = 3.00) and those from rural areas (Mdn = 4.00), in the sense that students from rural areas lived in a room with more people (z = −4.728, p = 0.000). Significant differences were also registered regarding the occupancy of the room where the students live; thus, unlike the students who shared a room with someone (Mdn = 3.00), students who lived alone (Mdn = 2.00) considered that online courses were not so interesting (z = −3.384, p = 0.001).The correlation analyses showed that there were positive correlations between students’ age and certain behaviors. We identified that as they got older, students consumed more energy drinks (r = 0.116 *, p = 0.044). Age correlated positively with their degree of satisfaction regarding their academic results during the pandemic, in the sense that the advanced year of study students were in, the more satisfied they were with the results they obtained (r = 0.137 *, p = 0.017).Regarding the online development of teaching activities, the results showed that the older the students, the better they could focus on the new information received (r = 0.348 **, p < 0.001) and the more efficiently they could organize their daily schedule (r = 0.288 *, p < 0.001). At the same time, the more advanced year of study students were in, the more they believed that, during the pandemic, the time they spent on studying increased (r = 0.272 *, p < 0.001).Additionally, age correlated positively with students’ feelings of fear, in the sense that as they got older, they were more frightened that they might become infected with COVID-19 during practical college activities (r = 0.193 **, p = 0.001) or from patients with whom they came into contact during practical internships (r = 0.152 *, p = 0.008). We identified negative correlations between the age and physical and mental health of students, in the sense that the older the students, the more they noticed that during the pandemic they had fewer appetite disorders (r = −0.118 *, p = 0.039), felt less depressed (r = −0.128 *, p = 0.026), and felt less anxious (r = 0.145 *, p = 0.011).We identified that cigarette consumption correlated positively with alcohol consumption (r = 0.328 **, p < 0.001) and the number of people in the household (r = 0.209 **, p < 0.001), in the sense that—in the case of smoking students—the more the students smoked, the more alcohol they consumed, and the more people that were in the house, the more students tended to smoke. A negative correlation was identified between cigarette consumption and the amount of time spent on recreational activities during the pandemic (r = −0.119 *, p = 0.039), in the sense that the more students smoked, the less time they spent on hobbies.Our results also proved a positive correlation between alcohol consumption and energy drink consumption (r = 0.132 *, p = 0.021), in the sense that alcohol consumption might increase in direct proportion to energy drink consumption. Moreover, alcohol consumption correlated positively with the degree of satisfaction regarding mental health (r = 0.121 *, p = 0.036) and the degree of satisfaction regarding relationships with family members during the pandemic (r = 0.118 *, p = 0.040), meaning that students with higher alcohol consumption had better relationships with family members and were more satisfied with their mental health. A strong negative correlation was identified between alcohol and drug use (r = −0.193 **, p = 0.001).Strong negative correlations were revealed between the degree of satisfaction with their physical health and several items related to the physical and mental state students felt during the pandemic, such as sleep disorders (r = −0.235 **, p < 0.001), eating disorders (r = −0.227 **, p < 0.001), depression (r = −0.227 **, p < 0.001), anxiety (r = −0.231 **, p < 0.001) and boredom (r = −0.166 **, p = 0.004). These results proved that the more satisfied the students were with their physical condition, the less depressed, anxious, or bored they were, and the fewer sleep disorders or appetite disorders they had. The degree of satisfaction with their mental health correlated positively with the degree of satisfaction regarding their relationships with friends (R = 0.457 **, p < 0.001), colleagues (r = 0.373 **, p < 0.001), and family members (r = 0.359 **, p < 0.001) during the COVID-19 pandemic, in the sense that the better the relationships with colleagues, friends, or family members, the more satisfied students were with their mental health. Furthermore, the degree of satisfaction with their mental health correlated positively with the degree of satisfaction regarding the academic results students obtained during the pandemic (r = 0.371 **, p < 0.001), as well as their degree of concentration on the new information received in online activities (r = 0.179 **, p = 0.002), in the sense that the higher students’ satisfaction with their mental health, the better they could concentrate on online activities and the more satisfied they were with their obtained academic results.Regarding the online academic activity, the correlation analyses showed that the degree of satisfaction regarding the academic results of the students obtained during the pandemic correlated positively with students’ full attendance of their courses or laboratory instruction in the online environment (r = 0.244 **, p < 0.001) and the students’ preference to stay connected with video cameras during the online activity (r = 0.149 **, p = 0.009), in the sense that the more satisfied they were with the results they obtained, the more they tended to stay connected through video during the online activity and to participate fully in the courses. We identified a positive correlation between the degree of satisfaction regarding the academic results obtained and the students’ preference for online activity to the detriment of the traditional instruction carried out within the institution (r = 0.285 **, p < 0.001), in the sense that the more satisfied the students were with their results, the more they preferred to stay at home and attend online courses. At the same time, significant correlations were revealed between the degree of satisfaction with the academic results obtained by the students during the pandemic and the opinion that practical activities (r = 0.317 **, p < 0.001) and courses (r = 0.346 **, p < 0.001) were more interesting in the online format.The results also proved a positive correlation between the fear of COVID-19 infection during practical activities carried out in the college setting and the fear of infection from patients (r = 0.881 **, p < 0.001) or from colleagues (r = 0.688 **, p < 0.001), in the sense that the more that students were scared of contracting the virus at their college, the more they feared that they could become infected from patients or colleagues with whom they came in contact during their teaching activities at the faculty. Moreover, a positive correlation was identified between the fear of infection with the new virus during practical activities at college and the preference for online classes (r = 0.508 **, p < 0.001), in the sense that the more scared students were of COVID-19 infection, the more they preferred to participate in online teaching activities from home.Medicine, regardless of specialization, can only be learned through direct contact with patients. Therefore, during the COVID-19 pandemic, a significant problem that many university centers faced was that clinical internships in hospitals and dental clinics being deeply affected, students were not allowed to enter hospitals because university management did not want to take the risk of exposing students to the disease. Therefore, in many medical specializations, including dentistry, the activity took place in a hybrid manner or completely online.During the period of online activity, we identified that most students connected from home using their laptops, and the majority of the students utilized their audio connection; only an exceedingly small percentage connected using a video camera. This practice limited student–teacher interaction and did not stimulate teamwork or active involvement in the virtual activity. We found that 50.60% of the students attended the course/lab in its entirety when they were connected online, meaning that not using a video connection allowed students to feel free to not always be involved in online discussions. Similar conclusions were pointed out by Jiang et al. [12] in a study conducted among Chinese dental students. The authors showed that during online classes, dental students preferred lecture-based learning and case-based learning more than problem-based learning (PBL), team-based learning, or research-based learning, showing that students preferred a more traditional teaching process and passive participation during online classes. In congruency with these results, Mukhtar et al. also recommended online modalities with a clear lesson plan that reduced cognitive load and increased interaction between students or between students and teachers [34]. However, the pandemic forced the revolutionization of the dental education system through the use of technology, as teachers were determined to find ways to help students adjust to the online learning challenges.We found that students were concerned about the impact of online classes on their level of knowledge and practical skills. More than two-thirds of the respondents considered that their practical training and dentistry skills were affected due to the online activity, and they considered that not all subjects can be taught online, especially during the practical stages of their program. Students in the last years of their study expressed more serious interest in maintaining their practical skills. We identified that students in the clinical years of their program preferred online activities and considered them to be more interactive, more interesting, and more helpful for them in terms of being able to organize their time better, and allowed them to better focus on the new information received in their instruction. We found that the older the students were, the better they could focus on the new information received, the more efficiently they could organize their daily schedule, and the more they believed that the time they spent on studying increased during the pandemic.Even if online teaching is new in dental settings, the opinion of respondents proved that the majority agreed with the statement that the academic staff was well-trained to use the online platforms and audio–video tools. Studies investigating students’ opinions about how teachers mastered online teaching included a range of statements, from satisfaction with teachers’ skills to the opinion that staff could have more training in online learning. For example, Prieto et al. identified that dental students described teachers as showing good predisposition, flexibility, availability, and empathy through the course but also suggested better staff training in online learning [14]. A study conducted by Sarwar et al. [17] among dentistry undergraduate students from medical universities in Pakistan showed that the majority of respondents were dissatisfied with the institutional learning management system, the level of teachers’ training for online lectures, and the quality of the available learning resources. The authors identified that the worst rating was reported for items inquiring about the effectiveness of online classes; freshmen reported the poorest interaction with teachers and a strong disagreement with the effectiveness of online classes. Similar findings were described by Varvara et al. [35], who conducted a cross-sectional survey among Italian dentistry students across all years of study. The results proved that students appreciated the new methods and their teachers’ efforts during online meetings, and mentioned that the lack of practical training was a significant problem.The results of the present study showed that students were concerned about unethical behaviors during online classes and unethical practices during online evaluation. More than 45% of the respondents agreed that the exclusive online assessment of students encouraged unethical behaviors. Similar findings were pointed out by Mukhtar et al. [34]. The authors conducted a survey among undergraduate dentistry students and identified that apart from the advantages, such as remote learning, comfort, and accessibility, teachers who used online learning were aware of the ineffectiveness and the level of difficulty involved in maintaining academic integrity. Ethical concerns are a major problem in medical sciences, with this problem also being identified in studies conducted within other medical specialties. Exam dishonesty appears as one of the major challenges faced with remote e-exams, as it was mentioned by more than one-third of questioned students in a survey conducted by Elsalem et al. [36].Social distance and working from home resulted in a lot of changes regarding lifestyles and social relationships. We identified that, during the first seven months of the COVID-19 pandemic, students were more satisfied regarding their relationships with family members but less satisfied with their relationships with colleagues or friends, and that they had a higher level of concern regarding their mental health than they did for their physical health. Our results showed that students suffering from chronic diseases noticed that, during the pandemic, they had more sleep disorders and believed that the amount of information taught in the online activity was higher when compared with clinically healthy students.The results of the present study showed that our respondents were more satisfied with their physical status than their mental status. After the first seven months of the COVID-19 pandemic, students self-rated as being more anxious, being more depressed, and having more sleep-related problems. Moreover, a large majority of students (over 80%) declared that they were concerned about their practical skills.The correlation analyses proved that the more satisfied the students were with their physical condition, the less depressed, anxious, or bored they were, and the fewer sleep disorders or appetite disorders they had. The degree of satisfaction with their mental health correlated positively with the degree of satisfaction regarding their relationships with friends, colleagues, and family members during the COVID-19 pandemic. We also found that living in an apartment was correlated with a higher consumption of alcohol and cigarettes and less time spent on hobbies and relaxing activities. Similar findings were identified by Prieto et al. [14]. The authors showed that Chilean dental students reported more depression and anxiety. Machado et al. [15] showed that Brazilian students were found to have positive impressions despite technical problems and related stresses. Students mentioned, to a great extent, that the time available for recreational activities/hobbies had increased, and their time devoted to studying had decreased. Similar findings were presented by Prieto et al. [14] who showed that Brazilian dental students had the opportunity to spend more time with families, but that staying at home for an extended period of time made them more stressed and anxious. Among other negative aspects, the authors identified tiredness, loneliness, and the need for an appropriate workspace. Similar results were presented by Škrlec et al. [37], who identified that more than half of the Croatian dental students reported increased stress levels, decreased physical activity, increased depression, and increased anxiety, and that one-third of students were insufficiently active during the second COVID-19 lockdown.We found that the higher satisfaction with their mental health was, the better students could concentrate on online activities and the more satisfied they were with the academic results they obtained. We can conclude that during the COVID-19 pandemic, the strong relationship between mental health, academic results, and the level of satisfaction with academic activity was very well highlighted. Thus, during the lockdown, students who found satisfaction in studying considered themselves physically and mentally healthier and found strategies to better cope with stress.The present research presents a broad description of the student reality in a pandemic context, which could be relevant for detecting problems and making methodological and curricular decisions within the context of the training of future dentists. The present results are useful for students to understand the particular context in which they had to struggle, and also for university teachers in order to adapt their curricula and methods to meet students’ expectations as well as their own standards of teaching.There are some strengths of the present research. The number of participants was beneficial for cross-sectional study, and important conclusions by gender, environment, year of study, and level of satisfaction with physical and mental health were able to be generated. Additionally, the results are important for both students and teachers working in dental institutions for across various issues related to online activity: academic tasks, evaluation methods, learning, relationships, mental and physical status, ethical concerns, and social aspects. Therefore, from this point of view, the present study presents important results for all actors in the academic field (i.e., students, teachers, and administrators).The limitation of this research is due to the regional context, meaning that the results could be influenced by the number of infections in the area and the severity of the lockdown restrictions. So, when results are considered in comparative analyses, researchers must also take into consideration the type of activity (i.e., partially online or totally online) when they proceed in drawing conclusions.The COVID-19 pandemic can be seen as a kickoff point for the motivation to invest in the technological innovation necessary to deliver the best possible education to our future dentists—even during the post-pandemic period—to ensure a higher level of satisfaction and stronger confidence in the preparedness of dental students for their future profession. The psychological impact of the online activity must not be neglected; teachers and psychologists must be aware of the strategies that have to be applied during online relationships in order to stimulate communication and increase the active participation of students.Conceptualization, R.I. and M.I.; methodology, M.I.; formal analysis, M.I.; investigation, R.I.; resources, L.-M.P.; data curation, R.I.; writing—original draft preparation, M.I. and L.-M.P.; writing—review and editing, R.I. and L.-M.P.; supervision, M.I. and R.I. All authors have read and agreed to the published version of the manuscript. All authors equally contributed to this manuscript.This research received no external funding.This study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Faculty of Medicine, University of Oradea, Romania, with the registration No. 12/20.11.2020.Informed consent was obtained from all subjects involved in the study.The data presented in this study are available on request from the corresponding author.The authors declare no conflict of interest.Study profile.Sociodemographic and medical characteristics 1.1 Number of answers (N) and corresponding percentages (%)/means and standard deviations (M ± SD).Self-rated items regarding psychological status 1.1 Number of answers (N) and corresponding percentages (%)/means and standard deviations (M ± SD); 1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree, and 5—strongly agree.Satisfaction regarding relationships with family, friends, and colleagues and academic results 1.1 Number of answers (N) and corresponding percentages (%)/means and standard deviations (M ± SD); 1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree, and 5—strongly agree.Dentistry students’ opinions regarding online activity 1.1 Number of answers (N) and corresponding percentages (%)/means and standard deviations (M ± SD).; 1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree, and 5—strongly agree.Self-rated items regarding students ‘opinions related to online evaluation 1.1 Number of answers (N) and corresponding percentages (%)/means and standard deviations (M ± SD); 1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree, and 5—strongly agree.Items investigating the fear of COVID-19 infection 1.1 Number of answers (N) and corresponding percentages (%)/means and standard deviations (M ± SD); 1—strongly disagree, 2—disagree, 3—neither agree nor disagree, 4—agree, and 5—strongly agree.Opinions of dental students regarding online activity. Differences by preclinical versus clinical years 1.1 Mann-Whitney analysis results.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06042.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Padel is becoming one of the most widespread racket sports that may have potential health benefits. Considering that several myokines mediate the cross-talk between skeletal muscles and the brain, exerting positive effects on brain health status, this study was designed to evaluate the responses of brain-derived neurotrophic factor (BDNF), leukemia inhibitory factor (LIF), and irisin (IR) to padel competition in trained players and to determine whether these responses were sex-dependent. Twenty-four trained padel players (14 women and 10 men with a mean age of 27.8 ± 6.3 years) participated voluntarily in this study. Circulating levels of BDNF, LIF, and IR were assessed before and after simulated padel competition (real playing time, 27.8 ± 8.49 min; relative intensity, 75.2 ± 7.9% maximum heart rate). Except for BDNF responses observed in female players (increasing from 1531.12 ± 269.09 to 1768.56 ± 410.75 ng/mL), no significant changes in LIF and IR concentrations were reported after padel competition. In addition, no sex-related differences were found. Moreover, significant associations between IR and BDNF were established at both pre- and post-competition. Our results suggest that while competitive padel practice stimulates BDNF response in female players, padel competition failed to boost the release of LIF and IR. Future studies are needed to further explore the role of these exercise-induced myokines in the regulation of brain functions and to identify the field sports that can contribute to myokine-mediated muscle–brain crosstalk.Padel is a particular racket sport that has grown in popularity and is currently being practiced by millions of people worldwide [1]. Padel is played in doubles on a rectangular court (20 × 10 m) divided into two fields by a central net that is lower than the tennis one. Moreover, this court is completely surrounded by a suitable combination of concrete walls and fence wire, which prevent the ball from exiting the playing area [2,3].As in other racket sports, several studies have been focused on the assessment of players’ court-movement patterns and the physiological demands of padel competition. Cardiorespiratory responses (oxygen consumption and mean heart rate) as well as perceived exertion rates are similar to those previously found in tennis. However, they are lower compared to those assessed in squash and badminton [1]. More specifically, and considering that padel practice is characterized by alternated intervals of intense and moderate-low exercise intensity, the mean VO2 measured during an official competition (lasting around 1 h) reached values below 50% of maximum VO2 (VO2max), whereas the mean HR represented approximately 74% of maximum HR (HRmax) [2]. Thus, moderate energy expenditure (with aerobic metabolism as the main energy source) and an easy and accessible technique seem to be the two key factors behind the extended practice of padel.However, although padel practice keeps increasing there remains a lack of information about its impact on players’ health. According to the reviewed scientific literature, there are not many studies dealing with the physiological, health-related effects of high-level padel competition [2]. To our knowledge, only one very recent study has been aimed at analyzing the changes in hematological and biochemical parameters induced by competitive padel practice. As can be observed, high-level padel competition provokes a significant increase in muscle damage biomarkers (e.g., creatine kinase) as well as remarkable decreases in blood electrolytes concentrations [4]. Nevertheless, apart from these intensity-related effects it would be necessary to conduct studies focused on determining the stimulating effects of competitive padel practice on different health-related benefits through specific biomarkers.Brain-derived neurotrophic factor (BDNF) is an important neurotrophin that plays important roles in the plasticity of several regions of the central nervous system (CNS) during development, adulthood, and aging [5]. Nevertheless, it has been suggested that BDNF is expressed in non-neurogenic tissues (including skeletal muscle), so that BDNF may play a role not only in CNS plasticity but also as a metabolic regulator of skeletal muscle (enhancing glucose consumption and fat oxidation). In any case, of all neurotrophins, BDNF seems to be the most susceptible to regulation by exercise and physical activity [6]. In fact, the response of BDNF to acute exercise has been investigated using different exercise protocols and, consequently, different results have been reported (from a lack of response to increases anywhere between 11.7 and 410.0% with respect to basal levels) [7].The skeletal muscle acts as a secretory organ that, in addition to BDNF, produces cytokines and other muscle fiber-derived peptides called myokines [8]. In general, myokines are muscle-derived molecules that exert physiological functions on maintaining systemic homeostasis. Thus, myokines regulate whole-body metabolism, bone growth, satellite cell proliferation, and muscle hypertrophy in an autocrine, paracrine, or endocrine manner [9,10]. Irisin (IR) and leukemia inhibitory factor (LIF) are two novel myokines that are associated with brain and muscle adaptations. A recent study has reported that IR is involved in whole-body metabolism regulation by stimulating FFA, oxidation, and lipolysis and inducing fat browning. Moreover, IR stimulates glucose uptake and regulates muscle growth [11]. Although the primary source of IR is skeletal muscle, another source of this myokine is the brain. Therefore, IR could play roles in mediating the effects of physical activity on the brain [12]. On the other hand, LIF is produced by skeletal muscle and affects intact muscles as well as isolated muscle cells. Among its various roles, the most important role of LIF in muscle satellite cell is the proliferation of proper muscle hypertrophy and regeneration [10,13].However, considering the beneficial effects of BDNF, LIF, and IR on brain health status, there is a lack of information about the magnitude of their responses to field sports activities [14]. As it has been reported in a recent study, the amount of evidence for the effects of exercise on the blood concentration of BDNF is moderate [15]. Moreover, studies on the response of IR to exercise have not been conclusive [16], and although LIF seems to play an autocrine role within the skeletal muscle tissue, it has been difficult to determine changes in their circulating levels in response to exercise.Thus, taking into account the increasing popularity of padel and the potential brain health effects of BDNF, LIF, and IR, this study was designed to evaluate the responses of these myokines to padel competition in trained players and to determine whether they are sex-dependent.A total of 24 trained padel players (14 female and 10 male young-adult players) with more than five years of experience in the professional circuit World Padel Tour participated voluntarily in this study. The characteristics of the subjects can be seen in Table 1. All of the participants gave their informed consent for inclusion prior to participation. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the Department of Health and Consumption of the Government of Aragon, Spain (code: 21/2012; date: 19 December 2012).The calculations for sample size and power were based on BDNF responses to moderate and vigorous exercise reported by a previous study [17]. Considering the large effect sizes (ES) shown by these authors (e.g., Cohen’s d range of 0.63–1.16), the a priori sample size calculation (G*Power v.3.1) with ES = 0.64 established that a sample of 22 would be sufficient to obtain a statistical power of 0.8 (p < 0.05). Therefore, our sample size of 24 allowed us to overcome a power of 85%.Participants were involved in two separate testing sessions with at least seven days between them (Figure 1). In the first session, subjects’ body composition analysis was assessed (bioelectrical impedance, TANITA BC–418MA, Amsterdam, The Netherlands) and a graded exercise test was also performed. The second session consisted of participating in a simulated competition following the International Padel Federation rules.For each session (conducted between 9:00 a.m. and 12:00 a.m.), participants were instructed to avoid strenuous physical activity during the previous 24 h and to abstain from food (overnight fasting), caffeine, and alcohol 12 h before testing. Padel matches were held on outdoor courts with a relative environmental humidity and temperature of 45.7 ± 7.3% and 24.1 ± 7.1 degrees Celsius, respectively.Players’ maximum oxygen consumption (VO2max) and maximum heart rate (HRmax) were assessed in the first session using an incremental running test on a treadmill (Pulsar HP Cosmos, Nussdorf, Germany) equipped with a gas analyzer (Oxycon Pro. Jaegger, Germany) and heart rate monitor (Cosmos, Nussdorf, Germany). After a warm-up period of 5 min of brisk walking (6 km/h), the initial speed was set at 8 km/h, increasing by 1 km/h every minute until volitional exhaustion. The treadmill slope was kept at 1°. VO2max was defined following the ACSM criteria [18], whereas HRmax determined as the highest HR value observed during the running test.All matches were played to the best of three sets. If the situation of six equal games was reached, a tie breaker was played. Before each match, players performed a standardized 15 min warm-up. Total playing time (TPT, full time of the match from the beginning to the end, considering the periods of game and rest), total resting time (TRT, sum of periods in which the ball was not in play), and real playing time (RPT, total playing time minus total resting time) were measured (s) for each match. Moreover, players’ HR was continuously recorded during the competition (Polar Team System, Kempele, Finland) as average values over 5 s.Blood sampling. In the second session, two 5-mL blood samples (pre- and post-padel competition) were drawn from the antecubital vein of each participant. Blood samples were collected in Vacutainer tubes (BD Vacutainer, Plymouth, UK) containing ethylenediaminetetraacetic acid (EDTA) as an anticoagulant. The first sample was collected 90 min before the competition (fasting conditions), and the second was collected within 10 min after the matches. Blood samples were immediately put on ice and transferred to a laboratory for processing.Hematological and biochemical assessment. Hematological parameters (red blood cells, hematocrit, and hemoglobin) were determined using the Coulter counter model Ac·T diff (Beckman Coulter Inc., Brea, CA, USA). BDNF concentrations were measured using the BDNF Exam Immunoassay System kit (R&D Corp., Minneapolis, MN, USA) according to the manufacturer’s instructions. Whole blood LIF concentrations were determined by enzyme-linked immunosorbent assay (ELISA) using a human LIF ELISA kit (PromoKine, Heidelberg, Germany) following the manufacturer’s instructions. On the other hand, IR blood levels were measured using a commercial kit (Wuhan Fine Biological Technology Co., Ltd., Wuhan, China). Absorbances for LIF and IR were measured in duplicate using a spectrophotometric microplate reader at a wavelength of 450 nm (Biotek, Winooski, VT, USA).Lastly, considering that the players were allowed to hydrate ad libitum during the competition (with bottled mineral water), changes in plasma volume were calculated using hematocrit and hemoglobin values following the methods of Dill and Costill [19]. Moreover, blood levels of BDF, LIF, and IR were individually corrected according to the formula described by Matomäki et al. [20].The data are expressed as mean ± standard deviation (sd). The Shapiro–Wilk test was applied to test for a normal distribution of variables. Parametric and non-parametric tests (ANOVA-time × sex interaction-, Wilcoxon signed-rank, and Mann–Whitney tests) were used where appropriate to determine both intragroup and intergroup differences between the pre- and post-competition time points. Moreover, effect size (ES) was calculated using both partial eta squared (η2p) and the d-value proposed by Cohen [21]. Thus, the ES was interpreted as trivial when η2p < 0.01 and d < 0.19; small when η2p = 0.01 and d = 0.20; medium when η2p = 0.06 and d = 0.50; and large when η2p = 0.14 and d = 0.80. Bivariate correlations were also performed using both Pearson’s r and Spearman’s rho, which was set at 0.500 for a positive correlation. For all tests, a p-value of <0.05 was used to indicate statistical significance.Body composition variables showed significant sexual dimorphism. Moreover, VO2max was significantly higher in male than in female players (Table 1).As it can be seen in Table 2, one of the most peculiar characteristics of padel competition was the 1:1.5 ratio established between TPT and TRT. Although TPT and TRT showed sex-related differences (both variables were higher in males), no differences were found in RPT.On the other hand, HRmean and HRmax during padel competition were similar in both males and females. Likewise, considering that HRmax measured during maximal exercise test did not report sex differences, the percentage of HRmean on reference HRmax (graded exercise test) did not show any statistical significance. Regarding PV changes, a slight but not statistically significant increase (+1.1 ± 2.3% for the total group; CI-95%: −0.6 to 3.1) was observed (Table 2).Although no sex-related differences were found, padel competition induced a significant increase in circulating BDNF levels in female players (from 1531.12 ± 269.09 to 1768.56 ± 410.75 ng/mL; CI-95%: −507.20 to 32.32; Z = −2.27, p < 0.05, d = 1.527). In contrast, BDNF concentrations measured in males after exercise were lower than those found before (1523.01 ± 307.10 and 1295.51 ± 288.88 ng/mL, respectively). However, no significant differences were observed (CI-95%: −52.61 to 507.61; Z = −0.866, p = 0.186, d = 0.476; Figure 2).Regarding LIF responses, no significant differences were observed when sex (F(3,43) = 0.590; p = 0.447, η2p = 0.014), time (F(3,43) = 0.004; p = 0.952, η2p < 0.01), and sex × time interaction (F(3,43) = 0.318; p = 0.576, η2p < 0.01) analyses were performed. Nevertheless, post-exercise LIF concentrations showed a slight decrease in females (from 8.48 ± 5.25 to 7.28 ± 3.76 ng/mL after competition; CI-95%: −2.35 to 4.75) but a nonsignificant increase in males (from 5.91 ± 3.95 to 6.88 ± 4.46 ng/mL, respectively; CI-95%: −4.93 to 2.99) (Figure 3).On the other hand, intergroup analysis revealed no significant sex-related differences for IR responses (CI-95%: −158.64 to 111.78, Z= −0.205, p = 0.837, d = 0.084 and CI-95%: −142.48 to 92.88, Z = −0.312, p = 0.755, d = 0.096 before and after padel competition, respectively). Moreover, unlike BDNF and LIF, the same decreasing trend of IR levels was observed in both female and male players after padel competition. Nevertheless, no significant differences regarding pre-competition levels were found (CI-95%: −89.34 to 118.49, Z = −0.594, p = 0.552, d = 0.322 for females, and CI-95%: −142.36 to 168.68, Z = −0.051, p = 0.959, d = 0.027 for males; Figure 4).Finally, a correlation analysis reported significant associations between BDNF and IR for the entire group in both before (rho = 0.461; p = 0.024), and after competition (rho = 0.665; p < 0.001).The main aims of this study were to evaluate the responses of BDNF, LIF, and IR to padel competition in trained players and to determine whether these responses were sex-dependent. To our knowledge, there have not been many studies evaluating the responses of neurotrophic factors and specific myokines to competitive sports practice. As has been indicated, padel is increasingly practiced by more people, so it is important to define all of its potential health benefits.According to previous findings, as was expected, the padel players evaluated in our study showed sexual dimorphism in body composition [22] and cardiorespiratory variables [23]. Moreover, sex-related differences were also observed in padel competition characteristics. TPT and RPT were higher in male than in female players. However, RPT did not show any differences between groups. While these sex-related effects were in accordance with the findings of a very recent study [4], they were contrary to those reported by Torres-Luque et al. [24], who found higher TPT and RPT in female players. In any case, TPT in our study was established approximately between 55 and 80 min, which is in line with these previous reports.On the other hand, we assessed cardiovascular responses to padel competition using HRmean during matches and its percentage on HRmax measured during a graded exercise test. Our results are similar to those previously described by Castillo-Rodriguez et al. [1], who reported a HRmean equivalent to 77% of HRmax during padel competition. Only one previous study [4] has considered the impact of padel practice on the hemodynamic variables that could modulate other physiological and/or biochemical responses in padel players. Accordingly, it seems necessary to check PV changes associated with intense outdoor exercise, especially if athletes need to hydrate to replace the water lost through sweating and to maintain an adequate thermoregulation. Thus, taking into account that our padel players kept hydrated during the competition, it was necessary to calculate PV changes to avoid any hemoconcentration or hemodilution that could affect the results. Nevertheless, although a mean increase of 1.1% in PV was measured after padel competition, biochemical data were individually corrected.Biochemical analysis was focused on BDNF, LIF, and IR, a group of myokines that, among others, could explain the underlying biological mechanisms of neuroprotective and regenerative potential effects of exercise in both CNS and the periphery [25]. In fact, increased levels of multiple myokines that have beneficial endocrine effects play crucial roles in the interactions between skeletal muscle and other organs in response to exercise [11]. Nonetheless, our study showed that, with some exceptions, padel competition fails to induce remarkable changes in circulating levels of these biomarkers. Moreover, it is important to note that their responses were characterized by large interindividual variability.The response of BDNF to acute exercise has been investigated by several authors using different exercise protocols and, consequently, different results have been reported [7]. In our study, padel competition induced only a slight but significant increase in the circulating BDNF concentrations of female players. This attenuated response is in line with those observed in previous studies in which low or moderate-intensity exercise was used [26,27]. The magnitude of BDNF increase seems to be exercise intensity-dependent, since high-intensity exercises induced huge BDNF responses [7,28], whereas low or moderate-intensity exercises were insufficient to do so [26]. Thus, padel competitions that consisted of 55–80 min of discontinuous exercise (RPT from 22 to 33 min) performed at 72–77% of HRmax may fail to stimulate large BDNF responses.Although LIF is a myokine mainly associated with muscle regeneration, various studies have demonstrated the importance of LIF at various stages of neurogenesis and its involvement in both neuronal cell differentiation and neuritic outgrowth [29]. Previous studies demonstrated that aerobic exercise induces the expression of LIF in human skeletal muscle [13,30]. However, our results reported no changes in LIF circulating levels after padel competition. Moreover, we did not observe differences regarding the players’ sex. Similar results were found by Donnikov et al. [31], who reported a slight decrease in LIF concentrations in a group of athletes after a six hour marathon ultra-race. This lack of LIF responses to exercise could be explained by many different hypotheses. First, as it has been previously observed, exercise-induced LIF responses are characterized by a remarkable interindividual variability. In fact, both increases and decreases in circulating post-exercise LIF levels have been measured in athletes [31]. Second, LIF seems to be muscle-specific, as LIF was undetectable in plasma. It is possible that LIF is secreted to the interstitial space between muscle fibers and does not easily reach circulation. Third, considering the short blood half-life of LIF (6–8 min), it could be difficult to detect accumulated circulating levels of LIF protein during prolonged exercise [32]. Finally, it seems that resistance exercises (mainly eccentric muscle contractions) regulate LIF secretion better than aerobic exercises [13], which would also explain the attenuated LIF responses observed here.Previous studies have confirmed that IR may stimulate both neuronal proliferation and differentiation [33,34]. Furthermore, IR contributes to the neuroprotective effect of exercise against brain disorders [35]. Although the primary source of IR in humans is skeletal muscle, which produces over 70% of the total circulating level of IR, other non-muscle sources of this myokine (such as the brain) play roles in mediating the effects of physical activity on the brain [12]. In any case, previous results reported that plasma IR levels were elevated in humans following exercise [36]. Thus, it was to be expected that circulating IR levels were increased after padel competition. However, we observed no effect of exercise on IR blood concentrations, since post-exercise values were similar to those measured before padel matches. At the same time, no sex-related differences were detected when both pre- and post-exercise IR levels were contrasted between male and female players. These results are in line with those obtained by Pekkala et al. [37] and Tsuchiya et al. [38], who did not find changes in IR concentrations after 1 h of aerobic cycling at intensities of 50% and 65% VO2max, respectively. In this sense, it seems that exercise-related IR response is intensity-dependent, since various studies have demonstrated significant IR acute responses when both trained and untrained subjects performed high-intensity exercises [39,40]. Moreover, after comparing resistance exercise with interval- and endurance-type exercise protocols, Huh et al. [41] reported that, as occurs with LIF, strengthening exercise induced a greater IR response than continuous cardiovascular ones.Nevertheless, contrary to the findings of Briken et al. [25], correlation analyses revealed significant associations between IR and BDNF in both pre- and post-exercise evaluations. On the basis of these associations, it is important to consider that exercise-related IR response is stimulated by an increased expression of PGC1-α, a regulator of mitochondrial biogenesis. PGC1-α is an inducer of fibronectin type III domain-containing protein 5 (FNDC5) expression, a single-pass membrane-spanning protein. Upon exercise, the ectodomain of FNDC5-IR is released into the bloodstream. Interestingly, FNDC5 was shown to mediate beneficial CNS effects of endurance exercise by upregulating BDNF expression [42]. In any case, future studies should explore this type of relationship and its potential benefits for brain health.Lastly, as with the majority of studies performed on sports competition, there are some limitations inherent to the design used that should be noted. First, simulated padel competition could be quite different from real padel competition. Second, taking into account the interindividual variability shown by the outcome variables, the sample size was relatively small. However, considering the recruiting difficulties inherent to competitive athletes, the participants in our study were homogeneous in terms of padel competitive category and training status.Our results suggest that competitive padel practice induces a slight but significant response of BDNF in female players. However, padel competition failed to stimulate the release of LIF and IR. Padel competition characteristics and relative playing intensity could explain this lack of stimulating effect. In addition, IR and BDNF showed an interesting association that needs to be studied in future research. Nevertheless, with respect to practical applications the findings of our study suggest that padel could be included as a part of programs promoting brain health, especially for women.Conceptualization, F.P., M.P.C., I.C.M.-D. and L.C.; methodology, F.P., M.P.C., I.C.M.-D. and L.C.; investigation, F.P., M.P.C. and I.C.M.-D.; formal analysis, F.P., M.T.N., L.C., M.P.C. and I.C.M.-D.; writing—original draft preparation, F.P., M.P.C. and I.C.M.-D.; writing—review and editing, F.P. and M.P.C.; supervision, F.P. and L.C. All authors have read and agreed to the published version of the manuscript.This research received no external funding.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Department of Health and Consumption of the Government of Aragón, Spain (code: 21/2012).Informed consent was obtained from all subjects involved in the study.The data presented in this study are available upon request from the corresponding author.The authors gratefully acknowledge the technical support of the Sport Medicine Lab, Huesca, Government of Aragón.The authors declare no conflict of interest.Study protocol.Blood BDNF concentrations (ng/mL) measured before and after padel competition. * p < 0.05.Blood LIF levels (ng/mL) measured before and after padel competition.Blood IR concentrations (ng/mL) measured before and after padel competition.Padel players’ characteristics.CI, confidence interval; BMI, body mass index; VO2max, maximum oxygen consumption; HRmax, maximum heart rate measured in the graded exercise test. Numbers in brackets represent the mean 95% CI of the mean. Italics are used to highlight statistical significance.Characteristics of simulated padel competition.CI, confidence interval; TPT, total playing time; RPT, real playing time; TRT, total resting time; HRmean, mean heart rate assessed during padel competition; HRmax, maximum heart rate assessed during padel competition; %HRmax, percentage of HRmean on reference HRmax (graded exercise test); PV, plasma volume. Numbers in brackets represent the 95% CI of the mean. Italics are used to highlight statistical significance.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06043.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: There is limited data on the association between diarrhoea among children aged under five years (U5D) and water use, sanitation, hygiene, and socio-economics factors in low-income communities. The study investigated U5D and the associated risk factors in the Zeekoe catchment in Cape Town, South Africa. Methods: A cross-sectional study was conducted in 707 households in six informal settlements (IS) two formal settlements (FS) (March–June 2017). Results: Most IS households used public taps (74.4%) and shared toilets (93.0%), while FS households used piped water on premises (89.6%) and private toilets (98.3%). IS respondents had higher average hand-washing scores than those of FS (0.04 vs. −0.14, p = 0.02). The overall U5D prevalence was 15.3% (range: 8.6%–24.2%) and was higher in FS than in IS (21.2% vs. 13.4%, respectively, p = 0.01). Water storage >12 h was associated with increasing U5D (OR = 1.88, 95% CI 1.00–3.55, p = 0.05). Water treatment (OR = 0.57, 95%CI 0.34–0.97, p = 0.04), good hand-washing practices (OR = 0.59, 95%CI 0.42–0.82, p = 0.002) and Hepatitis A vaccination (OR = 0.51, 95%CI 0.28–0.9, p = 0.02) had significant preventing effects on U5D. Conclusions: The study highlights that good hygiene practice is a key intervention against U5D in informal settlements. The promotion of hand-washing, proper water storage, and hygienic breastfeeding is highly recommended.In 2017, diarrhoea was among the leading cause of death among all ages globally (more than 1.7 million deaths) and the fifth leading cause of death among children younger than five years with more than 0.44 million deaths) [1]. According to the World Health Organization (WHO) 2017 report, more than a quarter (26.9%) of diarrhoea deaths occurred among children younger than five years, and about 90% of diarrhoea deaths occurred in South Asia and sub-Saharan Africa [2]. Statistics South Africa (Stats SA) estimates that 20% of under-five deaths may be attributed to diarrhoea [3], with 31,436 diarrhoea cases recorded in 2016 [4], while other sources estimate the mortality rate between 8%–24% [5,6].There is an unequal geographical distribution of diarrhoea burden among children younger than five years of age across Africa [6]. Children in low- and middle-income countries (LMICs) are the most vulnerable to diarrhoea. More than half of all the diarrhoea-related deaths among children in Africa were estimated to occur in 7.0% of the first-level administrative subdivisions (i.e., states, regions or provinces depending on the country) on the continent [6]. Diarrhoea is closely associated with environmental and socio-economic conditions, with the impoverished communities being most affected [7]. In 2016, over 75% (101/133) of diseases or disease groups listed in the Global Health Observatory had significant links with the environment, and more than one-quarter of the 6.6 million under-five child deaths were associated with environment-related causes and conditions [8]. Among diseases contributing to the environmental burden of diseases, diarrhoea diseases count for 22% and parasitic and vector-borne diseases for 12% [9].Informal Settlements (IS) in Africa are synonymous with public health challenges as a result of dense population and poor sanitation [10,11,12]. The urban population growth rates in Africa are the highest in the world, and it has been projected that by 2050, Africa’s cities will be home to an estimated 950 million people [13]. In South Africa, 3.5 million out of the 17 million households (20%) were living in rural and informal settlements. In 2013, the Cape Town metro alone had 204 informal settlements with more than 200,000 households estimated to be living in conditions where fire and sanitation are among major health hazards, and this number increases each year [14].In 2016, the population of Cape Town was estimated at 4,617,560 inhabitants with a 2.40% annual growth rate, in which 17% of the population was located in informal dwellings [12]. Many of these informal settlements are located on low-lying, flood-prone areas or steep slopes with poor water and sanitation services. Generally, those informal settlements are characterized by dense populations, unhygienic conditions, improper disposal of waste, and poor sanitation. With such living conditions, citizens who live in those areas are particularly vulnerable to diarrhoea. Additionally, a lack of access to potable water supplies, contamination risks associated with the water sources, poor sanitation, poor drainage, presence of contaminated grey water, and inadequate stormwater management, among other factors, may increase the risks for diarrhea—especially for children under five. During 2012–2015, there were an estimated 25,000–30,000 diarrhoea cases each year in children aged less than five years [15]. In 2017, Nerse et al. found the one and two-week prevalence of diarrhea among under-five children in informal settlements surrounding Cape Town were 13.22% and 16.9%, respectively [16].Diarrhoea is a major public health issue in the Western Cape, South Africa, with a seasonal variation that has a peak between March and June every year. However, the sources and burden of early childhood diarrhoea are multidirectional and difficult to truly ascertain [17,18], and further, less is known about the situation in informal settlements. In this study, we set out to assess under-five childhood diarrhoea and its risk factors in the informal settlements of the Zeekoe catchment area. The specific objectives were: (i) To determine the risk factors for diarrhoea in informal settlements; (ii) To identify the distribution of diarrhoea in children under five years in this setting; and (iii) To recommend prevention and control measures for children’s diarrhoea in the low-lying informal settlements in Cape Town.The study protocol was approved by the “Ethikkommission Nordwest—und Zentralschweiz (EKNZ)” in Switzerland (EKNZ BASEC 2016-00304) and by the “Human Research Ethics Committee (HREC)” of the University of Cape Town in South Africa (HREC REF 248/2016). All the interviews were conducted after the head of the household agreed to participate in the survey and informed consent forms were signed.The study was conducted in six informal and two formal settlements of Cape Town, located in the Zeekoe Catchment in the Cape Flats (a low-lying area of Cape Town). The map of the catchment study area is represented in Figure 1a–c. The catchment area overlays the Cape Flats Aquifer (CFA), a potential groundwater resource, and is marked by various land uses, precarious socio-economic conditions, black and grey water flows from settlements and industries, and a wastewater treatment plant (WWTP) that discharges treated effluent to the sea (Figure 1).The Zeekoe Catchment area consists of a wide range of communities, with many formal settlements blended into a poor informal settlement and sharing the same polluted premises [19]. Six informal settlements in the catchment selected in a previous phase of the study were also selected for this study, namely: Barcelona, Lotus Park, Phola Park, Sweet Home, Weltevrede, and Pelican Park (Figure 1c). In addition, we selected nearby formal houses from Gugulethu and Manenberg. The selection of these sites was based on the following criteria established with the civil society actors working in the areas: (i) Accessibility of the area; (ii) Interviewer’s availability for the survey in the settlement; and (iii) Security issues. Of note, all the settlements from the catchment considered for inclusion in the study were high-risk crime areas, and thus, it was important to minimize the threat to the safety of the field staff. Areas were visited if site facilitators were available to accompany field staff. Houses in the selected areas were selected opportunistically.The sample size was chosen for use in a Monte Carlo simulation. A sample size of approximately 90 households per each settlement was selected, assuming that the assumed prevalence of childhood diarrhoea in study sites was 40% with a standard deviation of 5%, and the ratio between exposed and non-exposed groups was 1:1. ‘Diarrhoea’ was defined as “having three or more loose or liquid bowel movements over a 24-h period, as reported by the mother or caregiver of the child, at any point during the seven days preceding the interview” [7].Data were collected from March to June 2017. The questionnaire was developed in English and translated into Xhosa and Afrikaans for appropriate communication within the local community, given that many residents are not fluent in the English language. A pilot study was carried out three days before the beginning of the household survey to identify possible shortcomings of the questionnaire and to verify the anticipated duration of the interview. The questionnaire covered the following sections: (i) basic demographic information, (ii) socio-economic status, (iii) water source, (iv) sanitation and hygiene, and (v) child health status. The survey questions were addressed to the head of the household or its legal representative at the time of the interview. The questionnaires were administered by six groups of trained interviewers consisting of two persons per group. The anonymity of the respondents was preserved throughout the course of the research.Statistical analyses were carried out using Stata14 software (StataCorp LP, Lakeway, TX, USA). Using factor analysis, a wealth index was created based on possessions of household items, and a hand-washing index was created based on hand-washing. All variables were screened for irregularities and outliers. Basic descriptive statistical analyses, such as median, mean, interquartile ranges, standard deviation (SD), and confidence interval (CI), were obtained for the relevant continuous variables, and a frequency table was generated for other variables.Due to the hierarchical structure of collected data where child’s data being nested in the household and household data in communities, multi-level logistic modelling was performed [20]. Mixed-effects logistic regression modelling was used to examine the effects of child-related variables (level 1) and household-related variables (level 2) on childhood diarrhoea.A total of 707 households participated in the survey conducted in six informal settlements and two nearby formal settlements. There was a total of 778 children aged 0–5 years old in the recruited households. The demographic and health-related information of those children stratified by living location is shown in Table 1. The mean age of the children was 28.8 months (SD = 17), and 51.9% were female. While the gender of children was distributed equally in the IS, more females than males participated in FS (58.2% vs. 41.8%, respectively). The vast majority of caretaker respondents (>90%) were female. The median number of people per household in FS was 5 (SD: 2.2, range 1–18). It was also observed that residents of FS stay longer in their current address relative to those from the IS with an average duration of (23 years vs. 8 years, p = 0.005). Although FS families had a statistically significantly higher mean wealth index than those living in IS (p = 0.05), there was not much difference in the proportion of children going to daycare (47.1% vs. 47.3%, p = 0.99).The majority of children in participating households had access to potable water (>97%) for domestic use. In IS, 74.4% of households used public taps as the main source of drinking water (Figure 2a). All residences of IS had access to water predominantly from a public tap, with the exception of Pelican Park that had piped tap water inside the yard as the main water resource (Figure 2b). In contrast, piped tap-water on the premises (89.6%) was the most common water resource in formal areas (Figure 2c).Most IS residents reported storing water either in plastic bottles or containers (96.3%), with the water storage duration exceeding 12 h in 13.1% of households. FS residents were more likely to store water <12 h than those in IS (91.3% vs. 76.2%, Chi-square χ2 =26.9, p < 0.001). Only a small proportion (0.6%) of the caretakers, all from IS, reported that they need over 30 min to get to the water source. A general overview of access to water resources of participating households during the study is represented in Table 2.Most households (98.3%) reported access to latrines, but only 150 households (21.2%) had access to a flush toilet system. The proportion of IS households having access to an improved sanitation facility (modern toilet, potable water, waste disposal) was considerably lower than those of FS households (5.0% vs. 75.0%, Chi-square test, p < 0.001). Shared facilities were the most common in all IS (93%), while a private toilet was the most common among FS households (75%) (Figure 3a). The mean number of households that shared a sanitary facility in most informal households was 4 (range 1–11). Most of the sanitary facilities were located less than 50 m from the houses (94.8%). While flush toilets were the most frequently reported facility in formal areas (>96%) (Figure 3c), facility types varied across informal sites, with the flush toilet the most common in Pelican Park (96.6%) and Phola Park (97.7%); the portable toilet system is most common in Barcelona (71%) and Shengu (chemical toilets) and flush toilets in Lotus Park, Weltevrede, and Sweet Home (Figure 3b).With regard to hygiene, the participants were asked, “When do you usually wash your hands?” with options: before eating, after eating, after defecation, after latrine use, before feeding child, after handing rubbish, after handling child’s diaper/feces, before food preparation, after touching animals. The questionnaire also included the hygiene question, “Have you seen any rats/cockroaches around your house in the last 7 days.” Additionally, interviewers recorded the availability of soap in the house using an observation checklist.The respondents reported whether they washed hands before eating (69.9% IS vs. 95.7% FS), after eating (61.1% IS vs. 55.5% FS), after defecation (51.2% IS vs. 37.2% FS), after latrine use (62.6% IS vs. 59.8% FS), before feeding child (61.9% IS vs. 61% FS), after handling rubbish (58% IS vs. 42.7% FS), after handling child’s diaper/feces (52.9% IS vs. 34.2% FS), before food preparation (58.9% IS vs. 64.6% FS), and after touching animals (12.5% IS vs. 12.2% FS). The results of an analysis using a hand-washing index derived from the above indicators using principal component analysis showed that respondents from IS had better hand-washing practices than FS (t-test, p = 0.02). The availability of soap at hand-washing places reported in both formal and informal settlements are similar (28.5% vs. 26.2%) and not statistically significantly different. With regard to wastewater disposal, most IS respondents reported that wastewater was disposed of onto open ground (63.8%), while most FS respondents reported that they disposed wastewater into the toilet (51.2%) and onto open ground far away from their houses (18.3%). Pest (rats, cockroaches) invasion 2 weeks prior to the survey occurred more frequently in IS compared to FS (87.0% vs. 46.0%, p = 0.00); however, using pesticides inside the house during 2 weeks prior to the survey was less frequently reported in IS than FS households (44.0% vs. 57.0%, p = 0.004). Hygiene practices of respondents from informal and formal households are presented in Table 3.The health-related indicators for participating children include vaccines monitored by the South Africa Expanded Programme on Immunisation (EPI-SA). Children from FS were more frequently reported to be vaccinated than IS children: 74.1% vs. 50.9% for Hep B vaccination and 76.2% vs. 64.2% for Rotavirus, respectively (Chi-square test, p < 0.001). Interestingly, for HepA vaccination, which is not sponsored under EPI-SA, children living in informal households were reported to get injected more frequently than those living in formal households (45.0% vs. 33.9%, Chi-square test, p < 0.001). Overall, 12.3% of all children were breastfed exclusively for 6 months after birth. More than two-thirds of studied children (76.4%) reportedly had no breastfeeding at all. There was no statistically significant difference in breastfeeding rates between FS and IS children (21.1% vs. 24.5%, Chi-square test, p = 0.33).The prevalence of diarrhoea among children under five years old was calculated based on the total number of cases that reported having diarrhoea during the 7 days prior to the interview in all the participating households. The distributions of diarrhoea cases by study site, age and gender distribution are illustrated in Figure 4.Among a total of 778 recruited children, diarrhoea occurrence was recorded in 119 children (15.3%). The period prevalence was highest in Gugulethu (24.2%), followed by Pelican Park (19.4%), Manenberg (17.8%), Sweet Home (15%), Barcelona (14.4%), Lotus Park (11.1%), Phola Park (11.1%), and Weltevrede (8.6%) (Figure 4b). The prevalence of diarrhoea cases in this study was higher in formal settlements compared to informal settings (21.2% vs. 13.4% respectively, Chi-square test, χ2 = 6.6, p = 0.01) (Figure 4a). As seen in Figure 4c, the age group of 6–11 months had the highest prevalence of diarrhoea (26%), following by the group of 24–35 months (18%); conversely, the diarrhoea prevalence in other age groups was lower, ranging from 12%–13%. Diarrhoea prevalence was higher among girls than boys in all age groups, except in children aged 12–23 months. However, the gender difference was not statistically significant (Chi-square test, χ2 = 1.5, p = 0.22) (Figure 4d).The results from the multivariate analysis show that the caregiver’s education level was positively associated with the diarrhoea occurrence in children under five years old (OR =1.59, 95% CI 1.06–2.40, p = 0.03). There was no significant association between economic status, based on the wealth index and under-five-year child diarrhoea (OR = 0.67, 95% CI 0.38–1.17, p = 0.16). Usage of water stored >12 h showed a statistically significant positive association with diarrhoea (OR = 1.90, 95%CI 1.02–3.79, p = 0.05). Children from households who treat water before drinking were less likely to experience diarrhoea compared to children from households who do not treat water (OR = 0.57, 95%CI 0.34–0.97, p = 0.04). None of the drinking water sources were statistically significantly associated with diarrhoea. Interruption of water supply was not found to be statistically significantly associated (OR = 1.21, 95%CI 0.69–2.09, p = 0.5) with diarrhoea. Additionally, sanitation characteristics and ownership status, when summarized as a continuous variable, showed a significant association with under-five diarrhoea (OR = 0.47. 95% CI 0.26–0.84, p = 0.12). Interestingly, a higher number of households sharing the same facility was associated with a lower risk of diarrhoea among study participants. There was a significantly lower risk of diarrhoea among households sharing the facility with more than three other households compared to those using non-shared facilities (OR = 0.35, 95%CI 0.15–0.85, p =0.02). Dysfunctional toilet facilities showed a significant positive association with childhood diarrhoea (OR = 1.88, 95%CI 1.00–3.55, p = 0.05). An increased hand-washing index was found to be strongly associated with a lower risk of diarrhoea (OR = 0.61, 95%CI 0.47–0.79, p < 0.001). There was no significant association between under-five diarrhoea, gender, and age. The duration of breastfeeding showed a weak and statistically non-significant positive association with the risk of diarrhoea (OR = 1.11, 95%CI 1.00–1.23, p = 0.07). The results in Table 4 shows Hepatitis A vaccination to be a statistically significant protective factor (OR = 0.51, 95%CI 0.28–0.90, p = 0.02) while rotavirus vaccination was not statistically significantly associated with under-five diarrhoea (OR = 1.62, 95%CI 0.86–3.08, p = 0.14).This study revealed that socio-economic factors were important predictors of under-five diarrhoea. Children living in the formal settlements showed greater evidence of access to socio-economic facilities (accommodation, education and health) relative to those living in the informal settlements. It is important to note, however, that living in a formal settlement does not necessarily suggest better hygiene practice. People living in formal settlements have a surprisingly lower index of hand-washing compared to people living in informal settlements. The combination of poor living conditions, poor hygiene practices and contaminated environment in the neighbourhood contributed to diarrhoea cases in children under five in both formal and informal households. The literature suggests that the caregiver’s increasing level of education has been considered as a protective factor against diarrhoea [21,22,23]. However, the results in this study show that the higher education levels of caregivers are associated with a consistently increased risk of diarrhoea in children under five. The paradox has been found in some other studies reporting that there were no differences in knowledge of adequate dietary practices between different maternal education levels [24,25]. In addition, higher education could lead to earlier weaning which can increase episodes of diarrhoea in children [26].Water access and resources is a crucial factor that serves as an indicator of community health. The majority of the investigated households in the formal settlements have access to the municipal water supply. Various factors relating to water access do not allow for any concrete conclusions about the safety of the system delivering water in informal households in this study. Inhabitants in informal settlements are not allowed to have private water connections to their homes; however, some make illegal connections without any supervision or instruction from local authorities, thus creating a path for potential pollution. There is seldom any system for draining the grey water generated by the in-house supply from the informal settlement [27].Our findings show that sharing sanitation facilities with others was associated with an increased risk of diarrhoea relative to non-sharing facilities. Contamination of drinking water resources was higher in households with poor sanitation compared to those with adequate sanitation, and this may explain the increased risk of diarrhoea in the former. However, there are different types of shared sanitation services in informal settlements, varying from a self-maintained facility that is shared between a few nearby households to publicly run facilities (some might be maintained better than others). Although our data suggested that shared sanitation does not pose higher risks than private access, other factors related to sanitation could be more important in diarrhoea transmission. For instance, IS are constrained both with respect to space and water supply, and the majority of communal sanitation facilities are not linked to a sewage system. Stagnant water is often observed in such dense informal communities, which poses a big risk of spreading diarrhoea pathogens into the environment. Previous studies demonstrated children would face a high risk of diarrhoea from playing in soil and surface water outside the household [28,29]. Reinforcing the findings from other studies, we also found that dysfunctional toilet facilities to be significantly associated with childhood diarrhoea in this study. This study found that hygiene practices were associated with a reduced risk of childhood diarrhoea, which is consistent with the literature [30,31].It is well-known that good hygiene could enhance healthy living, and treating point-use water reduces the risk of getting diarrhoea disease [30,32]. Our findings show that children from households with good hygiene are less likely to experience diarrhoea than those who did not. This study is consistent with similar studies elsewhere [33]. Hygiene practices, such as water purification methods, including water-boiling and filtration, could reduce the pathogen load in water if the process is carried out consistently and correctly according to standard techniques [30,34]. Another study also suggested that not all water treatment methods are effective if carried out in an unhygienic manner; improper infiltration could cause by unsafe storage water leading to recontamination after treating [35]. Our data shows that drinking water stored longer than 12 h could contribute to the potential risk of childhood diarrhoea. Similar results were found in other studies conducted in developing countries [36,37,38]. Our findings suggested that sharing a facility with more than four households significantly lower the risk of diarrhoea compared to those using a non-shared facility. Similar studies conducted in Tanzania showed that shared latrines contained lower concentrations of Escherichia coli compared to private latrines [39]. Both results are unexpected. Other studies and meta-analyses found that shared sanitation is generally dirtier and associated with an increased risk of diarrhoea compared to private sanitation [40,41,42,43], suggesting the need for further investigation.In this study, the highest diarrhoea prevalence occurred in children aged 06–11 months. This finding is consistent with that of previous research that suggests diarrhoea is more likely to happen to younger children under 12 months, and the risk reduces with increasing age [44,45,46]. Ordinarily, breastfeeding should increase the children’s immune system and thus protect them against diarrhoea, but this was not found to be the case for those aged under five in this study. This could be due to exposure to contaminated water lying around the informal settlements, increased exposure of the toddler to contaminated water and/or lack of personal hygiene. Additionally, practicing breastfeeding may contribute to childhood diarrhoea if the nursing mothers engage in unhygienic practices such as dirty breasts. Findings also revealed that children who took the HepA vaccine are at lower risk than those who were not HepA vaccinated. The prevalence of Hep A infection in the study is likely due to exposure to poor water and sanitation conditions. Serological studies detected HepA in fecal samples of children with acute diarrhoea in developing countries [47,48,49]. Furthermore, a number of studies have reported the detection or isolation of Hep A virus from water resources, with the detection rate ranging from 76% in surface water to 37% in wastewater plants and irrigation water [50,51]. However, there was no available data on the Hep A prevalence in children under five years in South Africa, and the role of Hep A virus as an element in diarrhoea has not been documented. The results from this study stress the need for further investigation with routine tracking of Hep A infection as a contributory cause of childhood diarrhoea in sub-urban areas in Cape Town.The strengths of this study include the opportunity to estimate the effects of multiple environmental factors and diarrhoea burden using individual-level data from a community-based survey and using multi-level analyses to consider the effects of individual-level and household-level factors in one analysis. Among the limitations of this study, we did not explore the comparison of the ways sanitation facilities are shared in informal settlements with private access facilities. Most households in informal settlements had access to only shared sanitation facilities. There are limited data on how shared sanitation facilities in informal settlements were managed and maintained. In addition, documentation of how stagnant water and leakages were contained and managed was not explored by our study. Those factors may be important sanitation-related confounders on the effects of sanitation on childhood diarrhoea. We were unable to check our data on diarrhoea prevalence against that collected by the local health authorities owing to the difficulty in obtaining the necessary approvals within the limited time frame of the study. A standard limitation is that diarrhoea was reported by respondents with no clinical confirmation. This method is widely used, although clinical confirmations make the disease classification more reliable. The study also did not track the casual, social, and permanent movement of people from informal settlements to nearby formal houses with exposure to different types of water sources and sanitation facilities, which might contribute to the risk of getting diarrhoea. Moreover, the cross-sectional study design only allows us to describe associations but does not allow us to infer causality. Finally, the study did not involve any biomedical tests which could provide stronger information on diarrhoea aetiology.The study found several factors that were significantly associated with the increased risk of diarrhoea among children under five years old living in informal and formal settlements of the Lotus River catchment area in Cape Town. Namely, these factors are caregiver’s education level, storing water longer than 12 h, not using proper water treatment, sanitation facility shared with more than four households, dysfunctional toilet facility, using pesticides inside the houses, poor hand-washing practices, longer duration of breastfeeding, and not having Hepatitis A vaccination. It was clear that good hygiene practices are key in the prevention of diarrhoea. Based on those findings, the promotion of hygienic practices is highly recommended. Caregivers should have more knowledge about and exercise better the practices of hand-washing, hygienic breastfeeding, and food preparation. Proper and healthy water storage is necessary, covering water containers, and basic water treatment techniques, such as boiling, filtration, and disinfection, should be observed. Local authorities should support improvement on the adequate sanitation and drainage service provision in both formal and informal settlements. A periodic water quality monitoring campaign at the communal level will be necessary to identify early enough possible water contamination problems, which can lead to diarrhoea. Finally, further studies on diarrhoea aetiology and transmission pathways in the informal urban context are required to get a better understanding of diarrhoea causality in these communities.Conceptualization, T.Y.C.N., G.C., C.S., M.R., N.P.A. and M.A.D.; Data curation, T.Y.C.N., G.C., M.R., and M.A.D.; Formal analysis, T.Y.C.N. and C.S.; Funding acquisition, G.C., C.S., M.R., and M.A.D.; Investigation, T.Y.C.N., G.C., M.R., and M.A.D.; Methodology, T.Y.C.N., G.C., N.R., S.F., C.S., M.R., and M.A.D.; Project administration, T.Y.C.N. and G.C. and M.A.D.; Supervision, G.C., C.S., M.R., and M.A.D.; Validation, G.C., C.S., M.R., and M.A.D.; Visualization, T.Y.C.N.; Writing—original draft, T.Y.C.N. and B.O.F.; Writing—review and editing, T.Y.C.N., B.O.F., G.C., N.R., S.F., J.O., C.S., M.R., N.P.A., K.C., and M.A.D. All authors have read and agreed to the published version of the manuscript.This study is imbedded within the South African–Swiss Bilateral SARChI Chair in Global Environmental Health of Mohamed Aqiel Dalvie, the Centre for Environmental and Occupational Health Research, the University of Cape Town, and Martin Röösli, the Swiss Tropical and Public Health Institute. This chair was formed in 2015 with funding sources from the South Africa National Research Foundation (NRF) SARChI (grant number 94883), Swiss State Secretariat for Education, Research and Innovation, University of Basel and the Swiss TPH.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by was approved by the “Ethikkommission Nordwest—und Zen-tralschweiz (EKNZ)” in Switzerland (EKNZ BASEC 2016-00304) and by the “Human Research Ethics Committee (HREC)” of the University of Cape Town in South Africa (HREC REF 248/2016).Informed consent was obtained from all subjects involved in the study.The data presented in this study are available from the corresponding author on reasonable request. The data are not publicly available due to the presence of confidential participants’ information.The authors would like to acknowledge and express thanks to all individuals and families who participated in the study. Special mention is needed for local teams of Edith Stephens Nature Reserve and A.L. Abbott & Associates, especially Dale Issac, who greatly supported the study and helped with logistical arrangements, translation, as well as community engagement during data collection.The authors declare no conflict of interest.Map of South Africa (a), Western Cape (b), and the catchment area in Cape Town showing the six informal settlements in the (Zeekoe catchment) and different land use (c).Types of drinking water sources (a) in each settlement type; (b) in six informal settlements (c) in two formal settlements.Types of sanitation facilities available (a) in each settlement type; (b) in six informal settlements; and (c) in two formal settlements.(a) The prevalence of diarrhoea cases by gender in each settlement type; (b) The prevalence of diarrhoea cases by gender in each study site; (c) The distribution of diarrhoea cases by gender in each age group; and (d) The prevalence of diarrhoea cases by house types in each age group. Note: BP (Barcelona Park), LP (Lotus Park), PP (Peligan Park), PH (Phola Park), WR (Weltevrede), SH (Sweet Home), GG (Gugulethu), and MB (Manenberg).Demographic and health-related characteristics of children in participating households.n = Number of children of participating households.Access to water resources of participating households from informal and formal settlements.The hygiene in participating households from informal and formal settlements.The multi-level multivariate mixed-effects logistic regression analysis of risk factors associated with diarrhoea.N.B: Results of multi-level multivariate analyses of risk factors associated with diarrhoea are presented, including the following groups of variables: socio-demographic (caretaker education level, having other under-5 children in the household and wealth index), water-related variables (drinking water sources, water storage >12 h, water treatment options and water interruption during within the period of seven 7 days prior the survey), sanitation and hygiene-related variables (facility access and sharing, reported a problem with toilet facility, disposal of child’s feces, hand-washing index), and children-specific variables (child’s age, gender, breastfeeding duration and immunization history of Hepatitis A and Rotavirus vaccine).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06044.txt ADDED Viewed

	@@ -0,0 +1 @@

+ The objective of this study was the translation and validation of a health consciousness scale in order to provide an economically and empirically confirmed measurement of health consciousness, which is associated with health-related behavior. We evaluated this translation on the basis of psychometric testing in a German convenience sample. A cross-sectional online survey (n = 470) was carried out using a translated version of the health consciousness scale, oriented on the basis of international guidelines. As previous studies have not consistently confirmed the factorial structure of the health consciousness scale, we conducted a Confirmatory Factor Analysis to verify its factorial structure. Furthermore, we cross-validated the questionnaire with other scales in order to verify convergent and discriminant validity. The results indicated a two-factor solution for the Health Consciousness Scale-German (HCS-G). The criterion validity was confirmed on the basis of a significantly positive correlation between the HCS-G and health literacy. Furthermore, strict measurement invariance was able to be verified, indicating that the HCS-G is an applicable measurement, regardless of gender. In practical research, this questionnaire can help to assess health consciousness and its influence on health-related constructs. Future studies should consider possible mediating variables between health consciousness and health outcomes.Health-related or preventive behaviors are an omnipresent topic in research. Meta-analyses as well as reviews have shown their influence on physical and mental health [1,2]. For example, reviews and different studies have found correlations between healthy diets and cardiovascular, metabolic (e.g., Diabetes mellitus), or oncological diseases [3,4,5]. Health-related and preventive behaviors such as physical activities are also recommended by international societies and studies, which have shown positive effects on disease prevention [6,7,8,9]. Furthermore, during the Covid-19 pandemic, preventive behavior seems key to protection from infection with the Covid-19 virus [10,11]. Different determinants of these health-related or preventive behaviors have been identified in studies [12,13,14,15]. These studies show that attitudes toward one’s own health or the perception of one’s health status are possible determinants of health-related or preventive behaviors. Constructs such as health consciousness [12], health orientation [16], health conception [17] and health locus of control [18] are psychological constructs that have been explored in health science that influence these behaviors. The construct of health consciousness, especially, focusses on one’s own perception of health status, and is described by Gould [12] as a self-awareness regarding one’s own health. Gould [12,19] distinguished between four different dimensions focused on an individual’s psychological inner state towards health-related awareness: health self-consciousness, health alertness, health self-monitoring and health involvement. Health consciousness and its interrelations with different health-associated outcomes has already been verified in various studies, such as with respect to different healthy activities [20], prevention of illness and promotion of long-term health [21], frequency of doctor visits [22] and less smoking/drinking, increased engagement with health information, and taking vitamin supplements [12]. Even though these outcomes were measured by self-reported questionnaires, relevant associations of health consciousness and health-associated outcomes were found. Furthermore, studies have shown that higher states of health consciousness are correlated with the likelihood of purchasing organic food and green furniture [23,24,25,26,27]. In this context, Shimoda and colleagues [28] additionally showed an association between health consciousness and pro-environmental behavior indicating that health consciousness is not only likely to promote one’s individual health but can also have positive environmental effects.At this point it is important to distinguish the definition of health consciousness by Gould [12] from that of Jayanti and Burns [14]. Jayanti and Burns [14] define health consciousness as health concerns and their integration in daily behaviors or activities of people. This is a more action-oriented definition than the definition of Gould [12], who assumes health consciousness rather as a mindset influencing health-related and preventive behaviors. Furthermore, self-awareness is a key variable in the definition of health consciousness. Self-awareness is a self-directed attention to oneself and is associated with self-consciousness [12,29]. These constructs are the core of health consciousness. Thus, from our perspective, the definition of health consciousness and its operationalization according to Gould [12] offer the greatest opportunities to focus on self-awareness regarding one’s own health. The constructs self-awareness and self-consciousness are part of health consciousness and influence subsequent health-related actions and health behaviors [12,30]. Health consciousness, as a psychological determinant of health-related and preventive behavior, reflects a concrete attention towards one’s own health status and symbolizes a conscious treatment with one’s own health. Thus, it is a main variable influencing health-related and preventive behavior.Gould’s scale [12,19] has already been used in many studies to measure health consciousness. However, the factorial structure has not been verified, and different factorial structures have been reported in several studies (e.g., [22,31,32]). According to our research, the factorial structure of Gould [12,19] has been confirmed in only one study with a first- and second-order factor structure [22]. However, looking at intercorrelations between second-order factors in Gould’s [12] factorial structure, it became evident that these factor loadings are too high to reflect independent factors. Here, a Confirmatory Factor Analysis (CFA) of the scale is needed to confirm the construct health consciousness in its subdimensions. Furthermore, the scale has not been used in German samples with a translation according to common scientific recommendations [33,34] and has not been validated in a German sample. It is important to ensure the construct validity of existing questionnaires. Therefore, the convergent as well as the discriminant validity of a questionnaire have to be verified. Another methodological approach focused on the verification of measurement invariance of the health consciousness scale is to confirm that the instruments measure the construct equally for both genders. To our knowledge, scientific evidence is still missing to assume measurement invariance of health consciousness between women and men. Nevertheless, statistical differences between different groups of people are not interpretable unless measurement invariance between these groups is confirmed [35]. Rather, a lack of measurement invariance of instruments may lead to misinterpretations of gender differences, as differences may be caused by the instrument itself [36]. Thus, Gould’s analysis of gender differences is not interpretable [12]. In fact, determinants of health-related and preventive behavior have been identified in different studies [12,13,14,15]. Thus, it is a key variable in the context of measuring one’s self-awareness towards one’s health status. However, a health consciousness scale in the German language does not exist and needs to be validated to focus on its impact on health-related and preventive behavior in the German population. Furthermore, different studies (e.g., [22,31,32]) using the health consciousness scale have shown ambiguous factorial structures, which have seldom been validated, while only rarely representing the assumed factorial structure of Gould [12]. In summary, we are appealing to different aims by conducting this validation study of a psychometric instrument measuring health consciousness. Due to the practical and scientific demands and the described limitations, we focused on four objectives. First, we wanted to develop a validated and a translated version of the health consciousness scale in German language according to common scientific recommendations [33,34] to ensure content validity. Second, we aimed to verify a valid factorial structure of this new instrument. For this purpose, we performed Confirmatory Factor Analysis (CFA) to examine the factorial structure of the translated instrument. Additionally, to confirm construct validity, we scrutinized convergent and discriminant validity. To verify convergent validity, we assumed a positive correlation between health consciousness and personality trait conscientiousness, because both constructs describe a person’s tendency to be aware of and responsive to one’s surroundings in a thoughtful way. We assumed that personality trait openness would be associated with health consciousness, because both constructs symbolize a person’s interest in searching for and discovering changes and improvements in one’s life with respect to health status and life situations (e.g., [37]). Moreover, as health consciousness focuses on a mindful handling of one’s body, we expected a negative interrelation with impulsivity. To verify discriminant validity, we looked at interrelations with different personality traits such as extraversion and neuroticism, assuming no significant correlations, as there were no content-related overlaps with health consciousness. Third, we wanted to ensure the criterion validity of the adapted instrument. For this purpose, we asked participants about their health literacy and the frequency with which they had used medical or therapeutic help within the previous twelve months. As in the original study of Gould [12], there were, unexpectedly, no significant interrelations found between health consciousness and physical or mental health, we captured these constructs as well. Fourth, we wanted to perform an analysis of measurement invariance of the Health Consciousness Scale—German (HCS-G) regarding gender. To our knowledge, there is no study examining measurement invariance of health consciousness between women and men. Nevertheless, a lot of studies have shown differences between women and men regarding their health and preventive behavior, such as their use of health services [38], their attitude towards seeking psychological help [39], and health status [40]. However, interpretation of gender differences should be made with caution when the measurement invariance of the applied instruments has not been confirmed. Therefore, it is essential to ensure measurement invariance between women and men, as increasingly frequently examined in recent studies (e.g., [41,42,43]).We conducted the translation following the guidelines for translation of academic literature to ensure content validity [33,34]. The translation was conducted in four steps. In a first step, the items were translated into German by two authors, and these translations were merged into a first translation proposal. In the second step, a systematic expert panel consisting of the two translators and two psychologists discussed the merged items. The two psychologists were experts within the context of healthcare. In a third step, the resulting items in the German language were translated back into English by an English native speaker, confirming that their meaning was consistent with the original items. In a fourth step, cognitive interviews were conducted to ensure easy understanding, and inoffensive and non-discriminatory phrasing of the items. The resulting final version of the translated questionnaire consisted of nine items. The original items, as well as the translated items, are displayed in Table S1. We used a 5-point Likert scale from 1 = “strongly disagree” to 5 = “strongly agree”. A sample item is “I’m very involved with my health”. We clustered the items according to the assumed factorial structure of Gould (1990): health self-consciousness, health alertness, health self-monitoring and health involvement [12]. Our cross-sectional online survey was conducted between October 2020 and November 2020 in a German-speaking convenience sample. Participants were recruited via online social networks like Xing, LinkedIn, and Facebook. All data were collected anonymously. We considered completed data sets only. We conducted the study according to the guidelines of the Declaration of Helsinki and the Ethics Committee of the Medical Faculty of the University of Duisburg–Essen approved our survey (approval number 20-9592-BO).524 participants completed our questionnaire which represents a completion rate of 32%. Participants who took less than 5:34 min (5% percentile) or more than 25:45 min (95% percentile) to complete the survey were excluded from the analysis. We decided not only to exclude extreme fast responders but also extreme slow responders as we wanted to ensure that our analysis was based on an average sample where possible biases in response behavior were minimized. Furthermore, we excluded one participant for being under 18 years of age. One person indicated their gender as “diverse”, which we excluded to perform the analysis of measurement invariance of gender. The resulting sample consisted of n = 470 participants, which is in accordance with the sample size recommendation for test validation [44,45].The following measurements were captured to cover the study objectives. Prior to the statistical analyses, inverted items were recoded. The Health Literacy Questionnaire (HLS-EU-Q16) measures health literacy with 16 items [46]. A sample item is “How easy/difficult is it to find information on treatments of illnesses that concern you?”. Health literacy was measured on a two-point scale (easy/hard), resulting in a sum-score of health literacy between 0 and 16.The Impulsive Behavior-8 Scale (I-8) measured impulsivity using eight items (e.g., “Sometimes I spontaneously do things that I should not have done” [47].The BFI-10 [48] measures personality traits (extraversion, neuroticism, openness, conscientiousness, agreeableness), which are each assessed by two items. A sample item for neuroticism is “I get nervous and insecure easily”. The reliabilities of conscientiousness and agreeableness were low, as the items capturing these constructs represent different contents of these traits. One item of conscientiousness measures if one tends to laziness, while the other measures if one completes tasks thoroughly. As we judged both aspects of conscientiousness to be highly relevant in the context of health consciousness, we decided to use the single items in further analyses. As agreeableness was not regarding as contributing to the validation of health consciousness, it was excluded from further analyses.Physical health and mental health were measured using self-developed single items on an 11-point Likert scale (0 = “very bad health” to 10 = “very good health”). Furthermore, we asked participants about their frequency of use of medical or therapeutic help within the previous twelve months from 1 = “never” to 6 = “every week”.Additionally, sociodemographic variables (age, gender, marital status, educational level, financial situation, and residential area) were assessed.All statistical analyses were performed with R, RStudio and additional packages [49,50,51,52,53,54,55]. We first examined sample characteristics and item statistics of the HCS-G. To investigate the factorial structure of the health consciousness items, we conducted several Confirmatory Factor Analyses. We considered the recommendations of Hu and Bentler [56] for model evaluation. We examined Comparative Fit Index (CFI), Tucker Lewis index (TLI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Residual (SRMR), as well as the factor loadings of items, to determine which model fits best to the data. As we collected survey data, we assumed data to be ordinal level rather than interval level. The use of Weighted Least Square Mean and Variance Adjusted estimator (WLSMV) is recommended for ordinal data rather than the Maximum Likelihood Robust estimator (MLR) in terms of the robustness of estimations of factor loadings and model fit indices [57]. Therefore, we conducted the CFA using the WLSMV estimator. However, WLSMV tends to overestimate inter-factor correlations, which will be taken into account in the discussion section [57,58]. Furthermore, we calculated Cronbach´s alpha (CA) with 95% confidence interval boundaries (CI) to report the reliability of the HCS-G, as well as the scales for construct and criterion validation. To scrutinize the convergent, discriminant, and criterion validity of health consciousness, we performed two-tailed Pearson correlations and considered correlations to be significant at p < 0.05. We examined measurement invariance to test whether the measurement of health consciousness was equally appropriate for both genders. The test of measurement invariance follows four sequential steps. Each step assumes a more constrained model. In step one, the number of factors and the factor-to-indicator relationship were considered to be equal between genders, which describes configural invariance. In step two, we scrutinized metric invariance by additionally keeping factor loadings equal between genders. Moreover, in step 3 (scalar invariance), intercepts are kept equal. Finally, the most constrained model 4 (conservative invariance) assumes the equality of residual variances. A more constrained model can be accepted in cases where the differences between it and the less constrained model are low [59,60]. The WLSMV estimator is also used for measurement invariance in order to compare the different models. As the interpretation of results of measurement invariance using the WLSMV estimator is limited, especially regarding changes in Chi2, we focus on changes in CFI and the other fit indices [61,62,63].After the described case exclusion, average completion time of our survey was 11:32 min (SD = 4:24 min, Median = 10:38 min). The mean age of our participants was 37.2 (SD = 13.4, Median = 33, Minimum = 18, Maximum = 82). Table 1 shows detailed sociodemographic characteristics of the study sample.Table 2 depicts the item statistics of the nine items of the preliminary HCS-G. The last characters indicate the subscales as proposed in the original study (C = self-consciousness; I = involvement; A = alertness; M = self-monitoring).All item distributions are slightly negatively skewed. Items hcon3C (health self-consciousness), hcon5A and hcon6A (both health alertness) showed high values of kurtosis. Response distribution and SD indicated that these items were poor in generating variance between participants.In our first model, all nine items loaded on a single factor. Model 2 considered the proposed subscales of the original study from Gould, 1990 [12]. Due to poor item statistics of three items (hcon3C, hcon5A and hcon6A) and their low factor loadings, we conducted a third model as a single factor model with remaining items. We verified a fourth model as a three-factor model with intercorrelated factors. Since in this model the subscales of health self-consciousness and health involvement were highly correlated (r = 0.93, p < 0.001), we carried out a fifth CFA as a two-factor model with items of health self-consciousness and health involvement considered as one factor. A second-level factor of health consciousness led to convergence problems just as Gould reported in his study [12]. Therefore, we assumed health consciousness to consist of two intercorrelated constructs which measure different aspects of health consciousness. Table 3 shows the results of performed CFA.In the three-factor (short) model, reliabilities for health self-consciousness, health involvement, and health self-monitoring were 0.74 (CI = 0.69–0.79), 0.80 (CI = 0.76–0.83), and 0.84 (CI = 0.81–0.87), respectively. Within the two-factor (short) model, items of health self-consciousness and health involvement were considered to load on a single factor, which reached a reliability of 0.85 (CI = 0.83–0.88). Considering these results, the two-factor (short) model was assumed to best represent the construct of health consciousness. In fact, the three-factor (short) model had partly better fit indices than the two-factor (short) model. Nevertheless, the correlation of 0.93 between self-consciousness and involvement strongly indicated that those two subscales were measuring the same content. Moreover, the two-factor (short) model was a less constrained model, representing a more applicable and parsimonious model.Figure 1 shows the final two-factor (short) model with factor loadings as well as inter-factor correlation.As described, our CFA process found the two-factor (short) model with intercorrelated factors to best fit the data. Therefore, we conducted the following analyses, considering the two identified factors self-consciousness (M = 3.81, SD = 0.77) and self-monitoring (M = 3.40, SD = 0.89) as independent variables.To test whether health consciousness differs between genders, two-sided t-tests were conducted. There was no significant difference between men and women in self-consciousness (t = 1.37, p = 0.17) or self-monitoring (t = 0.30, p = 0.77). Further interrelations between health consciousness and sociodemographic variables are shown in Table 4.The results show that there are no significant correlations between self-consciousness or self-monitoring and sociodemographic variables, except for educational level, which correlated significantly negative with self-monitoring (r = −0.13, p = 0.006).Table 5 depicts Pearson correlations of health consciousness scales and conscientiousness (single items), openness (M = 3.61, SD = 0.99, CA = 0.62, CI = 0.54–0.68), impulsivity (M = 2.78, SD = 0.59, CA = 0.72, CI = 0.68–0.76), extraversion (M = 3.30, SD = 1.04, CA = 0.79, CI = 0.76–0.83), and neuroticism (M = 3.08, SD = 0.97, CA = 0.66, CI = 0.60–0.72). Looking at convergent validity, both HCS-G scales correlated significantly positively with openness and significantly negatively with impulsivity. Furthermore, both items of conscientiousness were significantly interrelated with self-consciousness, but not with self-monitoring. Considering discriminant validity, correlations of both HCS-G scales with personality traits of extraversion and neuroticism were not significant.Regarding criterion validity, correlations between the health consciousness scales and the health-related construct of health literacy (M = 12.63, SD = 2.99, CA = 0.79, CI = 0.76–0.81), as well as the frequency of use of medical or therapeutic help (M = 3.80, SD = 1.50) are shown in Table 6.As in the original study by Gould [12], we additionally examined interrelations of health consciousness scales and physical health (M = 7.37, SD = 1.58), as well as mental health (M = 7.27, SD = 1.90). Neither self-consciousness (r = −0.02, p = 0.74 and r = 0.05, p = 0.32) nor self-monitoring (r = 0.00, p = 0.97 and r = 0.01, p = 0.83) correlated significantly with physical or mental health.The tests of measurement invariance regarding the gender of participants were conducted on the final two-factor model. Table 7 shows the results of these analyses.At all levels of measurement invariance, only small changes in CFI and the other fit indices could be observed. Improvement of RMSEA can be explained by the sensitivity of this fit index to small degrees of freedom [64].The most essential strengths of our study are due to the high methodological and psychometrical standards applied to the validation and adaptation of the HCS-G. In this regard, this study confirmed its content, construct, and criterion validity. Another main strength of this study is the confirmation of measurement invariance with respect to gender. To our knowledge, no other study has examined the measurement invariance of the health consciousness construct or revealed methodological insufficiencies.Our first aim was to develop a valid version of the health consciousness scale in the German language and to confirm the content validity of our new instrument, which we achieved by strongly following the scientific recommendations implied by the development of the HCS-G. To achieve our second aim, we validated our new instrument by performing several Confirmatory Factor Analyses to examine the factorial structure. Within the correlational analyses, we verified convergent and discriminant validity with respect to construct validity. We followed a stepwise CFA approach in order to acquire the best-fitting and most parsimonious model. Therefore, we excluded single items, resulting in a six-item model that represents health consciousness on two subscales. To our knowledge, this is the first study to critically examine the assumed factorial structure according to Gould [12]. A two-factorial structure best represented the construct of health consciousness, while also solving the convergence problems of Gould’s original instrument [12] and confirming a valid factorial structure which has been neglected in other studies so far [21,32,65,66,67]. Our final model achieved good fit indices and factor loadings of the items representing the factors self-consciousness and self-monitoring. A slight to high value of RMSEA can be explained by its sensitivity to small degrees of freedom, which was true for the tested model [64]. Both factors were significantly positively interrelated, but represented different aspects of health consciousness, as confirmed by our correlational analyses. Self-consciousness can be understood as the awareness of one’s health in the long term, whereas self-monitoring focuses more on daily routines of self-observation.As expected, the HCS-G scales were significantly positively correlated with consciousness and openness, as well as being significantly negatively correlated with impulsivity, indicating convergent validity. This indicates that a greater score in the HCS-G scales is associated with greater conscientiousness and openness, and lower impulsivity. Furthermore, the HCS-G scales did not correlate with constructs that were different from health consciousness (extraversion and neuroticism), confirming discriminant validity. Surprisingly, HCS-G (sm) was significantly negatively correlated with educational level. This result replicates the findings of Gould [12], who suggested that people with lower educational levels may spend more time on their health, resulting in a higher health consciousness. Nevertheless, a valid explanation of this relationship is still lacking. One could also think that conceptions regarding one’s health and health consciousness might carry different meanings in different educational levels, resulting in different relational patterns. However, as our sample mainly consisted of higher-educated people, it was not possible to administer an in-depth analysis within a subsample of lower-educated people. Hence, future studies should examine this relationship in lower-educated subsamples or from a causal viewpoint.Regarding our third aim, the criterion validity of the HCS-G was confirmed by verifying a positive correlation between both scales of health consciousness and health literacy. This indicates that a greater score in the HCS-G scales is associated with greater health literacy, which is in agreement with the results of Gould [12]. It is possible that people with greater health consciousness examine their health more precisely and look for health information from different sources in order to assess their own state of health. Furthermore, as health literacy can be understood as a competence in dealing with one’s health issues, which is associated with many health-related outcomes (e.g., [68,69,70]), health consciousness is a highly relevant construct in the understanding of a holistic concept of people’s health. However, only HCS-G (sc) was significantly correlated with the use of medical or therapeutic assistance, indicating that highly self-conscious people use medical or therapeutic help more often. HCS-G (sm) was not significantly correlated with the frequency of medical or therapeutic help, which could be explained by the fact that self-monitoring items tend to focus on daily routines, while self-consciousness items tend to be oriented towards long-term aspects. Unexpectedly, health consciousness was not correlated with possible outcome variables of physical and mental health, which is in agreement with the results of Gould [12]. We assumed health consciousness to be related to higher physical and mental health statuses. To explain these findings, one could think that people with lower subjective health statuses may be forced to become more concerned with their health, leading to higher scores in health consciousness. Furthermore, several studies have shown that conceptualizations, attitudes, and intentions are not causally linked to actual health behavior [71,72,73]. In consequence, the relationship between health consciousness and health-related outcomes may be mediated by other variables. Future studies should investigate possible mediating variables between health consciousness and health outcomes.In the case of the Covid-19 pandemic, health consciousness could be a relevant psychological construct for preventing further Covid-19 outbreaks. In our study, health consciousness and health literacy are associated constructs. Thus, one could assume that health-conscious people will be more likely to be self-aware and vigilant with respect to diseases and their influences on the body. Furthermore, health-conscious people will be more likely to understand Covid-19 interventions and to comply with the regulations, resulting in a reduced spread of Covid-19. The compliance with the regulations [74] is a relevant indicator for the prevention of an infection. Thus, health-conscious people could be more likely to successfully avoid Covid-19 infection, e.g., because of a higher adherence to safety behaviors [75]. Future studies should take the associations between health consciousness and infection prevention, especially infection with Covid-19, into account precisely because health consciousness could be a relevant mediator in the association between prevention knowledge and prevention behavior in Covid-19 infection [76].Our fourth aim contributed an examination of the measurement invariance of health consciousness between women and men. The measurement invariance of the HCS-G with respect to gender was confirmed, verifying that the HCS-G is a reliable instrument for measuring health consciousness and interpreting gender differences. Measurement invariance is an essential requirement when measurements are used to interpret differences between groups. Through the confirmation of the measurement invariance of the HCS-G with respect to gender, the interpretation of differences in gender can be regarded as valid, i.e., the differences found are not a result of the measurement method itself. Therefore, our instrument provides researchers as well as practitioners the opportunity to capture and interpret gender differences with respect to health consciousness, as well as to develop and evaluate interventions in consideration of gender. In conclusion, the HCS-G is an ideal instrument for practical applications, as the interpretation of the results is carried out independently of gender. As health consciousness is a relevant construct with respect to both maintaining health and influencing health-related behavior, the HCS-G can be used during the diagnostic process as one of the resources for patients. By considering two aspects of health consciousness, the HCS-G can also be used to develop and evaluate psychological training with the aim of improving health consciousness.Nevertheless, limitations have to be considered in this study. One limitation is related to the cross-sectional study design. Thus, correlations are not causally interpretable, but show relationships between the different measured constructs. A second limitation with respect to our study design is related to the method of data collection being via an online survey, which could lead to a selection bias. It is likely that only participants who were familiar with the use of the Internet took part in our study. A third limitation is related to our distribution of educational levels: a high proportion of people that took part in our study held a university degree, limiting the representativeness with respect to educational level. Even though it was our goal to collect data on a convenience sample representing the German population, the study sample did not reflect a balanced gender distribution (71% female participants).Health consciousness is a highly relevant psychological construct in the context of several health-related outcomes and in examining people’s health. We conducted our study with a focus on the highest methodological and sychometrical standards in order to provide a valuable opportunity to measure health consciousness, especially when it comes to the interpretation of gender differences. Until now, a deeper understanding of the factorial structure, as well as the validity, of the measurement of health consciousness has been missing. The HCS-G fills this important gap and enhances the field of application by confirming the measurement invariance. Furthermore, the HCS-G is the first instrument to measure this highly relevant construct in the German language.The following are available online at https://www.mdpi.com/1660-4601/18/11/6044/s1, Table S1: Original and translated items of the health consciousness scale.Conceptualization, M.M., G.E. and A.B.; Methodology, M.M., G.E. and A.B.; Project administration, M.M., G.E. and A.B.; Supervision, E.-M.S., M.T. and A.B.; Visualization, M.M.; Writing—original draft, M.M. and G.E.; Writing—review & editing, M.M., G.E., E.-M.S., M.T. and A.B. All authors have read and agreed to the published version of the manuscript.The APC was funded by the Open Access Publication Fund of the University of Duisburg–Essen.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the medical faculty of the University of Duisburg–Essen (protocol code 20-9592-BO, 03.11.2020).Informed consent was obtained from all subjects involved in the study.The data presented in this study are openly available in FigShare: Dataset: https://doi.org/10.6084/m9.figshare.14502267 (accessed on 30 January 2021). R-Syntax: https://doi.org/10.6084/m9.figshare.14502276 (accessed on 30 January 2021).Daniela Geiß and Maria Spank provided steady support and help during the conception and conducting of the study.The authors declare no conflict of interest.Two-factor (short) model of health consciousness. Square boxes represent items. Circles represent factors. Unidirectional arrows indicate factor loadings. Bidirectional arrow indicates inter-factor correlation.Detailed sociodemographic characteristics of the study sample.Item statistics for all items of the health consciousness scale.Results of Confirmatory Factor Analysis examining the factorial structure of the Health Consciousness Scale—German (HCS-G).df = degree of freedom. CFI = Comparative Fit Index. TLI = Tucker Lewis index. RMSEA = Root Mean Square Error of Approximation. SRMR = Standardized Root Mean Square Residual. short = models considering six items.Person correlations of health consciousness scales and sociodemographic variables.HCS-G (sc) = Health Consciousness Scale—self-consciousness. HCS-G (sm) = Health Consciousness Scale—self-monitoring. ** p < 0.01, *** p < 0.001.Pearson correlations of health consciousness scales and convergent and discriminant scales.HCS-G (sc) = Health Consciousness Scale—self-consciousness. HCS-G (sm) = Health Consciousness Scale—self-monitoring. consc.1 = conscientiousness item 1 (recoded): “I am comfortable, prone to laziness” (M = 3.18, SD = 1.10). consc.2 = conscientiousness item 2: “I complete tasks thoroughly” (M = 4.01, SD = 0.80). * p ≤ 0.05, ** p < 0.01, *** p < 0.001.Pearson correlations of health consciousness and criterion validity scales.HCS-G (sc) = Health Consciousness Scale—self-consciousness. HCS-G (sm) = Health Consciousness Scale—self-monitoring. Frq. Med. Help = frequency of medical or therapeutic help. * p ≤ 0.05, ** p < 0.01, *** p < 0.001.Measurement invariance of the final HCS-G model regarding gender.df = degree of freedom. CFI = Comparative Fit Index. TLI = Tucker Lewis index. RMSEA = Root Mean Square Error of Approximation. SRMR = Standardized Root Mean Square Residual. Δ CFI = change in CFI compared to less constrained model.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06045.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Reducing the treatment delay by organizing delivery of care on a regional basis is a priority for improving the quality of ST-segment elevated myocardial infarction (STEMI) care. This study aimed to evaluate the impact of the combined measures on quality metrics of healthcare delivery in Suzhou. The data were collected from the National Chest Pain Center (CPC) Data Reporting Database. 4775 patients were recruited, and after propensity-score matching, 1078 pairs were finally included for analysis. We examined the changes in quality metrics of care including prehospital and in-hospital processes, and clinic outcomes. Quality improvement (QI) implementation improved most process indicators. However, these improvements did not yield decreased in-hospital mortality. The door-to-balloon and the FMC-to-device time decreased from 85.0 and 98.0 min to 78 and 88 min, respectively (p < 0.001). Cases transferred directly via EMS had a greater improvement in most of process indicators. The proportion of patients transferred directly via EMS was 10.3%, much lower than that of self-transported patients at 58.3%. Tertiary hospitals showed greater performance improvement in process indicators than secondary hospitals. The percentage of cases using EMS remained low for suburban areas. The establishment of coordinated STEMI care needs to be accompanied with solving the fragmented situation of the prehospital and hospital care, and patient delay should be addressed, especially in suburban areas and on transferred-in inpatients.ST-segment-elevation myocardial infarction (STEMI) is the deadliest acute cardiac event and requires rapid coordination of care beginning at the time of symptoms’ onset, and percutaneous coronary intervention (PCI) within 120 min of onset is the typically recommended treatment [1]. Despite the widespread promulgation and endorsement of the guidelines, their translation into clinical practice remains suboptimal. The time from onset to PCI is approximately 290 min, and only 7% of patients receive timely PCI therapy, contributing to increased mortality from cardiovascular disease, which is the leading cause of death in China [2]. The times from call-to-emergency medical services (EMS) to the scene and the door-to-balloon time are about 21 and 94 min respectively, which is longer than the average time of 7 and 59 min in some developed countries [3,4]. Clinical studies have shown that for every 30 min that treatment is delayed, the one-year mortality rate after STEMI increases by 7.5% [5,6]. Therefore, reducing the treatment delay by organizing delivery of care on a regional basis is a priority for improving the quality of STEMI care in China.Hospital-based clinical registries and the related quality improvement (QI) initiatives can facilitate the delivery of effective healthcare [7,8]. This belief is supported by several prior efforts in many developed countries [9,10,11]. The development of a chest pain center (CPC) is a foundation for establishing delivery systems of healthcare for acute cardiac events. However, there has been no national unified registry for garnering hospital participation, standardizing clinical practices, and facilitating QI initiatives in China [12]. To address this need, with the support of the National Health and Family Planning Commission of China (NHFPC, renamed to the National Health Commission or NHC in 2018), the Chinese Society of Cardiology (CSC) officially issued the China CPC Accreditation Criteria in 2013. Afterwards, the NHFPC enacted the “Notice on Strengthening the Capacity of Healthcare Delivery for Acute Cardiac Events” in 2015. The official notice called for establishing regional collaborative healthcare networks through integrating community-, prehospital- and hospital-care for managing acute cardiac events. Under these directives, hospital-based CPCs were quickly developed throughout China, forming regional networks of multidisciplinary specialized cardiac care centers.Suzhou was the first city to respond to the national call, and has implemented the multifaceted QI initiatives that can facilitate the delivery of effective healthcare since 2016. The QI initiatives include: (i) accreditation of hospital-based CPCs, (ii) establishing a unified hospital-based clinical registry for quality monitoring and assessment, (iii) providing ongoing quality reviews and feedback, and (ⅳ) organizing education and training activities aimed at healthcare professionals. Suzhou was the first city to establish the Medical Priority Dispatch System (MPDS) to guide standardized prehospital EMS at a regional level.Moreover, the EMS in Suzhou has taken the lead in establishing an information sharing system by linking the MPDS and the hospital-based CPC registry, to integrate the prehospital and hospital care, and to facilitate the coordination of care at the time of entering the EMS system.The EMS system in China includes prehospital emergency centers that provide prehospital care, and hospital emergency departments and the intensive care units that provide hospital care [13]. In most parts of China, the prehospital and in-hospital processes are separately managed by emergency centers and hospitals [14]. Hence, connecting prehospital and in-hospital processes is crucial for STEMI treatment, and includes patterns of transport to the hospital, the transfer from secondary hospitals (with basic CPCs) to tertiary hospitals (with combined CPCs), and transfer from suburban to urban hospitals. The EMS in Suzhou is unique in that it has focused on the establishment of regional systems of STEMI care by facilitating the coordination between prehospital emergency centers and hospital emergency departments, and the hierarchical delivery between secondary and tertiary hospitals.Therefore, it is warranted to evaluate the implementation of the combined measures in Suzhou, including the combined QI initiatives and the establishment of the information sharing system between the MPDS and the hospital-based CPC registry. Building on efforts in establishing a unified hospital-based clinical registry, we developed the first prospective study in China: (1) to evaluate the impact of the combined measures focusing on establishment of regional systems of STEMI care on quality metrics of healthcare delivery in terms of the care processes and clinical outcomes; and (2) to compare the impacts between secondary and tertiary hospitals, among patients with different modes of transfer, and between urban and suburban areas.The study was conducted at all of the 40 accredited hospital-based CPCs in the nine districts of Suzhou. Suzhou is located in the east of China, has a population of 10.72 million, ranking 13th in China in terms of population size, with a gross domestic product per capita of ¥173,456.4 yuan ($25,161.8 USD), ranking 5th among the total 661 cities in mainland China in 2019. The operation of the QI initiatives was carried out and managed by the Management Board established by the CSC under the authorization of the NHC nationwide. The Data Management Committee, one of the committees of the Management Board was responsible for evaluating and monitoring the QI initiative. All the accredited CPCs in Suzhou were instructed to submit consecutive eligible patients to the China CPC Data Reporting Database (http://data.chinacpc.org/, accessed on 6 September 2019), a national surveillance system for monitoring the characteristics, treatments, and outcomes of patients diagnosed with STEMI. Each hospital is responsible for its own data collection.Data on all patients older than 18 years with a final diagnosis of STEMI at discharge in Suzhou were drawn from the China CPC Data Reporting Database. 4775 patients were recruited consecutively in the 40 hospital-based CPCs from 1 April 2016 to 31 March 2019. The combined measures including QI initiative implementation and accreditation were applied at the hospital level after hospital accreditation. So the pre-combined measures were cases included in the hospital which had not been accredited before the combined measures, and the post-combined measures were cases included in the hospital which had been accredited after the combined measures. 513 patients were excluded if they died before or within 10 min of hospital arrival [15,16], or if a contraindication was documented as the reason for withholding the program; thus, 4262 patients were included in the study. After propensity-score matching (PSM) by controlling the confounding factors (Supplementary Table S1) [17], 1078 pairs were finally included for analysis pre- and post-combined measures.Quality metrics included prehospital processes, in-hospital processes, and outcome indicators (Supplementary Table S2), which were used as the core set of quality indicators for measuring CPC performance in the quarterly and annually benchmarked reports, developed by the China CPC Headquarters. The quality metrics were key performance indicators, based on the ACC/AHA Performance Measures and Class I Recommendations from the most updated ACC/AHA clinical practice guidelines. Accredited CPCs should continuously report data for monitoring and feedback to the China CPC Data Reporting Database. Improvements in adherence to the guideline recommendations are facilitated through monthly and quarterly hospital-specific performance feedback reports.We assessed the changes in the quality metrics of STEMI care pre- and post-combined measures. We also compared the changes in quality metrics between secondary and tertiary hospitals, among patients who had different modes of transfer, and between suburban and urban areas. Changes in quality metrics were assessed using univariate analyses, including the Kruskal–Wallis test, chi-square test, t-test and one-way analysis of variance. Fisher’s exact test was used to compute 95% confidence intervals for each quality metric. p-values < 0.05 were considered statistically significant. All statistical analyses were conducted in R software (R Foundation for Statistical Computing, Vienna, Austria, and Version 3.6.3).Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.Before the PSM, patients enrolled after the combined measures (post-combined measures patients) were relatively younger (60.40 ± 14.65 years vs. 61.91 ± 14.97 years, p = 0.004) than those enrolled before the combined measures (pre-combined measures patients). Post-combined measures patients were less likely to have sustained chest pain (77.2% vs. 77.3%, p = 0.001) or intermittent chest pain (22.2% vs. 14.6%, p < 0.001), and relief of chest pain had a higher percentage (4.4% vs. 2.5%, p = 0.001). Regarding vital signs, the heart rate per minute was slightly lower (78.56 ± 18.23 vs. 77.07 ± 19.61, p = 0.021) in the post-combined measures patients. After the PSM, the mirrored histograms before and after matching are shown in Figure S1 to present the propensity score distribution. the standard deviation of all covariates was <2%, indicating no significant differences between the two groups in demographics, chest pain symptoms, vital signs or Killip grading (Table 1).Figure 1 shows the changes in reperfusion time and in-hospital mortality before and after the combined measures. The median total time (q1, q3) from symptom onset to PCI did not significantly change from 212.5 (150.8, 325.5) to 213.0 (142.0, 372.0) minutes. However, the door-to-balloon and the median FMC-to-device (q1, q3) time decreased from 85.0 (67.0, 108.0) and 98.0 (78.0, 132.8) minutes to 78 (61.5, 92.0) and 88 (71.0, 124.5) minutes, respectively (p < 0.001). The percentage of patients meeting guideline goals increased significantly, except for the percentage of patients arriving at the first hospital by ambulance (14.6% vs. 12.2%, p < 0.001), the proportion of call-to-EMS time ≤15 min for ambulance-transported cases (45.6% vs. 47.8%, p = 0.622) and the proportion of EMS-to-first electrocardiogram (ECG) time ≤ 10 min for ambulance-transported cases (75.0% vs. 91.5%, p = 0.171). The in-hospital mortality persisted before and after the combined measures (2.9% vs. 2.9%, p = 1.000).We compared the changes in reperfusion time and in-hospital mortality before and after the combined measures by tertiary and secondary hospitals (Table 2). For prehospital processes, although most indicators of the secondary hospitals did not show improvement after the combined measures, the tertiary hospitals exhibited significant increases in some indicators, including the rate of prehospital ECGs (21.4% vs. 27.1%, p = 0.006), the proportion of onset-to-FMC (EMS arrival or walk-in to emergency department) time ≤ 60 min (27.8% vs. 32.3%, p = 0.047) and the proportion of ambulance ECG-to-door time ≤ 15 min for ambulance-transported cases (2.4% vs. 12.2%, p = 0.001).For in-hospital processes, most indicators of tertiary hospitals improved significantly after the combined measures, including the proportion of door-to-balloon time ≤ 60 min (18.9% vs. 24.2%, p = 0.029), the proportion of FMC-to-device time ≤ 90 min (44.8% vs. 55.8%, p < 0.001), the door-to-balloon time (85.0 [66.8, 108.3] min vs. 78.0 [61.3, 92.0] min, p < 0.001), the FMC-to-device time (98.0 [78.0, 132.3] min vs. 88.0 [71.0, 124.8] min, p < 0.001), and the PCI rate (70.8% vs. 76.9%, p = 0.003).Regarding outcome indicators, the in-hospital mortality decreased in both the tertiary and secondary hospitals, although not significantly. However, the heart failure incidence rate increased significantly in the tertiary hospitals (13.4% vs. 21.9%, p < 0.001).We compared the changes in quality metrics before and after the combined measures among walk-in patients, in-hospital patients, those transferred directly via EMS, and those transferred from other hospitals (Table 3). For walk-in patients, most indicators improved significantly after the combined measures, including the proportion of onset-to-FMC time ≤ 60 min (22.6% vs. 29.7%, p = 0.004), the proportion of door-to-balloon time ≤ 60 min (8.8% vs. 17.2%, p = 0.001), the proportion of FMC-to-device time ≤ 90 min (55.0% vs. 69.8%, p < 0.001), and the proportion of onset-to-device time ≤ 120 min (12.3% vs. 19.8%, p = 0.005), as well as the door-to-balloon time (89.0 [75.0, 120.0] min vs. 82.0 [68.0, 99.3] min, p < 0.001), the FMC-to-device time (88.00 [73.0, 119.0] min vs. 80.00 [66.0, 98.0] min, p < 0.001), and the PCI rate (52.7% vs. 78.2%, p < 0.001).For patients transferred to the hospital directly via EMS, most indicators improved significantly, including the proportion of onset-to-FMC time ≤ 60 min (39.6% vs. 62.9%, p = 0.002), the door-to-balloon time (87.0 [72.0, 109.0] min vs. 71.0 [60.0, 88.0] min, p = 0.005), the FMC-to-device time (102.5 [82.3, 128.3] min vs. 85.0 [68.0, 105.0] min, p = 0.004), and the onset-to-device time (175.0 [140.0, 254.0] min vs. 141.5 [115.3, 175.8] min, p = 0.010).For transferred-in patients, the onset-to-device time (240.0 [171.0, 347.0] min vs. 277.00 [193.0, 465.0] min, p = 0.004) increased significantly after the combined measures. The door-to-balloon time (65.0 [47.8, 84.5] min vs. 71.5 [49.0, 87.0] min, p = 0.158) and the FMC-to-device time (121.5 [98.3, 161.8] min vs. 136.0 [99.0, 185.0] min, p = 0.055) also increased, but not significantly.We compared the changes before and after combined measures between urban and suburban areas (Table 4). Pre-combined measures data showed that all in-hospital indicators were better in urban hospitals than in suburban hospitals, except the rate of intensive statin use within 24 h. In both urban and suburban hospitals, most indicators improved significantly after the combined measures, including the rate of prehospital ECGs, the proportion of onset-to-FMC time ≤ 60 min (EMS arrival or walk-in to emergency department), the proportion of FMC-to-device time ≤ 90 min, the door-to-balloon time, the FMC-to-device time, and the PCI rate. However, for urban areas, the rate of intensive statin use within 24 h (67.7% vs. 54.3%, p < 0.001) decreased, and the heart failure incidence rate (14.2% vs. 24.5%, p < 0.001) increased. For suburban areas, the percentage of patients arriving at the first hospital by ambulance (10.7% vs. 5.5%, p < 0.001) decreased significantly.Many studies have shown that the transmission of prehospital response is an important base for bypassing the emergency department and CCU. It is also an important way to reduce the in-hospital delay and is of great significance in shortening the treatment time, which is consistent with the Suzhou results [7,18]. The EMS of Suzhou represents the combined measures for translating efficacy into effectiveness as narrowing the evidence-based gap in the treatment of STEMI. The EMS of Suzhou focuses on the establishment of regional systems of STEMI care that can be adapted to the context of China’s health system. This study, to our knowledge, is the first retrospective study to examine the impacts of the regionalization of STEMI care, including the combined QI initiatives and the establishment of the information sharing system between the MPDS and the hospital-based CPC registry. The findings have implications for further promoting the measures, with the ongoing engagement of QI efforts and for their potential implementation in low- and middle-income settings where the burden of STEMI is increasing at an unprecedented rate [19].We found that implementing the combined measures improved many process indicators significantly. However, these improvements did not translate into better clinical outcomes. The results are consistent with findings from the recent systematic review that analyzed 32 studies (randomized controlled trials and nonrandomized quasi-experimental studies) from 5858 records for QI interventions on process of care measures and clinical outcomes [20]. Although there is substantial heterogeneity in the QI interventions, a range of studies show that certain interventions are associated with better processes of delivering STEMI care [21,22,23]. In our study, the most important finding is that there was a significant improvement in the in-hospital service quality. We found that although the in-hospital mortality rate and onset-to-device time did not change significantly, the time from FMC to device was significantly shortened. The achievements can be attributed to coordination of care at a regional level, which has gained support and been incorporated into national guidelines for CPC accreditation.Generally, it is important for us to coordinate emergency services systems and population campaigns to raise awareness about STEMI management and avoid preventable delay. The key elements of our efforts included broad regional leadership, highly-developed eHealth technology [24], and well-organized coordinators. First, a nationwide collaborative network which led the implementation and supervision of the QI initiatives is well organized by the CSC under the authorization of the NHC. The operational structure consists of the Management Board and the Executive Board, between which the the China CPC Headquarters is the link, located in Suzhou. The China CPC Headquarters is committed to monitoring the implementation of the QI initiatives through continuous audit and feedback on the China CPC Data Reporting Platform. The China CPC Headquarters creates practical tools for sites to help improve the quality of data reporting and consistency across the registered hospitals. Examples of these tools include guidelines for data reporting, a set of internal quality assurance tools, and a yearly data audit program. Secondly, with the support of the NHC of Suzhou, the online software, mobile APPs and other eHealth technology were developed for real-time data reporting, to improve data quality and reporting efficiency. Suzhou took the lead in establishing the MPDS and the information sharing system by linking the MPDS and the hospital-based CPC registry, to facilitate the coordination of care at the time of entering the EMS system. STEMI patients can be transferred to an appropriate hospital based on disease conditions with the support of the MPDS, and can be guided through first aid by the first bystander before the ambulance arrives by the tele-guide of the dispatcher. The prehospital ECGs on the ambulance can be transmit to the hospital through the online software, thereby allowing transfer to the Cath Lab, bypassing the emergency department and coronary care unit. Third, there are well-organized coordinators affiliated with each of the hospital-based CPCs. The coordinators are responsible for the coordination and monitoring of the QI activities, including maintaining care coordination between multidisciplinary clinical services for operational integrity with respect to patients with STEMI, conducting performance appraisal and feedback, and training targeted to healthcare professionals and towards health education among community residents. Prior studies have shown that a dedicated coordinator in charge of implementing systematic improvements within hospitals and EMS agencies could play a critical role in maintaining coordination of care [25,26].Although the NCPCP has made achievements in reducing the reperfusion times, the total delay time (from symptom onset to reperfusion) was about 212.5 min, much longer than the guideline recommends [27]. Onset-to-FMC time accounted for about 53.9% of total delayed time. Patient delay (from onset to when patients enter the EMS system) still appears to hinder the timely delivery of STEMI care. These results may be attributable to low population awareness of signs and symptoms of STEMI or the option of calling EMS, or lack of appropriate EMS response. For a better understanding of coordinated care, we compared the quality of STEMI care by transfer mode. We found that cases that went directly via EMS had a higher increase in most process indicators than cases who self-transported to hospitals. Nevertheless, the proportion of patients presenting directly to PCI-capable hospitals via EMS was much lower than that of self-transported patients, and the proportion declined after the QI initiatives. Therefore, our findings serve to emphasize that we should target specific strategies at the population level, including improving early recognition of STEMI symptoms and awareness of the option of calling EMS among the public, to increase the use of prehospital services and prehospital activation of PCI. To further improve the implementation of the combined measures, EMS should focus more on prehospital processes.Whereas the specific goals of the NCPCP have been focused on process of care measures, the ultimate goal is to improve clinical outcomes. In the examined time-period, the adjusted in-hospitals mortality increased, despite the improvements in the reperfusion time after the combined measures. This finding can be explained by three factors. First, the major reason for these results was the inclusion of all hospitals, regardless of the duration or the progress of the implementation of QI activities. By providing aggregate results, the absolute potential of the QI activities may not be reflected because some hospitals that participated late in the program were unable to implement improvements within a short timeframe. Second, the program was made available to all secondary and tertiary hospitals, and hospitals continue to join the program in a staggered manner. At the beginning of the program, the majority of registered hospitals were tertiary hospitals with PCI capability. While with the promotion of the program nationwide, an increasing number of secondary hospitals without PCI capability participated in the program, which may lead to worse performance on delivering care and poor improvements in performance. We found that secondary hospitals had longer onset-to-device times than secondary hospitals. Thus, coordinated and hierarchical care between secondary and tertiary hospitals should be enhanced to reduce system delay (from when patients enter the EMS system to reperfusion). Third, the increase of adjusted in-hospital mortality could also be attributable to the enrollment of consecutive patients presenting increasing severity of symptoms, which was documented to be related to increased mortality risk for STEMI. We detected increased rates of cardiac arrest and cardiogenic shock, and an increase in patients with Killip class IV, after the QI initiatives. Before the combined measures, many deaths may have occurred before admission to the hospital of a sudden and unexpected nature, which can be associated with delays in seeking care. With the increasing number of cases entering hospitals since the program, cases with lengthy delays and severe symptoms were recorded on the China CPC Reporting Platform.We further found regional variations in the quality of STEMI care. Generally, the process of care was better for well-developed urban areas than less-developed suburban areas. Most of the top tertiary hospitals in Suzhou are located in the center of the urban area, and STEMI patients in the suburbs have difficulty transferring to these hospitals in a timely manner. However, the percentage of patients arriving at the first hospital by ambulance after the QI initiative decreased significantly, making it more difficult for patients to reach urban hospitals in time for treatment. Therefore, attention should be paid to increasing the distribution of ambulances in suburban areas, and to improving the prehospital first-aid system in suburban areas. Short-term goals should focus on promoting health education on STEMI rescue and improving public knowledge on the use of EMS at the community level, such us organizing community health missions and training or putting up STEMI rescue related posters. Long-term goals should focus on improving prehospital medical resources to meet the needs of the local population for more effectively mobilizing prehospital resources in the suburban areas. The outcomes of this study could be translated into a systematic solution for improving the quality of STEMI care, generating knowledge about process outcomes and core components that is transferrable, and where local adaptation is needed for replication in other settings. This actionable knowledge is also critical for implementors of scale-up activities in low- and middle-income settings.The variability in improvement was related primarily to the speed with which districts could implement effective regional systems of STEMI care. A large proportion of work during the implementation of the program was to persuade the majority of health bureaus of districts to put forward a systemic design for establishing a system of coordinated care. The biggest resistance lies in the fragmented financing and supervision for prehospital and hospital care. This highlights the challenge of pursuing such a large-scale implementation during a relatively short time period. Nevertheless, the results from the most improved districts indicated that it was possible to improve reperfusion time step by step according to our approach. From our experience, districts most able to improve reperfusion time had common characteristics, including EMS leadership concentrated to a few dominant agencies, and active engagement by coordinators. Future studies should focus on regional disparities in the performance of the program, to assess the factors among myriad local socioeconomic and political factors causing disparities in quality of care. In addition, more detailed analyses and case studies with low quality of care are needed to identify the interventions that may lead to better outcomes, and that could be applied to the other local settings.This study had several limitations. First, the study was limited to Suzhou, which limits the generalizability of the results because Suzhou differs from other cities in terms of implementation of the QI initiative, socioeconomic development, population health, and the EMS system. Extrapolation of these results requires further study. Data from other cities should be further explored and verified to improve the accuracy and adaptability of these conclusions. Second, this study was a retrospective study; therefore, we could not establish a causal relationship. Even with PSM, unobserved variables may have biased the results. Follow-up studies with prospective randomized studies are needed. Third, while the implementation of complex strategies that require different steps will become more efficient after a long time of implementation efforts, the lack of mixed analysis of long-term intervention contexts together with recently included hospitals is still a limitation. Follow-up long term analysis is needed in the future.This study showed that the combined measures, including the multifaceted QI initiatives and the establishment of an information sharing system, can improve quality of care and showed potential for improving clinical outcomes among STEMI patients. For further promoting the measures, patient delay should be addressed, to reduce the delay for entry into the EMS system, especially in suburban areas, and on transferred-in inpatients. The regional disparities in performance improvement can be related to the speed with which districts could implement effective regional systems of STEMI care. The establishment of coordinated care needs to be accompanied with solving the fragmented situation of the prehospital and hospital care which should be designed specifically to fit into the health systems on a regional basis. The consecutive recruitment of accredited hospitals warrants more efforts to enhance the implementation of the QI initiatives.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116045/s1, Supplement Table S1: Definition and measures of quality metrics. Supplement Table S2: Definition and measures of confounding variables for propensity score matching. Supplement Figure S1: Mirrored Histogram before (A) and after (B) propensity score matching. X axis is the number of patients in each group. Y axis is the propensity score. The blue bar presents the pre-combined measures group and the red bar for the post-combined measures group.J.M. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Design of the study: Y.J., Z.-J.Z. Collection, management, analysis, and interpretation of the data: Y.J., J.M. Preparation, review, or approval of the manuscript: Y.J., Z.-J.Z. Decision to submit the manuscript for publication: J.M., X.D., Y.J., Z.-J.Z. We thank the China Cardiovascular Association for providing the database, and Suzhou Emergency Center for comments on the study’s findings. All authors have read and agreed to the published version of the manuscript.This paper was supported by the National Natural Science Foundation of China (No. 71904004). The study sponsor has no role in study design, data analysis and interpretation of data, the writing of manuscript, or the decision to submit the paper for publication.The study was conducted according to the guidelines of the Decla-ration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Peking University First Hospital Biomedical Research Ethics Committee (2020Research242).Not applicable.The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the license of National Chest Pain Center (CPC) Data Reporting Database.The authors declare no conflict of interest.No financial disclosures were reported by the authors of this paper.Comparison of pre- and post-combined measures after propensity-score matching (* For ambulance transported cases). (A): Time change of pre- and post-combined measures; (B): Rate change of pre- and post-combined measures.Characteristics of patients with discharged diagnosed STEMI Before and After PSM.* Mean(SD).Changes in reperfusion times and in-hospital mortality among STEMI patients pre- and post-combined measures quarter by hospitals.Changes in reperfusion times and in-hospital mortality among STEMI patients pre- and post-combined measures quarter by transfer mode.Disparity between urban and suburban area in reperfusion times and in-hospital mortality among STEMI patients.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06046.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Diabetic peripheral neuropathy (DPN) is a common complication of type 2 diabetes mellitus (DM). DPN causes a decrease in proprioception, which could reduce balance ability. We investigated the association of impaired vibration sense, based on vibration perception threshold (VPT), with assessments of balance and other factors affecting balance impairment and fear of falling in patients with type 2 DM. Sixty-three patients with DM aged >50 years were categorized as having normal vibration sense (NVS; n = 34) or impaired vibration sense (IVS; n = 29) according to a VPT value of 8.9 μm. The following parameters were evaluated for all patients: postural steadiness through the fall index using posturography, functional balance through the Berg Balance Scale (BBS), the Timed Up and Go test (TUG), and fear of falling through the Falls Efficacy Scale-International (FES-I). The IVS group showed a significantly greater balance impairment in fall index, BBS, and TUG, as well as greater fear of falling on the FES-I than the NVS group. The linear regression analysis showed that the fall index was associated only with the VPT, whereas BBS, TUG, and FES-I were associated with the VPT, age, and/or lower extremity muscle strength. VPT, age, and/or muscle strength were identified as predictors of balance and fear of falling in patients with type 2 DM. Therefore, along with age and lower extremity strength, the VPT can be useful for balance assessment in patients with type 2 DM.Diabetes is a serious and chronic condition that negatively affects human life and burdens the health care system. The global prevalence of type 2 diabetes mellitus (DM) is increasing annually and was estimated at 6059 cases per 100,000 in 2017, and the number is predicted to rise to 7079 cases per 100,000 by 2030 [1]. The global diabetes-related health expenditure was estimated to increase to USD 760.3 billion in 2019, and is predicted to increase to USD 845.0 billion by 2045 [2]. Diabetes can cause serious damage to the heart and blood vessels, eyes, kidneys, nerves, and teeth, which can lead to serious health problems [3]. Diabetic peripheral neuropathy (DPN) is one of the most common complications of type 2 DM [4,5]. Adults with type 2 DM have about 40% prevalence of DPN [6]. DPN is associated with foot ulceration, amputation [7], and balance impairment in patients with type 2 DM [8].The proprioceptive feedback of the lower extremities, in combination with visual and vestibular senses, provides the main sensory information to maintain and control posture. DPN causes a decrease in proprioception, such as a sense of position, decrease in the sense of vibration, loss of somatosensory feedback, and inappropriate motor response, which can reduce the ability to control balance [9,10]. In fact, patients with DPN are at a high risk of falling [11,12,13,14,15,16]. Falling has destructive consequences, including a decline in mobility, activity avoidance, institutionalization, mortality, and higher social and economic burden [8,17].DPN can be assessed in several ways using different types of symptom scores and several kinds of measurements [18]. The vibration perception threshold (VPT) with a quantitative sensory testing (QST) approach was demonstrated to be a useful screening tool in many studies [19,20,21,22]. Because vibration sense is often impaired in large myelinated nerve fiber deficits such as in DPN, the VPT is widely used for screening for large fiber function in DM [23,24]. High VPT values are associated with DPN symptoms and predictive of serious complications, such as foot ulcers and amputation [25]. VPT is painless [26] and practical to implement in routine clinical care. Information on the VPT and predicting balance impairment would be helpful for identifying patients who need balance assessments and providing patient education to prevent falls in outpatient clinics [27]. Although several studies have evaluated balancing ability in patients with DPN, the assessment tools for DPN are very diverse across studies [10,13,16,20,28,29]. Few studies have comprehensively evaluated the association between the VPT and various balance assessment methods.Identifying factors associated with balance impairment is another essential part of fall risk assessments of patients with type 2 DM. Although previous studies have reported that DPN is associated with an impairment of balancing ability [11,12,13,14,15,16], it can be expected that several factors cause balance impairment in diabetic patients either alone or in combination with DPN. However, studies on factors related to balance and fall risk in patients with type 2 DM are still insufficient, and the results are unclear [15]. Functional balance plays an important role in performing many complicated tasks during daily life [14,16]. Since the body tries to find additional ways to reduce postural instability at the functional level [14,15], related factors may differ for postural steadiness and functional balance. In addition to balance impairment, psychological factors, such as a fear of falling, with or without a history of falling, are risk factors for falls. A fear of falling is a psychological obstacle to the performance of physical activities and is predictive of future falls [30]. Both physical and psychological approaches will enable more focused fall risk assessment. Thus, it is necessary to evaluate factors related to balance ability through both physical and psychological assessments, including postural steadiness, functional balance, and the fear of falling.The objective of this study was: (1) to investigate the association between VPT and various balance measures, including postural steadiness, functional balance, and fear of falling, and (2) to identify significant factors affecting balance and fear of falling in patients with type 2 DM.In this cross-sectional study, a total of 63 patients with type 2 DM were recruited as subjects among patients who visited the diabetes clinic in the Department of Endocrinology and Diabetes of the University Hospital for the purpose of controlling type 2 DM. All subjects were over 50 years of age, medically stable, understood the instructions (Mini Mental State Examination score >24), and were able to stand independently. Subjects were excluded if they had musculoskeletal disorders (e.g., spinal radiculopathy, foot ulcer, lower limb amputation), rheumatic disease, surgery (e.g., spinal surgery, replacement of lower limb joint), history of trauma of the lower extremity that could impair balance ability and lower extremity function, neurological disorders (e.g., stroke, Parkinson’s disease, multiple sclerosis, spinal radiculopathy), visual impairment, or any other vestibular disease. All subjects provided written informed consent, and this study was approved by the Institutional Review Board of our hospital (EMCS 2020-04-023).The VPT was measured using a vibratory sensory analyzer VSA-3000 (Medoc Ltd., Ramat Yishai, Israel), a quantitative sensory testing (QST) computerized device. The examiner, who was not involved in administering the balance outcome measures, explained VPT test methods to the patients. Before the test trial, three practice trials were conducted to allow the examiner to explain the procedure and to help the patient become acquainted with the test. The patients were asked to sit on a chair. The VPT was assessed at four sites of the bare testing sole (big toe, 1st metatarsal, 5th metatarsal, and heel) [31] by delivering a 100 Hz vibration stimulus. The method of limits was used for the vibratory thresholds [32]. The intensity of the stimulus was increased at a rate of 0.8 μm/s from the baseline (0 μm). As soon as the patient detected the stimulus, a button was pressed to stop the machine from delivering it, and thereafter the next trial was initiated from the baseline. The average result of three trials was considered as the VPT for each foot site. Finally, the VPT value was taken as the average VPT from the eight site measurements. Impaired vibration sense was detected using the threshold value of VPT >8.9 μm, which was reported to be the best cut-off value for detecting DPN [20]. Subjects were classified into normal vibration sense (NVS) and impaired vibration sense (IVS) groups, based on the VPT cut-off value of 8.9 μm.The nerve conduction study (NCS) was performed to assess DPN severity by the physician experienced in the electrodiagnostic study. The composite score was calculated to confirm the significant difference in DPN severity between the two groups based on VPT, and the correlation between VPT value and composite score. The composite score consisted of peroneal motor distal latency, peroneal motor amplitude, peroneal motor conduction velocity, tibial motor distal latency, and sural sensory amplitude. The peroneal motor nerve was stimulated at the ankle and fibular head, with recording over the extensor digitorum brevis muscle. The conduction velocity was measured between the ankle and fibular head. The tibial motor nerve was stimulated posterior to the medial malleolus, with recording over the abductor hallucis muscle. The sural nerve was stimulated at 14 cm proximal to the lateral malleolus in the posterior calf line, with recording posterior to the lateral malleolus. Each component was transformed to an abnormal percentile relative to the distribution of the values of normal NCS (<95th = 0; ≥95th–99th = 1; ≥99th–99.9th = 2; ≥99.9th = 3). The points of the five components were totaled, divided by the number of attributes with available values, and multiplied by five [33].Postural steadiness was evaluated by static posturography using the Tetrax® (Sunlight Medical Ltd., Ramat Gan, Israel) with the supervision of a physical therapist. The device has four individual plates (A-B-C-D, A: left heel; B: left toes; C: right heel; D: right toes) for measuring the vertical pressure fluctuations with a single axis load cell, which is one of the load sensor types using the strain gauge [34]. The vertical pressure distributed across each force plate was measured while standing in an upright position for 32 s. The data of vertical pressure fluctuations were computerized and analyzed by the Tetrax® software to obtain the fall risk index, stability index (ST), Fourier index (FI), weight distribution index (WDI), and synchronization index. The patients stood in the center of four force plates and were instructed to maintain eight different postures, such as head rotation, upward, and downward, with eyes open, eyes closed, and on unstable surfaces. In this study, the postural steadiness was estimated with reference to the fall risk index, ST, and FI for a medium to high frequency of 0.5–1 Hz [11,34,35,36,37].Fall risk index: The fall risk index is globally calculated using the data of ST, FI, WDI, and synchronization index, considering the oscillation velocities computed by the Tetrax® program [34,36,37]. It is numerically expressed from 0 to 100. A value of 0–36 indicates low risk, 37–58 indicates moderate risk, and 59–100 indicates high risk. A higher fall risk index indicates a greater risk of falling [34,35,36,37].Stability index (ST): The ST indicates the patient’s overall stability and ability to control postural change. The amount of postural sway is analyzed using the square root of the sum of the squared differences between consecutive pressure signals. Then, it is divided by the patient’s body weight [34,35,37]. A higher index score indicates more postural instability [34,35,36,37].Fourier index (FI): The FI is a regression analysis of the intensity of postural sway mathematically calculated using a Fourier transformation. It represents the frequency of postural sway within a variable spectrum between 0.01 and 3.0 Hz. Fourier power values are compared with the regression curve calculated by the software, evaluating the difference of coefficient between a graph of measured data and theoretically ideal regression [37]. The FI value for medium to high frequency of 0.5–1 Hz (F 5–6) is high when the patients present with somatosensory dysfunction of the lower extremities and spine [34,38]. A higher index score for frequency represents greater postural instability [35,36,37].The functional balance assessments were conducted by a physical therapist with 30 min of rest period after a postural steadiness assessment. The general instruction of functional balance tests was explained before the actual examination.Berg Balance Scale (BBS): The BBS is a tool for evaluating the functional balance required for movements related to actions encountered in daily life. It is composed of 14 tests, with each test scored from 0 to 4 points. The maximum score is 56 points, and a higher score reflects better functional balance [14,39,40].Timed up and go (TUG) test: The TUG test was performed 30 min after the evaluation of BBS. The TUG test is a dynamic functional balance test, which measures the time it takes to stand up from the chair, turn around at a turning point at a 3 m distance at ordinary walking speed, and sit back on the same chair. The TUG test is useful for assessing a person’s functional mobility and the risk of falls in community-dwelling older people. In older adults, if the TUG test requires more than 12 s, the risk of falls is high [41,42].Falls Efficacy Scale-International (FES-I): The FES-I is a widely used self-report questionnaire tool for evaluating concerns about falls while carrying out routine activities. The FES-I is simple and easy to administer, and its score categories cover both easy and difficult physical and social activities inside and outside the house [43]. It is composed of 16 items that evaluate basic activities of daily living, with each item scored from 1 to 4 points. The total score ranges from 16 to 64 points, and a higher score reflects a greater fear of falling when performing daily activities [44,45].To confirm the differences in fall index, BBS, TUG, and FES-I between the two groups (NVS and IVS) in this exploratory study, the sample size was calculated based on effect size = 0.8 (the large effect by Cohen), alpha = 0.05, power = 0.8, by an independent sample t-test using G-power 3.1.2 (Heine Heinrich University, Dusseldorf, Germany). The minimum required sample size was 26 in each group. Therefore, it was intended to enroll more than 26 patients per group. The baseline characteristics were analyzed using an independent sample t-test for continuous variables and a chi-squared test for nominal variables. An independent sample t-test was performed to assess the difference in postural steadiness and functional balance capabilities between the two groups. A Pearson’s correlation analysis was used to assess the correlation between balance parameters and VPT. Finally, a stepwise multiple linear regression was used to evaluate factors that affect postural steadiness and functional balance, respectively. Independent variables included age, sex, height, weight, body mass index (BMI), DM duration, glycated hemoglobin levels (HbA1c), Medical Research Council (MRC) score, presence of hypertension, chronic kidney disease, retinopathy, and the value of the VPT. To determine the final regression model, a variable was added to or removed from the model according to the repeated stepwise method algorithm that included variables for which the p-value was <0.05, and removed variables for which the p-value was >0.1. All statistical analyses were performed using SPSS version 22.0 (IBM, Armonk, NY, USA). Statistical significance was set at p < 0.05.Sixty-three subjects were recruited for this study and completed all the tests. Based on the VPT value of 8.9, the patients were classified into two groups: NVS (n = 34) and IVS (n = 29) groups. Demographic and clinical differences between the two groups are shown in Table 1. The IVS group was associated with higher VPT values and longer diabetes duration than the NVS group (p < 0.01). The composite score in the NCS was significantly higher in the IVS group than in the NVS group (p < 0.001). In the correlation analysis, there was a significant correlation between VPT and composite score (r = 0.673, p < 0.05). The differences in the other studied parameters (i.e., group, sex, age, height, weight, BMI, HbA1c, MRC score, hypertension, DM nephropathy, retinopathy, chronic heart disease, and chronic lung disease) were not significantly different between the two groups (p > 0.05). Regarding the MRC score, five patients had lower limb weakness in the IVS group even though they had no musculoskeletal or neurological disorders belonging to the exclusion criteria of this study.Table 2 and Table 3 show a comparative analysis of the postural steadiness parameters: fall risk index, ST, and FI 5–6 for both groups using Tetrax®. The fall risk index in the IVS group was significantly higher than that in the NVS group (39.1 vs. 82.9%, p < 0.001; Table 2). The ST and FI 5–6 of the IVS group were significantly higher than those of the NVS group, regardless of postural conditions (p < 0.05, Table 3).Table 4 shows a comparative analysis of functional balance parameters and the fear of falling in both groups. Categorization in the IVS group was associated with a decreased BBS (52.0 vs. 43.7 points, p < 0.001), with increased TUG test scores (8.8 vs. 12.0 s, p = 0.001) compared to the NVS group. The FES-I score was significantly higher in the IVS group than in the NVS group (20.2 vs. 26.0 points, p < 0.001).Using Pearson’s correlation, the VPT value was strongly correlated with postural steadiness parameters of the fall risk index (r = 0.730, p < 0.001; Table 5). The VPT value was significantly correlated with the functional balance parameters of BBS (r = −0.680, p < 0.001) and TUG test (r = 0.417, p = 0.001; Table 5). The VPT value was also significantly correlated with the FES-I (r = 0.423, p = 0.001; Table 5).A multivariate regression analysis was performed using the significant factors in univariate regression analysis to investigate the factors related to postural steadiness, functional balance, and fear of falling. The fall risk index was significantly associated only with VPT values after adjusting for covariate factors. Not shown in the table, VPT was the only significant predictor of the STs and FIs (F 5–6) for eight postural conditions.Table 6 displays the final multivariate linear regression model for the fall risk index, BBS, TUG, and FES-I. In the multivariate analysis of fall risk index, a higher VPT was associated with a higher fall risk index with a beta coefficient (95% CI) of 2.472 (1.881–3.064), p < 0.001. In the multivariate analysis of BBS, a higher VPT and older age were associated with lower BBS with beta coefficients (95% CI) of −0.412 (−0.529–−0.294), p < 0.001; and −0.264 (−0.369–−0.160), p < 0.001, respectively. In the multivariate analysis of TUG, a higher VPT, older age, and lower MRC score were associated with a higher TUG with beta coefficients (95% CI) of 0.103 (0.017–0.188), p = 0.020; 0.209 (0.134–0.283), p < 0.001; and −1.848 (−3.133–−0.562), p = 0.006, respectively. In the multivariate analysis of FES-I, higher VPT, older age, and lower MRC score were associated with higher FES-I with beta coefficients (95% CI) of 0.209 (0.050–0.369), p = 0.011; 0.207 (0.067–0.347), p = 0.004; and −2.871 (−5.270–−0.472), p = 0.020, respectively.These results indicate that patients with DM with IVS, as indicated by the VPT, exhibited decreased postural steadiness and functional balance and increased fear of falling compared to those with NVS. The current study also suggests that the VPT value, as well as the presence of IVS, were significantly associated with postural steadiness, functional balance, and fear of falling. Diminished vibration sense contributes to balance deficits and increased fall risk [46,47]. In Hafström’s study [46], decreased vibration sense had negative effects on perceived and functional balance in relatively healthy older adults. In Bergin’s study [47], vibration perception was increased in the patients with peripheral neuropathies, and there was a significant correlation between vibration sense and body sway. The results of our study are consistent with the results of previous studies for diabetic patients, which suggested that DPN assessed with various kinds of measurements was associated with postural steadiness and functional balance [14,29,48,49,50].To date, the NCS is the most common method for the evaluation of DPN. However, it is painful, expensive, requires specific training, and quantitative assessment of the severity of DPN is more difficult compared to the use of VPT [26]. The VPT test can compensate for these limitations of the NCS. The VPT test is practical for clinical settings because it is non-invasive, painless, faster, and easier to perform compared to the NCS [18]. In our study, the VPT value showed a strong correlation with the composite score of NCS reflecting DPN severity. As a higher VPT value is associated with balance and fear of falling, this convenient [19] and reliable [21] quantitative assessment may represent a useful addition to the standardized evaluation of patients with type 2 DM.We identified differences in factors related to postural steadiness and functional balance capability in patients with type 2 DM. In terms of postural steadiness, computerized posturography using Tetrax® showed that only the VPT value influenced postural steadiness. Emam et al. reported that poor glycemic control, as assessed by HbA1c, along with DPN diagnosed by Hoffman’s reflex and electromyography, influenced postural instability using computerized posturography [51]. Guy et al. reported that quantitative sensory measurement of DPN and age were correlated with postural instability [10]. Differences across studies in factors reported affecting postural steadiness may be attributed to differences in the tools used to measure postural steadiness. In our study, the VPT value was the only factor that affected postural steadiness and showed an important correlation with postural balance using Tetrax® posturography assessment. This result suggests that in cases where it is necessary to evaluate the balance deficit exclusively due to DPN, postural steadiness assessment may be required.The results of this study showed that the VPT score and age had a significant association with the BBS and TUG test scores, which act as indicators of functional balance. This finding is consistent with previous studies that evaluated DPN by other methods [29,52]. Bogdan et al. reported that the severity of DPN as quantified using the Michigan Neuropathy Screening Instrument (MNSI), age, and depression symptoms were associated with impaired balance [29]. Further, Renata et al. reported that advanced age, absence of proprioceptive sense assessed by a 128 Hz tuning fork, disability, and absence of step strategy were associated with abnormal balance and mobility in elderly patients with type 2 DM [52]. Although there is a debate over whether the strength of the lower extremities affects balance ability in older people [53,54,55], lower extremity muscle strength of diabetic patients also affected the TUG test results in our study. The loss of muscle strength in these patients may have a significant impact on TUG test results, not on BBS. It can be speculated that the TUG test is affected by lower extremity muscle strength and the mobility aspect to a greater degree than the BBS tool. Moreover, the fall risk assessed by BBS may be underestimated due to a ceiling effect in patients with relatively higher levels of physical performance enrolled in this study.It was identified that muscle strength is a factor that influences functional balance in patients with type 2 DM in this study. The muscle strength is related to the severity of DPN and may be caused by neurogenic atrophy due to axonal degeneration of the motor fibers [56]. The muscle strength was an independent predictor of functional balance measured by the TUG test after adjusting for other covariate factors. These results suggest that both VPT evaluation and lower extremity muscle strength should be regularly monitored for balance assessment in patients with type 2 DM.The patients with IVS had a greater fear of falling than patients with NVS. This result is consistent with previous studies [29,57]. In Bogdan’s study [29], the FES-I score was significantly higher in the DPN group compared to the without DPN group based on MNSI. In Riandini‘s study [57], a fear of falling measured by FES-I showed a significant positive correlation with DPN severity quantified using MNSI. Our findings indicated that the VPT, age, and lower extremity strength had a statistically significant association with the fear of falling. With regard to the FES-I findings, patients with IVS may have a concern about inappropriate somatosensory feedback from the lower limbs due to DPN. Patients with DPN who have a history of falls tend to have a fear of falling and a cautious gait, as indicated by a decrease in walking speed, reduction in step length, and an increase in step width [58,59]. Especially in older patients with DPN, a high concern for falls may lead to changes in postural control strategies, such as rigid compensatory strategies [28]. However, an excessively rigid compensatory strategy can further increase the risk of falls. In addition, older patients with IVS who are afraid of falls may lose confidence, resulting in limitations of physical activity, physical frailty, and falls. This may lead to a vicious cycle that accelerates the progression of diabetes, which further increases the severity of DPN. Therefore, evaluation of VPT is necessary for providing proper interventions to restore confidence and prevent falls.This study had some limitations. First, because the follow-up evaluation was not performed, which is a limitation of cross-sectional studies, we could not confirm any association between changes in the VPT and changes observed in balance. A prospective follow-up study is required in the future. Second, for evaluation of Tetrax® posturography, patients with a relatively high level of functioning who could stand independently were enrolled in this study. However, it is meaningful that the VPT is useful for assessment of balance ability even in these patients who may not be suspected of poor balance in an outpatient setting. Using the advantage of VPT as a simple screening method, further research is needed in patients with type 2 DM with more severe disability and limited mobility.This study demonstrated significant associations between VPT and several measures of balance ability in patients with type 2 DM. Postural steadiness was mainly affected by the VPT; however, age and/or muscle strength in the lower extremity were also related to functional balance and fear of falling. Therefore, along with age and lower extremity strength, the VPT might be useful for balance assessment, thus serving as a marker for fall risk in patients with type 2 DM.Conceptualization, Y.-J.K. and H.C.; Data curation, J.J. and M.-G.K.; Formal analysis, J.J. and H.C.; Project administration, J.J. and H.C.; Investigation, J.J. and M.-G.K.; Resources, K.M. and K.-A.H.; Writing—Original Draft, J.J.; Writing—Review and Editing, H.C.; Supervision: H.C. All authors have read and agreed to the published version of the manuscript.This research received no external funding.This study was approved by the Institutional Review Board of Nowon Eulji University Hospital (EMCS 2020-04-023).All subjects gave their informed consent for inclusion before they participated in the study.The data presented in this study are available on request from the corresponding author. The data are not publicly available because they represent information that could compromise the privacy of the study participants.We would like to thank the physical therapists Sung Jae Lee and Sun Ik Cho for their efforts and during the study.The authors declare no conflict of interest.Baseline characteristics of patients with type 2 DM.Data are presented as number (n) and percentage or mean ± standard deviation. * p < 0.05 (calculated through independent sample t-test). DM: diabetes mellitus; NVS: normal vibration sense; IVS: impaired vibration sense; BMI: body mass index; HbA1c: glycosylated hemoglobin; MRC: Medical Research Council; VPT: vibration perception threshold; NCS: nerve conduction study.Postural steadiness in patients with DM with NVS vs. IVS according to VPT value.Data are presented as mean ± standard deviation. * p < 0.05 (calculated through independent sample t-test). VPT: vibration perception threshold; DM: diabetes mellitus; NVS: normal vibration sense; IVS: impaired vibration sense.Stability index and Fourier indices in eight positions in patients with DM with NVS vs. IVS according to VPT value.Data are presented as mean ± standard deviation. * p < 0.05 (calculated through independent sample t-test). VPT: vibration perception threshold; DM: diabetes mellitus; NVS: normal vibration sense; IVS: impaired vibration sense.Functional balance and fear of falling in patients with DM with NVS vs. IVS according to VPT values.Data are presented as mean ± standard deviation. * p < 0.05 (calculated through independent sample t-test). VPT: vibration perception threshold score; NVS: normal vibration sense; IVS: impaired vibration sense; DM: diabetes mellitus; BBS: Berg Balance Scale; TUG: Timed Up and Go test; FES-1: Falls Efficacy Scale-International.Correlations between VPT and postural steadiness, functional balance, and fear of falling in patients with type 2 DM.* p < 0.05 (refers to the p-value of Pearson’s correlation coefficient). VPT: vibration perception threshold score; BBS: Berg Balance Scale; TUG: Timed Up and Go test; FES-1: Falls Efficacy Scale-International.Factors associated with postural steadiness, functional balance, and fear of falling in patients with type 2 DM.* p < 0.05. DM: diabetes mellitus; BBS: Berg Balance Scale; TUG: Timed Up and Go test; FES-1: Falls Efficacy Scale-International; VPT: vibration perception threshold; MRC: Medical Research Council.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06047.txt ADDED Viewed

	@@ -0,0 +1,2 @@


1	+ These authors contributed equally to this work.Background: Indonesia ranks 7th highest in the world for the number of deaths caused by tobacco use including those caused by Chronic Obstructive Pulmonary Disease (COPD). The purpose of this study was to determine the influence of initial smoking age and habit on the incidence of COPD. Methods: This research was a case-control study. The sampling in this research took a systematic random sampling method. The samples of this study were 56 respondents of a case group and 56 respondents of a control group. This study was conducted at Ngudi Waluyo Hospital, Wlingi, Blitar from October to November 2017. Results: The factors that influenced the incidence of COPD were being male (p = 0.00; OR = 6.333; 95%CI = 2.776–14.450), a smoker (p = 0.00050; OR = 5.1318; 95%CI = 1.9004–13.8958), initially smoking at <15 years old (p = 0.00; OR = 11,769; 95%CI = 4.086–33.903), initially getting into a smoking habit at the age of <15 years old (OR = 12; CI = 1346–106,950), initially getting into a smoking habit at the age of ≥15 years old (OR = 3647; CI = 1625–8183) and having smoked for ≥30 years (OR = 8857; CI = 3298–23,787). Conclusion: There are three factors of smoking behavior that influence COPD: smoking habit, initial smoking age and smoking duration. Of all factors, forming a smoking habit at the age of <15 years old has the biggest risk (OR = 12; CI = 1346–106,950).Chronic obstructive pulmonary disease (COPD) is a non-communicable disease with an incidence worldwide that increases from year to year [1]. COPD is a major source of morbidity, mortality and cost in the Western world [2]. Currently, the number of people with COPD globally reaches 384 million; it is estimated that this number will continue to increase, with COPD becoming the third-leading cause of death in the world by 2030 [3]. The burden of chronic respiratory diseases is generally increasing across the globe, and COPD is one of the main causes of mortality and morbidity [2].The main risk factor for COPD is smoking [3]. Cigarette smoking has been shown to be the leading cause of COPD in the United States [4]. The pathogenesis of smoking-related COPD includes the protease, anti-protease and oxidant-antioxidant hypotheses and abnormal repair processes [5]. COPD caused by smoking is usually initiated by injury to the lungs. A smoking habit is the cause of eight out of ten cases of COPD [6]. More than 75% of COPD cases result from lung injury caused by a long period of smoking [7].Other risk factors that can influence COPD are environmental exposure to biomass fuels and air pollution, and host factors such as age, sex or socioeconomic status [3,8]. Age is often listed as a risk factor for COPD as all vital organs lose their function with age, hence the decline in lung function, which occurs progressively after about 25 years of age. In the case of COPD, the age factor plays a role in increased cell aging, stem cell fatigue, increased oxidative stress, changes in the extracellular matrix and a reduction in endogenous anti-aging molecules and protective pathways such as autophagy [9]. In the past, most studies have shown that the prevalence and mortality of COPD are higher in men than women, but recent studies in developed countries have reported that the prevalence of COPD in men and women is almost the same. This may be due to changes in smoking behavior patterns [10].COPD is the first cause of disability in the world [11]. Limited ability to perform daily activities occurs in three out of four cases of COPD [12]. The disease also limits a person’s ability to climb the stairs [13]. Early retirement occurs in 40% of patients with COPD [14]. In Europe, COPD accounts for 50% of total health funds each year and causes an annual productivity loss of €48.4 billion [15]. In the United States, it is estimated that COPD costs $50 billion each year, consisting of direct costs of as much as $30 billion and further indirect costs reaching $20 billion [16]. It is estimated that 12 million people suffer from COPD and that 120,000 die of the condition each year [17]. In Indonesia, COPD results in a total daily loss of productivity of 901,744 h and medical expenses of IDR 1,294,165,188,810; it is the sixth-leading cause of death nationally [18]. As such, it is necessary to make a preventive effort to reduce the prevalence of COPD. With that aim in mind, the purpose of this study is to determine the influence of initial smoking age and habit on the incidence of COPD.Health is a human right mandated by the 1945 Constitution of the Republic of Indonesia; therefore, there is a need tobacco control for protection against the effects of cigarette smoke, especially for people who are not active smokers, in accordance with Health Law No. 36 of 2009 article 115 concerning the establishment of a No Smoking Zone (KTR) policy as part of government efforts to protect the public in the public environment. Based on 2013 health data, out of 38 districts/cities, only nine had implemented the KTR policy in East Java Province [19].Blitar, a district in East Java, had still not implemented this regulation in 2018, despite one of the diseases caused by smoking, namely COPD, having a fairly high prevalence (3.7%) in the district compared to other diseases [19]. Based on a recapitulation of patient data from Ngudi Waluyo Wlingi Hospital, Blitar District, in 2016, the morbidity rate of COPD inpatients was ranked tenth with 226 patients, while for outpatient care, COPD was ranked thirteenth with 953 patients [18]. The main risk factor for COPD is exposure to cigarette smoke. Of the total population aged ≥10 years in Blitar District, 36.3% have a history of smoking habits, with 30% still actively smoking. In Blitar District, the average number of cigarettes smoked per person per day in 2013 was 9.7 cigarettes [19].This study took place in Ngudi Waluyo Wlingi Hospital, Blitar District, Indonesia. Worldwide, Indonesia has the 7th highest number of deaths due to COPD, while Blitar district is one of the districts in East Java, Indonesia that does not have regulations regarding smoking-free areas. As a result, the prevalence of smoking-related diseases is high in Blitar District, one of which is COPD. The prevalence of COPD in Blitar District, East Java, Indonesia is 3.7% [19]. Ngudi Waluyo Wlingi Regional Hospital was designated as a Regional Public Service Agency (BLUD) Hospital in Blitar. It is a teaching hospital that has 16 accredited services and, as such, a suitable place for research. Based on a recapitulation of patient data from Ngudi Waluyo Wlingi Hospital, Blitar District, in 2016, the morbidity rate of inpatient COPD was ranked tenth with 226 patients, while for outpatient care, COPD was ranked thirteenth with 953 people [20].This research was an analytical observational study using a case-control design. This research was conducted in April–December 2017 at Ngudi Waluyo Hospital in Wlingi, Blitar, East Java, Indonesia. The case group was people who had been diagnosed as suffering from COPD, while the control group was people who had never been diagnosed as suffering from COPD.The sampling used in this study was systematic random sampling. The case population in this study was patients with COPD aged ≥30 years at Ngudi Waluyo Wlingi Regional Hospital, Blitar District, East Java. The diagnosis of COPD was determined based on the doctor’s diagnosis on the patient’s medical record. Meanwhile, the control population in this study were all patients aged ≥30 years who did not have a history of COPD based on medical records from Ngudi Waluyo Wlingi Regional Hospital, Blitar District, East Java. The sample size was calculated using Lemeshow’s comparison case, which generated 56 samples in each group (1:1). The samples were divided into case and control groups. The case group consisted of samples with COPD, while the control group consisted of samples without COPD. The cases and controls were selected randomly using a systematic random sampling technique. The control group was selected based on the current age of the respondent, namely, at least 30 years old. Systematic random sampling was carried out in both the case and control groups based on the serial numbers of patients seeking treatment at Ngudi Waluyo Hospital. This was based on the number of patients receiving treatment divided by the number of samples needed; the sample was taken based on the multiple of the results of the distribution in the sampling frame.The dependent variable studied was COPD, while the independent variables studied were divided into two groups: characteristics of respondents (age, sex, job and education) and smoking behavior (smoking habits, early smoking age and duration of smoking). Determination of COPD status was made in the case group using medical records and in the control group using a questionnaire. The age variable was determined based on the interview date and the respondent’s date of birth, and grouped into two categories: adult (≥30–<65 years old) and elderly (≥65 years old). The sex variable was divided into two categories: male and female. The education variable was divided into two categories: low education (from no school up to junior high school) and higher education (from senior high school to college).The smoking habit variable was divided into three variables: non-smokers (never smoked), smokers (people who have smoked at least 100 cigarettes in their lifetime but in the environments where they live and work there are no smokers) and smokers who were exposed to secondhand smoke (smokers who are exposed to cigarette smoke itself and from other people). The variable of initial smoking age was divided into three groups: non-smokers, smokers who initially smoked at <15 years old and smokers who initially smoked at ≥15 years old. The variable of smoking duration was divided into three groups: non-smokers, smokers who smoked for <30 years and smokers who smoked for ≥30 years.Data collection was carried out using a questionnaire and secondary data, i.e., patient medical records. The questionnaire used in this study was a modification of that for basic health research and non-communicable disease research. The questionnaire was on participants’ smoking habits and statuses in relation to COPD. The questionnaire had been tested on 20 respondents to ensure its validity and reliability.The collected data were then analyzed descriptively and analytically with computer assistance. Data processing was carried out with bivariate analysis using the confidence of interval of 95% (α = 0.05).All participants were provided with written informed consent approved by the Ethics Commission of the Faculty of Public Health, Universitas Airlangga (certificate number: 536-KEPK).Table 1 presents the distribution of respondent characteristics and their effects on the incidence of COPD, showing that of a total of 112 respondents, the majority of the COPD patients were male (75%), elderly (53.57%) and had a low educational background (67.86%). Significant results showed that males had a 6.33 times greater risk than females (p = 0.00; OR = 6.333; 95%CI = 2.776–14.450) and that the elderly group had an 11.769 times greater risk than respondents in the adult group (p = 0.00; OR = 11,769; 95%CI = 4.086–33.903).Table 2 presents the distribution of respondents’ smoking behavior and its effect on the incidence of COPD, showing that of a total of 112 respondents, 55 were smokers and 57 were non-smokers, with 29 of the latter exposed to secondhand smoke.Significant results showed that smokers had a 5.1318 times greater risk of developing COPD compared with non-smokers (p = 0.00050; OR = 5.1318; 95%CI = 1.9004–13.8958). Yet, secondhand smokers had no influence on the incidence of COPD (p = 0.236; OR = 1.5278; 95%CI = 0.5028–4.6418). Significant results also showed that smokers who initially smoked at <15 years old had a 12 times greater risk of developing COPD compared with non-smokers (p = 0.026; OR = 12; 95%CI = 1.346–106.95), while smokers who initially smoked at ≥ 15 years old had a 3.647 times greater risk of developing COPD than non-smokers (p = 0.002; OR = 3.647; 95%CI = 3.298–23.787). A smoking period of fewer than 30 years did not influence the incidence of COPD (p = 0.881; OR = 1.091; 95%CI = 0.350–3.401), while a smoking period of more than 30 years was influential. Smokers who smoked for more than 30 years had an 8.857 times greater risk of developing COPD compared with non-smokers (p = 0.000; OR = 8.857; 95%CI = 3.298–23.787) (Table 2).The factors that influenced the incidence of COPD were being male (p = 0.00; OR = 6.333; 95%CI = 2.776–14.450), elderly (p = 0.00; OR = 11,769; 95%CI = 4.086–33.903), a smoker (p = 0.00050; OR = 5.1318; 95%CI = 1.9004–13.8958), developing an initial smoking habit at the age of <15 years old (p = 0.026; OR = 12; 95%CI = 1.346–106.95), developing an initial smoking habit at the age of ≥15 years old (OR = 3647; 95%CI = 1625–8183) and smoking for ≥30 years (OR = 8857; 95%CI = 3298–23,787).The current study showed that the OR value of the elderly group was higher than that of the adult group (p = 0.00; OR = 11,769; 95%CI = 4.086–33.903). This is in line with previous studies in Sousse, Tunisia in 2013 on elderly patients >70 years (p = 0.007; OR = 10,403; 95% CI = 2072–52,222) and in west and northern Sweden over 2008–2012 on elderly patients ≥60 years (OR = 8.40; 95%CI = 3.70–19.1), as well as based on the results of Indonesian health data from 2013 for elderly patients ≥60 years (OR = 4.5; 95%CI = 7.6–8.2) [21,22,23]. The three previous research studies have similar results, namely, that the elderly group has the highest risk of suffering from COPD versus other age groups.The current study showed that the OR value of the male group was higher than for the female group (p = 0.013; OR = 3.273; 95%CI = 1.291–8.2999). This is in line with previous studies in Sousse, Tunisia in 2013 on a male group (p = 0.010; OR = 0.198 95%CI = 0.062–0.635) and in Anhui, China on a male group (OR = 2.01; 95%CI = 1.22–3.33), along with Indonesian health data from 2013 for a male group (OR = 1.3; 95%CI = 4–4.3) [21,22,24,25]. The four previous research studies have similar results, namely, that males have a higher risk of suffering from COPD than females. The differences in the ORs in each study, both for age and sex, could be caused by several factors, such as differences in the methods used and numbers of samples taken, and bias resulting from other influential risk factors.The current study shows that the OR values for smokers are higher than those for other groups (p = 0.00050; OR = 5.1318; 95%CI = 1.9004–13.8958). This is in line with a previous study in 2019 in the Abeshge District, Southern Ethiopia on a smokers group (OR = 4.19; 95%CI = 2.59–6.78) [8]. Two previous studies in Sousse, Tunisia and Anhui, China, which described the smoking group in more detail, showed that current smokers have a higher risk than former smokers [21,25]. When compared with three previous studies, the OR value of a smoking habit in this study was higher than in those before. This may be because the methods used in this study were different and the number of participants was smaller. However, there was a similarity in that the smoker group had a greater risk of suffering from COPD than the non-smoker group. Smoking cigarettes over a long period disrupts ciliary movement, inhibits alveolar macrophage function and ultimately, leads to hypertrophy and hyperplasia of the mucus excreting gland [26]. Cigarette smoke contains many species of reactive oxygen (free radicals), which deplete the antioxidant mechanism, resulting in tissue damage and potentially leading to COPD [27].Developing a smoking habit at the age of <15 years old presents the biggest risk, more so than any other (OR = 12; 95%CI = 1346–106,950). There is a significantly higher risk of COPD when starting smoking at <15 years of age (3.647–12%) than ≥15 years old as smoking in childhood and adolescence can slow the growth and development of the lungs, thereby increasing the risk of incidence of COPD during adulthood. The lungs still grow and develop until the age of 20 years; therefore, smoking prior to the age of 20 years causes one’s lungs to develop sub-optimally both in terms of ability and function. These kinds of lungs fail to work properly, produce breath that tends to be short and will cause problems when used to exercise or perform physical activity. Although the health of people who quit smoking will increase dramatically compared with their health when they smoked, some cases of lung damage at an early age due to cigarettes are irreversible [28]. Therefore, it is necessary to implement policies regarding age restrictions on the purchase of cigarettes, especially for mobile cigarette sellers, as well as fostering cross-sector cooperation between health, education and other government sectors to spread the message about the dangers of smoking from an early age.Smokers who smoked for more than 30 years had an 8.857 times greater risk of developing COPD compared with non-smokers (p = 0.000; OR = 8.857; 95%CI = 3.298–23.787). Yet, the results for smoking <30 years were not in line with other studies, which showed that smoking has an effect, i.e., that someone who has had a smoking habit for >20 years is three to four times more likely to develop COPD compared to someone whose has had a smoking habit for ≤20 years [29]. The result of this study may have been influenced by exposure to cigarette smoke from other people, as of the 55 respondents who smoked, 26 were also exposed to cigarette smoke from other people, which could have disturbed the results of the study. Nonetheless, the results for smoking ≥30 years are in line with other literature stating that the longer the smoking habit, the greater the risk of developing COPD [30].A starting age for smoking of less than 15 years old has the highest risk (OR = 12; 95%CI = 1346–106,950) for COPD. The other factors that influenced the incidence of COPD were being male (p = 0.00; OR = 6.333; 95%CI = 2.776–14.450), elderly (p = 0.00; OR = 11,769; 95%CI = 4.086–33.903) and with a smoking duration ≥ 30 years (OR = 8857; 95%CI = 3298–23,787).W.S. carried out the data analysis, drafted the article and approved the publication; K.D.A. supported the data interpretation, revised important content and approved the publication; C.-Y.L. led the data collection, data interpretation and revising of important content, and approved the publication; S.M. designed the study, led the data interpretation, acted as corresponding author, led the revision process and approved the publication. All authors have read and agreed to the published version of the manuscript.This research was funded by Demographic Institute of the Faculty of Economics and Business, University of Indonesia in collaboration with JOHN HOPKINS Bloomberg School of Public Health, grant number 06/UN2.F6.D2.LDM/HKP/2017” and The APC was funded by Ministry of Research, Technology and Higher Education of Republic of Indonesia (Kemenristekdikti).This study was approved by Ethics Commission of Faculty of Public Health Universitas Airlangga (certificate number: 536-KEPK).Informed consent was obtained from all subjects involved in this study.The data presented in this study are available on request from the corresponding author. The data are not publicly available due to some of the data taken contains the respondent’s personal information which cannot be disseminated publicly.The authors would like to thank the patients as respondents for this research, the TCSC (Tobacco Control Support Center) organization, Indonesia Public Health Association (IPHA) East Java, Blitar District Health Office, Ministry of Research, Technology and Higher Education of Republic of Indonesia (Kemeristekdikti), Demographic Institute of the Faculty of Economics and Business, University of Indonesia, JOHN HOPKINS Bloomberg School of Public Health and the entire staff of Ngudi Waluyo Hospital in Wlingi, Blitar, East Java for their support of this research.The authors declare no conflict of interest.The following abbreviations are used in this manuscript:
2	+ COPDChronic Obstructive Pulmonary DiseaseUSDUnited States DollarBLUDRegional Public Service Agency (in English)OROdds RatioCIConfidence IntervalCharacteristics of Respondents.Smoking Behavior of Respondents.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06048.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: During the COVID-19 pandemic, the number of admissions to the emergency department (ED) due to a primary diagnosis of atrial fibrillation (AF) has decreased when compared to pre-pandemic times. The principal aim of the study was to assess the frequency of SARS-CoV-2 infections and sinus rhythm restoration among patients who arrived at the ED with AF. Secondary aims included determining whether patients arriving at the ED principally due to AF delayed their presentations and whether the frequency of successful cardioversion for AF was decreased during the pandemic period. Materials and Methods: A retrospective analysis of medical records of patients admitted to two hospital EDs due to AF during July–December 2019 (pre-pandemic period) versus July–December 2020 (pandemic period) was performed. Results: During the study periods, 601 ED visits by 497 patients were made due to the primary diagnosis of AF. The patients were aged 71.2+/−13.5 years and 51.3% were male. The duration of an AF episode before the ED admission was 10 h (4.5–30 h) during the pandemic period vs. 5 h (3–24 h) during the non-pandemic period (p = 0.001). A shorter duration of the AF episode before ED admission was associated with the successful restoration of the sinus rhythm. During the pandemic period, among patients with short-lasting AF who were not treated with Phenazolinum, the restoration of the sinus rhythm was more frequent in the Copernicus Memorial Hospital than in the University Hospital (p = 0.026). A positive SARS-CoV-2 test was found in 5 (1%) patients, while 2 other patients (0.5%) had a prior diagnosis of COVID-19 disease noted in their medical history. Conclusions: 1. The number of AF episodes treated in these two EDs was lower during the pandemic than non-pandemic period. 2. The patients with AF appeared at the ED later after AF onset in the pandemic period. 3. Successful cardioversion of atrial fibrillation was more frequent during the pre-pandemic period in one of the two hospitals. 4. A difference of approaches to the treatment of short-lasting AF episodes between EDs during the pandemic period may exist between these two EDs. 5. The patients with SARS-CoV-2 infection during the second wave of the COVID-19 pandemic constituted a small percentage of the patients admitted to EDs due to an AF episode.The global death toll due to the pandemic of the coronavirus disease 2019 (COVID-19) soon exceeds 3.3 million persons [1]. Most COVID-19 disease deaths are caused by severe acute respiratory infections by the coronavirus known as SARS-CoV-2 [2]. Nosocomial SARS-CoV-2 infection among hospitalized patients is common [3]. Patients’ awareness of this threat could make individuals reluctant to come to hospitals, not only for diagnostic tests but also for treatment [4]. Atrial fibrillation (AF) is the most common arrhythmia treated in an Emergency Department (ED), and prompt treatment can make successful cardioversion more likely. AF occurs in up to 10% of emergency department admissions and is the primary diagnosis for 1% of patients admitted to the ED [5]. The rate of hospital admissions for patients with a primary diagnosis of AF was 27.8% in the Blitz-AF study. This rate exceeded the overall hospital admission rate observed among all patients who presented to the ED for treatment [6]. AF may occur in up to 7.5% of COVID-19 patients [7]. Contrary to an expectation that the incidence of AF would increase during the pandemic, Schnaubelt et al. observed fewer visits by patients with paroxysmal and persistent atrial fibrillation during the pandemic period than during the corresponding months of the years before the pandemic [8]. Provisional recommendations for the treatment of patients with AF include pharmacological cardioversion, avoidance of electro-cardioversion, and the use of transoesophageal echocardiography [9]. During the current pandemic, less is known about the effectiveness of therapeutic management in patients admitted to the ED with a diagnosis of AF. The principal aim of the study was to assess the frequency of SARS-CoV-2 infection and sinus rhythm restoration among patients admitted to an ED due to AF. The secondary aim was to evaluate for possible delays of seeking treatment by patients with recent-onset AF and whether such delay, if present, was associated with a lower frequency of rapid and successful cardioversion.This is a retrospective analysis of medical records of the patients admitted to the ED of two hospitals, during two corresponding periods: The “non-pandemic period” (NPP), from July through December 2019, versus the “pandemic period” (PP), from July through December 2020. The ED at the University Hospital in Wrocław is one of four EDs in Wroclaw and is the Regional Trauma Center. Wroclaw is a principal city of the Lower Silesia voivodeship and has approximately 3,000,000 inhabitants. Its ED admission rate was approximately 3000 per month, with an average hospital inpatient admission rate of 25%. A total of 28% of its ED admissions arrived via emergency medical services (EMS). The 30-day mortality was 3.3% [10]. The ED has a “catchment area” of approximately 200,000 inhabitants. The ED at the Copernicus Memorial Hospital in Lodz is the Regional Trauma Center for the Lodz voivodeship. Lodz is the principal city of the Lodz voivodeship and has approximately 680,000 inhabitants. The Copernicus Memorial Hospital is the principal hospital for Lodz. Its ED admission rate was approximately 2000 per month, with an average hospital inpatient admission rate of 14%. The 30-day mortality was 3.2%. The COVID-19 infection was first detected in Poland on 4 March, 2019 [11]. As of December 31st, the total of confirmed cases of 2020 in Lower Silesia was 88,176, and the total of confirmed cases in the Lodz voivodeship was 84 760 [12,13]. The population on 30 June 2020 were 2,898,500 in the Lower Silesia voivodeship and 2,448,713 in the Lodz voivodeship [14,15].At the beginning of the pandemic, COVID-19 patients were preferentially routed to Infectious Disease Hospitals [16,17]. Beginning in September 2020, this routing was stopped, and most patients requiring inpatient care were admitted to the nearest hospital [18]. With time, each hospital was ordered to make inpatient beds available for COVID-19 patients. Of these, only those patients requiring tertiary care procedures were to be referred to the designated and more extensively resourced “COVID hospitals”.The ICD-10 code I48 was used to identify patients diagnosed with atrial fibrillation as a cause of ED admission. The study group included only patients with paroxysmal or persistent AF in whom the AF diagnosis was the primary one, as verified by the review of each subject’s medical record. The following data were collected for each patient: age, sex, estimated time since AF onset, treatments administered in the ED, the result of the swab test for SARS-CoV-2 infection, whether there was a history of prior COVID-19 disease, and whether there was a restoration of sinus rhythm in the ED. If the duration of the AF episode was recorded as “unknown” or if it was non-recorded, it was designed as “unknown”.Data are presented as mean ± standard deviation (SD) for normally-distributed data and as the median and interquartile range (IQR) for ordinal or non-normally-distributed data. The Student‘s t-test and the Mann–Whitney U test were used for statistical inference from these data. Statistical inference from nominal data expressed as rates or frequencies were compared with the Chi-square test. Classification and regression trees (CART) analyses were performed in both study periods (PP and NPP) to explore for possible associations in both EDs between sinus rhythm restoration at the ED and:patient age (years) dichotomized as ≥65 years or <65 years;duration since AF episode onset trichotomized as up to 6 h, 7–43 h, or longer than 43 h (or unknown, as appropriate);sex.patient age (years) dichotomized as ≥65 years or <65 years;duration since AF episode onset trichotomized as up to 6 h, 7–43 h, or longer than 43 h (or unknown, as appropriate);sex.Global Cross 102 Validation (CV) cost and its standard deviation were calculated. p less than 0.05 was regarded as significant.The study group consisted of 497 patients aged 71.2 ± 13.5 years (range 21–97). 255 (51.3%) were male. A total of 310 patients were treated at the University Hospital in Wrocław, and 187 patients were treated in the Copernicus Memorial Hospital in Lodz. The total number of ED visits due to AF was 601, with 389 visits to the University Hospital and 212 visits to the Copernicus Memorial Hospital. Between July and December 2019, there were 18,937 admissions to the University Hospital. Between July and December 2020, there were 16,435 admissions. At the Copernicus Memorial Hospital, the numbers of ED admissions were 11,713 and 9221, respectively. At the University Hospital, there were 232 AF episodes in 193 patients treated in 2019 and 157 AF episodes in 128 patients treated in 2020. These constituted 59.6% of the episodes of evaluations in the ED for AF in 2020 and 40.4% of all episodes of AF in 2019 (p < 0.001). At the Copernicus Memorial Hospital, there were 117 AF episodes among 102 patients treated in 2019 and 95 AF episodes in 85 patients treated in 2020. These constituted 55.2% and 44.8% of all assessed AF episodes, respectively, at that hospital (p = 0.032). These data, along with SARS-CoV-2 data, are also presented in Table 1, Table 2 and Table 3. The sum of the number of patients in 2019 and 2020 does not equal the total number of patients across both years because some patients were admitted to an ED both in 2019 and 2020. The sex and age distribution in the pandemic period and the non-pandemic period are presented in Table 1. These did not differ significantly between the two EDs. However, in the Copernicus Memorial Hospital, the patients were older. Overall, the mean (and IQR) of the time since the AF onset until ED admission was 10h (4–48 h) during the pandemic period and 5.5 h (3–23 h) in the non-pandemic period, which was significantly shorter (p = 0.019). This shorter time during the NPP between AF onset and ED admission was also individually observed at both EDs. Successful restoration of the sinus rhythm in the ED in the University Hospital was greater during the non-pandemic period. However, at the Copernicus Memorial Hospital ED, there was no significant difference between time periods.A total of 5 patients had positive swab tests for SARS-CoV-2 infection, while 2 patients had a prior history of SARS-CoV-2 infection. The patients with new or prior SARS-CoV-2 infection constituted less than 2% of all studied patients.The restoration of the sinus rhythm occurred more frequently at the University Hospital during the non-pandemic period, whereas, at the Copernicus Memorial Hospital, there was no significant difference between the periods.The CART analysis for sinus rhythm restoration is presented in Figure 1. Global CV cost = 0.26; s.d. CV cost = 0.019.The CART analysis revealed that the most important factor predicting a restoration of the sinus rhythm in the ED is the time interval between AF onset and admission to the ED. A shorter time is more favorable. The treatment of patients during the pandemic era was associated with a lower rate of restoration of the sinus rhythm at the University Hospital, but not at the Copernicus Memorial Hospital. We were not able to address the principal aim of the study. The presence of a current or recent SARS-CoV-2 infection was very low in the cohort studied, making any statistical inferences about the restoration of the sinus rhythm in patients with versus without current or recent SARS-CoV-2 infection impossible. Further, the percentage of the patients who attended the ED with a primary diagnosis of AF was approximately 1% during both periods. This finding is consistent with the reported incidence of primary AF diagnosis in a large observational study [6], and it supports the accuracy of the obtained results but illustrates the difficulty of a study of this matter as a single-site study. Regarding the secondary aims, there was clear statistical evidence that patients with AF delayed their presentations to the ED during the pandemic period when compared to the non-pandemic period. This may have been related to reluctance on the part of the AF patients to attend the ED because of fear of nosocomial COVID-19 disease. This delay in presentation for AF during the pandemic period was associated with a decreased rate of success of chemical or electrical cardioversion back to sinus rhythm at one but not both of the hospitals in which ED patients were studied. This finding might have many causes. Patients with a greater propensity for restoration of spontaneous sinus rhythm may have had a spontaneous return to sinus rhythm at home, and, thus, they did not choose to visit the ED. The tactic of delaying the reporting to the ED in the case of an attack of hemodynamically stable atrial fibrillation represents a new therapeutic option [19,20]. This option was tested in the Rate Control versus 193 Electrical Cardioversion Trial 7-Acute Cardioversion versus Wait and See (RACE 7 194 ACWAS) [21]. This underused, usual option could allow a rapid and spontaneous resolution of many AF episodes. This option may have been unconsciously chosen by AF patients due to fear of nosocomial COVID-19 disease. If this occurred, the consequence could have been that attendance to the ED was done by patients with AF who are more likely to be resistant to treatment, with a lower rate of the restoration of the sinus rhythm rate in the ED. Attendance at the ED was less during the pandemic period than during the non-pandemic period. This is concordant with reports of the other authors and with data derived from the ED of the University Hospital from the early phase of the pandemic [4]. The total number of patients admitted to the ED at both the University Hospital and the Copernicus Memorial Hospital was lower in 2020 than in 2019, which is in line with the findings of other authors [8,22].The management of patients with AF episodes differed between the two EDs. This finding is consistent with other reports indicating significant variation in the emergency management of acute atrial fibrillation [23]. 1. The number of AF episodes treated in the ED was lower during the pandemic than during the non-pandemic period.2. During the pandemic period, the patients with AF arrived later at the ED, and they were less likely to be cardioverted back to sinus rhythm.3. Differences in approaches to the treatment of recent-onset episodes of AF between EDs during the pandemic period appeared to exist. 4. Patients with a current or prior SARS-CoV-2 infection during the second wave of the COVID-19 pandemic constituted a small percentage of the patients admitted to the ED because of an AF episode.Conceptualization, Ł.B. and D.Z.; methodology, D.Z.; software, D.J.; validation, R.K. and D.T.; formal analysis, D.Z.; investigation, Ł.B. and W.T.; resources, M.M., D.Z., J.W., and Ł.B.; data curation, Ł.B., K.B., W.T., D.T., and M.T.; writing—original draft preparation, Ł.B., W.T., D.Z., D.T., and M.T.; writing—review and editing, R.K. and D.T.; visualization, D.J., M.D., and P.R.; supervision, D.Z., D.T., and M.M.; project administration, K.B. and M.T.; funding acquisition, D.Z., M.T., and M.M. All authors have read and agreed to the published version of the manuscript.This research was financially supported by the Ministry of Health subvention according to number STM A280.20.037 from the IT Simple system of the Wrocław Medical University and it was conducted within the EU-financed InterDoktorMen project (POWR.03.02.00-00-I027/16; Medical University of Lodz).The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Wroclaw Medical University (protocol code KB-426/2021 17 May 2021).Not applicable—Retrospective study.The data presented in this study are available on request from the corresponding author.This research is conducted within the EU-financed InterDoktorMen project (POWR.03.02.00-00-I027/16; Medical University of Lodz). This research was financially supported by the Ministry of Health subvention according to number STM A280.20.037 from the IT Simple system of the Wrocław Medical University.The authors report no conflicts of interest to disclose.The CART analysis. The dependent variable is the restoration of the sinus rhythm. The independent variables include sex (male or female); age <65 years or ≥65 years; AF duration since onset of 0–43 h vs. >43 h or unknown; AF duration 0–6 h vs. >6 h or unknown; University Hospital vs. Copernicus Memorial Hospital; pandemic period vs. non-pandemic period; electrical cardioversion versus chemical cardioversion using Phenazolinum, Amiodarone, Propafenone, or beta blockers.Patients with AF as the primary diagnosis in the non-pandemic period (NPP) and the pandemic period. (PP). Further, the number of COVID-19 positive and convalescent COVID-19 patients are presented.* p < 0.05 vs. University Hospital in the corresponding period.The distribution of the AF episodes, time since AF onset to the ED admission, the duration of ED stays, and sinus rhythm restoration in the studied hospitals during non-pandemic and pandemic periods.N*—the number of patients with available data. # p < 0.05 vs. corresponding period in the University Hospital. $ p = 0.007 vs. non-pandemic period. @ p < 0.001 vs. non-pandemic period. ^ p < 0.05 vs. non-pandemic period.The management of AF episodes in the studied EDs during non-pandemic and pandemic periods.* p < 0.001 vs. University Hospital in period.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06049.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Our aim was to evaluate clinical, biochemical and microbiological markers related to dental caries in adults. A sample that consisted of 75 volunteers was utilized. The presence of caries and the presence of plaque and gingival indices were determined. Unstimulated salivary flow, pH, lactate, Streptococcus mutans and Streptococcus dentisani were measured in the participants’ plaque and saliva samples before and after rinsing with a sugar solution. Lactate in plaque was found to be significantly related to age, gender, tooth-brushing frequency, the presence of cavitated caries lesions and plaque and gingival indices (p < 0.05). The levels of S. dentisani in plaque increased significantly with tooth-brushing frequency (p = 0.03). Normalized plaque S. dentisani values and the percentage of S. dentisani were slightly higher in patients with basal lactic acid levels ≤ 50 mg/L. After rinsing with a sugary solution, the percentage of S. mutans levels in plaque were higher in patients with lactic acid levels > 350 mg/L (p = 0.03). Tooth-brushing frequency was the factor which was most associated with oral health. Women reflected better clinical and biochemical parameters than men. Low pH and high lactic acid levels tended to be associated with high caries rates. No association was found between bacteria levels and caries indices.Dental caries is defined as a “biofilm-mediated, diet modulated, multifactorial, non-communicable, dynamic disease resulting in net mineral loss of dental hard tissues”. It is determined by biological, behavioral, psychosocial and environmental factors. As a consequence of this process, a caries lesion develops [1]. The transition from health to disease occurs when there is a disturbance that modifies the conditions of the oral environment and favors the development of a more acid-producing and acid-tolerant microbial community [2]. This potential for tolerance and production of acids cannot be attributed to a single group of microorganisms, but to a bacterial consortium that interacts in a complex manner and that, under certain conditions, would increase in proportion or activity, to the detriment of other bacteria whose metabolic output would be less acidogenic [3]. When the environmental conditions change, the oral microbiota can also change and reverse the effect.In the biofilm of a caries lesion, Streptococcus mutans is not the most numerous species; there are many other microorganisms with aciduric and acidogenic potential that are also present [4]. However, S. mutans has been extensively studied as it is an important producer of extracellular matrix and can rapidly modulate the formation of a cariogenic biofilm when aided by the presence of fermentable carbohydrates in the diet. Sucrose is considered the most cariogenic sugar because, in addition to being fermented by oral bacteria, it is an excellent substrate for the synthesis of extracellular (EPS) and intracellular (IPS) polysaccharides [5,6]. In contrast, the presence of microorganisms with arginolytic and ureolytic activity in the biofilm can buffer the pH and present a high cario-preventive potential and this is due to the production of ammonia as a final metabolite [7]. The presence of sucrose, a low buffering capacity of saliva, and a low pH have been shown to be important factors that hinder the production of alkali from arginine [8]. Thus, the final cariogenic potential of an oral biofilm is not only determined by the presence of sugar-fermenting and acidogenic organisms but also by the levels of protective and pH-buffering bacteria in the oral cavity. However, the simultaneous quantification of bacterial biomarkers of both types of organisms is rarely performed.The microbial diversity of a caries lesion is considerably lower than that of dental biofilms [9]. Within the complex microbiota that has been identified in genomic studies, it has been observed that a large part of the microorganisms involved are commensal. Many have not yet been cultivated and have not even been assigned to scientific nomenclature yet [10,11]. It has been possible to identify a series of microbial species in the metagenomic studies of dental plaque, which are more compatible with dental health than others [12]. In 2014 a new species of Streptococci from the mitis group was described and named as Streptococcus dentisani [13]. When this bacterium was isolated in pure culture and grown in the presence of cariogenic organisms, such as S. mutans or Streptococcus sobrinus, it was observed that the bacterium inhibited or killed pathogens by means of bacteriocins [14]. It has also been demonstrated that S. dentisani has the capacity to metabolize arginine into ammonia, which neutralizes dental plaque pH. Therefore, S. dentisani would provide a double anticariogenic mechanism; it would inhibit the growth of acidogenic bacteria, stimulate the formation of ammonium which results in a more favorable pH for dental health and it could act as a biomarker for beneficial oral bacteria [14].Dental plaque constitutes the habitat in which microbial metabolic activity takes place and where both pathogenic and protective processes occur, which affects the development of the lesion. Urea, nitrate and arginine are the three main sources of alkali generation in plaque and saliva. Ammonia produced as a result of their metabolism can be an important endogenous inhibitory factor of the acidogenic microbiota and caries development by neutralizing acids and stabilizing the oral microbiota [15,16,17,18,19]. Thus, the levels of ammonia, pH and organic acids, such as lactate, could act as biomarkers to indicate the acidogenicity of oral communities and may be related to the risk of developing caries lesions. However, it is unclear whether the levels of these compounds have a better diagnostic and predictive value when measured in dental plaque or in saliva.As early as 1940, Stephan reported that between 2 min and 15 min after rinsing with a sugary solution, the pH of dental plaque drops, lactic acid being mainly responsible for this drop, and then returns to its basal level around the 40 min mark [20]. This is possible thanks to the implementation of buffer mechanisms among which bicarbonate, urea and arginine stand out [18,21,22,23]. Thus, not only is the basal pH important as a measure of acidogenicity but also the level achieved after fermentation takes place. Accordingly, in the current manuscript pH and lactate levels before and after the pH drop caused by sugar exposure were measured.The responsiveness of saliva and dental plaque when subjected to acid stress will be an indicator of the ability to compensate and balance acids in the oral environment either through the salivary compensatory mechanisms or the already described mechanisms of action of microorganisms such as S. dentisani [14]. To the authors’ knowledge, there are currently no studies that evaluate the possible association between the response to acid stress in saliva and plaque with the levels of S. mutans and S. dentisani, simultaneously, and its association with previous clinical records of caries and biochemical salivary and plaque parameters. Therefore, the present study aims to simultaneously analyze clinical, biochemical and microbiological factors potentially related to caries experience in a sample of adults. The following elements were assessed: plaque and gingival indices; salivary flow; salivary and plaque pH; and salivary and plaque lactate before and after sugar exposure. Additionally, the presence of S. mutans and the presence of S. dentisani in both plaque and saliva were quantified by quantitative PCR. Finally, the associations of all these parameters with dental caries experience, gender and tooth-brushing frequency were evaluated in order to better understand their relationships relative to one another and to identify potential biomarkers and risk predictors of the disease.This study is part of two CECT 7746 clinical trials carried out in collaboration between the FISABIO Foundation and the Lluís Alcanyis Foundation of the Universitat de València. It was approved by the DGSP-CSISP Ethics Committee with project codes ABB-Sdent-Colonization and ABB-dentisani 2015. The study began in 2016 and was completed in 2018. Compliance of the protocol and surveillance of the study was performed by the external CRO Effice (Madrid, Spain).In order to carry out the present study, 184 volunteers were recruited. Informed consent was obtained from all subjects involved in the study. Inclusion criteria for the selection of participants were: age between 18 and 65 years; basal salivary pH (after brushing with water) ≤ 7; at least 21 teeth present in the oral cavity; previous caries experience and/or presence of active caries lesions; and the absence of other oral diseases. Patients with chronic diseases; use of medication or previous procedures, such as head and neck radiotherapy or pathologies that reduce salivary flow; basal pH at the time of recruitment > 7; and/or absence of previous caries experience; and/or active caries lesions were excluded. Individuals taking antibiotics during the previous 3 months or regular use of oral antiseptic mouthwashes during the previous week were also excluded. After applying the inclusion and exclusion criteria, a total of 75 participants were selected.After being included in the study, each participant received an appointment to which he/she had to attend without having brushed his/her teeth since the night before. In this appointment, which was carried out between 4:00 p.m. and 7:00 p.m., frequency of daily tooth-brushing was recoded. A sample of unstimulated basal saliva obtained by drooling was taken for five minutes to determine salivary flow by considering values from 0.25 to 0.3 mL/min as the normal secretion rates (sample 0). From these salivary samples, the buffer pH was determined.Subsequently, an oral examination was performed and the plaque and gingival index of Silness and Löe and Löe and Silness [24], respectively, were determined. Additionally, the supragingival biofilm of all teeth surfaces was collected in two 1.5 mL Eppendorf tubes using a sterile dental excavator. Biofilm from quadrants 1 and 3 were deposited in 100 µL phosphate buffer solution (PBS) for its conservation until the subsequent extraction of bacterial DNA. Biofilm from quadrants 2 and 4 were deposited in 100 µL sterile H2O (pH = 7) for the subsequent determination of pH and lactic acid content after sugar exposure.After the plaque sample collection, patients were instructed to brush their teeth 2 min with water using a manual toothbrush. The presence of caries was then assessed using the ICDAS II criteria by the same experienced explorer in all participants [25].Next, an additional sample of unstimulated saliva obtained by drooling (sample 1) was collected, in which the pH was determined. The participants then rinsed with a 10% sugar solution for 1 min and 10 min after rinsing a final salivary sample was collected (sample 2).Salivary pH was determined from samples 0, 1 and 2 by means of a reflectometer (Reflectoquant; Merck, Darmstadt, Germany) which was calibrated with the corresponding pH strips (reference: 116996).For the determination of biofilm lactate content, 30 µL of the initial sample (100 µL H2O + biofilm) was deposited onto the test strips. After 7 min, they were introduced into the reflectometer for their measurement (t0) after calibration with the lactic acid strips (reference: 116127). For the measurement of biofilm pH, 30 µL of the same sample was deposited onto the corresponding test strips and, after 10 s, they were introduced into the reflectometer. Then, 40 µL of 20% sucrose solution was added to the initial sample, thus obtaining a final sucrose concentration of 10%. Immediately afterwards, the samples were introduced in a laboratory oven at 37 °C for a 10 min incubation period. After the incubation period, the pH and lactic acid (t10) were measured again.The microorganism count was performed in plaque and saliva samples: 250 µL of basal saliva (sample 0) and the plaque samples from quadrants 1/3 resuspended in 100 µL of PBS buffer were used. First, the DNA was extracted in an automated process, using the MagnaPure LC JE379 equipment and the MagnaPure LC DNA Isolation Kit that are both from Roche (Basel, Switzerland), after an enzymatic lysis [26]. A fluorometric method (Quant-iT PicoGreen dsDNA Assay, Invitrogen) was used to quantify the extracted DNA. The reactions for the quantification of S. mutans, S. dentisani and total bacteria were carried out by means of qPCR (quantitative Polymerase Chain Reaction) with the LightCycler 480 equipment and the LightCycler 480 SYBR Green I Master Mix kit (Roche, Basel, Switzerland).The specific primers used for the quantification of S. dentisani (CkSdF and CkSdR) [14] amplify a 77 base pair region of the carbamate kinase gene. For the quantification of S. mutans, primers which were already reported in the literature were used which amplify a 415 base pair fragment of the glycosyl transferase gene [27]. In the case of total bacteria, the target gene was the 16S rDNA ribosomal gene that is highly conserved in the Bacteria Domain. The primers used (515F-789R) amplify a region of 274 base pairs [28,29].All amplification reactions were carried out in a final volume of 20 µL containing 1 µL of template DNA (5–22 ng/µL), 10 µL of LightCycler 480 SYBR Green I Master Mix, 0.4 µL of each primer and 7.2 µL of nuclease-free water. The thermocycling program used is described as follows: an initial denaturation step at 95 °C for 5 min, 40 cycles of 10 s at 95 °C, 20 s at 65 °C (for total bacteria it was reduced to 58 °C) and 25 s at 72 °C. All reactions were performed in duplicates as well as their corresponding positive and negative controls.The ICDAS II values were recalculated as Decayed, Missing and Filled Teeth (DMFT) values. In component D, cavitated caries (caries ICDAS codes 3–6 and restauration ICDAS codes 7 and 8) on the one hand and cavitated and not cavitated caries on the other (caries ICDAS codes 1–6 and restauration ICDAS codes 7 and 8) were recoded. In components M and F, ICDAS code 97 and restoration ICDAS codes 3–6, respectively, were included [30].Using the Kolmogorov–Smirnov test, it was determined if the quantitative variables did not follow a normal distribution. Non-parametric tests were used if this was the case. The Kruskal–Wallis test was used to compare tooth-brushing frequency and quantitative variables. The Mann–Whitney U test was used to compare gender with the quantitative variables. Chi-squared test was used for comparison between gender and tooth-brushing frequency. Box plot graphs representing median and Inter quartile range (IQR) were used for presenting data. Finally, correlation analyses were performed between the different study parameters using the Spearman correlation coefficient. In all cases, a significance level of 95% was used.The sample consisted of 24 men (32%) and 51 women (68%) with a mean age of 34.72 ± 10.84 years.Analyses by gender revealed that caries indices and their components were higher for men than for women, except for the filled teeth that were slightly higher for women. After analyzing their components, missing teeth values for men were significantly higher than for women (p = 0.01). Plaque index was significantly higher in men than in women (p = 0.01). Men brushed their teeth significantly less frequently than women (p = 0.02, Chi-squared test) (Figure 1A). The pH of saliva or of plaque at different time points were similar in men and in women (Figure 1B). Plaque lactate levels at both t0 and t10 were significantly higher for men (Mann–Whitney U test) (p < 0.01) (Figure 1C). The levels of S. mutans and S. dentisani were obtained in saliva normalizing by volume (CFUs/mL) and in plaque after normalizing by the total bacterial DNA present (CFUs/ng). Values for S. mutans were significantly higher in men than in women (p < 0.01). Figure 1D shows median and IQR for the levels of S. mutans and S. dentisani transformed in log10 values.Participants who brushed their teeth two or three times a day presented lower values of DMFT (with or without non cavitated lesions) than those that brushed only once a day. Components D and M were also lower in these patients (p < 0.05). Plaque and gingival indices were significantly lower in participants who brushed their teeth three times a day compared to those that brushed once per day (p < 0.05) (Figure 2A). Lactate values at t10 decreased significantly when the tooth-brushing frequency increased (p < 0.05) (Figure 2C). The pH values and normalized levels of S. mutans and S. dentisani in saliva (CFUs/mL) or in plaque (CFUs/ng) did not vary significantly with tooth-brushing frequency (Figure 2B,D).Table 1 shows the correlation analysis of the different clinical and biochemical variables analyzed (Spearman’s correlation coefficient). A positive significant correlation was found between plaque index and gingival index and plaque pH and lactate levels at t0 and t10 (p < 0.05). The basal saliva pH (sample 0) was significantly positive correlated with saliva pH after brushing (sample 1) and after sugar rinse (sample 2), respectively (p < 0.05). Plaque pH and lactate levels at t0 and t10 were also significantly positive correlated (p < 0.05).The levels of S. mutans and S. dentisani in saliva (CFUs/mL) and plaque (CFUs/ng) were determined after normalizing the quantification according to the total bacterial DNA present. The percentage of both microorganisms was also calculated according to the total number of bacteria present in the plaque material. The levels of S. dentisani in plaque increased significantly when tooth-brushing frequency increased (7.71 × 102/tooth-brushing once a day, 1.42 × 103/tooth-brushing twice or three times a day) (p = 0.03) (the same trends were observed in the percentages of S. dentisani in plaque) (0.29 %/tooth-brushing once a day, 0.42 %/tooth-brushing two or three times a day) (p = 0.04) (Kruskal–Wallis test). Values of S. dentisani (1.38 × 103/1.07 × 103) and percentages of S. dentisani in plaque (0.42%/0.32%) were higher in patients with DMFT levels ≤ 8. Levels of S. mutans in saliva were higher in patients with baseline salivary pH below 6.4 (5.29 × 103/1.57 × 104). Plaque normalized S. dentisani values and the percentages of S. dentisani were slightly higher in patients with lactic acid levels ≤ 50 mg/L. After the sugar rinse, percentages of S. mutans levels in plaque were higher in patients with lactic acid levels > 350 mg/L (p = 0.03) (Mann–Whitney U test).Table 2 shows the correlations between microbiological parameters and clinical and biochemical parameters (Spearman correlation analysis). Age was found to be significantly correlated with S. dentisani levels in saliva and plaque. Plaque index was observed to be significantly related to S. mutans and S. dentisani levels in saliva and the percentages of S. dentisani in plaque. The levels of S. mutans and S. dentisani in saliva were significantly associated, as were S. mutans levels in plaque and their percentages. Basal pH in saliva (sample 0) was significantly and inversely associated with S. dentisani levels. The pH in plaque at t0 was found to be significantly and inversely associated with S. dentisani levels in saliva and plaque and with S. mutans in plaque. pH at t10 was found to be significantly and inversely associated with microbiological levels in plaque and saliva. Lactate values at t0 and t10 were significantly correlated with microbiological levels in saliva and S. mutans levels in plaque.The mean age of the study sample corresponds to the group of young adults proposed by the WHO, which includes the population between 35 and 44 years old. If that population group is taken as a reference in the data obtained in the last national survey carried out in 2020 then the average DMFT index was 7.40 ± 4.86, which is very similar to our sample where the mean DMFT values were 7.45 ± 4.83. The mean values of cavitated and filled lesions were similar to those from the national study for the young adult population, while the mean values of missing teeth were lower: 0.73 vs. 1.92 [31]. In the study of Eustaquio M.V. et al. in 2010 on the oral health of the population of the Valencian Region (Spain), similar DMFT values were observed in young adults [32] to those of the present study (7.45 vs. 7.64).With regards to the association between gender and tooth-brushing frequency, it was observed that men brushed less frequently than women. It was also observed that men had higher DMFT values and its components, with the exception of the restoration index which was higher among women. However, a significant difference was only found in component M (missing to caries). These data are consistent with those found in the 2020 national epidemiological study [31]. In different studies, the direct relationship between tooth-brushing habits and the previous experience of caries has been verified [33,34,35,36]. Regarding the plaque and gingival indices, these were also higher in men although the difference was only significant for the plaque index. These data are consistent with those found in the literature [31,36,37]. However, a limitation of the present study could be not having taken into account other relevant factors in caries etiology such as diet, social and behavioral factors as well as the difference in sample size between men and women. Thus, although the microbiological and biochemical features measured in the current study correspond only to 75 individuals, this agreement in clinical data with the national survey suggests that we are working on a representative set of individuals.Additionally, significantly higher levels of S. mutans were found in plaque and in saliva in men than compared to women. These data are correlated with the results obtained regarding the plaque index. Good oral hygiene habits control the oral microbiota at levels compatible with health [38] and our data suggest better oral hygiene parameters in women, which is translated in better clinical and microbiological features. Likewise, lactate levels were higher in men both before and after the sugar rinse. This can be explained because higher concentrations of lactate are related to a greater amount of plaque with acidogenic properties. It is well established that acidogenic and aciduric biofilms are associated with an increased risk of disease [39]. In a recent study that microbiologically evaluated 268 young patients in the Netherlands, a strong gender pattern was also observed in microbial composition and metabolism and more cariogenic microbiota was observed in men resulting in significant pH differences [40].With the data obtained in the present study, it can be highlighted that factors associated with gender influences the incidence of dental caries; it is observed that men have poorer oral health and that women take greater care of their oral hygiene both individually (tooth-brushing) and professionally (restorations). Nevertheless, other factors such as dietary habits or sociocultural and behavioral factors should also be considered when estimating the risk of developing caries.After analyzing the tooth-brushing frequency with the quantitative variables, it was observed that the participants who brushed their teeth once a day had worse oral health than those who did so 2–3 times a day. In addition, participants who brushed their teeth three times a day had a significantly lower gingival index and lower plaque lactate content. The beneficial effect of tooth-brushing on the risk of caries has been demonstrated in numerous studies and is the most effective method for removing plaque [41] and the best fluoride delivery system [42]. Consequently, it can be concluded that brushing three times a day reduces the risk of tooth decay and that brushing frequency is predictive of caries risk. In agreement with this, it is interesting that the levels of the cariogenic bacterium S. mutans corelated negatively with tooth-brushing frequency but positively with the health-associated species S. dentisani.When correlating the clinical and laboratory variables, it was observed that DMFT increases with age, that tooth absences are significantly related to plaque and gingival indices and that plaque and gingival indices have a significant association with pH levels and lactate content before and after the sugar rinse. In 1987, Firestone et al. verified how the maturity of dental plaque considerably increases the drop in pH after a meal ingestion [43]. This fact supports the correlation observed between plaque index, pH and lactate content of the plaque itself. Experimental studies have also demonstrated the association between certain microbial components of mature plaque and the ability to modify pH, lactic acid production and the presence of enzymes with high pathogenic potential [3].Among the factors associated with microbiology, lactate is the most reliable marker of caries. Given that this marker was measured in plaque and that S. mutans levels were better correlated with caries parameters when this bacterium was quantified in plaque, the present work supports the use of dental plaque to measure microbiological parameters with greater reliability than saliva, which showed erratic and inconsistent trends in the different factors measured. Despite this fact, it is interesting that the levels of both S. mutans and S. dentisani in plaque were significantly correlated with the values of these bacteria in saliva. This is possibly due to the fact that both species are mainly inhabitants of hard tissues and, therefore, their higher levels in plaque imply a greater release in saliva. Other streptococci which are found at high levels in soft tissues, such S salivarius, may not correlate between these two niches.Microbiological data support that, despite the fact that S. mutans is a minor inhabitant of dental plaque representing less than 1% of the total microbiota, its presence is correlated with lactate production and is therefore an important bacterium in acid generation [12]. However, the correlations of this bacterium with caries levels were neither positive nor significant, neither with pH values nor with caries levels or caries indices. Future studies should include the levels of other acidogenic bacteria in order to establish the microbiological risk of caries more accurately. It was observed, for example, that the measurement of S. mutans and Lactobacillus in a combined manner correlates more with caries parameters than each one separately [44]. Other authors are trying to incorporate other bacterial species in the map of the prediction of the risk of suffering from the disease [45].The levels of S. dentisani were higher in individuals without restorations, which could indicate a negative association with the history of caries, while the opposite pattern was detected in S. mutans. However, these data were inconsistent with the levels of active caries or cavitated caries and thus there is no clear protective relationship between the presence of S. dentisani and caries in individuals who already have the disease. Given that there are higher levels of S. dentisani in individuals without caries [13], it is possible that the beneficial effect of S. dentisani requires minimal levels of the bacteria to be effective although it cannot be ruled out that said protective effect depends on the strain of S. dentisani present in each individual.Unlike what occurred with S. mutans, S. dentisani levels did not correlate with lactate production in plaque, although the percentage of S. dentisani in plaque was higher in patients with lower lactate levels both before and after the sugar rinse. However, the levels of both bacteria before and after the sugar rinse were negatively related to plaque pH. This suggests that pH is related to the total amount of plaque or bacterial load, regardless of whether or not the protective effect of S. dentisani was present.In in vitro experiments, it has been shown that the buffer effect of S. dentisani through the arginine pathway requires several hours [14] and therefore would not be appreciated in the data of the present work. A second explanation for this lack of correlation between pH and levels of S. dentisani would be given by the absence of individuals without caries [DMFT = 0] in the study since it is these individuals who have shown higher levels of S. dentisani and of arginolytic activity in general [18,46]. Therefore, the data would support the idea that once the disease is present, the levels of S. dentisani in these individuals are not sufficient to prevent either S. mutans levels or lactate production or to maintain a more alkaline basal pH. In fact, it is surprising that despite the substantial inhibition produced by this bacterium against cariogenic organisms under laboratory conditions, there is no negative correlation between the presence of S. dentisani and the levels of S. mutans. Future work should therefore establish this correlation in individuals who have never suffered caries.In a recently published article, it has been shown that the administration of a bioadhesive gel, applied with individualized trays for a week, containing the probiotic S dentisani causes an increase in salivary pH, a reduction in the production of lactic acid up to 30 days after its application, a reduction in the colonization of S mutans and an increase in the presence of S dentisani up to 14 days after its application [47]. In a second clinical trial where the probiotic was applied for a month, improvements in plaque and gingival indices, as well as in salivary ammonia and pathogen levels were observed [48] and this suggests, again, that there is minimum level of this organism that is necessary to observe an improvement in clinical and microbiological features.Meanwhile, in patients who already have the disease, our data strongly confirm that tooth-brushing is the most effective method to improve caries risk and the development of cavities [41,42]. In this sense, the significant correlation that exists between tooth-brushing and S. dentisani levels is interesting and it indicates that this bacterium is favored by correct oral hygiene. This suggests that it is an early colonizer of the plaque and/or that it is favored by the lower frequency of a mature and acidogenic plaque. Therefore, a synergistic effect could occur between tooth-brushing and the presence of S. dentisani as it has been observed after tongue-brushing for other beneficial organisms such as Rothia [48]. Given that S. dentisani uses arginine for its pH buffering action [45], it would be interesting to test whether the use of arginine toothpastes and the administration of S. dentisani as a probiotic can have a synergistic effect in individuals with cavities.Tooth-brushing frequency was the most decisive factor associated with oral health. Low pH and high lactic acid levels tended to be associated with higher caries rates. However, no association was found between the levels of specific bacteria and dental caries and this suggests that a single bacterial species cannot be used as a valid biomarker of the disease.Conceptualization, A.M. and C.L.; methodology, M.D.F. and A.L.L.; formal analysis, A.M. and C.L.; investigation, S.P., M.D.F. and A.L.L.; data curation, A.M. and C.L.; writing—original draft preparation, M.M.; writing—review and editing, J.L.S.; supervision, A.M. and C.L.; funding acquisition, A.M. All authors have read and agreed to the published version of the manuscript. This research received funding by grant RTC-2015-4292-1 from the RETOS-Colaboración Call of the Spanish Ministry of Health and Competitiveness and by grant RTI2018-102032-B-I00 from the Spanish Ministry of Science, Innovation and Universities.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of FISABIO foundation (CEIC DGSP-CSISP).Informed consent was obtained from all subjects involved in the study. The data presented in this study are available upon request from the corresponding author.The authors declare no conflict of interest.Clinical, biochemical and microbiological variables by gender. The graphs show medians and IQR values for the different variables. (A) DMFT = ICDAS codes 3–6 + restoration ICDAS codes 7 and 8 + missing teeth due to caries + filled teeth due to caries. DMFT with non-cavitated lesions = DMFT + ICDAS codes 1–6 + restoration ICDAS codes 7 and 8. (B) pH_saliva (sample 0) = basal salivary pH. pH_saliva (sample 1) = salivary pH after tooth-brushing. pH_saliva (sample 2) = salivary pH after sugar rinse. pH plaque t0 = plaque pH before sugar rinse. pH plaque t10 = plaque pH after sugar rinse. (C) Lactate plaque t0 = plaque lactate levels before sugar rinse (mg/L). Lactate plaque t10 = plaque lactate levels after sugar rinse (mg/L). (D) Log SM saliva = Log10 transformation of normalized CFUs/mL of S. mutans in saliva, Log SD saliva = Log10 transformation of normalized CFUs/mL of S. dentisani in saliva. Log QSM plaque = Log10 transformation of normalized CFUs/ng of S. mutans in plaque, Log QSD plaque = Log10 transformation of normalized CFUs/ng of S. dentisani in plaque. The expected loss in statistical power due to the difference in male:female ratio is <0.05.Clinical, biochemical and microbiological variables by daily tooth-brushing. Graphs show medians and IQR values for the different variables. (A) DMFT = ICDAS codes 3–6 + restoration ICDAS codes 7 and 8 + missing teeth due to caries + filled teeth due to caries. DMFT with non-cavitated lesions = DMFT + ICDAS codes 1–6 + restoration ICDAS codes 7 and 8. (B) pH_saliva (sample 0) = basal salivary pH. pH_saliva (sample 1) = salivary pH after tooth-brushing. pH_saliva (sample 2) = salivary pH after sugar rinse. pH plaque t0 = plaque pH before sugar rinse. pH plaque t10 = plaque pH after sugar rinse. (C) Lactate plaque t0 = plaque lactate levels before sugar rinse (mg/L). Lactate plaque t10 = plaque lactate levels after sugar rinse (mg/L). (D) Log SM saliva = Log10 transformation of normalized CFUs/mL of S. mutans in saliva, Log SD saliva= Log10 transformation of normalized CFUs/mL of S. dentisani in saliva. Log QSM plaque = Log10 transformation of normalized CFUs/ng of S. mutans in plaque, Log QSD plaque = Log10 transformation of normalized CFUs/ng of S. dentisani in plaque.Analysis of the correlations (Correlation Coefficients (CC) and p-values) between the quantitative variables assessed. Significant correlations are bolded.1 D_cavitated = cavitated caries. 2 D non cavitated = non-cavitated caries. 3 M = missing teeth due to caries. 4 F = filled teeth due to caries. 5 DMFT = ICDAS codes 3–6 + missing due to caries + filled due to caries teeth. 6 DMFT + non cavitated lesions = DMFT + ICDAS scores 1–2. 7 PI = plaque index. 8 GI = gingival index. 9 pH_saliva (sample 0) = basal salivary pH. 10 pH_saliva (sample 1) = salivary pH after tooth-brushing. 11 pH_saliva (sample 2) = salivary pH after sugar rinse. 12 pH_plaque_t0 = plaque pH before sugar rinse. 13 pH_plaque_t10 = plaque pH after sugar rinse. 14 Lactate_plaque_t0 = plaque lactate levels before sugar rinse. 15 Lactate_plaque_t10 = plaque lactate levels after sugar rinse. CC: correlation coeficient. Values in bold are statistically significant (p < 0.05).Correlation analysis (Correlation Coefficient, CC and p-value) between the levels of S. mutans and S. dentisani in saliva or plaque with different caries indices and biochemical parameters.1 PI= plaque index. 2 GI = gingival index. 3 D_cavitated = cavitated caries. 4 D_non cavitated = non-cavitated caries. 5 M = missing teeth due to caries. 6 F = filled teeth due to caries. 7 DMFT = ICDAS scores 3–6 + missing due to caries + filled due to caries teeth. 8 DMFT + non cavitated lesions = DMFT + ICDAS scores 1–2. 9 pH_saliva (sample 0) = basal salivary pH. 10 pH_saliva (sample 1) = salivary pH after tooth-brushing. 11 pH_saliva (sample 2) = salivary pH after sugar rinse. 12 pH_plaque_t0 = plaque pH before sugar rinse. 13 pH_plaque_t10 = plaque pH after sugar rinse. 14 Lactate_plaque_t0 = plaque lactate levels before sugar rinse. 15 Lactate_plaque_t10 = plaque lactate levels after sugar rinse, CC: correlation coefficient. Values in bold are statistically significant (p < 0.05).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06050.txt ADDED Viewed

	@@ -0,0 +1 @@

+ The overuse of synthetic pesticides in plant protection strategies has resulted in numerous side effects, including environmental contamination, food staff residues, and a threat to non-target organisms. Several studies have been performed to assess the pesticidal effects of plant-derived essential oils and their components, as partially safe and effective agents, on economically important pests. The essential oils isolated from Satureja species are being used in medicinal, cosmetic, and food industries. Their great potential in pest management is promising, which is related to high amounts of terpenes presented in this genus. This review is focused on the acute and chronic acaricidal, insecticidal, and nematicidal effects of Satureja essential oil and their main components. The effects of eighteen Satureja species are documented, considering lethality, repellency, developmental inhibitory, and adverse effects on the feeding, life cycle, oviposition, and egg hatching. Further, the biochemical impairment, including impairments in esterases, acetylcholinesterase, and cytochrome P450 monooxygenases functions, are also considered. Finally, encapsulation and emulsification methods, based on controlled-release techniques, are suggested to overcome the low persistence and water solubility restrictions of these biopesticides. The present review offers Satureja essential oils and their major components as valuable alternatives to synthetic pesticides in the future of pest management.Although synthetic chemicals have been considered as the pest management strategy so far, their overuse has led to several side effects. These include soil and groundwater pollution, toxic residues on the food stuffs, pest resistance, outbreak of secondary pests, and harmful effects on non-target organisms such as fish, bees, predators, and parasites [1,2,3,4].The plant essential oils as low-risk agents are recommended alternatives to chemical pesticides [5,6]. Essential oils are complex mixtures of aromatic and aliphatic compounds, which mainly consist of hydrocarbon monoterpenes, monoterpenoids, hydrocarbon sesquiterpenes, and sesquiterpenoids, and can be made by all plant parts, such as flowers, seeds, leaves, stems, and bark [7]. Essential oils are composed by plants as secondary metabolites with anti-herbivore activity, resulted in critical defense strategies against herbivorous pests along with other significant roles, such as allelopathic plant–plant interactions and attraction of pollinators [8]. Hence, the possibilities of pest resistance to plant-derived essential oils is very low [9]. Along with multiple modes of action and efficiency against a wide range of arthropod pests, essential oils also exhibit comparative lower toxicity on non-target organisms, such as mammals and beneficial insects compared to chemicals [10]. Additionally, with about 24–48 h half-lives, they are degraded quickly by natural degradation mechanisms and considered as biodegradable agents [9]. The pesticidal effects of essential oils isolated from several species of plant families, such as Lamiaceae, Asteraceae, Myrtaceae, Apiaceae, Cupressacae, and Rutaceae, against diverse groups of agricultural pests have been well-endorsed in recent years [11,12,13]. Along with the toxicity of plant essential oils to arthropod pests, there are promising findings against pathogenic nematodes [14,15].The genus Satureja belongs to the Lamiaceae family, Nepetoidae subfamily, and the Mentheae tribe, that includes about 200 species of aromatic herbs and shrubs. They are broadly distributed in America, the Mediterranean area, Middle East, North Africa, and West Asia [16]. Several species from this genus, conventionally known as savory, especially summer savory (Satureja hortensis L.), are cultivated in various countries [17]. These aromatic plants possess a high content of essential oil (even about 4%) located in their leaves, stems, and flowers [18]. Numerous medicinal properties, including reduction of blood pressure, joint pains, rheumatic pains, stomachache, toothache, fever, diarrhea, dyspepsia, gastrointestinal bloating, influenza, colds, scabies and itching, eye strengthening, antioxidant, antidiabetic, and antimicrobial properties, of Satureja species, especially their extracted essential oils, are well-documented in the literature [16,19,20,21].The present review aimed to update the current knowledge on the essential oils extracted from different Satureja species in controlling economically damaging insects, mites, ticks, and nematodes. Thus, vast amounts of individual research have been gathered from scientific databases, including Scopus, Web of Science, PubMed, and Google Scholar. Our main aim was to introduce a novel, safe, and efficient bio-rational agent(s), as alternatives to the detrimental chemicals. The search also considers the sub-lethal and biochemical changes after application of these compounds in order to obtain a thorough insight into their mode of action.The great potential of several species from the Satureja genus, including S. aintabensis Davis, S. bachtiarica Bung, S. cilicica Davis, S. cuneifolia Ten, S. hellenica Halásky, S. hortensis L., S. intermedia C. A. Mey, S. isophylla L., S. khuzestanica Jamzad, S. montana L., S. parnassica Heldr & Sart ex Boiss, S. parvifolia (Phil) Epling, S. rechingeri Jamzad, S. sahendica Bornm, S. spicigera Boiss, S. spinosa L., S. thymbra L., and S. wiedemanniana (Avé-Lall) Velen, has been reported in the insects, mites, ticks, and nematodes’ management. As shown in Table 1, the efficiency of Satureja essential oils was assessed against a diverse group of insects from Coleoptera to Diptera, Hemiptera, Homoptera, Lepidoptera, Phthiraptera, and Thysanoptera orders, and similarly, on other arthropods, including mites and ticks, and plant pathogenic nematodes.The pesticidal effects of Satureja essential oils can be considered from two viewpoints, i.e., lethal and sub-lethal. For example, along with acute fumigant toxicity of S. thymbra essential oil against the adults of Acanthoscelides obtectus, Ephestia. kuehniella, and Leptinotarsa decemlineata, its repellent effect on Aedes albopictus was also reported [22,23,24]. In general, there are several sub-lethal bio-efficiencies of Satureja essential oils, including repellent and antifeedant activities and adverse effects on fecundity, fertility, and life cycle. Some of these studies have also considered the biochemical mode of action in pests such as general esterase, acetylcholinesterase, and cytochrome P450 monooxygenases [25,26,27]. The studies include different developmental stages of pests, from eggs to larvae, pupae, and adults. Among the large species of Satureja studied, the essential oils of S. hortensis, S. montana, and S. thymbra are considered as the most promising in pest management (Table 1). Another prospective is the possibility of using Satureja essential oil along with other pest control agents, such as entomopathogenic fungi. For example, Hosseinzadeh et al. [28] indicated that the essential oil of S. sahendica had a significant synergistic effect with entomopathogenic fungus Beauveria bassiana against the cowpea weevil, Callosobruchus maculatus (Fabricius).Reported acaricidal, insecticidal, and nematicidal effects of the essential oils isolated from different Satureja species.Furthermore, as shown in Table 1, in addition to agricultural pests, the acute toxicity and repellent action of Satureja essential oils against larvae and adults of blood-sucking mosquitos that carry pathogenic agents were also approved. For example, high susceptibility of the Asian malaria mosquito (A. stephensi) and the filariasis vector mosquito (C. quinquefasciatus) to the essential oil of S. bachtiarica was reported, in which 100% larval mortality of both insects was attained by the concentration of 160 ppm after 24 h exposure time [30].The major compounds of essential oils of different Satureja species’ insecticidal, acaricidal, and nematicidal activities are depicted in Table 2. Some compounds such as γ-terpinene, borneol, carvacrol, p-cymene, and thymol were identified in many species. For example, thymol with high percentage is the main compound of S. aintabensis, S. bachtiarica, S. cilicica, S. intermedia, S. isophylla, S. montana, S. parnassica, S. sahendica, S. spinosa, S. thymbra, and S. wiedemanniana essential oils. However, some compounds, such as estragole, piperitenone, piperitenone oxide, α-terpineol, β-caryophyllene, and β-myrcene, were recognized in a species: estragole in the S. hortensis, Piperitenone and piperitenone oxide in S. parvifolia essential oil, and β-myrcene in S. isophylla essential oil (Table 2).The identified compounds in the essential oils of Satureja species are categorized in the monoterpene hydrocarbon, monoterpenoid, sesquiterpene hydrocarbon, sesquiterpenoid, and phenylpropanoid groups (see Table 3). Indeed, the majority of recognized compounds are in the monoterpene group, with lower molecular weight than others, and only three compounds belong to other categories. There is sufficient evidence that the monoterpenes, especially monoterpenoids, have high pesticidal properties, and some novel and reliable outcomes in this field are shown in Table 3. For example, the toxicity of thymol, as one of main components in several species of the Satureja genus, was reported against the African cotton leafworm (Spodoptera littoralis Boisduval), the bed bugs (Cimex lectularius L.), the Colorado potato beetle (Leptinotarsa decemlineata Say), the granary weevil (Sitophilus granarius (L.)), the green peach aphid (Myzus persicae (Sulzer)), and the root-knot nematode (Meloidogyne javanica (Treub) Chitwood) [73,76,77]. It can be concluded from these studies that the presence of higher total monoterpenoid content of essential oils had a positive correlation with their pesticidal activity [78,79,80,81]. Thus, the acaricidal, insecticidal, and nematicidal effects of Satureja essential oils may be related to the high amounts of compounds listed in Table 3. It was also demonstrated that the phenolic monoterpenoids such as thymol with CH(CH3)2 functional group displayed significantly higher pesticidal effects compared to other terpenes, such as carvacrol and eugenol with CH3 and OCH3 functional groups, respectively [82,83]. However, the synergistic acaricidal, insecticidal, and nematicidal effects of minor components such as α- and β-pinene, camphor, menthol, sabinene, and thujene should also be considered [84,85,86,87]. For instance, the synergistic insecticidal action of terpenes that have methyl functional groups such as p-cymene and limonene with borneol is another consideration already reported by Pavela [83].The acetylcholinesterase (AChE) is actively involved in metabolic conversion of ‘acetylcholine’ in the synaptic cleft of arthropods and has two catalytic and peripheral target sites. The insect-specific cysteine residue positioned at the acetylcholinesterase active site is a proposed target site for developing insecticides to reduce off-target toxicity [94]. On the other hand, inhibition of pest-specific acetylcholinesterase will decrease the risk of utilized pesticides on non-target organisms, such as mammals [94]. Some essential oils and compounds are reported to bind with these target sites to inhibit the AChE action [95,96,97]. Park et al. [26] revealed that the essential oil of S. montana had significant AChE inhibitory activity against the fruit fly (Drosophila suzukii (Matsumura)), along with high toxicity. The inhibition of AChE leads to acetylcholine accumulation, hyperactivity, paralysis, and death of the pest. Along with terpenes, the well-known phenylpropane estragole has also shown AChE inhibitory effects [98,99]. It should be noted that the AChE inhibition can occur in both contact and fumigation methods of used essential oils [100,101]. Octopamine, as a neurotransmitter, neuromodulator, and hormone, is one of the important biogenic amines in invertebrates and is released at times of high energy demands [102]. Octopamine receptor alteration is considered as another mode of action of essential oils or their components [103]. The blockage of gamma-amino butyric acid (GABA) and nicotinic acetylcholine (nAChR) receptors has also been documented in some studies [97,104].Beside the neurotoxic modes of pesticidal action of essential oils and compounds, there are several studies indicating enzymatic and non-enzymatic effects. The destructive effects of essential oils and their compounds on esterases and glutathione S-transferases (GSTs) as imperative detoxifying enzymes in arthropod pests are reported [88,105,106]. Disruption of the function of detoxifying enzymes may reduce the probability of pest resistance [107], and this has been clearly depicted by essential oils and their components. Farahani et al. [27] showed that the essential oil of S. khuzestanica had adverse effects on cytochrome P450 monooxygenases (P450, responsible for the oxidative metabolism of a variety of xenobiotics and endogenous compounds) function of two spotted spider mites (Tetranychus urticae Koch), along with toxic and repellent activities. The adverse effects of these agents on digestive enzymes such as lipases, proteases, α-amylases, α-glucosidases, and β-glucosidases were also reported [106], which can be very effective in reducing the nutritional efficiency of pests. Effects on energy reservoirs of the pest by decreasing the protein, glucose, and triglyceride contents and disrupting the action of immunological and hematological parameters are the other reasons to approve the multiple modes of action of these eco-friendly bio-pesticides [108,109].Although great potential for acaricidal, insecticidal, and nematicidal activity of Satureja essential oils and compounds have been reported, limitations such as susceptibility to light, moisture, oxygen, and temperature may restrict their application in the pest management strategies [5]. Indeed, the use of essential oils and their components in non-crop agriculture in the management of stored product pests, flies, and cockroaches is effective [110]. Additionally, the larvicidal activity of essential oils by treating standing water and waterways and their repellent effects on adults may be useful in mosquito management (See Table 1 and Table 3 for examples). Due to the disadvantage of low persistence in environmental conditions, the application of essential oils in crop agriculture can be limited [6]. Soft body and sucking pests (viz., aphids, thrips, and mites) are usually controlled by essential oils on crops, particularly under low pest pressure [110]. For example, Western flower thrip and green peach aphid were successfully controlled by the essential oil-based insecticide Ectrol (EcotecTM, California, USA) on lettuce and strawberry. However, partial efficiency was achieved against larger chewing insect pests, such as coleopterans and lepidopterans [110].Nanoencapsulation based on the controlled release technique has been offered to overcome the lack of persistence restriction of bio-pesticides [111]. In the nanoencapsulation process, the active agent as a solid, liquid, or gas is surrounded by a thin layer of natural or synthesized polymer or a membrane to keep the core active agent from harmful environmental factors [112]. Generally, reducing the amount of active ingredients and minimizing evaporation and its controlled release are main advantages of nanoencapsulation [111]. However, along with above-mentioned advantages, expensive and difficult processes of the creation of nano-formulations should be considered. In the study of Ahmadi et al. [65], encapsulation of S. hortensis essential oil in chitosan-tripolyphosphate nanoparticles improved its ovicidal and adulticidal toxicity against T. urticae. Along with high toxicity, nanoencapsulation of S. hortensis essential oil in chitosan-tripolyphosphate nanoparticles enhanced its persistence so that 80% and 15% mortality was achieved for nano-encapsulated and pure essential oil formulation after 14 days. Usha Rani et al. [113] evaluated the antifeedant activity of pure and silica nanoparticles-based capsulated α-pinene and linalool against the tobacco cutworm (Spodoptera litura F.) and the castor semi-looper (Achaea janata L.). Although both terpenes had significant antifeedant effects, nano-capsule formulation augmented their effectiveness up to 10 and 25 times for A. Janata and S. litura, respectively. The same results regarding the enhancing toxicity and persistence of other essential oils by encapsulation in polymeric and non-polymeric materials, such as poly(ethylene glycol), myristic acid-chitosan, and mesoporous material, were also documented [114,115,116]. The preparation of nano-emulsions is another applicable method to solve the solubility restriction of essential oils in water and is more effective with minute quantities of toxic substances, both in medicinal and agricultural pest management prospects [117,118]. Further, the combination of essential oils with other protectants such as microbial agents may enhance their effectiveness. For example, the combination of S. sahendica essential oil with entomopathogenic fungus Beauveria bassiana augmented its toxicity against cowpea weevil, and insect pest mortality increased from 50% after a 1-day exposure time to 80% after 7 days [28].Along with antibacterial, antifungal, antiviral, and general importance in medicinal, food, and cosmetic industries [119,120,121], the essential oils isolated from different species of Satureja genus could have great potential in the management of detrimental mite and tick Acari, insects, and nematodes. Pesticidal effects of Satureja species essential oils, which may be commonly related to their main terpenes [67,83,86], were reported as lethal contact and fumigant toxicity to sublethal repellent action, developmental inhibitory effects, adverse effects on the feeding, life cycle, oviposition, and egg hatching, and biochemical disturbances, such as reduction in general esterase content and inhibition of acetylcholinesterase and cytochrome P450 monooxygenases functions (see Table 1 and Table 3). Such multiple modes of action of essential oils and their compounds, in addition to reducing pest resistance, can affect a wide range of pests [5,9]. Despite all of the mentioned advantages, high volatility or lack of persistence and insolubility in water are the main restrictions in the commercialization and extensive application of these compounds [110]. Accordingly, their application is principally focused against indoor non-crop pests such as storage pests, flies, and cockroaches [96,114]. Further, the acute toxicity against larvae and repellent activity on the adults of mosquitos that carry pathogens and suck blood were also documented in Table 1 and Table 3. However, with micro- and nano-encapsulation on the basis of controlled release techniques, their persistence can be increased [122]. Although nano-emulsification is also a suitable way to dissolve essential oils in water [123,124], it is possible to increase their effectiveness by combined application with microbial control agents, such as entomopathogenic fungi [28]. These less-toxic substances may help in agriculture and environmental protection and can be proposed to countries that apply extreme amounts of synthetic pesticides. However, effects on beneficial and non-target organisms, residues on food products, and more importantly, considering a method for lower cost of Satureja essential oils and their components, should also be investigated in future research.Conceptualization, A.E.; methodology, A.E.; investigation, A.E., J.J.S., and M.Z.; resources, A.E., J.J.S., and M.Z.; writing—original draft preparation, A.E. and P.K.; writing—review and editing, A.E., J.J.S., M.Z., and P.K. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.This study received financial support from the University of Mohaghegh Ardabili, which is greatly appreciated. The publication of this review was financially supported by Chiang Mai University, Thailand.The authors declare no conflict of interest.Main components of the Satureja species essential oils documented as promising insecticidal, acaricidal, and nematicidal agents.Characteristics and pesticidal activities of main components identified in Satureja species.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06051.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Background: Peer leader interventions are effective strategies for promoting prevention behaviors in communities at risk for HIV, yet little is known about their effects on the social and behavioral dynamics of peer leaders themselves. Methods: Using data from PrEP Chicago, an RCT PrEP for prevention intervention for young Black MSM (YBMSM), we apply stochastic actor-based models to longitudinally model the impact of study participation on the online friendship and PrEP adoption dynamics among a network of peer leaders (n = 174) and a network of control group counterparts (n = 166). Results: Peer leaders assigned to the same leadership training workshop were more likely to form new Facebook friendships with one another, whereas control participants assigned to the same attention control workshop were no more or less likely to form new friendships. Further, peer leaders with greater PrEP intentions and those living with HIV were more active in forming new friendships with other peer leaders, effects not found in the control network. PrEP adoption was not influenced by network dynamics in either group. Conclusions: The implications of these findings are discussed through the lens of community-capacity building and the role that peer leader interventions and the networks they engage can impact public health.Despite the clear efficacy of Pre-Exposure Prophylaxis (PrEP) in preventing HIV transmission, meaningful uptake in populations experiencing high HIV incidence, most notably young Black gay, bisexual, same gender loving and other men who have sex with men (hereafter YBMSM), has yet to occur [1,2]. In a population-based sample of YBMSM in 2014 for example, only 41% had ever heard of PrEP and 4% ever used it [3], despite this same population experiencing some of the highest rates of HIV in the United States [4]. Likewise, in PrEP demonstration and implementation projects in Washington, DC [5] and New York City, NY, USA [6], less than 15% of PrEP clients identified as Black, while only 31% of clients identified as Black in the CDC’s Sustainable Health Center Implementation PrEP Pilot (SHIPP) [7]. Although individual factors like low awareness, misperceptions about suitability, and concerns about side effects are common first-order barriers to PrEP uptake [8,9,10], deeper and more complicated social obstacles are also increasingly observed. For example, stigma associated with being perceived as HIV positive or sexually promiscuous have been linked to YBMSMs’ reluctance to consider PrEP [11]. Additionally, as intersectional racial and sexual minority identities [12], YBMSM must cope with homophobic and racist discrimination in multiple contexts, including clinical spaces, which can complicate their willingness to engage with traditional sources of HIV prevention messaging such as primary and sexual health care providers and public health researchers [13,14]. Given these challenges, innovative implementation strategies are clearly needed that can reach greater portions of most impacted communities like YBMSM and provide alternative paths to prevention services that are perceived by YBMSM as more accepting and less stigmatizing.Decades of social diffusion research underscores the role that personal relationships, and the trust they engender, play in the adoption and spread of novel ideas and behaviors in a population [15]. More recently, social network interventions—intentional efforts to leverage network structure and peer influence processes to accelerate the diffusion process—have been advanced from a public health framework [16]. Perhaps the most intuitive network intervention is the peer leader intervention, where members of the prioritized population (i.e., peers) are positioned in the role of health educators who disseminate information about a health innovation through their personal networks [16,17,18,19]. Applied to the challenge of advancing PrEP, peer leader interventions offer the opportunity to reach larger portions of communities at risk for HIV seroconversion by treating social networks as opposed to individuals in isolation [20] while also privileging endogenous, community-based systems of communication and influence over institutionalized ones. Indeed, efforts to leverage peer leaders to promote PrEP awareness and early linkage among Black MSM, although few in numbers, show promise in this regard [21,22,23]. Unsurprisingly, the impact of peer leader interventions tends to be measured on the basis of observed changes in the health behaviors of network associates with whom peer leaders are trained to interact, not on the basis of change in the behaviors of the peer leaders themselves. This is because peer leaders are a network intervention’s active ingredient, not the focus of change. Although cogent arguments have been made that peer leader interventions have valuable secondary effects on the attitudes and behaviors of peer leaders themselves [24], rarely have these effects been rigorously evaluated (for exceptions see [25,26,27]). As a consequence, we know little about the way in which peer leader interventions impact the network and health behavior dynamics among the peer leaders themselves.To understand the significance of learning how the evolving social and behavioral dynamics of peer leaders are impacted by their involvement in an intervention, it is vital to see peer leaders, collectively, as an investment in community capacity. Theories of community development underscore the importance of building capacity in communities facing adversity, meaning that community members’ abilities to become active agents (rather than objects) of change must be nurtured [28]. Two fundamental components of capacity building are: (1) investment in the development of local leaders who are empowered to help the community make good decisions, and (2) nurturing the formation of social networks that facilitate flows of information and support [28]. Endogenous leaders and the networks they form are, therefore, a public good, embedded with knowledge, experience, and social capital from which the community as a whole can benefit [29]. From this perspective, a peer leader intervention is more than just a means to diffuse information about an innovation like PrEP through the networks of YBMSM: it is also a means to strengthen the capacity and resilience of a marginalized community through its activation and nurturing of a network of young community health leaders [30,31,32]. The degree to which the intervention nurtures this system of human and social capital, however, depends on the collective capacity that the peer leaders generate among themselves through their own network-building [33,34,35] and their own buy-in with regard to the behavior they are being asked to promote (i.e., PrEP adoption). The objective of this study, then, is to ascertain whether the training and support provided by a PrEP peer leader intervention is indeed an engine of social and behavioral change among YBMSM peer leaders or whether these dynamics are more attributable to factors outside the scope of the intervention, for example individual characteristics or structural features of their naturally-evolving organic networks. To these ends, we draw on a novel dataset of social network and behavioral data collected from a large cohort of peer leaders in a community-based PrEP for prevention intervention for YBMSM living in Chicago, IL, USA. Using stochastic actor-based models (SABMs) [36,37,38], we longitudinally model the intervention’s impact on both friendship formation and PrEP adoption among peer leaders during the first year of their study enrollment. Further, given that the social and behavioral dynamics of peer leaders are likely to be interdependent, these models also allow us to simultaneously test the effects of behavior on friendship selection (i.e., when PrEP adoption informs friendship formation among peer leaders) and the effects of friendship on rates of behavior change (i.e., when friendships influence PrEP adoption). We anticipate that the results of our analysis will help identify where improvements to peer leader training and engagement may be needed to enhance and strengthen their capacity as a cohort of community health leaders. Data for this study were collected from March 2016–March 2018 as part of a randomized controlled trial peer leadership intervention among 423 YBMSM living primarily on the south and west sides of Chicago. Participants were considered eligible if they met the following criteria: (1) 18–35 years of age, (2) identified as Black or African American, (3) assigned male sex at birth, (4) had sex with a man in the past 12 months, and, because the intervention emphasized social media as a communication tool, (5) had an active Facebook profile. All data collection implicated in this study received ethics approval from the University of Chicago School of Medicine, Biological Sciences Division and from NORC at the University of Chicago.Participants were recruited using respondent-driven sampling (RDS), a procedure well suited for identifying members of “hard-to-reach” populations like MSM [3]. A variant of snowball sampling [39,40], RDS draws on peer referral chains, beginning with a set of initial “seeds” that meet study eligibility. Because seeds should have large social networks (i.e., are popular) and have ties to a diverse array of people belonging to different subpopulations [40,41,42], we selected YBMSM seeds based on their central or boundary spanning positions (i.e., structural signatures of popularity and diversity, respectively) within a previously derived Facebook friendship network among the focal population [43]. Once a seed was enrolled and completed their baseline assessment, they were instructed to recruit up to six peers (or “sprouts”) who also met the eligibility criteria. Following enrollment, sprouts were also instructed to recruit peers, and the process continued until the recruitment target was reached. Participants received a $20 cash incentive for each peer whom they successfully referred into the study.The study design and data collection approach have been previously published in Young and Schumm [44]. To summarize, participants were assigned randomly to one of two treatment sequences: (1) receives the peer leader training in year 1 of the study (Year 1 intervention arm) or (2) receives peer leader training in Year 2 (Year 1 attention control arm). Here, we focus on Year 1 of the study, as the availability of an intervention network and control network allows us to compare the evolving network and behavioral dynamics in each group and ascertain whether the peer leader training itself (i.e., the treatment) impacts those dynamics. Once participants were randomized, they were scheduled for a baseline visit. All participants provided written consent during that baseline visit.The peer leader training adapted the peer educational and mentoring program developed as part of the HIV Prevention Trials Network [45,46] and was conducted in small groups (6–10 people) in a single half-day workshop. The training curriculum was designed to develop an individual’s PrEP knowledge and their communication skills for engaging network associates in PrEP-related conversations. Participants not assigned to the peer leader training in Year 1 were assigned to a attention control condition that reproduced the nonspecific procedures used to engage with intervention participants (e.g., small group, half-day workshops led by study staff) without including its specific content [47] (see Young, Schumm [44] for more details). Data used in this study were collected at Baseline and 12-months. Collection modalities included: (1) a computer-assisted self-administered survey capturing information about PrEP knowledge and attitudes, sexual health behaviors, psychographics, and demographics; (2) biomedical HIV and STI testing; and, to evaluate the relationship between social connectivity and intervention outcomes, (3) a manual download of participant’s Facebook friendship data. A waiver of consent from the IRB for third party (non-participant) network members was obtained given the minimal risk to these individuals. Data protections to secure third party identities (e.g., hashing, numeric de-identification) were also established [20].With the Facebook friendship lists we acquired from consenting participants at baseline and 12-months, we constructed two sets of unweighted undirected edge lists—one set that represented Facebook friendship ties among intervention participants at each time point and another set representing Facebook friendship ties among control participants at each time point. In both sets of edge-lists, all ties to third-parties (i.e., non-participants) were excluded. PrEP use was measured on the basis of a participant’s self-report of being on PrEP at the time of the baseline and 12-month assessment, respectively. From these self-reports, we created a binary PrEP use variable (1 = currently taking PrEP; 0 = not currently taking PrEP). To account for how study participation influences the network and behavioral dynamics in each group, we include a dyadic measure of being in the same training cohort. Specifically, this represents having been assigned to the same peer leader training cohort (for participants assigned to the intervention arm) or the same risk assessment cohort (for participants assigned to the attention control arm). We interpret this particular dyadic covariate in each model as an effect of study participation on the ongoing social dynamics among study participants.Three actor attributes believed to be associated with PrEP adoption and/or the formation of Facebook friendships are also included in our analysis. First, we account for a participant’s age (measured at baseline), which has been shown to be related to Facebook connectivity in previous work [48] and that has also been linked to willingness to adopt PrEP [49]. Second, we include a measure of HIV status (1 = HIV positive; 0 = HIV negative) as this, too, has been linked to increased Facebook connectivity among YBMSM [48] and is an explicit eligibility criteria for PrEP adoption. HIV status was measured using biomedical lab testing (i.e., blood tests) or self-reports if lab tests were not available. Third, to account for the theorized relationship between behavioral intentions and behavioral adoption [50,51], we account for PrEP intentions [52] measured as the perceived likelihood that a participant would take PrEP in the next six months (1 = probably/definitely would not take PrEP; 2 = might take PrEP; 3 = probably/definitely would take PrEP).To account for expected relationships between offline social relationships and Facebook friendship dynamics, we also examine the effect of an offline physical-world dyadic attribute representing being tied to another participant through a referral connection. Although referrals are directed relationships, we include them here as non-directed ties to represent whether or not two actors have a physical world connection outside the study. The analytic sample is a subset (n = 340) of the study participants derived from who among the 423 enrolled participants consented to the Facebook data collection at both their baseline and 12-month assessments. Specifically, of the 423 YBMSM study participants at baseline, 346 were retained at 12-months. Of the 346 participants who were retained at 12-months, six either did not consent to the data collection or experienced technical difficulties when downloading their data at the time of data collection. This yielded a total of 340 participants for whom we have Facebook friendship data at both waves. The analytic dataset was then sub-divided into two sub-samples, one comprised of participants assigned to the intervention arm (n = 174) and the other comprised of participants in the control arm (n = 166). No significant differences were found between those who were retained and consented to the Facebook data collection (n = 340) and those who were not retained or did not consent (n = 83). To model interdependencies between changes in Facebook friendships among the YBMSM study participants and the rate of PrEP adoption among members of each sub-sample, we applied an extension of stochastic actor-based models (SABMs) [37,38] called the actor-based model for diffusion of innovations in dynamic networks [53]. All analyses were implemented using ‘RSiena’ (Simulation Investigation for Empirical Network Analysis) version 1.2-23 for the statistical system R version 4.0.2 [54]. A description of the foundational premise and underlying logic of these models have been published previously by the first author [48] and elaborated extensively by Snijders and Van de Bunt [36] and Greenan [53]. The foundational premise of these models is that an innovation (e.g., PrEP adoption) is not only dependent on the social network to which an individual belongs, but also on the changes in that social network [53]. As such, they model the co-evolution of social networks and behaviors. A more detailed explanation of the logic behind these models is available in Appendix A.Given that intervention participants were intentionally motivated to think and talk about PrEP and to engage with one another as a cohort of peer leaders, while control participants were not, we modeled the co-evolving network and behavioral dynamics of each subsample independently from one another, using the same model specifications. This allowed us to compare model results across treatment and control conditions to see how study enrollment differentially impacted each group and to better understand potential mechanisms of capacity building among peer leaders. The actor-based model for the treatment and control samples includes two sub-models that are estimated simultaneously: a network dynamics sub-model to predict changes to network members’ Facebook friendships, and an adoption process sub-model to predict the rate of PrEP adoption. Facebook friendship ties were operationalized as symmetric (or non-directed) connections and modeled with the assumption that ties are unilaterally initiated and reciprocally confirmed, while confirmation is not required for tie dissolution [55]. These assumptions correspond to how friendships are formed on Facebook: friendship ties are formed when one user initiates a friendship request to another user and the recipient of that request confirms the request, while “de-friending” can be done unilaterally.Guided by these assumptions, the network dynamics sub-model includes a rate function, capturing the speed of change in the Facebook network, and a set of evaluation effects that represent the mechanisms by which the dependent behavior (PrEP adoption), study participation (i.e., being a part of the same training cohort), other actor and dyadic covariates, and the network itself govern the formation of Facebook ties (see Table 1). Two evaluation effects tested the impact of the behavioral dependent variable (PrEP adoption) and each constant actor covariate (age, HIV status, and PrEP intentions) on changes to Facebook friendship ties: (1) the effect of either the behavior or covariate attribute on an actor’s propensity to form Facebook friendships (egoPlusAltX), and (2) the effect of assortativity on the behavior or covariate attribute, where actors are more likely to form friendships with other actors who share the same behavior and/or attribute (sameX). Additionally, we included a constant dyadic covariate term to represent the effect of a study-specific relationship between participants assigned to the same intervention or control training group. We also controlled for a dyadic covariate that represents the effect of a study referral relationship. Finally, we also controlled for several structural effects that represent the way in which Facebook friendships are formed in response to the presence or absence of other ties in the network. Specifically, we included: (1) a required degree effect that models the overall tendency for actors to form Facebook friendships (density), (2) an effect that represents the tendency to have network closure in Facebook friendships (geometrical-weighted edgewise shared partnerships (gwesp)), (3) a term representing the preference to form Facebook friendships with highly connected network members (i.e., actors with high Facebook degree) (degPlus), (4) an effect representing assortativity on Facebook degree (degree assortativity), and (5) an effect that models the tendency for network isolates to remain isolated (outIso). Structural parameters were chosen on the basis of theoretical considerations and the results of overall goodness of fit tests.As demonstrated by Greenan [53], we model the PrEP adoption process as a proportional hazards model [56], meaning that we model at any given point in time the risk of a single actor adopting PrEP for the first time, conditional on the current state of the dynamic network. We consider three types of adoption effects in our adoption process sub-model (see Table 1). Total Exposure (totExposure) captures social influence conveyed through overt exposure and is measured by the total number of network contacts that are PrEP users [57]. Infection by degree (infectDeg) is a measure of how influential an actor’s PrEP use is on the rest of the system [58], where influence is operationalized in terms of their popularity on Facebook (i.e., Facebook friendship degree) [53]. Finally, we also consider how intrinsic characteristics affect an actor’s propensity to adopt PrEP, irrespective of the PrEP use of other network members, by considering the effects of an actor’s HIV status and their PrEP intentions.A summary of descriptive statistics for the behavior (PrEP adoption) and each constant attribute covariate for the intervention and control arm sub-samples is presented in Table 2. Results of tests of difference (not shown) revealed no significant differences between the intervention and control participants on these attributes.Structural characteristics of each Facebook friendship network at baseline and 12-months and their dynamic features are summarized in Table 3. First, in both the intervention and control networks, an increase in Facebook connectivity among study participants during the first 12 months of the study is evident. Study participants in both arms of the study gained on average four friendships with other study participants between baseline and 12-month observations, which corresponded to a 0.02 increase in network density in both arms. Second, there were also slight increases in friendship closure among study participants in each study arm, made apparent by changes in transitivity: transitivity increased from 27% to 30% in the intervention arm and 25% to 31% in the control arm. Finally, although the global centralization of the friendship network is relatively low at baseline (16% concentration in the intervention arm and 18% concentration in the control arm), there was a 7% and 5% increase in friendship concentration around “hubs” in the intervention and control arm networks, respectively. When we examine each network over time at a more granular tie-level, the network changes become more obvious. Between Waves 1 and 2, intervention participants added 555 new ties to their Facebook friendship network and control participants added 533 new ties to their Facebook friendship network. This equated to about 3.2 new friendships per person in each network. At the same time, 198 (17%) of the 1140 baseline friendships among intervention arm participants were dissolved, while 200 (25%) of the 813 baseline friendships among control arm participants were dissolved. In total, 942 (83%) and 613 (75%) of the baseline Facebook friendships observed among intervention and control arm participants, respectively, were maintained. Figure 1 shows the Facebook friendship networks among intervention and control arm participants at baseline and 12-months, with actor nodes colored by their PrEP adoption status at each wave.In the intervention arm (Table 4, Model 1), the network dynamics sub-model reveals that the intervention arm participants were more likely to form and maintain Facebook friendships with one another if they were also co-members of the same peer leader training cohort (b = 0.92, p < 0.001). This lends support to the idea that the group training context can be an important ingredient in increasing the collective capacity of candidate peer leaders.Although we observed an 8.5% increase in PrEP adoption among intervention participants between baseline and 12-months, results show that changes in PrEP adoption had no effect on the formation of Facebook friendships among intervention participants. Rather, actor covariates like age, HIV status, and PrEP intentions were more important mechanisms of network change. Specifically, intervention participants who were living with HIV (b = 0.19, p < 0.01) and who had greater intentions to adopt PrEP (b = 0.13, p < 0.01) were more active in forming friendships with other network members irrespective of the HIV status and PrEP intentions of those members. Intervention participants were also more likely to form friendships on the basis of age (b = 1.30, p < 0.001) and HIV status (b = 0.18, p < 0.05) similarities. Having an existing offline relationship in the form of a study referral tie (b = 0.67, p < 0.10) also influenced the formation of new Facebook friendships, although the significance of that effect was marginal.Finally, the formation of Facebook friendships among intervention participants was also governed by the structure of the network itself. The formation of friendship ties was not done arbitrarily (negative degree term), but they were more likely to form friendships with the friends of their friends, ensuring network closure (positive gwesp term (b = 0.26, p < 0.05)), and more likely to form friendships on the basis of mutual popularity (positive outInAss term (b = 0.06, p < 0.01)).The PrEP adoption process sub-model featured in model 1 shows that the rate of PrEP adoption among intervention participants was not influenced by their Facebook friendships; neither their connections to friends who were PrEP adopters (totExposure term (b = 0.13, p = n.s.)) nor their connections to influential PrEP adopters (infectDeg term (b = −0.01, p = n.s.)) impacted their PrEP adoption. Unsurprisingly, the rate of PrEP adoption was significantly predicted by the intrinsic effect of being HIV positive (b = −1.34, p < 0.05), which we included as a control variable. PrEP adoption was not, however, influenced by participants’ PrEP adoption intentions (b = 0.51, p = n.s.).In the control arm (Table 4, Model 2), the network dynamics sub-model shows that being a part of the same training group had no effect on the formation and maintenance of Facebook friendships among control arm participants (b = 0.34, p = n.s.). This makes sense given that the attention control workshop was not designed to encourage participants to think of themselves as a collective or to build connections with one another.Similar to the intervention group, changes in PrEP adoption among control arm participants had no effect on their friendship dynamics: PrEP adopters were no more or less likely to form friendships with other network members (b = 0.11, p = n.s.) nor were they more or less likely to form friendships with other PrEP adopters (b = 0.26, p = n.s.). Instead, older participants were more likely to form new Facebook friendships with other network members (b = 0.02, p < 0.01) and, like the intervention arm, new friendships were more likely to form among control arm participants who were similar in age (b = 0.78, p < 0.001). However, unlike intervention participants, friendships among control participants were not influenced by their HIV status. Namely, control participants who were living with HIV were no more or less likely than HIV negative participants to form friendships with other network members (b = −0.10, p = n.s.), and they were no more or less likely to sort based on HIV status similarities (b = −0.005, p = n.s.). Further, control participants were more likely to form new Facebook friendships if they had an existing offline relationship in the form of a study referral tie (b = 1.46, p < 0.001), unlike their intervention counterparts.Finally, the structure of the network itself also influenced network dynamics among control participants. In line with results of the intervention model (Model 1), the formation of new friendship ties in the control network were more likely to ensure network closure (b = 0.58, p < 0.001) and were more likely to form on the basis of degree similarity (b = 0.06, p < 0.01). Further, the effect of network isolation was positively significant (b = 2.69, p < 0.001), indicating a positive tendency toward network isolation.Modeling the PrEP adoption process among control participants yielded similar results to those from the intervention model. The rate of PrEP adoption among control participants was not influenced by their Facebook friendships, neither through their exposure to friends who were PrEP adopters (b = 0.41, p = n.s.) nor through their connections to influential PrEP adopters (b = −0.02, p = n.s.). Likewise, the rate of PrEP adoption was significantly predicted by the intrinsic effect of HIV status (b = −1.99, p < 0.05) and marginally influenced by PrEP adoption intentions (b = 1.12, p < 0.10). Goodness of fit results for both models are available in Appendix B (see Figure A1). The purpose of this study was to evaluate the impact of a PrEP peer leader intervention on a large cohort of YBMSM peer leaders, with a specific interest in understanding whether and how the activation of study participants as peer leaders altered their online network and PrEP behavior dynamics. Our analysis showed that online tie formation among participants in both arms of the study increased during the first 12-months of the intervention and, for the intervention participants, this increase was partially attributed to their participation in the peer leadership training program. Specifically, candidate peer leaders who participated in the same small-group peer leader training workshop at the onset of the intervention were more likely to form new Facebook friendships with one another during their 12-month enrollment. Our findings also suggest that the increase in connectivity among peer leaders can also be attributed to characteristics of the peer leaders themselves, namely their HIV status, their PrEP intentions, and their age. Peer leaders who were living with HIV were more active in forming new friendships with other peer leaders and were more active in forming new friendships with one another. Further, HIV negative peer leaders who had greater PrEP intentions were also more likely to form new connections with other peer leaders during the study, and new friendships were more likely to emerge between peer leaders who were similar in age. Although our analysis shows that PrEP adoption did increase among both sets of participants during the first 12-months of the study, those increases were seemingly unrelated to their network dynamics. To rule out the possibility that our findings regarding the absence of social influence effects on PrEP adoption were the result of having too few PrEP adopters in either arm, we performed supplementary analysis (not shown here) of the PrEP adoption process in the unstratified sample using the same model specifications applied to the stratified samples. Results of this supplementary analysis (not shown here) reveal similar results: neither social influence term (total exposure or infection by degree) played a significant role in PrEP adoption. Our findings have implications for future peer leadership interventions and their role in community capacity-building. To begin, we are encouraged by the fact that the peer leadership training, a seminal component of the implementation of the intervention, played a significant role in encouraging the formation of new ties among newly activated peer leaders. Although we cannot say for sure that peer leaders perceived an increase in their individual and collective capacities as a result of their involvement in the training session, the fact that new online friendships were forged among members of the same training cohort after their training suggests that the relational momentum behind collective capacity was initiated. We interpret the social effect of the training to be related to three aspects of its implementation. First, the training curriculum itself was explicitly designed to empower peer leaders to see themselves as a collective. That the trainings were conducted in small group settings was intentional, as it encouraged a cohort mentality and helped create the sense of a shared experience. Second, a Facebook group was created as part of the study to serve as a PrEP information repository and communication channel for peer leaders during their enrollment in the study. Participants were informed about this group during their training session. Although participants were not required to participate in group conversations or connect with other group members, it is likely that having access to the Facebook group enabled these connections and communication exchanges to emerge voluntarily. Third, several study events were held for peer leaders during their 12-month enrollment that brought training cohorts together to celebrate their work and to exchange experiences engaging with peers about PrEP. These, too, provided additional opportunities for connection and potential capacity-building. For these reasons, it seems clear that organizing additional social opportunities for peer leaders, where they can connect and support one another, will help nurture their collective capacity as community health leaders. Another set of noteworthy findings pertain to the role that PrEP itself played in the formation of new friendships among PrEP peer leaders. Although 15 peer leaders adopted PrEP during their tenure in the peer leader role, their changes in behavior did not influence their friendship dynamics with other peer leaders. In other words, we did not observe PrEP adopting peer leaders playing a more active role in strengthening the social fabric among peer leaders. We take this as an indication that being an engaged peer leader is not necessarily contingent on taking PrEP. It has been surmised that drawing on community members who have personal experiences with the behavior of interest is an optimal strategy for selecting candidate peer leaders, as this may increase their enthusiasm and engagement in the study [59,60], as well as their self-efficacy [61]. However, findings from previous work and the current study suggest otherwise. In prior work, we learned that PrEP adopters were no more or less likely to recruit others into the study or to complete check-in calls with study staff [20], and in the current study we learned that this extends to their likelihood of connecting with other peer leaders. That said, controlling for peer leaders who had already adopted PrEP and who were living with HIV, we learned that peer leaders who had greater intentions to adopt PrEP at the start of the intervention were significantly more likely to form new relationships with other peer leaders. As such, it seems as though having greater interest in taking PrEP in the near future may be a critical motivation for being more engaged in the study and wanting to build community with other peer leaders, as these particular peer leaders could relate to the intended audience of the intervention (i.e., YBMSM who are good candidates for PrEP). With this in mind, it may be wise to recruit candidate peer leaders who demonstrate greater interest in taking PrEP themselves. Additionally, these findings also suggest that more effort could have been made to motivate PrEP-adopting peer leaders to take a more active role in strengthening the capacity of other peer leaders, for example by asking them to share their experiences being on PrEP with other peer leaders and to take on a leadership role within the peer leader cohort.Also worth highlighting is the role that people living with HIV played in building community capacity among peer leaders. Despite not being eligible for PrEP themselves, we surmised that people living with HIV could be a powerful voice in bringing attention to a biomedical tool that can prevent the transmission of HIV and liberate status discordant couples. Our findings provide some preliminary evidence for this intuition: peer leaders living with HIV were more likely to contribute to the increased connectivity of the peer leader network and were more likely to do so as a cohort, as evidenced by the fact that they were more likely to form new ties with other participants living with HIV. That this dynamic was only evident in the peer leader network (as opposed to the control arm network) suggests that the intervention itself may have been a motivating factor. Pragmatically speaking, the tendency for PrEP peer leaders who are living with HIV to forge new friendships among themselves could be leveraged in the implementation of the intervention, for example by helping them coordinate their outreach efforts as a cohort.Our results also point to a structural effect that has implications for community capacity-building. In both sub-samples, we found significant positive effects of network closure (i.e., the tendency to form friendships with the friends of your friends) on friendship formation. Given Facebook’s “People You May Know” Recommender, which suggests people you should connect with based on mutual friendships, it is unsurprising that network closure was a significant positive predictor of network change among study participants in both the intervention and control arms. In the context of a peer leadership intervention, network closure can have advantages and disadvantages. On one hand, network closure can generate bonding social capital [62] by nurturing trust, support, and solidarity among peer leaders. Research has shown that these social assets can be critical for capacity-building, specifically for creating AIDS-resilient communities [63]. Therefore, future research and implementation planning should be directed toward devising strategies to leverage the bonding capital that often emerges between peer leaders toward sustaining their coordinated outreach and engagement in the focal community. On the other hand, too much network closure, particularly if it occurs among peer leaders who are more trepidatious or less effective as peer leaders, could close these individuals off from new perspectives that could increase their confidence in the role. For this reason, supporting bonding social capital should not come at the cost of nurturing bridging capital, especially when it brings together peer leaders with different skill sets and different levels of confidence.Finally, it is also worth discussing what we did not find. Namely, the increase in PrEP adoption between baseline and 12-month observation points in both intervention and control arms was seemingly unrelated to social influence processes as observed on Facebook. In many ways these null results are unsurprising given the nature of the network and the scope of the larger information environment in which the study took place. With respect to the network, peer influence on an individual’s PrEP decision making may have been more likely to occur in the context of more intimate physical world relationships that were unobserved, for example in the context of confidant relations or sexual partnerships. As we were unable to account for the potential influence of online peers who were not in the study, it is also possible that our Facebook network may have been missing some of its more influential actors and ties. This study should be interpreted in the context of several limitations. First, previous research has shown that offline and online relationships, particularly Facebook friendships, have a tendency to overlap [64,65]. However, the effort required to form online connections and, therefore, the meaning of those relationships make them notably different from offline relationships. Precisely how community leaders’ online relationships and the communication that occurs within them contribute to community capacity is an open question that requires more attention. Second, our singular focus on the network dynamics among peer leaders presents only a partial picture of the relational infrastructure from which community capacity is built. Although it was beyond the scope of this study, future research should be directed at understanding the effect of peer leader interventions such as PrEP Chicago on the formation of relationships between newly activated peer leaders and members of the larger community to which they belong (e.g., YBMSM peers, community leaders and organizations). Third, so that we could effectively compare how study participation differentially influenced the network and behavioral dynamics of intervention and control arm participants, we chose to treat participants in each arm as two mutually exclusive sub-groups. However, this forced us to remain agnostic to the fact that online friendships also existed across groups. Whether and how the intervention influenced the network and behavior dynamics between conditions is a question that needs further exploration and which has implications for our understanding of the intervention’s impact.Despite these limitations, this is the first study to our knowledge that fully considers the effect of a peer leader intervention on the co-evolving relationships and behaviors of the peer leaders themselves and to unpack those dynamics in terms of their implications for community capacity-building. To these ends, we applied novel longitudinal social network models to determine the relationship between the online friendship networks of a cohort of PrEP peer leaders and their personal PrEP adoption behaviors, followed by a comparison analysis of the same dynamics among a cohort of control participants. Peer leaders actively shaped their online social environment by forming friendships with other members of their training cohort and who were similar in age and HIV status. Although PrEP adoption did not motivate the formation of new friendship ties, having greater PrEP intentions did. In comparison, online friendships among control participants were unaffected by their co-participation in a training session, their HIV status, PrEP adoption, and PrEP intentions. Although our findings are in part specific to the PrEP Chicago intervention, our goal was to articulate and apply a joint theoretical and analytic framework that help us see and evaluate peer leadership interventions and their social and behavioral effect on peer leaders as critical ingredients of longer-term capacity-building efforts in communities that are undergoing social and behavioral change. Conceptualization, L.E.Y. and J.A.S.; Methodology, L.E.Y.; Formal Analysis, L.E.Y.; Writing—Original Draft Preparation, L.E.Y.; Writing—Review & Editing, J.A.S. and L.E.Y.; Visualization, L.E.Y.; Funding Acquisition, J.A.S. and L.E.Y. All authors have read and agreed to the published version of the manuscript.This work was supported by NIH grants R01AI20700 and R00HD094648.This study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Boards of the University of Chicago School of Medicine, Biological Sciences Division and the National Opinion Research Center (NORC) at the University of Chicago.Informed consent was obtained from all participants included in the study.The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethics requirements.This study was conducted under the auspices of the PrEP Chicago study team. We would like to thank intervention staff and our partners at the National Opinion Research Center (NORC) at the University of Chicago for their invaluable support. We also would like to thank study participants for their time and commitment to the study.The authors declare no conflict of interest. SABMs and their variants are actor-oriented models, which means they conceive the co-evolution of networks and behaviors as happening at the hands of individual actors who make decisions to change their relationships or their behaviors in the name of optimizing their position in the network [37]. Changes between each wave of observed data are modeled using continuous-time Markov chains to determine the most likely series of unobserved micro-steps taken by actors when changing their network ties or their behaviors. As a continuous-time Markov chain process, the micro-changes to network ties and behaviors are assumed to occur sequentially. Further, the model assumes that actors can change a network tie or change their behavior (or make no change), and are assumed to react to changes made by other network actors [36].Social and behavioral changes that occur between each wave of observed data are captured in two components of the model: the rate function captures the speed by which the dependent network or dependent behavior changes; and the evaluation function determines the “rules” that motivate these changes [36]. These rules are parameterized in the model as model effects, the estimates for which allow us to infer which rules were most likely to guide the unobserved micro-steps that led to larger observed changes to the dependent variables [36]. These effects can be purely structural, whereby individual changes are made in response to the presence or absence of ties around them, or attribute-based.To determine whether the selected models for the intervention and control arm samples provide a good fit to the observed network dynamics, we assess the fit of each model with respect to degree distribution and geodesic distribution. In RSiena, the goodness of fit (GOF) function operates by comparing the observed values at the end of a period with the simulated values for the end of a period. The differences are assessed by combining the auxiliary statistics using the Mahalanobis distance [66]. A model is considered an acceptable fit to a particular auxiliary statistic if the p-value of the Mahalanobis distance is larger than the conventional threshold of α = 0.05. Goodness of fit statistics for the (A) intervention arm and (B) control arm models.The Facebook friendship networks among intervention (n = 174) and control arm (n = 166) participants at baseline (T1) and 12-months (T2), with information about PrEP adoption status. Each circle (node) represents one study participant in either the intervention or control arm sub-samples. Circles are colored by their PrEP adoption status at each wave: grey denotes a study participant who was not a PrEP adopter, dark blue denotes a study participant who reported being on PrEP at baseline (T1), and aqua blue denotes a study participant who adopted PrEP at 12-months (T2). Visualizations were created in Python.Description of the effects included in the network and adoption process sub-models.Characteristics of YBMSM study participants, stratified by intervention and control group assignment in Year 1.Structural properties of the intervention and control arm Facebook friendship networks at baseline and 12-months.a The Jaccard index measures the amount of change between observed waves, and indicates whether the data collection points are not too far apart. Values greater than 0.3 are desired to meet assumptions that the network change process is gradual [36].Significance of parameter estimates of the Facebook network and PrEP adoption process sub-models.Note: Convergence t-ratios < 0.07 and 0.05 and overall maximum convergence ratio = 0.17 and 0.20 for intervention and control arm models, respectively. † p < 0.10, two-tailed; * p < 0.05, two-tailed; ** p < 0.01, two-tailed; *** p < 0.001, two-tailed.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06052.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Natural resource extraction projects are often accompanied by complex environmental and social-ecological changes. In this paper, we evaluated the association between commodity extraction and the incidence of diseases. We retrieved council (district)-level outpatient data from all public and private health facilities from the District Health Information System (DHIS2). We combined this information with population data from the 2012 national population census and a geocoded list of resource extraction projects from the Geological Survey of Tanzania (GST). We used Poisson regression with random effects and cluster-robust standard errors to estimate the district-level associations between the presence of three types of commodity extraction (metals, gemstone, and construction materials) and the total number of patients in each disease category in each year. Metal extraction was associated with reduced incidence of several diseases, including chronic diseases (IRR = 0.61, CI: 0.47–0.80), mental health disorders (IRR = 0.66, CI: 0.47–0.92), and undernutrition (IRR = 0.69, CI: 0.55–0.88). Extraction of construction materials was associated with an increased incidence of chronic diseases (IRR = 1.47, CI: 1.15–1.87). This study found that the presence of natural resources commodity extraction is significantly associated with changes in disease-specific patient volumes reported in Tanzania’s DHIS2. These associations differed substantially between commodities, with the most protective effects shown from metal extraction.Implementation of natural resource extraction projects often triggers a series of complex environmental and social–ecological changes [1,2,3,4]. These changes include increased population growth and urbanization, infrastructure improvements, movement and installation of heavy machinery, changes in land use, increased business and economic opportunities, and household resettlement [5,6,7]. Such changes can positively or negatively affect the health status of communities in proximity of resource extraction activities [8,9,10,11,12]. Studies have linked activities of the resource extraction with increased incidence of respiratory diseases [13,14]; sexually transmitted infections, including HIV/AIDS [15]; malnutrition [16]; vector-related diseases [17,18,19]; mental health [20,21]; and cancer diseases [9]. On the other hand, resource extraction projects can positively contribute to population health through improved labor market opportunities and corporate social responsibility (CSR) activities that support local health systems or contribute to general health and education programs [22,23,24,25].Even though a growing literature has highlighted the importance of comprehensive health impact assessments (HIA) of mining projects [26,27,28], evidence on the impact of large-scale mining projects on population health remains limited. One of the key challenges for such projects is the lack of reliable data sources for local population health [4,29,30]. Due to major international efforts, such population-level health data are increasingly becoming available at the health facility level in low- and middle-income countries, often supported by new digital systems. DHIS2 (District Health Information Systems) is one such web-based software application that can be integrated into the national health management information system (HMIS) to facilitate data collection, data use, data management, and archiving of routine data [31]. DHIS2 has been installed in over 73 countries worldwide [32], allowing researchers and policy makers to collect and aggregate data across the health system [33,34,35].In this study, we used the data from Tanzania’s DHIS2 to evaluate the association between the presence of different types of natural resource extraction projects and disease incidence at the district level.Tanzania has an estimated total population of 57 million people [36] and ranks 163 out of 189 countries and territories with a human development index value of 0.529 [37]. Tanzania has a decentralized health system with administrative units in the mainland area extending to 26 regions, 184 districts and municipal councils, and 8941 health facilities (including both private and public health facilities). According to the Institute of Health Metrics’ (IHME) Global Burden of Diseases estimates, the top three Disability Adjusted Life Years (DALYs) per 100,000 population in Tanzania between 2015 and 2019 were consistently shown as (1) neonatal disorders, (2) lower respiratory infections, and (3) HIV/AIDS [38].Tanzania has a long history of resource extraction activities [39] and was one of the earliest DHIS2 implementers after the platform was developed in 2007 [40,41].This is a multi-year cross-sectional study that aims to assess the associations between the presence of district-level commodity extraction and disease diagnoses reported in Tanzania’s DHIS2.Disease diagnoses data come from the DHIS2 outpatient department (OPD) dataset indicators and one additional indicator (i.e., total clients tested and found positive for HIV) from the HIV testing and counselling dataset. As part of national HMIS, disease diagnoses made during health facility visits are captured and summarized in HMIS books, and summary information is entered into the DHIS2. We extracted annual aggregated data at the district level covering the period from 2015 to 2019. The data contain counts of disease diagnoses from health facility visits (both public and private). We excluded laboratories, maternity homes, and all facilities that do not provide general OPD services. We excluded the region of Dar-es-Salaam (with 5 administrative districts) due to the high density of population and health facilities and the absence of large-scale resource extraction projects. We merged data from the Ifakara town district and Kilombero District to accommodate available shapefiles in the geographical information system (GIS), whereby Ifakara town District was formally part of Kilombero District until 2016. Our final dataset contained a total of 178 districts.Data on natural resource extraction activities were accessed via the mineral occurrence map of Tanzania [42], which was based on the Geological Survey of Tanzania (GST) conducted in 2015 [43]. The GST lists a total of 480 resource extraction projects with location coordinates and commodity types. In this list, 42% of the projects are marked prospective, 20% active, and 38% inactive. We restricted our analysis to natural resource extraction projects that were classified as active in the GST as of 2015 (at the beginning of our sample period).The primary exposure variable of interest was the presence of a given commodity extraction in the district. We considered three types of commodity extraction: (1) construction materials, (2) gemstone, and (3) metals. As shown in Figure 1, most districts contain only one type of extraction project; multiple types of extraction projects were found in only seven districts. Given that coal was identified in only one district and hydrocarbons are mostly located offshore, these two groups were excluded from our analysis. A complete list of the specific commodities extracted is available in Appendix A (Table A1).We selected and grouped disease indicators based on the environmental health areas (EHA) framework defined in the HIA guidelines developed by the International Finance Corporation (IFC) [44]. The EHA framework provides a conceptual linkage between resource extraction activities and potential community-level impacts, incorporating a variety of biomedical and key social determinants of health [45]. For this study, we divided all reported health programs in the DHIS2 into disease groups based on the EHA framework. These groups, as well as the OPD indicators falling into each group, are summarized in Table 1.To assess the association between the presence of commodity extraction projects and disease incidence, we used Poisson regression models controlling for period (in years), population, region fixed effects, and number of health facilities by time in the district. We assumed all commodity extraction projects to remain active over the full sample period. We employed cluster-robust standard errors to correct for residual correlation at the district level over time as well as over dispersion. We report incidence rate ratios (IRR) and corresponding 95% confidence intervals (CI) from the final models. All analyses were implemented using the STATA 15.0 statistical software package [46].This study obtained ethical approval from Ifakara Health Institute Review Board and the National Institute for Medical Research (NIMR) in Tanzania; the Ethics Committee of Northwestern and Central Switzerland (Ethikkommission Nordwest- und Zentralschweiz, EKNZ); and the institutional review board of the Swiss Tropical and Public Health Institute (Swiss TPH) in Switzerland. A summary overview of the number and type of health facilities and commodity extraction projects in districts in Tanzania is shown in Table 2. On average, a district contained 1 hospital, 5 health centers, 36 dispensaries, and 1 health clinic. A total of 2 districts were exposed to all 3 types of commodity extraction, 5 districts were exposed to 2 types of commodity extraction, and 38 districts exposed to only 1 type of commodity extraction.Extraction of construction materials occurred in 14% (N = 25) of all districts, metals in 12% (N = 22), and gemstone in 4% (N = 7). A total of 1% (N = 2) of the districts had all of the three types of commodity extraction, and 4% (N = 7) had at least two types of commodity extractions. A total of 45 districts (25%) were exposed to at least one type of commodity extraction. The distribution of health facilities by type of extractive activity is similar across the three types of commodity extraction (Table A2 in the Appendix A).Table 3 summarizes the annual number of disease diagnoses per 100,000 inhabitants for different groups of diseases. Out of the main disease categories analyzed, the most common were respiratory infections and malaria, with an average of 52,704 and 33,981 cases across the five years, respectively. The least commonly diagnosed health problems were cancer and tuberculosis, with an average of 127 and 266 cases per year, respectively.Table 4 shows the association between the type of commodity extraction and disease groups. In the fully adjusted models, we found that the presence of construction material extraction was associated with an increased incidence of chronic diseases (IRR = 1.4795% CI = 1.15–1.87). The presence of metals extraction was associated with a reduced incidence of chronic diseases (0.61, 0.47–0.80), mental health disorders (0.66, 0.47–0.92), undernutrition (0.69, 0.55–0.88), parasitic infections (0.84, 0.72–0.98), sexually transmitted diseases (0.85, 0.74–0.97), and diarrhea diseases (0.88, 0.77–0.10). No association was found between gemstone extraction and any of the main disease categories. The outputs of the crude model are available in Table A3.Table 5 shows the results for disaggregated disease categories. Extraction of construction materials was associated with an increased incidence of hypertension (IRR = 1.49, CI: 1.11–2.01), neoplasms/cancer diseases (IRR = 1.46, CI: 1.07–1.99), bronchitis asthma (1.38, 1.14–1.66), and other cardiovascular diseases (1.72, 1.27–2.32). Gemstone extraction was associated with a significant increase in the incidence of malaria (blood slide microscopy method) (2.08, CI: 1.06–4.07), non-severe pneumonia (1.40, 1.04–1.87), diarrhea with some dehydration (IRR: 1.38, 1.11–1.73), and severe pneumonia (1.25, 1.01–1.54). Protective effects were found for metal extraction projects in the majority of disease-specific indicators.In this paper, we evaluated the relationship between the district-level presence of natural resource extraction activities and the disease-specific patient volumes reported across all health facilities in Tanzania in a given district and year.Our results show that the districts’ health infrastructure across the three types of commodity extraction is relatively the same (Table A3). In addition, the presence of commodity extraction projects is associated with significant changes in the incidence of diseases reported in the routine national HMIS of Tanzania. The changes in disease incidence appear to differ substantially by the type of commodity extracted; in districts where metal resources are extracted (most typically Gold in Tanzania), we observed a significant decrease in the incidence of chronic diseases (−39%), mental health disorders (−34%), undernutrition (−31%), parasitic infections (−26%), sexually transmitted diseases (−15%), and diarrheal diseases (−13%). On the other hand, the presence of construction material extractions was associated with a significant increase in the incidence of chronic diseases (+47%), including hypertension (+49%), neoplasms/cancer (+46%), bronchitis asthma (+38%), and other cardiovascular diseases (+72%). Gemstone extraction was not associated with any of the aggregated disease categories but showed positive associations with severe and non-severe pneumonia (+40% and +25%, respectively), as well as diarrhea with dehydration (+38%).The consistent inverse association between the metal extraction industry and disease incidence was rather striking and could be due to several factors. The observed decreased incidence of chronic diseases can partially be explained by the in-migration of young and healthy mining workers and other job seekers [47,48,49], who are at lower risk for chronic diseases, such as hypertension and diabetes mellitus [50]. In addition, the majority of large-scale metal industries in Tanzania are led by transnational corporations [51], which are more likely to adhere to international environmental and social standards (e.g., Performance Standards of the International Finance Corporation (IFC) [52]) and recommended industry practices (e.g., International Council on Mining and Metals (ICMM) [53] and Extractive Industries Transparency Initiative (EITI) [54]). Hence, it seems plausible that metal extraction projects are, on average, more able to support local health systems and community-level interventions [55] or socio-economic development programs [56]. This may also explain why, in contrast to previous studies that reported elevated levels of sexually transmitted diseases in regions where metals are extracted [15,18,57,58], our results show that districts that are exposed to metal extractions activities generally have lower incidences of sexually transmitted diseases, including HIV/AIDS, even though these differences were not statistically significant.It is also important to highlight that our study covered the period 2015–2019, which followed two major national-level government interventions. The first intervention was the Tanzania Mineral Policy 2009 Revision, which intended to increase the mineral sector contribution to income generation [59], and the second was the re-enactment of the Mining Act 2010, which requires mining companies to have a plan to increase Tanzania’s citizens’ participation in mining activities through employment and expatriate succession plans, as well as a plan for procurement of goods and services through the available local market in the United Republic of Tanzania [60]. It is likely that these measures had a particularly positive effect on the districts with the presence of metal extractions activities, which are, in general, larger operations than construction and gemstone commodity extraction projects.In contrast to metal extraction, increased levels of disease incidence were observed for districts with construction material or gemstone extraction activities. Extraction of construction materials covers commodities such as cement, silicate, carbonate hard rocks, decorative stones (i.e., dolomite and limestone), and industrial minerals (i.e., gypsum, clay, halite, phosphate, and mica). The observed increase in the incidence of chronic diseases appears consistent with the risk factors for such projects described by IFC, namely, high exposures to dust, vibration, noise, and unhealthy lifestyle (related to cigarette smoking and excessive alcohol intake) [61]. Studies have also shown that exposure to particulate matter (PM) from construction dust can significantly influence the appearance of diseases such as cancer and cardiovascular and respiratory diseases [62,63]. Studies from the construction and building materials literature report that there is a growing demand for construction material commodities, and they equally draw attention to the environmental and health effects this demand imposes on the livelihoods of affected communities [64,65,66]. In our study sample, 40% of the commodities extraction projects were categorized as construction materials, highlighting the need to further scrutinize this sector in terms of environmental and social responsibility, including the mitigation of potential adverse health impacts.To our knowledge, this is the first large-scale study of the relationship between mining and disease incidence in Tanzania using the complete national HMIS database. The dataset covers—by definition—the entire country and allowed us to analyze a range of different mines at the same time. Having five years of data also allowed us to reduce the risk of specific results being driven by misreporting in a given month or year. The study also has several limitations. First, DHIS2 and routine HMIS data have known data quality issues [67], including concerns regarding completeness, consistency, coverage, and disease coding [68,69,70,71]. In our analysis, we did not account for potentially different data quality across districts. If such differences in quality exist, they would only bias our results if reporting was systematically correlated with the presence of mines. This is possible in principle if mines directly contribute to the health system capacity and could potentially bias results against mines. We also did not distinguish between missing data and the true zero values in our analysis, which means that the reported case numbers may underestimate the true burden of disease and patient numbers. Due to the unavailability of good data, we did not account for district per capita income, nor for the number of job seekers who migrated into mining areas. The OPD diagnoses consist of both clinical and non-clinical indicators. Non-clinical indicators can be influenced by physician’s opinion and are subject to reporting bias. As long as these biases are not correlated with the presence of mines, this should not introduce any systematic bias to our analysis. In addition, we used district boundaries to link mining operations to patient volumes. Even though districts are relatively large, it is possible that some of the mines—particularly when located near district boundaries—also affect patient volumes in neighboring districts. This exposure misclassification would likely result in less precise estimates but could also lead to an underestimation of the true causal effects if there are systematic spillovers into districts considered unexposed in our analysis. Lastly, OPD indicators represent counts of people who sought care at health facilities and do not represent actual case numbers in the population.The results of this study suggest that the health impact of mining operations may not only depend on the specific health outcome but also on the types of mines. While metal extraction projects display consistent protective effects against a range of disease indicators, extraction of construction materials and gemstones is associated with increased incidence of chronic and diarrheal diseases, respectively. More research is needed to understand the underlying factors of the differences observed across the types of commodity extracted.Conceptualization: I.L., G.W.L., G.F. and M.S.W.; methodology: I.L., G.L., A.F., G.F. and M.S.W.; formal analysis: I.L.; writing—original draft preparation: I.L. and M.S.W.; writing—review and editing: all. All authors have read and agreed to the published version of the manuscript.This research received funding from the Swiss Programme for Research on Global Issues for Development (r4d Programme, www.r4d.ch (accessed on 10 February 2021)), which is a joint funding initiative of the Swiss Agency for Development and Cooperation (SDC) and the Swiss National Science Foundation (SNSF) (grant number 169461).The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the Ifakara Health Institute (No: 32-2018, Date: 28 September 2018), Tanzania National Institute for Medical Research — NIMR (ID: NIMR/HQ/R.8a/Vol.IX/2969, Date: 3 December 2018) and the Ethics Committee of the Northwest and Central Switzerland — EKNZ (EKNZ Req-2018-00386, Date: 23 May 2018). Not applicable.The data presented in this study are available on request from the Ministry of Health Community Development, Gender, Elderly and Children (MoHCDGEC). The data are not publicly available due to local restrictions. The authors thank the Tanzanian Ministry of Health, Community Development, Gender, Elderly and Children who allowed the use of DHIS2 data from Tanzania.The authors declare that they have no conflict of interest. The funders had no role in the design of the study, data collection, analyzes and findings´ interpretation, the writing of the manuscript, or in the decision to publish the results.List of commodities extracted in Tanzania.1 excluded in the analysis due to limited exposure.Distribution of health facilities by type of commodity extraction.Crude model of relationship between disease groups and type of commodity extraction.Note: estimates are adjusted for cluster robust; *** p < 0.01, ** p < 0.05, * p < 0.1.Location of selected commodity extraction projects in Tanzania.List of disease groups and included outpatient department (OPD) indicators from DHIS2.1 This indicator is obtained from HIV Counseling and Testing (HTC) unit. Notes: HIV stands for human immunodeficiency virus, mRDT stands for malaria rapid diagnostic test.Overview of the number and type of health facilities and commodity extraction projects per district in Tanzania.Total number of districts = 178.Number of disease diagnoses per 100,000 population by year.1 Includes all other diagnoses which are not featured in any of the disease groups used for this study.Relationship (adjusted model) between disease groups and type of commodity extraction.Note: estimates are adjusted for cluster robust; *** p < 0.01, ** p < 0.05, * p < 0.1.Relationship (adjusted model) between disease-specific indicators and type of commodity extraction.Note: estimates are adjusted for cluster robust; *** p < 0.01, ** p < 0.05, * p < 0.1; mRDT stands for malaria rapid diagnostic test, HIV stands for human immunodeficiency virus.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06053.txt ADDED Viewed

The diff for this file is too large to render. See raw diff

Med-MDPI/ijerph_7/ijerph-18-11-06054.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Equal Contribution.Membership of the Mental Health Group of the Fukushima Health Management Survey is provided in the Acknowledgments.Oral health is closely related to subjective general health and systemic diseases. This cross-sectional study aimed to identify the factors related to oral symptoms and their worsening in relation to psychosocial factors after the Great East Japan Earthquake. In this study, 64,186 residents aged 15–101 years old, who experienced the earthquake on 11 March 2011, were surveyed regarding their oral symptoms; psychological factors, such as post-traumatic reactions and psychological distress; and social factors such as evacuation, work change, and loss of a close person; history of systemic diseases; and lifestyle. Binomial logistic regression analysis was used to calculate odds ratios, and 95% confidence intervals were established for each factor associated with prevalent and exacerbated oral symptoms. The proportions of participants with prevalent and exacerbated oral symptoms were 10.3% and 1.6%, respectively. The multivariate odds ratios and 95% CI of psychosocial factors associated with exacerbated oral symptoms were as follows: post-traumatic stress disorder symptoms, 2.24 (1.64–3.06); work changes, 1.88 (1.34–2.65); history of dyslipidemia, 1.74 (1.27–2.39); and subjective current poor health condition, 2.73 (2.00–3.75). Psychological factors, social factors, and physical factors were associated with both prevalent and exacerbated oral symptoms.Disasters have a great impact on the lives and health of evacuees. Residents of Fukushima Prefecture were forced to evacuate for an extended time following the Great East Japan Earthquake on 11 March 2011, and the subsequent tsunami and nuclear accident at the Fukushima Daiichi Nuclear Power Plant. These disasters have hindered the evacuees not only at the time of the disaster but also afterward by affecting both their physical [1,2] and mental conditions [3,4,5]. Six to eleven months after the Great East Japan Earthquake, 42.6% of the evacuees had moderate or severe mental health problems [6]. Furthermore, the mental health problems and trauma reactions of the evacuees continued for several years after the disaster [7], and problems were manifested in the body, particularly for the evacuees and unemployed people [8].The oral cavity is involved in eating and communication. Oral health has been associated with subjective general health [9,10], and periodontal disease has been associated with systemic diseases such as diabetes, dyslipidemia, stroke, heart disease, and pneumonia [11]. Maintenance of good oral health is important because it prevents oral symptoms from worsening, improves subjective general health, and might prevent dyslipidemia and diabetes from worsening. The direct oral health problems observed immediately after a disaster include deterioration of oral hygiene caused by poor cleaning related to lack of water, as well as deterioration of oral functions due to loss of dentures [12]. Subsequent indirect problems include dental caries and periodontal disease related to unbalanced diets and the stress of prolonged evacuation. In a previous study in Otsuchi-cho, Iwate, Japan, Kishi et al. reported that during the Great East Japan Earthquake [9], age, loss or damage of dentures, having fewer than 20 teeth, untreated teeth, and tooth mobility were factors associated with lower oral-health-related quality of life (OHRQOL). In addition, oral symptoms, including toothaches, have been found to be related not only to actual caries and periodontal disease but also to various other factors such as psychological problems, environment, and lifestyle [13,14]. People who lived in temporary housing after the Great East Japan Earthquake had a significantly higher prevalence of toothache [15].Given the mental distress and decline in oral hygiene following large-scale disasters, as well as the association of subjective general health with oral health, it is very important to understand the relationship between psychosocial factors and oral symptoms following disasters such as the Great East Japan Earthquake. To date, however, there have been few large-scale studies that have focused on this relationship. The purpose of this study was to identify factors related to oral symptoms, such as toothache and gingival inflammation, and their worsening in terms of mental health factors such as disaster-related psychosocial and lifestyle factors, and the general medical history of residents of the evacuation zone in Fukushima Prefecture after the Great East Japan Earthquake. The findings of this study are likely to be useful for oral and mental health considerations following future disasters.In this cross-sectional study, the Mental Health and Lifestyle Survey was conducted as part of the Fukushima Prefecture Health Care Survey among 64,186 participants who were born on or before 1 April 1995, i.e., those who were approximately 16 years old or older at the time of the Great East Japan Earthquake. The details of this survey are described in previous studies [2,16]. Briefly, the survey was conducted by sending self-administered questionnaires to approximately 210,000 people who were registered as residents in evacuation areas at the time of the earthquake and divided into five age groups (0–3 years old, 4–6 years old, elementary school students, junior high school students, and people approximately 16 years old or older) according to their current mental and physical health, lifestyle, and living conditions. The survey was conducted from 18 January 2012, to 31 October 2012, approximately 10 months after the earthquake occurred on 11 March 2011. A total of 73,431 out of 180,604 people aged approximately 16 years and above responded (response rate, 40.7%) [17], of which proxy responses (n = 9245) were excluded. Of the respondents, 64,186 evacuees aged 15–101 years (mean age: 55.2 years) were included in the analysis (Figure 1).This study was conducted in accordance with the provisions of the Declaration of Helsinki. The study protocol was approved by the Fukushima Medical University Ethics Committee (approval number 1316), and a questionnaire was mailed to the participants stating the purpose of the study and that it would be used for analysis. By returning the questionnaire, participants were considered to have given their written consent to participate.All the participants were asked to answer the multiple-choice question (Q5 in Supplementary Material S1), “Have you felt unwell in the past several days due to illness or injury etc.?” If they answered yes, then we asked, “What kind of symptoms have you had? Please circle all applicable symptoms and draw a double circle around any symptoms that have worsened since the disaster.” Thus, answer choices were “none,” “yes,” and “yes + worsened since the disaster” for the symptoms of toothache or swollen/bleeding gums. If the participant circled for any of the oral symptoms, we assessed that the participant had oral symptoms. If the participant drew a double circle “worsened since the disaster” to any of the questions on oral symptoms, we assessed that the participant had worsened oral symptoms. Thus, the participants were classified into three groups: without oral symptoms (absent), with oral symptoms (prevalent), and worsened oral symptoms (exacerbated). The questionnaire for this study is available in Supplementary Material S1.Psychological factors that may influence oral symptoms were evaluated as psychological distress (K6: Kessler Psychological Distress Scale) [18,19] and post-traumatic stress disorder (PTSD: PCL-S; PTSD Check List Stressor-Specific Version) [20,21].The K6 scale was used to assess psychological distress [18]. This scale asks participants whether they have experienced any of the following six feelings during the past 30 days: “nervousness,” “hopeless,” “restless or inability to relax,” “inescapable sadness,” “worthless,” and “everything requires painstaking effort.” Each question was scored on a five-point Likert-type scale from 0 to 4. The scores ranged from 0 to 24, and higher scores indicated worse psychological distress. The Japanese K6 version has been validated and was used in this study [19]. Psychological distress was defined as K6 scores ≥13.The PCL-S [20] was used to assess PTSD caused by the experience of the Great East Japan Earthquake and the following accidents. The PCL-S is a 17-item self-report measure designed to detect PTSD, where each item is scored from 1 to 5, corresponding to “not at all,” “slightly,” “moderately,” “quite a lot,” or “very much,” respectively. The scores ranged from 17 to 85, and higher scores indicated worse PTSD symptoms The Japanese PCL-S version has been validated and was used in this study [21]. The presence of PTSD symptoms was defined as a PCL-S score ≥44.Information was also collected on potential confounders such as the experiences of evacuation, earthquake, tsunami, or hearing the sound of explosions from the nuclear power plant accident, job loss, work changes since the disaster, and loss of a close person (no/yes), and house damage (major/minor or none). Experienced evacuees were defined those whose addresses at the time of the disaster were in nine towns or villages (Hirono Town, Naraha Town, Tomioka Town, Kawauchi Village, Okuma Town, Futaba Town, Namie Town, Katsurao Village, and Iitate Village) or residents of four cities or town (Tamura City, Minamisoma City, Kawamata Town, and Date City) who lived in shelters or temporary housing after the earthquake. The question on house damage was scored on a five-point scale from 1 to 5, defined as follows: 1 = no damage, 2 = partial damage, 3 = partial collapse, 4 = partial but extensive collapse, and 5 = total collapse. If the participant answered 4 or 5, we assessed the damage as “major”; otherwise as “minor or none.”Information on history of mental illness, hypertension, diabetes mellitus, dyslipidemia (no/yes), and current health condition (good/normal/bad) was collected. Each participant’s current health condition was assessed as good if they answered “good” or “normal” and as poor otherwise.Results were collected on the following lifestyle factors: current smoking habit (no/yes), current drinking habit (no/yes), regular exercise habit (once a week or less/twice a week or more). Information on potential confounders, such as sex (men/women) and age category (49 years or younger/50–69 years/70 years or older), was also collected.The mean ages for the sex and age category and the proportions of nominal scale data were calculated. Two outcomes were considered: (i) prevalent oral symptoms (yes/no) was defined as endorsement of at least one oral symptom and did not include any oral symptoms also endorsed as “worsened since the disaster”; and (ii) exacerbated oral symptoms (yes/no) was defined as endorsement of at least one oral symptom as “worsened since the disaster.” Logistic regression models of each outcome were used to calculate odds ratios and their 95% confidence intervals (CIs) for the following factors: sex; age category; PTSD symptoms; experiences of evacuation, tsunami, hearing nuclear explosions, house damage, job loss, work change, and loss of a close person; histories of mental illness, dyslipidemia, hypertension, and diabetes mellitus; current health condition; smoking and exercise habits. Two models were used. Model 1 was run for each factor individually, while controlling for sex and age category. Model 2 had sex and age category as factors and all factors found to be statistically significant in Model 1; those factors not reaching statistical significance in Model 1 were not included in Model 2. The results of the multivariate adjustment analysis were presented using Model 2. As multicollinearity was found between the PTSD symptoms (from the PCL-S) and psychological distress (from the K6 scale), and the presence of PTSD symptoms was associated more with oral symptoms than with psychological distress, we presented the results of the multivariate adjustment analysis using the PTSD symptoms alone. We conducted stratified analysis by sex and age category (age 49 years or younger, age 50–69 years, age 70 years or older) with reference to previous literature [9]. SAS statistical software version 9.4 (SAS Institute Inc., Cary, NC, USA) was used for analyzing the data. P-values of less than 0.05 were considered significant.The mean age of the participants was 55.2 years. The proportions of the participants with the measured factors were as follows: PTSD symptoms, 21.5%; psychological distress, 14.6%; experiences of disaster (evacuation 46.8%, tsunami, 20.3%; hearing nuclear explosions, 52.6%; house damage—half or more destroyed, 15.9%; job loss, 20.9%; work change, 55.7%; loss of a close person, 20.0%); history of systemic diseases (mental illness, 5.0%; hypertension, 41.2%; diabetes mellitus, 19.5%; and dyslipidemia, 35.8%); current poor health condition, 18.5%; and current lifestyle (smoking habit, 21.1%; drinking habit, 46.1%; and exercise habit—once a week or less, 64.7%). The characteristics of the participants and their oral symptoms according to the sex and age category are shown in Table 1. The proportion of participants with oral symptoms (prevalent) was 10.3% (by sex: men, 9.9% and women, 10.6%; by age category: 49 years or younger, 8.8%; 50–69 years, 11.4%; and 70 years or older, 10.6%) and was slightly higher for women and those aged 50–69 years. The proportion of participants with exacerbated oral symptoms was 1.6% (by sex: men, 1.4% and women, 1.8%; by age category: 49 years or younger, 1.8%; 50–69 years, 1.7%; and 70 years or older, 1.1%) and tended to be higher among those aged 69 years or younger.For prevalent and for exacerbated oral symptoms, Table 2 shows the odds ratios and 95% CIs for factors considered in Model 1 and in Model 2. Each group of prevalent and exacerbated oral symptoms was multivariable-adjusted (Model 1 and Model 2) using a logistic regression analysis, and each reference had no oral symptoms (absent).After adjustment for sex and age category (Model 1), the following factors were significantly associated with prevalent oral symptoms: psychological factor of PTSD symptoms; experience of disaster, including work change, hearing nuclear explosions, loss of a close person, job loss, evacuation, tsunami, and house damage; histories of systemic diseases, including mental illness, dyslipidemia, diabetes mellitus, and hypertension; current poor health condition; current lifestyle, including exercise habit (once a week or less) and smoking habit; and being female. After the multivariate adjustment (Model 2), PTSD symptoms; experiences of disaster, including work change, hearing nuclear explosions, and loss of a close person; histories of mental illness, dyslipidemia, and hypertension; and current poor health condition were significantly associated with prevalent oral symptoms.After adjustment for sex and age category (Model 1), the following factors were significantly associated with exacerbated oral symptoms: psychological factor and PTSD symptoms; experiences of disaster, including work change, loss of a close person, job loss, evacuation, hearing nuclear explosions, house damage, and tsunami; histories of systemic diseases, including mental illness, dyslipidemia, and hypertension; current poor health condition; current lifestyle, including smoking and exercise habits (once a week or less); and being female. After the multivariate adjustment (Model 2), PTSD symptoms, work change, history of dyslipidemia, current poor health condition, and age categories (49 years and younger and 50–69 years) were significantly associated with exacerbated oral symptoms.For prevalent and for exacerbated oral symptoms, Table 3 shows the multivariate (Model 2) odds ratios and 95% CIs of the prevalent and exacerbated oral symptoms in the participants who fit for each factor as compared with those who did not fit each factor, obtained using a logistic regression analysis according to sex and age category. As a sensitivity analysis, the same analysis was performed by replacing PTSD symptoms with K6, and almost similar results were obtained.In both men and women, significant association with the prevalent oral symptoms was found as follows: having PTSD symptoms, work change, histories of mental illness, dyslipidemia, and current poor health condition. There were significant associations between the experience of evacuation among men and the loss of a close person among women. No statistical evidence showed that tsunami experience, house damage, histories of hypertension and diabetes mellitus, and current smoking habit were related to the prevalent oral symptoms in both men and women.Similarly, significant associations with worsened oral symptoms (exacerbated) were found in both men and women as follows: having PTSD symptoms, work change, history of dyslipidemia, and current poor health condition. No statistical evidence supported that experiences of evacuation, tsunami, job loss, and house damage; histories of mental illness, hypertension, and diabetes mellitus; and current exercise habit were related to the exacerbated oral symptoms for both men and women.In the age categories, significant associations with the presence (prevalent) of oral symptoms were found among all the age categories as follows: having PTSD symptoms, history of dyslipidemia, and current poor health condition. For ages 50 years and older, history of mental illness was associated with the prevalent oral symptoms. For ages 69 years and younger, work change was associated with the prevalent oral symptoms.In the age categories, significant associations were found between exacerbated oral symptoms for all age categories and PTSD symptoms and current poor health. For 69 years and younger, having work change, loss of a close person, and a history of dyslipidemia were associated with the exacerbated oral symptoms. For each of the age categories, there was no statistical evidence that exacerbated oral symptoms were related to experiences of tsunami, hearing nuclear explosions, and job loss; histories of mental illness and diabetes mellitus; and current smoking and exercise habits.This cross-sectional study of residents of the evacuation zone in Fukushima Prefecture after the Great East Japan Earthquake found an association between disaster-related psychosocial factors and oral symptoms. In particular, the factors associated with exacerbated oral symptoms were having PTSD symptoms; experience of evacuation, hearing nuclear explosions, work changes, loss of a close person; history of dyslipidemia; and subjective current poor health condition. The associations of PTSD symptoms and current poor health conditions with the exacerbated oral symptoms were particularly pronounced, even when stratified by sex and age category. Even 10 to 19 months after the earthquake in this study, the effects of the disaster were still seen in many cases, suggesting that these factors may have psychologically influenced the worsening of oral symptoms.In this study, the participants with PTSD symptoms had a higher prevalent and exacerbated oral symptoms than those without PTSD symptoms, and the association between PTSD symptoms and exacerbated oral symptoms was particularly strong. As no studies have reported the associations between psychological factors and oral symptoms after a large-scale disaster, the results of this study were difficult to compare with those of previous studies. However, a well-known fact is that stress influences the worsening of oral symptoms [14,22,23]. Chronic stress has been reported to cause high glucocorticoid levels in the blood, suppressed immunity [24], and a lower threshold for chronic pain perception [25]. In patients with stress reactions, periodontal tissue pain has also been reported to be caused by the application of excessive force to the teeth, such as clenching and grinding [26]. It is also well-known that psychological trauma and long-term stress can produce symptoms of temporomandibular joint dysfunction (TMD) such as temporomandibular joint (TMJ) noise, pain, and restricted opening. Ferreira et al. reported a bidirectional relationship between PTSD and painful TMD in their review [27]. Although TMJ symptoms were not investigated in this study, the fact that PTSD was associated with oral symptoms suggests that oral pain may have been affected by the presence of TMD. While approximately two-thirds of patients with PTSD recover over time, others may be more severely affected, and their post-traumatic state may persist, affecting their cognitive and behavioral functions for years or longer [28,29]. Therefore, the association between psychological factors and the exacerbated oral symptoms may be stronger than usual owing to the continuation of chronic stress after a large-scale disaster.Social factors can also worsen the oral environment [14,22,23]. Factors contributing to psychosocial stress related to the disaster in this study included evacuation, unemployment, and work changes. Greater psychological distress was observed in those who lived in temporary housing for long periods of time due to evacuation [30]. For example, the use of shelters could also cause psychological stress due to changes in available space and noise from neighbors [30,31]. The fact that evacuation was associated with exacerbated oral symptoms in participants suggests that psychological stress caused by evacuation was likely to worsen oral symptoms. Unemployed people have been reported to be more likely to be affected by psychological factors [32], but after adjustment by Model 2, no relationship with job loss was found in this study. It has been reported that after the earthquake, younger people had more negative attitudes toward work, and those whose work was affected in some way by the earthquake had more negative attitudes toward work [33]. In present study, workers under 70 years of age may have had more factors associated with oral symptoms, as changes in the work environment, personal relationships, and work content at the new evacuation site caused greater stress. The results of this study suggest that social factors such as life changes after the disaster, were potentially associated with psychological stress, and such stress might be associated with exacerbated oral symptoms. However, it is important to note that psychological stress may be related to other lifestyle factors such as sleep, social participation, and diet, although smoking and exercise habits were not associated with worsened oral symptoms after the multivariate adjustment in this study. However, it is important to note that psychological stress may also be related to other lifestyle factors such as sleep, social participation, and diet.Regarding physical factors, the participants with a history of dyslipidemia in this study were more likely to have prevalent and exacerbated oral symptoms than the participants without a history of dyslipidemia, and the association with exacerbated oral symptoms was particularly strong. Dyslipidemia causes abnormalities in immune system cells and wound healing, resulting in increased susceptibility to periodontitis and other infections [34]. In residents of the evacuation zone in Fukushima, lifestyle-related diseases such as obesity and dyslipidemia increased after the earthquake owing to dietary changes and chronic stress [35], which may have aggravated periodontal disease. Dutta-Roy (1994) also reported that dyslipidemia may be associated with the development of many diabetic complications rather than hyperglycemia [36]. In fact, the associations between history of diabetes mellitus and the prevalence and exacerbation of oral symptoms disappeared after adjustment for Model 2 in the present study, which supports the results of previous studies. On the other hand, considering that this study was a cross-sectional study, oral symptoms may have affected the development of dyslipidemia. Recently, dyslipidemia and periodontal disease have been reported to be interrelated [37]. Periodontal disease is associated with increases in the levels of lipopolysaccharides (LPS) and cytokines such as tumor necrosis factor alpha and interleukin 1 [37], and these cytokines may have a negative effect on lipid metabolism [38]. Moreover, periodontal infection (Porphyromonas gingivalis) has been associated with abnormal lipid metabolism, which is associated with the progression of atherosclerosis [39]. In addition, it has been reported that periodontitis induces dyslipidemia [40]. Therefore, the relationship between dyslipidemia and oral symptoms must be clarified by prospective studies in the future.Symptoms such as continuous pain, swelling, and discomfort in the oral cavity affect not only eating but also normal life. The more severe the caries and periodontal diseases, the stronger the subjective symptoms. Furthermore, the weaker the immune system, the lower the resistance to infection and the worse the oral symptoms [41]. Therefore, it is important to maintain oral health in disaster-affected areas where the oral environment is likely to deteriorate. OHRQOL is an index for evaluating oral health based on three factors: “pain and discomfort,” “physical functioning,” and “psychosocial functioning” [42]. There have been many reports on the association between a history of toothache and periodontal disease and OHRQOL [43], as well as the association between subjective general health status and OHRQOL [9,10,44]. In addition, tooth movement and the remaining teeth, which are characteristics of periodontal disease, have been associated with lower OHRQOL, and OHRQOL has been reported to be poor in middle-aged and elderly people in the areas affected by the earthquake [9]. According to results of the 2016 Dental Disease Survey conducted by the Ministry of Health, Labor, and Welfare with 6278 people in Japan, the subjective symptoms of toothache were 15%–20% in those aged 64 years or younger and less than 10% in those aged 65 years or older, while the subjective symptoms of gingivitis were around 15% in those aged 64 years or younger and around 10% in those aged 65 years or older. However, actual oral conditions and subjective symptoms, such as tooth decay and periodontitis, cannot be compared because they progress asymptomatically and there are no studies showing worsening during normal times. In this study, we stratified by age category and found that the middle-aged groups (age 50–69 years) had more factors associated with worsening oral symptoms than other age categories. This is most likely due to the fact that many middle-aged people have many remaining teeth, have periodontal disease, and are more vulnerable to social factors such as work and family. In this study, the proportions of all the participants with oral symptoms or exacerbation were 10.6% in those aged 49 years or younger, 13.1% in those aged 50–69 years, and 11.7% in those aged 70 years or older, which is not consistent with previous reports that people older than 70 years old reported the fewest oral symptoms.The strengths of this study are the following: (1) to the best of our knowledge, this is the first large-scale study to examine oral symptoms after an environmental disaster epidemiologically and in detail in relation to psychosocial factors, and (2) the recruitment and survey were systematically conducted after the disaster. With regard to the deterioration of the oral environment after a large-scale disaster, it is important for dentists to understand that there is a possibility that the cause is not only oral but also related to the health of the whole body. Furthermore, as a preventive measure against systemic diseases, maintaining the oral environment could prevent the worsening of systemic diseases, such as dyslipidemia.There are several limitations in this study. First, the overall response rate was low (40.7%), and there were proxy responses (12.6%), which resulted in a relatively small number of respondents for analysis. Moreover, there was a possibility that some of the non-evacuees might have actually been evacuated. It should be noted that the study may have underestimated the association between psychological factors, including experience of evacuation and oral symptoms, as non-respondents were reported to be more likely than respondents to be employed (p = 0.02), socially isolated (p = 0.047), and to have a high psychological stress response (p = 0.03) [45]. Furthermore, the participants in this study were residents of the evacuation zone and did not represent all residents of Fukushima Prefecture. Second, because we did not conduct actual oral examinations and evaluated the data subjectively, the results might differ from the actual oral condition, such as the presence or absence of dental treatment; the number of teeth; and the presence, damage, or loss of dentures and, therefore, could not be compared with actual oral changes or denture restoration. However, since 70% of the participants who had dentures had them restored by 9 months after the Great East Japan Earthquake, and there was no difference between the people of unrestored and restored dentures [9], and since subjective symptoms of caries and periodontal disease worsen as the severity of the disease increases and the immune system is weakened, we believe that similar results could be expected from a subjective assessment. Third, since this was a cross-sectional study, it is not possible to show causality. In the future, it is necessary to confirm the causal relationship by conducting a longitudinal study after further examining the conditions and survey methods.In this cross-sectional study of Fukushima residents after the Great East Japan Earthquake, it was found that psychological factors such as PTSD symptoms; social factors such as evacuation, work change, and loss of a close person; and physical factors such as current poor health condition and history of dyslipidemia were associated with exacerbated oral symptoms. Since worsening oral function is a factor in the future development of cardiovascular disease and nursing care requirement, it is necessary to take a preventive approach from each of the physical, psychological, and social aspects in order to prevent disease and need for nursing care in the post-disaster population.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116054/s1, Supplementary S1. Fukushima Health Management Survey Mental Health and Lifestyle Survey (for adults) 2011.Conceptualization, methodology, and formal analysis, N.F., A.T., E.E. and T.O.; validation, N.F., F.H., E.E. and T.O.; investigation and data curation, T.O., M.M., H.Y., S.Y. and K.K.; resources, T.O., M.M., H.Y., S.Y. and K.K.; writing and visualization, N.F., A.T., S.T., E.E. and T.O.; supervision, S.T., T.O. and E.E.; project administration and funding acquisition, T.O., M.M., H.Y., S.Y. and K.K. All authors have read and agreed to the published version of the manuscript.This survey was conducted as part of “Fukushima Prefecture’s post-disaster recovery plans” and was supported by the National Health Fund for Children and Adults Affected by the Nuclear Incident for design and conduct of the study.This study protocol was approved by the Fukushima Medical University Ethics Committee (approval no. 1316).A questionnaire was mailed to the participants stating the purpose of the study; by returning the questionnaire, the participants were considered to have given their written consent to participate.The datasets analyzed during the present study are not publicly available because the data from the Fukushima Health Management Survey belongs to the government of Fukushima Prefecture and can only be used within the organization.We would like to express our sincere gratitude to the chairpersons, other expert committee members, advisors, and staff of the Fukushima Health Management Survey Group, and the following members of the Mental Health Group of the Fukushima Health Management Survey: Masaharu Maeda, Atsushi Takahashi, Maho Momoi, Saori Goto, Tetsuya Ohira, Mitsuaki Hosoya, Akira Sakai, Hirooki Yabe, Kanae Takese, Itaru Miura, Hajime Iwasa, Shuntaro Itagaki, Mayumi Harigane, and Naoko Horikoshi. The present study was conducted by the Fukushima Medical University on consignment by the Fukushima Prefecture using the Fukushima prefectural health survey funds. Furthermore, the opinions presented in the report are those of the authors and not of the Fukushima Prefecture residents.The authors declare no conflict of interest.Flowchart of study participants after the Great East Japan Earthquake.Prevalence of oral symptoms and basic background information of 64,186 participants.Data are presented as means with the standard deviations or numbers with the proportions. SD: standard deviation, PTSD: post-traumatic stress disorder.Adjusted odds ratios and 95% CI for prevalent and for exacerbated oral symptoms.1 Model 1: adjusted for sex and age category. 2 Model 2: adjusted for sex and age category + all variables that were significant in Model 1. Bold considers statistically significant. CI: confidence interval, OR: odds ratio, Ref: reference, PTSD: post-traumatic stress disorder.Multiple adjusted odds ratios and 95% CI for prevalent and for exacerbated oral symptoms by sex and age category.1 Adjusted for age category + everything else. 2 Adjusted for sex + everything else. Bold considers statistically significant. CI: confidence interval, OR: odds ratio, Ref: reference, PTSD: post-traumatic stress disorder.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06055.txt ADDED Viewed

	@@ -0,0 +1 @@

+ The aim of our study was to explore whether energy drink consumption is associated with both emotional and behavioural problems and whether this association might be mediated by amount of sleep and breakfast consumption among adolescents. The nationally representative Health Behaviour in School-aged Children (HBSC) study, realised in 2018 in Slovakia in schools, was used to acquire needed data, with the research sample of 8405 adolescents from 11 to 15 years old (mean age = 13.43; 50.9% boys) who completed the questionnaires on their own in a presence of researchers and research assistants. Emotional and behavioural problems were assessed by a Strengths and Difficulties Questionnaire, while energy drinks consumption, breakfast consumption and sleep duration was assessed by questions in line with the HBSC study protocol. Linear regression models assessed the associations between energy drinks consumption and emotional and behavioural problems. Mediation by sleep duration and breakfast consumption was assessed with parallel mediation models. Energy drink consumption was significantly associated with emotional (p < 0.001) and behavioural problems (p < 0.001), with higher consumption of energy drinks leading to more emotional and behavioural problems. Results from a parallel mediation analysis indicated that energy drink consumption is indirectly related to both emotional and behavioural problems through its relationship with the amount of sleep and breakfast consumption. Parents and professionals working with adolescents should be aware that unhealthy dietary habits and lack of sleep might be related to emotional and behavioural problems.Emotional and behavioural problems, which occur mainly in childhood and adolescence, affect approximately 10–25% of child and adolescent populations [1,2] can have long-lasting consequences not only for adolescents but also for their families and society as a whole [3]. It is therefore of great importance to reveal factors associated with the increased likelihood of emotional and behavioural problems.Previous research has confirmed that regular energy drink (hereinafter referred to as EDs) consumption in adolescents might be considered as one of those potential risk factors, as it has been associated with a wide range of emotional and behavioural problems including depression, stress, anxiety, emotional difficulties, self-destructive, violent and risky behavior [4,5,6,7,8,9]. These adverse effects are related mainly to the consumption of a high amount of caffeine, which can exceed 500 mg in some EDs, whereas the safe dose is 200 to 300 mg [10,11]. Caffeine, a main ingredient in EDs, is the most commonly used psychoactive drug in the world, legally and is easily available for children [4], with negative influence on their physical and mental health status [12]. Moreover, the toxic effect of caffeine is potentiated by other ingredients in EDs, such as taurine [13], guarana and additives containing caffeine, including kola nut, yerba mate and cocoa. However, the amount of caffeine in these ingredients is not required to be listed on the ED packaging labels, thus, the actual caffeine amount is often higher than listed [14,15,16]. Consumption of energy drinks is popular among adolescents and has increased in the last 10 years [11,17,18]. Evidence shows that 12 to 35% of adolescents consume energy drinks at least once every week and that consumption is more frequent among males and older adolescents [6,9,19,20].High ED intake is also associated with unhealthy behavior, such as reduced and insufficient sleep and breakfast skipping in adolescents [7,17,21,22,23,24]. Evidence shows that 30 to 50% of adolescents sleep less than recommended 8 h [25,26,27], and the prevalence of sleep problems varies between 10 to 20% in adolescents [28,29,30]. According to the large studies conducted in Europe and other countries [31,32,33,34,35], a high proportion of adolescents were classified as breakfast-skippers, ranging from 40–50% (China, Austria, Slovenia) to 10–20% (Spain, Poland, Japan). Moreover, it has been confirmed that adolescents with insufficient sleep duration and poor sleep quality suffered from more emotional and behavioural problems [31,36,37,38,39]. Some studies have also found an association between breakfast omission and emotional and behavioural health problems, however, the quality and strength of the evidence was weak [7,31,40,41] and research indicates that the quality of breakfast might play an important role in this association [42]. Nevertheless, it remains unclear what role might amount of sleep and breakfast consumption play in the association between ED consumption and emotional and behavioural problems. Our study provides the opportunity to add to the existing knowledge by exploring one of the potential pathways between EDs and emotional and behavioural problems via the amount of sleep and breakfast consumption on a nationally representative sample.Therefore, the aim of our study was to explore whether ED consumption is associated with both emotional and behavioural problems and whether this association might be mediated by the amount of sleep and breakfast consumption among adolescents.To explore our aim, we used data from the nationally representative Health Behaviour in School-aged Children (HBSC) study realised in 2018 in Slovakia that is part of World Health Organisation collaborative, cross-national HBSC study of 11-, 13- and 15-year-old school-aged children from 50 countries and regions across Europe and North America. The study focused on health and health-related behaviours in their social context, with the aim to deepen the understanding of the mechanisms influencing differences and changes in the health and health-related behaviour of school-aged children [42]. More detailed information about the two-step sampling process and procedure used to acquire this nationally representative sample could be found in our previously published paper [26]. On the school level, we approached 140 randomly chosen schools from all regions in Slovakia (in rural as well as urban areas) with a response rate 77.9%. On an individual level, we collected questionnaires in Slovak (and in case of schools with a Hungarian minority, in Hungarian) language from 8405 adolescents from 11 to 15 years old (mean age 13.43; 50.9% boys). Pupils completed the self-reported questionnaires at school on their own, in the presence of researchers and research assistants.The study was approved by the Ethics Committee of the Medical Faculty at the P.J. Safarik University in Kosice (16N/2017). First, schools were contacted and were asked for participation in the study on a voluntary basis. Second, parents were informed about the study via the school administration and could opt-out if they disagreed with their child’s participation. Finally, pupils themselves were given the opportunity to not participate in a data collection even if their parents provided the consent. Participation in the study was therefore fully voluntary and anonymous, with no explicit incentives provided for participation on all levels of data collection.To cover energy drinks consumption, we asked the question: “How many times a week do you usually drink energy drinks, for example, Red Bull?” with answers: (1) “never”; (2) “less than once a week”; (3) “once a week”; (4) “2–4 days a week”; (5) “5–6 days a week”; (6) “once a day, every day” and (7) “every day, more than once” [8].To cover breakfast consumption, we asked the question: “How often do you usually have breakfast during weekdays (more than a glass of milk or fruit juice)?” with answers: (1) “I never have breakfast during a week”, (2) “one day”, (3) “two days”, (4) “three days”, (5) “four days” and (6) “five days” [42].To cover sleep duration, we computed the time between bedtime and wake-up time on school days. We covered bedtime and wake-up time with two questions: “When do you usually go to bed if you have to go to school the next morning?” and “When do you usually wake up on school mornings?” with answers ranging in half-hour intervals [42,43].Emotional and behavioural problems were measured with the Strengths and Difficulties Questionnaire (SDQ). This questionnaire consists of 25 items [44], of which we used the 20 items covering emotional and behavioural problems. Answers were: not true (0), somewhat true (1), and certainly true (2). We calculated emotional (internalising) problems subscale (score 0–20) and behavioural (externalising) problems subscale sum score (score 0–20) [45]. A higher score indicates more problems. We explored Cronbach’s alpha which was 0.73 for the whole scale and 0.71 for the emotional and behavioural problems subscales, respectively.Family affluence was measured using the Family Affluence Scale III (FAS-III), which consists of six questions: “Does your family own a car, van or truck?” (No/Yes, one/Yes, two or more), “Do you have your own bedroom for yourself?” (Yes/No), “How many computers does your family own?” (None/One/Two/More than two), “How many bathrooms (room with a bath/shower or both) are in your home?” (None/One/Two/More than two), “Does your family have a dishwasher at home?” (Yes/No), “How many times did you and your family travel out of your country for a holiday/vacation last year?” (Not at all/Once/Twice/More than twice). We computed the sum score, which we converted to a ridit score ranging from 0 to 1. We then created tertile categories of low (0 to 0.333), medium (0.334 to 0.666) and high (0.667 to 1) socioeconomic position [46].First, we described the background characteristics of the sample using descriptive statistics. Second, we performed a series of analyses to explore the associations of ED consumption, sleep duration and breakfast consumption with emotional and behavioural problems, using linear regression analysis. We repeated these analyses with adjustment for gender, age, and family affluence. Next, we conducted a final analysis on the parallel mediation by sleep duration and breakfast consumption of the relation between ED consumption and emotional and behavioural problems. We did so by assessing the mediation effect of all variables separately and then building a parallel mediation model using the PROCESS macro model 6 [30]. These analyses were all controlled for gender, age and family affluence, and all indirect effects were subjected to follow-up bootstraping analyses, with 5000 bootstrap samples. The indirect effect was calculated using the a*b product method, and bootstrapped 95% confidence intervals for the indirect effect of ab was provided as a test of the indirect effect [47]. Arrows were added to the parallel mediation models in Figure 1 and Figure 2 to illustrate the mediation, expected and tested direction of the associations presented. All analyses were performed in SPSS v. 23 for Windows (IBM Corporation, New York, NY, USA).Table 1 and Table 2 shows the basic descriptive statistics of the studied variables in our research sample.In the exploratory linear regression analyses presented in Table 3, we found significant association between ED consumption and both emotional and behavioural problems; a higher frequency of ED consumption was positively associated with both emotional and behavioural problems. We also found both assumed mediators to be significantly associated with both emotional and behavioural problems; higher amount of sleep and higher frequency of breakfast consumption were negatively associated with both emotional and behavioural problems even after adjustment for gender, age and perceived socioeconomic status of the family (Model 1). Association with emotional problems lost its significance after adding sleep duration and breakfast consumption, association with behavioural problems decreased but remain significant after adding sleep duration and breakfast consumption (Model 2).Results from a parallel mediation analysis indicated that ED consumption was indirectly related to both emotional and behavioural problems through its relationship with amount of sleep and breakfast consumption. As can be seen in Figure 1, we found that adolescents with higher ED consumption reported lower amount of sleep and lower frequency of breakfast consumption. The lower amount of sleep and lower frequency of breakfast consumption was subsequently related to more emotional problems, respectively. A 95% bias-corrected confidence interval based on 5000 bootstrap samples indicated that the indirect effects through sleep duration and breakfast consumption was entirely above zero), and therefore significant.We found similar findings also for behavioural problems, as could be seen in Figure 2. Adolescents with higher EDs consumption reported lower amount of sleep and lower frequency of breakfast consumption. A lower amount of sleep and lower frequency of breakfast consumption was subsequently related to more behavioural problems, respectively. A 95% bias-corrected confidence interval based on 5000 bootstrap samples indicated that the indirect effects through sleep duration and breakfast consumption was entirely above zero, therefore significant.The aim of our study was to explore whether ED consumption is associated with emotional and behavioural problems and whether this association might be mediated by the amount of sleep and breakfast consumption among adolescents.We found in our analysis that higher consumption of ED was associated with both emotional and behavioural problems. Our results are consistent with the results of other studies. Adolescents who regularly drink EDs reported more emotional difficulties and symptoms of depression, anxiety, nervousness and stress [6,7,9,48,49]. There is also evidence that the consumption of EDs among adolescents is linked to increased risk of wide range of negative behavioural outcomes such as substance use, binge drinking, aggressive, violent and self-destructive behavior, hyperactivity/inattention symptoms and sensation seeking behavior [5,8,11,18,19,49]. The adverse effects of EDs on emotions and behavior of adolescents can be explained by the consumption of high doses of caffeine [12] and other psychoactive agents such as taurine, guarana or ginseng [4,11]. Energy drinks therefore pose potential health risks when it comes to emotions and behaviours because of above mentioned stimulants content, and are inappropriate for adolescents [10]. However, adolescents are not aware about these potential risks, and on the contrary, they have expectations about the positive effects of EDs on mood, performance and alertness [17]. Moreover, they consider EDs as an easy source of energy that helps them to cope with stressful situations and experience positive emotions such as pleasure and excitement [11]. Although the evidence shows a positive association between the consumption of EDs and emotional and behavioural problems in adolescents, it is not clear whether this relationship is casual and what and how other factors might influence this association. We therefore examined the role of sleep duration and breakfast on the association between ED consumption and emotional and behavioural problems in adolescents.We found that sleep duration mediated the association between EDs consumption and both emotional and behavioural problems among adolescents. Adolescents who consumed a higher amount of EDs slept less and reported more emotional and behavioural problems. Previous studies confirmed our findings that excessive use of EDs was associated with reduced and insufficient sleep duration [20,21,23,50] or even insomnia [24]. The association between insufficient sleep duration and emotional and behavioural problems has also been confirmed by several cross-sectional, experimental studies, as well as by systematic reviews [25,26,37,38,39,51,52,53]. Although most of the studies were cross-sectional and used subjective measures for sleep duration, the evidence from the experimental studies and systematic reviews suggests that the relationship between lack of sleep and emotional and behavioural problems in adolescents is likely to be casual and thus supports our findings that the effect of ED consumption on emotional status and behaviour of adolescents is not direct, but mediated through sleep duration. Furthermore, the association between sleep duration and emotional and behavioural problems is likely to be bidirectional and more complex, with other factors contributing to this association [51,53]. In our study, we also focused on one aspect related to the dietary habits of adolescents, specifically on the frequency of breakfast consumption, whose mediating role on the association between EDs consumption and emotional and behavioural problems will be discussed next.Finally, we found that breakfast consumption mediated the association between ED consumption and both emotional and behavioural problems among adolescents. Adolescents who consumed a higher amount of EDs regularly skipped breakfast during weekdays and reported more emotional and behavioural problems. We have found just two studies focusing on the relation between EDs and breakfast consumption and their impact on mental health and behaviour of adolescents [7,22]. The studies have provided evidence that breakfast omission was associated with stress, anxiety and depression levels and poor academic performance in adolescents, and that this was largely observed for both those who frequently consumed EDs and those who did not. Energy drink consumption was associated with more frequent consumption of junk food and breakfast omission. However, it is important to notice that other factors might be responsible for breakfast skipping in addition to the ED consumption, e.g., absent family rules regarding breakfast consumption. The findings indicate that consumption of EDs is associated with negative mental and behavioural outcomes not directly, but rather being part of an unhealthy diet, including breakfast skipping.The major strength of our study is its large nationally representative sample and the use of validated measures as a part of a standard HBSC questionnaire in the confidential setting, with children filling out the anonymised questionnaires in the presence of a research assistant only. Some limitations regarding our study should be mentioned as well. First, the cross-sectional design limits our ability to establish causal relationships. It cannot be determined if ED consumption contributes to the poor mental health of young people as it is possible that young people may use energy drinks to manage symptoms related to poor mental health (such as feeling tired or poor concentration). Second, we used self-reported measures, which are not able to measure sleep and EDs and breakfast consumption as precisely. This issue, combined with recall bias, may influence the accuracy of data, with the possibility of over and underreporting. Nevertheless, this is common in questionnaire studies that cover a broad range of topics and a large sample of participants, such as the HBSC study, and previous research has shown them to be valid [26,54,55]. Third, the sleep duration was calculated as a difference between wake-up time and bedtime, which is the time spent in bed and not the time spent asleep. This could have caused the number of sufficient sleepers to be overestimated.Professionals working with children should be aware that unhealthy dietary habits and lack of sleep might be related to emotional and behavioural problems in adolescents. They should consider that consumption of high amounts of EDs, breakfast skipping and lack of sleep may serve as markers to identify adolescents at risk for emotional and behavioural problems [4,8,9]. They should educate parents and adolescents about the importance and health benefits of regular breakfast consumption and sufficient sleep duration based on existing international recommendations [56] and recommend avoiding the consumption of EDs. Public health strategies should focus on the restriction of EDs consumption among children and adolescents by the limitations of availability of Eds, including the implementation of age limits in stores and sales ban in schools [5,10,11]. To formulate the strategies and intervention to promote adolescents’ health behaviour and well-being, more emphasis should be put on regular breakfast consumption and sufficient sleep duration. Further research is needed to examine the other factors influencing the relationships between ED consumption and emotional and behavioural problems of adolescents. To establish the causality in relationships, longitudinal and randomised controlled studies are required to be conducted.Adolescents with higher consumption of EDs reported more emotional and behavioural problems in line with previously published findings. Our study provides additional information about one of the possible ways such associations might be maintained. We revealed that the associations between EDs and emotional and behavioural problems were mediated by the amount of sleep and breakfast consumption. These findings suggest that parents and professionals working with adolescents should be aware that unhealthy dietary habits and lack of sleep might be related to emotional and behavioural problems.Conceptualisation, Z.D.V. and M.K.; methodology, Z.D.V. and D.H.; investigation, Z.D.V., D.H. and M.K.; formal analysis, Z.D.V.; data curation, Z.D.V.; writing—original draft preparation, Z.D.V. and M.K. All authors have read and agreed to the published version of the manuscript.This research was funded by the Slovak Research and Development Agency under the Contract no.s APVV-15-0012 and APVV-18-0070, and by the Scientific Grant Agency of the Ministry of Education, Science, Research and Sport of the Slovak Republic and the Slovak Academy of Sciences, Reg. No. 1/0177/20.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the Medical Faculty at the P.J. Safarik University in Kosice (16N/2017, 1.12.2017).Informed consent was obtained from all subjects involved in the study.The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical issues.The authors declare no conflict of interest.Mediation by sleep duration and breakfast consumption of the relation between EDs and emotional problems. Notes: *** p < 0.001. All presented effects are unstandardised; an is the effect of ED consumption on mediators; bn is the effect of mediators on emotional problems; c’ is the direct effect of ED consumption on emotional problems, and c is the total effect of ED consumption on emotional problems.Mediation by sleep duration and breakfast consumption of the relation between EDs and behavioural problems. Notes: *** p < 0.001. All presented effects are unstandardised; an is the effect of ED consumption on mediators; bn is effect of mediators on behavioural problems; c’ is the direct effect of ED consumption on behavioural problems, and c is the total effect of ED consumption on behavioural problems.Descriptive statistics of the sample (HBSC-study, Slovakia 2018, 11–15 years old, N = 8405).HBSC-study, Health Behaviour in School-Aged Children study, N, number of respondents, FAS-III, Family Affluence Scale III, SD, standard deviation.Frequencies of everyday energy drinks consumption, everyday breakfast consumption and sufficient sleep duration stratified by age and gender (HBSC-study, Slovakia 2018, 11–15 years old, N = 8405).Associations between ED, sleep duration and breakfast consumption with emotional and behavioural problems based on linear regression analysis leading to regression coefficients (B) adjusted for gender, age and family affluence (HBSC-study, Slovakia 2018, 11–15 years old, N = 8405).Notes: *** p < 0.001; Model 1, univariate analysis adjusted for gender, age and family affluence; Model 2, multivariate analysis adjusted for gender, age and family affluence.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06056.txt ADDED Viewed

	@@ -0,0 +1 @@

+ These authors contributed equally to this work as co-first Authors.These authors contributed equally to this work as co-last Authors.Over the last decade, medical education changed from traditional teaching methods to telematic and networking scholar and e-learning approach. The objective of the present systematic review was to evaluate the effectiveness and teachers/student’s acceptability of e-learning applied to the field of orthodontics and paediatric dentistry. A database search of the literature was conducted on PubMed and Embase databases from January 2005 to May 2021. A total of 172 articles were identified by the electronic search, while a total of 32 papers were selected for qualitative analysis. Overall, 19 articles investigated the effectiveness of e-learning, and no difference of acceptability was reported between e-learning and traditional methods for a wide part of the articles selected. A total of 25 papers provided a satisfaction questionnaire for learners and all were positive in their attitude towards e-learning. The results showed that e-learning is an effective method of instruction, complementing the traditional teaching methods, and learners had a positive attitude and perception. The evidence of the present study reported a high level of acceptability and knowledge level of e-learning techniques, compared to frontal lecture methods, in the fields of orthodontics and paediatric dentistry.One of the most important developments in recent years is the evolution of technology, which has changed many aspects of our everyday life: our means of communication, information retrieval, even the way we spend our free time (e.g., computer games) [1,2] The importance of technology became even more evident during the COVID-19 pandemic that has had a massive impact on people’s lives and habits. Restrictions limited people’s mobility while remote working, e-learning, and online platforms started to grow, along with online leisure solutions, such as gaming and video streaming [3,4,5,6,7,8,9]. The COVID-19 pandemic debuted in December 2019, and since then, changed the lives of every person around the world [10,11,12,13,14,15,16,17,18,19,20,21,22,23]. During this pandemic, physical distancing measures were imposed, and consequently, the education field had to adapt and transition to online platforms because this type of learning allows participation from all over the world to a meeting, webinar, course, or class [24,25,26,27,28,29,30,31]. The educational sector globally has shifted from traditional classroom teaching towards e-learning since most countries around the world experienced the temporary closure of all educational institutions in order to contain the spread of the pandemic [32]. (Figure 1).Over the last decades, e-learning has rapidly expanded in medical education, health promotion, patients, and medical education that take advantage of a useful networking flow and flexibility of the communication system [33,34,35]. Moreover, healthcare education faces higher challenges determined by an increase in students’ access to post-degree courses and specialisation that requires novel strategies to improve the quality of the scholarship and the didactic level [35,36,37]. E-learning allows students to learn anywhere and anytime outside the classroom, overcomes the shortages of teachers, and promotes learner’s motivation, cognitive effectiveness, and flexibility, leading to a shift from passive, teacher-centred learning to active, student-centred learning. It is affordable, saves time, and reduces costs [27,38]. As stated by Zhang et al., the mass quarantine caused a feeling of fear [39], and Hasan et al. showed in their study that there can be a strong relationship between the e-learning-related breakdowns and the psychological status of the student [40]. In the past, education occurred by means of textbooks, handouts, and notes taken during courses. E-learning enables teachers to represent the information using media in the form of text, images, animation, video, and audio [41,42] (Figure 2).There are several important factors that need to be considered for the success of e-learning: human factors pertaining to the instructors, the instructors’ and students’ technical competency, the instructors’ and students’ attitudes, the level of collaboration, and the technical support [43]. E-learning is a generic term that refers to electronically supported learning and teaching. It includes a variety of modalities and terms such as web-based learning, online learning, computer-assisted instruction, internet-based learning, distance learning, and virtual learning [27,44,45]. E-learning can be synchronous or asynchronous. Synchronous e-learning requires participants to log on at the same time and allows students to interact with each other and their teachers during the lessons. Asynchronous e-learning refers to e-learning that is ‘pre-recorded’ or available to students at any time of the day, potentially from any place [46]. Numerous studies conducted on e-learning in medical education showed that participants considered e-learning as an effective reinforcing method for medical training, without missing the traditional style of teaching [47,48,49,50]. A combination of traditional face-to-face learning and e-learning is called blended learning. The main advantage of blended learning is that it integrates the strengths of synchronous traditional face-to-face teaching and asynchronous/synchronous web-based learning activities [1,2]. Blended learning increases the learning flexibility in a demand-driven educational environment while maintaining the personal contact of the traditional face-to-face teaching, enhancing the classroom experience, and improving effectiveness and efficiencies by reducing lecture time [51,52]. It has been suggested that blended learning, i.e., e-learning and virtual learning environments mixed with a traditional lecture style, improve competencies and core knowledge of students [53]. During the last decade, the large use of smartphones and the internet has fostered widespread use of social media. Social networks such as Facebook, Twitter, YouTube, Google Drive allow people from different backgrounds to communicate and collaborate with other users across the world [53,54,55]. The growing interest in social media among students and the ubiquitous distribution of portable electronic devices has led instructors to improve their teaching and learning through combining social media applications, online platforms, and mobile technologies (Figure 3) [56,57,58,59].Mobile learning occurs when the student is not in a permanent and fixed location or if he is using mobile learning technologies. It is considered a part of e-learning [60]. Mobile devices such as smartphones, tablets, and laptop computers enable users to learn at any place and time, in different contexts and situations, and by interacting with others [30,61,62,63,64,65,66,67]. Furthermore, the growing availability of smartphones and tablets allows the mobile use of augmented reality in medical education [68]. The development of technology has affected even the field of dentistry, and numerous studies have been conducted on digital development in dental education [69,70,71,72,73,74,75,76,77]. In recent years, the diffusion of social media activities and web-based technologies has potentiated the information flow shared in several medical contexts and also in dental field education [78]. This form of interaction is useful at many different levels, such as for the education of undergraduate students, to enhance the expertise of younger dentists, in addition to improving the learning processes of experienced clinicians (Figure 4) [78,79,80,81,82,83,84,85,86].This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [87]. The article screening, selection for eligibility, and qualitative analysis of the study data were conducted by two independent paired reviewers (A.P., F.I.). If any disagreement occurred and unresolved issues were solved by consulting a third reviewer (F.L.). The screening phase was conducted on electronic databases which evaluated the manuscript title and abstract. The full text was collected for all identified articles in order to evaluate the qualitative analysis eligibility.Articles in which the objective was to determine the effectiveness and acceptability of e-learning or to compare e-learning with conventional teaching methods were considered.The inclusion criteria were based on the PICOT question guidelines [88,89]:-Population: students from graduate and postgraduate courses in orthodontics or paediatric dentistry; university staff; dentists who used e-learning tools to update their knowledge and continuing formation;-Intervention: use of virtual environments for learning;-Comparison: traditional classroom learning; traditional methods of instruction through the lectures, the clinical or laboratory demonstration, tutorial, text-or note-based learning;-Outcome: effectiveness and acceptability of e-learning;-Types of study to be included: cohort, observational, retrospective, or prospective study with emerging effectiveness in the last 16 years.Population: students from graduate and postgraduate courses in orthodontics or paediatric dentistry; university staff; dentists who used e-learning tools to update their knowledge and continuing formation;Intervention: use of virtual environments for learning;Comparison: traditional classroom learning; traditional methods of instruction through the lectures, the clinical or laboratory demonstration, tutorial, text-or note-based learning;Outcome: effectiveness and acceptability of e-learning;Types of study to be included: cohort, observational, retrospective, or prospective study with emerging effectiveness in the last 16 years.The inclusion filters were cohort, observational, retrospective, or prospective studies regarding the e-learning and virtual learning performance of dentistry specialisation of student scholars.Reviews, letters, conference readings, editorial, personal opinion, and studies without abstracts were excluded. We limited the searches to articles that were published in the last 16 years.A systematic electronic search on PubMed and EMBASE databases was performed limited to English language articles published between January 2005 and May 2021. A preliminary search was conducted by the Pubmed MeSH terms function of medical subject headings to identify the most appropriate descriptors and qualifiers of the present research topic to use for the Boolean search. The EMBASE Boolean search has been conducted by Emtree search algorithm. We used the following keywords: orthodontics, pedodontics, paediatric dentistry, e-learning, distance learning, web-based learning, and virtual learning. The search algorithm was (orthodontics OR pedodontics OR paediatric dentistry) AND (e-learning OR distance learning [Mesh] OR web-based learning [Mesh] OR virtual learning [Mesh:NoExp]). The final search was run on 30 April 2021.The assessment of the risk of bias of the included studies was independent and in duplicate in accordance with the EPOC guidelines [90]. A contribution was considered at high risk of bias in case of high/unclear risk of the ‘random sequence generation’ criterion. The risk of bias assessment was performed by a special data form by the software package Review Manager RevMan V 5.1 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Danmark). Study selection was accomplished through three different levels as follows:(1)Screening: all articles retrieved from these initial search criteria were subjected to a screening process by reading titles and abstracts;(2)Eligibility: in a second phase, the eligibility criteria were applied to the full-text version of the selected articles;(3)Inclusion: the remaining articles were included in the qualitative synthesis.Screening: all articles retrieved from these initial search criteria were subjected to a screening process by reading titles and abstracts;Eligibility: in a second phase, the eligibility criteria were applied to the full-text version of the selected articles;Inclusion: the remaining articles were included in the qualitative synthesis.The following data were extracted from each study: year of publication, country and setting of the study, aims of the research, number of participants, e-learning teaching method, comparison with traditional teaching methods, effectiveness and acceptability to students, teachers, or private practitioners of e-learning.The article search identified 172 studies consisting of 83 papers from electronic database search and 89 contributions detected manually. After the initial screening identification process, a total of 54 articles met the inclusion criteria applied to the title and abstract assessment. The full texts were evaluated for the eligibility criteria, and a total of 32 papers were deemed suitable for inclusion in this review. The study selection process is illustrated in Figure 5. The included studies are summarised in Table 1.The studies were conducted in the United States (10), United Kingdom (8), Brazil (5), Germany (3), Australia (3), Greece (1), Iran (1), and Saudi Arabia (1). Among the selected studies, 20 analysed the use of the e-learning teaching method in orthodontic education, while 11 studies evaluated the efficacy of e-learning in paediatric dentistry. One study evaluated the effectiveness of web-based self-instruction in both orthodontics and paediatric dentistry. The sample size for the included studies ranged from 9 to 430 participants. Participants were mostly undergraduate and postgraduate students or faculty members. Three studies involved dentists working in PHC, and one study involved private orthodontists. Studies evaluated many different educational interventions of varying duration, frequency, and format. The delivery modes used to deliver the educational materials included CD-ROM, learning management systems (e.g., WebCT, Moodle, and Blackboard), DVD, web browsers, and virtual learning environments. The methods used for interaction between trainers were videoconference, telephone, internet chat, or e-mail. Additionally, 16 articles included a comparison between e-learning or blended learning and traditional teaching methods.The outcome of the risk of bias assessment of the included articles was reported in Figure 6 and Figure 7. The 48.38% of articled reported a low bias of randomisation protocols, while the similarity of outcome measurements and selective reporting of outcomes presented a low risk of bias. The blinding approaches presented an unclear risk of bias of the studies selected. In most of the articles, the contamination bias was low.Two outcome measurements were considered in this review: effectiveness and acceptability. The effectiveness of e-learning was investigated in 19 studies evaluating the quantity of knowledge gain using multiple-choice questionnaires, open-ended questions, or practical exams. A significant improvement in participants’ knowledge after web-based courses was reported in eight studies. Mulgrew et al. concluded that travel commitments for trainees have reduced as a result of introducing the web-based resource but not as expected [97]. Camargo et al. found that graduate students finished the course with better performance than undergraduate students [107]. Of the 32 studies, 16 compared e-learning with traditional learning. In the majority of the studies, no difference was observed in knowledge gained between the two methods, whereas two studies concluded that e-learning was more effective than traditional methods. Papadopoulos et al. found a statistically significant difference between the group that used a virtual patient and the control group showing a gain in knowledge in the simulation group [106]. Luz et al. assessed that the ICDAS e-learning programme was more effective than traditional learning in improving dental students’ ability to use ICDAS [108]. Bains et al. compared e-learning with blended learning and face-to-face learning and he found that e-learning was less effective, while blended learning was the most preferred [99]. The changes in performance following learning were evaluated in five studies. Schorn-Borgmann et al. evaluated the performance of students in the construction of orthodontic appliances, and no significant improvement in the practical result was identified [110]. Ludwig at. al also failed to identify significant differences between face-to-face learning and the use of cephalometric imaging software [111], while Al-Riyami et al. found no difference in student performance in diagnosing TMD after VLE learning or face-to-face learning [98]. Luz et al. evaluated students’ performance in detecting dental caries [108]. Students’ acceptability was considered as an outcome in 25 studies. Seven of these studies mentioned that student satisfaction was evaluated with a Likert-scale questionnaire. The other studies used different types of questionnaires or surveys without mentioning the use of the Likert scale. All these studies reported a positive response from students when using online learning. In six studies, the students viewed online learning helpful as a supplement to their learning rather than a replacement for traditional teaching methods. Linjawi et al. stated that students responded ‘very positive’ to ‘positive’ for orthodontic e-course design, course delivery, and course outcome, but the orthodontic e-course was considered by most students as an adjunct and not a replacement of the traditional teaching methods [96]. Asiry found that few students preferred the online teaching method, and fewer students agreed to replace traditional lectures and live demonstrations with online tutorials, while most students preferred a combination of these teaching methods [113]. Mulgrew et al. concluded that despite the popularity of web-based learning resources, trainees continue to value the opportunity to interact face-to-face with their teachers [97]. Zafar et al. found in their study that 80% of the participants disagreed that virtual reality should replace conventional simulation [119]. In another article, Zafar et al. assessed that the use of VR simulation can be an additional tool that enhances students learning experience, without replacing traditional training methods [122]. According to the majority of studies, online courses were easy to access, well constructed, and understandable. However, Klein et al. found that the logistics of scheduling distant seminars, and uneven quality of the audio and video recordings were the major concerns of participants. They also assessed that students’ perceptions of the quality of the learning material were influenced by the depth of their preparation [103]. In the article of Peterson et al., students preferred the online textbook to traditional textbooks, but they had technical problems associated with online use of computers running obsolete (internet) browser software [93]. Bednar et al. stated that acceptability of the distance seminars appeared to be influenced by the instructor’s personality and teaching style, the seminar subject, and the residents’ technological level [92]. Only two studies evaluated the opinions of faculty members that showed a positive attitude towards e-learning. Klein et al. concluded that faculty members were somewhat more enthusiastic about the experience than were residents, and they would like to use this approach to distance learning again [104]. Mulgrew et al. found that the trainers felt that teaching has been more interactive and enjoyable since the introduction of the web-based learning resource even if they stated that it has changed but not reduced teaching commitments [97].To the best of our knowledge, this is the first systematic review examining the use of e-learning in paediatric dentistry, while several reviews have been published in orthodontics [123]. This review showed that the use of e-learning has a positive impact on healthcare education. The rationale of the present investigation considered only the bodies of evidence on e-learning methods in the last 16 years in accordance with the first worldwide expansion of scholarship using social media platforms, while Facebook reported, on 1 October 2005, a total of 21 universities in the UK and others around the world use the platform. This evidence is commonly considered the beginning of the social media application in a scholarly environment.The limits of the present investigation regarded the several differences of learning methodologies, the wide heterogeneity of the study population (undergraduate students/specialisation-related courses/teachers), and the feedback measurements modalities of the acceptability and effectiveness level. According to these bias factors, a statistical consideration/meta-analysis approach was not applicable for the present investigation.On the contrary, the rationale of the present study design offered the widest possible level of scholars, from novice/undergraduate students to those with advanced levels of expertise, not dispersing the sensitivity of the study.Most studies reported a significant gain in knowledge after e-learning, which confirms that e-learning is effective in increasing knowledge after training in both orthodontics and paediatric dentistry. Studies that compared e-learning to traditional methods concluded that e-learning was at least as effective as traditional learning.These results agreed with those of Lima et al. [124]. In a review, they evaluated the impact of tele-education in the field of orthodontics and concluded that orthodontic distance learning is an effective but complementary element, with no significant differences from the traditional methods of learning [124]. Kumar found that e-learning classes are at least as good as and/or better than face-to-face classroom learning and the blended approach which combines both traditional face-to-face learning and e-learning is the best method of teaching and learning [125]. Our secondary aim was to assess the acceptability of e-learning from students and teachers. This topic was explored in the interviews and questionnaires. The majority of participants considered e-learning to be effective and easy to use. According to Bednar et al., there are two benefits from using distance learning. It can enhance the experience of residents by exposing them to a variety of different thoughts, ideas, and other residents and instructors, and it can alleviate problems associated with decreasing numbers of experienced full-time faculty [92]. Many studies underlined the importance of interaction with faculty members. According to Camargo et al., interaction with tutors should provide motivation, guidance, and support to students. Klein et al. found that 92% of the participating residents thought the post-seminar discussion was an important part of the learning experience [103]. Furthermore, Miller et al. stated that participants preferred post-seminar videoconference in comparison with audio-only or chatroom interaction [101]. The studies we reviewed suggest that students prefer that online modules are used as a support to learning, and they dislike the replacement of traditional lectures with online instruction. In fact, a blended approach, mixing person-to-person contact with e-learning methods, seems the most preferred. Possible explanations could be as follows: (1) compared with traditional learning, blended learning allows students to review electronic materials as often as necessary; (2) compared with e-learning, blended learning learners are less likely to experience feelings of isolation or reduced interest in the subject matter [126].A particular type of e-learning is virtual learning. Only four articles included in this review examined the use of virtual reality. Kleinert et al. described the use of an interactive, multimedia virtual patient module developed on compact disc (CD-ROM) to increase students’ competence in caring for children with developmental disabilities [95]. Papadopoulos et al. demonstrated that a paediatric dentistry virtual patient built in a virtual world offers significant learning potential when used as a supplement to the traditional teaching techniques [106]. This result agreed with Zafar et al. who assessed that the Simodont simulated learning environment could be used as an adjunct in training dental students for preclinical paediatric dentistry restorative exercises [119]. Finally, Zafar et al. [122] presented the use of a VR simulator tool for local anaesthesia teaching in paediatric dentistry.Other studies have demonstrated the consistent efficacy of virtual patient programmes in a variety of educational fields, including clinical training in healthcare professions [127,128,129,130,131,132,133,134,135].Our study has several limitations. There are many confounding factors in learning that were not controlled for in the studies, such as the level of motivation of the studies, previous knowledge, and teaching style of the educators. The protocol of the present review excluded studies with no abstract. Interventions, topics, durations, and settings were different for every study. Traditional evaluation methods such as written texts, questionnaires were used for evaluation knowledge gain. It is unclear to what extent these methods can measure the effectiveness of e-learning, and how they may have influenced the outcome.The number of studies published on the use of e-learning, in comparison with traditional learning methods, was relatively limited. Other limitations were found in the selected studies, especially due to the failure to define the content quality and type of specific e-learning intervention being analysed.Moreover, studies did not report motivations that led to choosing a specific teaching method.Furthermore, we observed that the impression of the educator was evaluated in few studies.The Sars-CoV-2 pandemic worldwide emergency produced deep modifications of the institutional educational system with increased use of the smart-working approach, e-learning platforms, and limited use of traditional methods of academic learning. Within the limits of the study, the effectiveness of the present investigation demonstrated that e-learning is effective as traditional classroom methods, and the learners in these studies reported positive attitudes about e-learning with a high level of efficacy and acceptability by the operators and students. More detailed studies are necessary to understand the integration of e-learning into the learning methods in academic institutions and the implementation of interactivity in learning environments of dental students with special attention to the practicing clinical decision-making skills and operative procedures.Conceptualisation, A.D.I., A.M.I., and A.P.; methodology, A.D.I., N.C.; software, I.R.B., A.P. and P.P.; validation, F.I., F.L., A.M.I., and A.S.; formal analysis, A.M.I., F.L., and I.R.B.; investigation, G.M. (Giuseppina Malcangi), G.D., A.D.I., F.L., and M.B.; resources, A.M.I., G.D., A.D.I., F.I., I.R.B., and G.M. (Giuseppina Malcangi); data curation, G.D., M.B., P.P., N.C., and G.M. (Grazia Marinelli); writing—original draft preparation, A.D.I., A.M.I., and F.I.; writing—review and editing, F.I., M.B., G.M.(Giuseppina Malcangi), N.C., and G.D.; visualisation, F.L., A.S., and I.R.B.; supervision, F.I., A.D.I., and I.R.B.; project administration, F.I., G.D., G.M.(Giuseppina Malcangi), A.P., and P.P. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.All experimental data to support the findings of this study are available contacting the corresponding author upon request. The authors have annotated the entire data building process and empirical techniques presented in the paper. The data underlying this article are not freely available by agreement with our partners to protect their confidentiality.The authors declare no conflict of interest.Transition from traditional learning to e-learning.Example of two of the most employed software to create educational content.E-learning combination between telecommunication and medical field.Webinars on multiple topics held online.Selection of studies for the systematic review.Graph of the risk of bias assessment of the included studies.General summary of the risk of bias assessment of the included studies.Summary of the studies included in the systematic review.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06057.txt ADDED Viewed

	@@ -0,0 +1,63 @@

+Patient-centered care (PCC) has the potential to entail tailored primary care delivery according to the needs of patients with multimorbidity (two or more co-existing chronic conditions). To make primary care for these patients more patient centered, insight on healthcare professionals’ perceived PCC implementation barriers is needed. In this study, healthcare professionals’ perceived barriers to primary PCC delivery to patients with multimorbidity were investigated using a constructivist qualitative design based on semi-structured interviews with nine general and nurse practitioners from seven general practices in the Netherlands. Purposive sampling was used, and the interview content was analyzed to generate themes representing experienced barriers. Barriers were identified in all eight PCC dimensions (patient preferences, information and education, access to care, physical comfort, emotional support, family and friends, continuity and transition, and coordination of care). They include difficulties achieving mutual understanding between patients and healthcare professionals, professionals’ lack of training and education in new skills, data protection laws that impede adequate documentation and information sharing, time pressure, and conflicting financial incentives. These barriers pose true challenges to effective, sustainable PCC implementation at the patient, organizational, and national levels. Further improvement of primary care delivery to patients with multimorbidity is needed to overcome these barriers.Patient (or person)-centered care (PCC) receives a great deal of attention and has been adopted widely in healthcare organizations throughout the world [1,2,3,4,5,6]. In the past two decades, many interventions have been implemented to make healthcare organizations more patient centered. Commonly implemented PCC interventions for patients entail patient empowerment, physical support, and information provision; those for healthcare professionals focus mainly on education and training and improvement of the continuity and coordination of care [6].With such efforts, most organizations claim to be patient centered; the reality, however, is more nuanced [7,8,9]. In theory, PCC should be delivered using a comprehensive approach, with multiple interventions tailored specifically to the needs of the most vulnerable groups in society (e.g., patients with less education, migration backgrounds, or low health literacy) [2]; in practice, achieving this goal remains a huge struggle [10,11,12]. This nuanced picture of PCC in practice is especially relevant for primary care delivery to patients with multimorbidity (two or more co-existing chronic conditions [13]), who are often considered to form one of the most vulnerable groups in society [14]. Globally, more than half of people aged >65 years have multiple chronic conditions, which are treated mainly in the primary care setting [15,16]. Patients with multimorbidity are often older, with lower socioeconomic status and fewer health literacy skills [16]. Multimorbidity is also more prevalent among patients with migration backgrounds than among those without migration backgrounds [17]. Furthermore, multimorbidity is often related to adverse patient outcomes, such as poor health, low quality of life, functional impairment, and a greater risk of mortality [15,18,19,20].Current primary care delivery is not optimally tailored to the needs of patients with multimorbidity; PCC has the potential to overcome this obstacle [21,22,23]. The Picker Institute developed an eight-dimension framework that describes all aspects of PCC [24] (Figure 1): (1) patient preferences, (2) information and education, (3) access to care, (4) physical comfort, (5) emotional support, (6) family and friends, (7) coordination of care, and (8) continuity and transition. According to this framework, PCC delivery to patients with multimorbidity requires, among other efforts, that healthcare professionals strive to support patients in the setting and achievement of treatment goals guided by patient preferences. Patients with multi-morbidity can be viewed as being experts on their diseases [25] who should be empowered by healthcare professionals to be in charge of their own care. To do so, healthcare professionals should provide information and education that is accessible and understandable to all, regardless of education, age, educational background, or health literacy. Furthermore, PCC emphasizes the need for good access to care, meaning, among other characteristics, affordability and the accessibility of buildings to all patients, including those with mobility limitations. Moreover, as having many chronic conditions is often accompanied by physical problems, and as the perceived quality of the physical comfort (e.g., spatial layout) offered in healthcare settings affects the perceived quality of care, attention should be paid to patients’ physical comfort (e.g., management of sleeping problems, pain, shortness of breath; provision of comfortable facilities) [26]. Having multiple chronic conditions impacts patients’ lives, social relations, and/or jobs, and is often accompanied by feelings of anxiety and depression [27,28]. Thus, to be patient centered, healthcare professionals should offer emotional support to patients. Furthermore, chronic illnesses affect not only patients, but also their family and friends [29]. With PCC, healthcare professionals should involve these individuals in the care process, as they also have roles in care delivery and support [30]. Finally, care delivery to patients with multimorbidity often involves multiple healthcare professionals, within organizations (coordination of care) and across healthcare disciplines (continuity and transition). To ensure PCC, all healthcare professionals involved in care delivery to a multimorbid patient should be well informed, which involves regular and adequate transfer of information, and care delivery should be aligned to avoid fragmentation [31,32].In practice, the Picker Institute’s framework [24] is often used for the development of PCC guidelines and interventions. An example of such interventions is the establishment of patient-centered medical homes, which serves as a model for high-quality primary care that is considered to be more effective than standard care for patients with chronic conditions [33]. A systematic review has shown that the organization of care according to these eight dimensions of PCC results in better organizational and patient outcomes [2].Although a clear vision of PCC for patients with multimorbidity has been developed [34], PCC implementation in practice is not always straightforward. Barriers occasionally hamper adequate PCC delivery or prevent PCC implementation entirely. Healthcare professionals in management positions frequently mention the lack of time and funding as obstacles [23]. Multimorbid patients often have complex problems and needs, which take much time and effort to identify [35]. The identification of the problems at hand and the care and support required is particularly difficult for patients with low health literacy and/or education levels [36,37]. In addition, patients with multimorbidity form a heterogenous population requiring more than one type of PCC delivery [38]. Furthermore, most healthcare systems remain single disease oriented, and thus not adequately responsive to the needs of patients with multiple chronic conditions [39], resulting in complications in practice [40]. This situation reflects the need for and added value of PCC, as well as the challenges faced in its implementation. Despite agreement about the importance of PCC for patients with multimorbidity in the primary care setting, the realization of PCC in practice remains difficult. Although healthcare professionals’ perspectives of primary care delivery for patients with multimorbidity have been investigated [40,41,42,43], evidence from healthcare professionals regarding the sources of difficulties with PCC implementation for these patients is scarce. Thus, the identification of barriers to such implementation is a first step toward further improvement in practice.To make primary care for patients with multimorbidity more patient centered, insight on perceived barriers to PCC delivery for this population is needed. Thus, the present study was conducted to investigate such barriers, as perceived by healthcare professionals in a primary care setting.This study was conducted using a constructivist qualitative research design [44]. Data from semi-structured interviews were analyzed to identify barriers to PCC delivery for patients with multimorbidity in the primary care setting, as perceived by healthcare professionals (general practitioners (GPs) and nurse practitioners (NPs)). Its methodology is described according to the consolidated criteria for reporting a qualitative research checklist (e.g., participant selection, setting, data collection, analysis) [45].All participating healthcare professionals, from seven GP practices in Noord-Brabant, the Netherlands, participated in a 1-year-long (2017–2018) PCC improvement program initiated by a regional cooperative of GPs (Zorggroep RCH Midden Brabant BV). The program’s aim was to improve primary PCC delivery to patients with multimorbidity. Participants attended meetings for the improvement of their knowledge about PCC and the sharing of their experiences with PCC implementation in practice. A toolbox of interventions for PCC improvement was provided, and participants were instructed in its use in several workshops (the PCC improvement program and intervention toolbox have been described in detail previously [34]). The first and third authors (SK and JC) were present at all program meetings.At the end of the program, interviews were conducted to identify perceived barriers to PCC delivery for patients with multimorbidity in the primary care setting. This approach is similar to that used in previous qualitative studies of barriers to primary care delivery [46,47]. Sampling was purposive, with the intent of interviewing at least one GP and one NP per practice. The practices selected healthcare professionals for participation. As three practices had the same healthcare team, 10 interviews were planned. One interview was cancelled due to the participant’s illness. Thus, nine healthcare professionals (four GPs and five NPs; one male and eight females), comprising 43% of PCC improvement program participants, agreed to participate and were interviewed. After these interviews, the authors presented the themes to the healthcare professionals, and together, the group decided that no additional interview was needed, as all themes were recognized and no additional theme emerged. For the same reason, no repeat interview was conducted.The medical ethics committee of Erasmus Medical Centre, Rotterdam, the Netherlands, determined that the rules stipulated in the Medical Research Involving Human Subjects Act did not apply to this study (protocol no. MEC-2018-021). In January and February 2018, the first author conducted semi-structured interviews lasting about 1 h each. Each interview was conducted at the GP practice of the interviewee, with only the researcher and participant present. All interviewees were familiar with the purpose of the research and with the interviewer, with whom they had established relationships during prior program meetings. During the interviews, the eight PCC dimensions were used as a guide for consistency. Open questions (without a predetermined set of questions) were used to investigate the interviewees’ conceptualizations of each dimension of PCC, and of what could be further improved. Verbal informed consent was obtained from all participants. With the participants’ permission, the interviews were recorded digitally. No fieldnotes were made during the interviews.Reflexive thematic analysis was applied to the data, based on the steps defined by Braun and Clarke [48], to identify patterns of meaning across the dataset. The authors analyzed the data inductively; coding and theme development were directed by its content. To identify patterns of meaning, six steps were defined for the analysis (Figure 2). First, all interviews were transcribed verbatim (~3.5 h per transcript), and the first author read the full transcripts to familiarize herself with the data. The respondents did not read the transcripts. Second, the first author coded the content using ATLAS.ti, (version 8.4.18; ATLAS.ti Scientific Software Development GmbH, Berlin, Germany). Third, all authors examined the codes and identified themes in each PCC dimension representing barriers to PCC delivery for patients with multimorbidity in the primary care setting identified by the respondents. Fourth, all authors reviewed and refined the themes, discussing their scope and names until agreement was reached (triangulation). Finally, to validate the findings, all themes were discussed during a meeting, with all 22 healthcare professionals participating in the PCC improvement program; the professionals recognized the themes raised, and no additional theme emerged during this meeting.Descriptive statistics of the participants are presented in Table 1. The healthcare professionals identified barriers (themes) in all eight PCC dimensions (Table 2). The barriers are presented by dimension, but described below in no specific order, as all of the dimensions are important for the improvement of PCC.The consideration of patients’ preferences, wishes, and needs in care delivery often requires a shift from paternalistic consulting toward a coaching role for healthcare professionals. According to the interviewees, this shift is not always easy. The assumption of this new role, and the exploration of patients’ preferences, take time.
+              I have been working as a practitioner for many years and I have my ways, so I also have to get used to a change and a new approach to healthcare delivery.
+            Moreover, this shift requires additional communication skills and techniques to enable healthcare professionals to explore patients’ preferences and support them in goalsetting. Not all healthcare professionals, however, have been trained or acquired these new skills, which makes PCC delivery challenging. Furthermore, not all healthcare professionals are willing to make this change.
+              I still get very easily into sending mode. Sometimes you just convey certain information without having properly tested where the patient’s needs lie.
+            For adequate PCC delivery, a mutual understanding of patients’ needs and priorities is crucial; the interviewees reported that achieving such understanding can be challenging. For example, the exploration of patient needs and preferences is more difficult when there is a language barrier or cultural difference.
+              Sometimes a language barrier or culture also makes it difficult. With a language barrier, patients do not always understand what is going on and that they have a say too. And culture also often does determine how people cope with their disease process. Often, they are used to me telling them what is wrong, what they have to do, and then they do it.
+            The exploration of patients’ wishes and needs is also more difficult when patients do not want to be actively involved in care delivery. Some patients have difficulties being proactive, sharing their perspectives, or setting goals concerning their care. They prefer care as usual, with goalsetting done mainly by their healthcare professionals. The receipt of care as usual can be considered as a patient preference, although healthcare professionals sometimes struggle with this factor.
+              It can also be that the patient comes to me with very different expectations and does not feel the need to express what he wants, but adopts more of a consuming attitude: “well, just tell me how the blood sugar is and whether the blood pressure is okay and I will be satisfied.” Then it is difficult to find out what people really want with their health.
+            The interviewees emphasized that the time needed to deliver PCC, especially to patients with multimorbidity, should not be underestimated. As NPs often have flexible consultation times, this barrier applies mainly to GPs. Most consultations with GPs last 10–20 min, which is a short period of time for patients with complex care needs. Blood pressure or glucose measurement and/or the discussion of other physical complaints often take up most of this time. Financial arrangements with healthcare insurers have restricted consultation durations, limiting multimorbid patients’ access to care. Spending more time with patients than agreed upon with health insurers is not rewarded. These agreements are thus perceived as barriers to PCC delivery, as the time pressure means that healthcare professionals cannot always discuss patients’ care preferences or set goals with them.
+              What I find very strange is that if you tailor your care to the needs of the patient, help and invest in them well, then you get penalized very badly financially for that.
+              If I have only ten minutes, I go much less deeply than if I have double the time. Then I can ask a lot more thoroughly what the patient means and list all the options. Sure, I always try to do that, but really teaching the patients to make and set their own goals goes a bit further than that.
+            Another example is the healthcare insurers’ predetermination of the number of follow-up visits for multimorbid patients. With PCC, this number should be determined according to patients’ preferences, but this is currently difficult, as the insurers take the performance of fewer follow-up visits to represent low-quality care delivery and do not provide reimbursement for visits beyond the number agreed upon.
+              Well, if the patient says “I like it so much here I will come back next week,” you also have a problem. Because then he comes next week and the week after that, but you only get paid for two or three contacts a year. And that, of course, averages out. The health insurance company only looks at the care that was delivered. And if you get paid twice and you see him ten times, they would rather see that, than if you get paid three times and you only see him once.
+            Healthcare professionals often use community support elements, such as taxi rides to GP appointments for patients with mobility limitations, as part of good PCC provision. However, these services are not always (financially) accessible for patients, as they are often not reimbursed.
+              Exercise programs can make a huge contribution to care. But people do not get reimbursed for it, and there is still a group of people with small budgets who cannot afford it themselves. In order to provide PCC, sometimes a bit of professional guidance to get and stay in motion is also very much needed. I think that is a real gap in the regulations.
+            The interviewees acknowledged the importance of offering physical comfort at GP practices, but noted that they struggle with what to provide and what is considered to be sufficient (i.e., what exactly is “comfortable”). Moreover, they sometimes have limited options for comfort provision. For example, space limitations can make the provision of adequate privacy via separate waiting rooms and a separate front desk difficult. Furthermore, some interviewees expressed awareness that physical comfort (e.g., swinging doors) was suboptimal at their practices, but had no concrete plan to solve this problem.
+              My consultation room is upstairs where you can only get to by stairs. That is not ideal for some patients. But the lack of space forces me to do this. Sometimes when people cannot manage it, I make house calls and some of the people we know about we try to schedule them for a day when we have a free consultation room downstairs. But this is becoming increasingly difficult because we are indeed short of space. I realize that we also have swing doors as a front door, which is not very handy with the wheelchair.
+            The interviewees stated that they struggle with the involvement of multimorbid patients’ family members and friends in care delivery, including consultations, because they are simply not used to doing so. In addition, not all interviewees were aware of the benefits of this practice in terms of patient outcomes.
+              Well we can always do better, but I do not know how. Then you have to learn yourself to bring up those kinds of things [private situations] more often. But I do not quite see how to do that in an ordinary consultation. I only do that in exceptional cases. I do not ask the standard diabetic patient how things are at home. I will bring it up, but not every three months, I think.
+            The GPs and NPs also stated that they often do not involve patients’ relatives due to the time required to do so and to pay attention to and address their needs and questions. As their consultation times for this patient population are often limited, GPs choose to pay more attention to other aspects of care delivery.
+              The time is too limited. And if there is a problem, you would like to do something with it. And wanting to do something with it means the more things you bring up, the more problems there are, the more time you need to find a solution for all those problems.
+            The interviewees explained that family members’ preferences sometimes contradict those of patients, which contributes to the difficulty of involving relatives in care delivery.
+              Involving family is sometimes difficult. Sometimes I do get phone calls from [patients’] children. Sometimes that is nice, sometimes it is not. If several children are involved in the care delivery, and all want something different, it sometimes creates difficult situations.
+            The interviewees recognized that not all patients think that their GP practice is the place to discuss emotional issues or the impacts of chronic diseases on one’s private life. Although some patients know that the exploration of such issues is the task of mental-health NPs, they do not believe it to be the task of GPs. This perspective may impede the provision of adequate emotional support to patients who need it.
+              Sometimes you also see that there is some doubt if they [patients] can say it here, because how will we [healthcare professionals] think of it [an emotional problem].
+            The interviewees acknowledged that they do not always address possible emotional problems accompanying multimorbidity.
+              Of course, I do not always ask about it [emotional problems]. Yes, if people start talking about it themselves, I do listen. I do my best with that, or I suggest the accessible mental healthcare nurse practitioner. But there is not always attention to emotional aspects. Someone with diabetes with good values is doing well. Then I am not going to actively ask whether he is also under stress.
+            Furthermore, not all interviewees felt comfortable discussing emotional aspects accompanying patients’ diseases, such as depressive feelings or anxiety.
+              Well, there will undoubtedly be intrinsic factors in myself as well, on account of which I may be more likely to discuss certain things rather than other topics. I also bring my own person into a conversation. So that can be a barrier.
+            The GPs interviewed also noted the lack of clear boundaries for the provision of emotional support, whether the recognition of problems is sufficient or more is needed. This factor is related to time pressure; the interviewees stated that they do not want patients to believe that they can make appointments solely to discuss emotional problems, as they feel that this is not their role and that time is limited.
+              I do not have time myself to talk for half an hour every week, but the mental healthcare nurse practitioner does. Some people do like that, other people say no I do not want that, I just want to talk about it here. And then I think, no way I am going to free up my schedule to talk for half an hour every week. We also have to set boundaries.
+              If a patient is very sad, you cannot say “well, the time is up.” You do not do that. So yes, that also makes the planning of the consultation hours difficult, because they come for something and if everything else comes along, which is quite often, then it runs late. And you cannot schedule everyone for half an hour, because even if you were to work twenty-four hours a day, you still would not have seen all the patients. So, you always have to choose and share. And that is just annoying. You can never do the best for everyone and that is very frustrating.
+            The interviewees emphasized the importance and difficulty of providing information specific to multimorbidity, as disease-specific information on comorbidities does not always exist.
+              I would like to give more psycho-education, so people get more specific information. But that is difficult to do for such a wide range of conditions. There are so many things that play a role in multimorbidity.
+            Not only healthcare professionals, but also patients, need to possess skills to explore their preferences. Patients need to have health literacy and communication skills to share preferences and information and set goals. Thus, the interviewees found the lack of such skills to impede PCC delivery.
+              You will see that patients with multimorbidity are often older people. And older people often look up to the doctor as well. And have a little less knowledge, they think, of all kinds of diseases, while of course that is not the case. Because they have been on Earth much longer than I have. But the elderly are more sensitive to it. The younger people can decide much easier, and often find a lot of information on the internet to make a targeted choice.
+            The interviewees noted that health literacy skills vary greatly in this patient population, making the adjustment of information provision to individual patients difficult. It can be difficult to recognize what patients need to gain better health literacy skills, and to determine whether patients have truly understood the information provided.
+              And as to low literacy, here in the village it is not too bad, but for someone who barely finished secondary school or did not finish it at all, it is obviously quite difficult to think about conditions, pills, solutions and options, to make a choice. And then it seems as if you have to be smart to make a good choice, but someone who is less educated can do that just as well. Provided that the information fits well. And of course, there is a barrier in that. Because as professionals we communicate on a completely different level. We use much more complicated words and terms that do not always come across.
+            The interviewees mentioned that the development and use of multiple resources (e.g., brochures) adapted to all education levels and language backgrounds would aid the provision of good information aligned with patients’ needs and characteristics. Although such materials exist, the interviewees did not use them often.
+              I could perhaps do more with the foreign people here in the district in terms of informational material. Because I do that a lot in Dutch now. Of course, they are often accompanied by someone who can speak Dutch, but then it all goes through an intermediary. And I think there are enough materials in other languages as well that are not yet available at the thuisarts website [which provides disease-specific information to patients].
+            According to the interviewees, adequate PCC delivery requires all practice team members to believe in the added value of this approach. They noted that the coordination of care differs between small and large teams in GP practices. For PCC, the same team should be involved in every instance of care delivery to a patient. However, coordination becomes more complex with the addition of team members (e.g., multiple assistants at the front desk, part-time workers).
+              We were looking at how to divide the patients among three nurse practitioners. At first we had one nurse practitioner, and then of course there was nothing to divide. But now we have more. And one works only so many hours part time and the other works only so many hours part time. So, it all just has to fit, but coordinating this can be quite a challenge.
+              For a patient, it is quite difficult. Having your own general practitioner and a nurse practitioner is manageable. But there are also eight assistants they have to deal with, and I think that can be confusing. That could be organized better.
+            The interviewees emphasized the importance of the team’s morale and atmosphere for the adoption of a new approach. When no safe environment to provide feedback and ask critical questions exists, improvement is difficult.
+              It is enjoyable to watch each other’s work and you can get a lot of tips and find many improvements by doing so. But feedback is sometimes given in such a way that makes it come across as hurtful or threatening. There must also be a sense of safety.
+            In many cases, healthcare professionals from diverse disciplines in various healthcare settings (e.g., primary, hospital, community, and social care) are involved in care delivery to patients with multimorbidity. The interviewees noted that this situation may hinder the continuity of care; longer chains of care are more vulnerable to disruption.
+              Because there are many healthcare settings involved, there are many links and each link is vulnerable. If I verbally pass something on to you and you pass it on to someone else and they pass it on to their colleague. After ten people, look what finally emerges.
+            To ensure the continuity of care, collaboration among healthcare settings is very important. The GPs interviewed stated that they tried to take leading roles in managing the continuity and transition of care, but emphasized that this was easier said than done. The part-time work schedules of many healthcare professionals render the continuity of care even more difficult, due to difficulties with the scheduling of meetings and alignment of advice. Furthermore, the interviewees stated that they did not always know the expertise of professionals in other disciplines, especially those outside of the healthcare setting (i.e., in the community or social domain), which makes the transition of information and referral difficult.
+              I think that as a GP I have a particular task when people see several specialists and those specialists are not always well informed about each other’s goals and treatments. Patients sometimes lose their way because of this, because they feel that there is not enough holistic collaboration. My job is to call or consult with the specialist or refer someone who is a bit older to a geriatrician. And then I sometimes ask specifically whether the geriatrician could take over the check-ups from the various specialists. But that is often not the case. If someone is a very specific rheumatologist or a patient has a cardiac or pulmonary condition, you do not let those specialists go easily. Then you sometimes have to call more often to get things coordinated. I think that takes a lot of energy. And it takes a lot of energy from the patient as well.
+              More and more people work part time. So, in any case you also get more and more people within the chain who are not always available at the time that you work.
+            The interviewees identified data protection laws as barriers to PCC, and in particular to the continuity and transition of care. Good, complete documentation shared among all healthcare professionals involved in a patient’s care is important, but these laws prohibit the sharing of some information with professionals in all disciplines, resulting in the loss of (relevant) information. Medical information may be transferred only between medical doctors, and cannot be shared with paramedics, who are members of multidisciplinary teams providing PCC. The laws also make information sharing during multidisciplinary team meetings difficult.
+              We have a pharmacy here in the building. I am not allowed to just hand over a list to the pharmacy saying these are all the people with heart failure, could you please check if the medication is okay. Because that is a data leak. So, I have to ask permission from each individual patient to tell the pharmacy that they have heart failure. And then if the patient says yes, then it is allowed. Otherwise it is not. So, you have to take a lot of steps to get there.
+              We are only allowed to transfer information to another physician. So, not all the allied healthcare professionals are allowed to have certain information, because that is all protected. We also have a chain information system, but everyone’s information is open to a limited extent. Most healthcare professionals involved really only get the referral and no additional information is allowed.
+            The data protection laws also complicate communication with healthcare professionals involved in a patient’s care, as the (unprotected) exchange of emails is not permitted. This situation often results in a loss of efficiency in seeking to achieve continuity of care.
+              Email traffic in primary care really needs to be implemented safely at breakneck speed, although it is apparently very difficult. This is really a shortcoming. This would allow us to communicate even better with the patient. For me as a NP, the GP is ultimately responsible, so I have to regularly consult with the GP and then call the patient back. The patient also has to stay at home especially for that phone call. With an email you can save a lot of time, but it will also help the patient since he can read everything back at leisure. If you start with medication, the patient has to pick it up at the pharmacy, take it at a certain time for a certain amount of time. That is a lot of information, and putting that in an email might be more convenient.
+            According to the interviewees, the data information systems used within the organization and for the entire care chain are not optimally designed to function concurrently. Given the use of two different systems, not all relevant information is transferred adequately to all professionals on multidisciplinary care teams. This situation complicates communication among all healthcare providers involved and may result in the fragmentation of care.
+              When I report on diabetes care, all the doctors involved can just see it in the chain information system [CIS]. But within the practice we work with a GP information system (GIS), but those two systems do not always work well together. For example, when patients last visited the optometrist. Nine times out of ten, the data is correctly processed in CIS but sometimes it does not come across well in GIS. So, for example, they go to their GP for an annual check-up and the GP asks when was the last time they saw the optometrist? Sometimes the patient cannot remember, so the GP looks in GIS and cannot find the report. Then they have to ask me to look in CIS to look it up. This is not very efficient.
+            This study was performed to investigate barriers to PCC delivery to patients with multimorbidity, as perceived by healthcare professionals in a primary care setting. Although the participating healthcare professionals acknowledged the value of PCC in this context, they identified barriers in all eight PCC dimensions. According to the study findings, healthcare professionals face difficulties in making the shift from a paternalistic consulting to a coaching role; the assumption of a new role takes time, and additional skills are necessary to, for example, thoroughly explore patient preferences. Such changes of mindset have been mentioned frequently as barriers to PCC implementation [49]. Furthermore, although patient-centered communication encompasses several skills, such as the expression of empathy and shared decision making [50], many healthcare professionals are not trained in such skills and do not realize that their possession could help them improve their patient-centered communication [50,51]. Communication training could achieve this goal [52], potentially enabling healthcare professionals to gain a better understanding of their patients’ conditions and care needs, in turn resulting in better treatment alignment [53]. Healthcare professionals also encounter barriers with regard to patient preferences (e.g., language barriers) when creating mutual understanding with their patients. Language barriers perceived by patients and healthcare professionals have been found to impede PCC delivery to immigrant and refugee women [10]. In addition, healthcare professionals who participated in this study reported feeling that not all patients want to be actively involved in their care and/or have difficulties with goalsetting. Patients have been found to differ in their proactivity and skills for active PCC involvement [38]. Although a patient’s preference for care as usual should be respected, we emphasize the need for thorough examination of whether the patient truly does not want to be in charge of his or her care, or whether the selection of care as usual is simply easier for him or her, as he or she may have difficulties with expressing his or her needs or preferences. The latter reflects the need for extra support from healthcare professionals to identify patients’ needs and preferences.In the access to care dimension, healthcare professionals reported the lack of reimbursement for care provided as a barrier to effective PCC implementation. PCC often requires that healthcare professionals spend more time and exert more effort during consultations and in additional training sessions and workshops, and that they collaborate with professionals in other healthcare disciplines. The lack of financial structures supporting such activities may hamper the sustainability and widespread embedding of PCC into care systems in the long term. Concerns similar to those identified in this study have been raised by many healthcare professionals participating in programs aiming to improve the quality of primary care (e.g., integrated primary care for community-dwelling frail older persons, interventions based on the chronic care model) [54,55]. Supporting financial structures are often described as prerequisites for the effective and sustainable implementation of healthcare delivery [56,57]. In addition, as the financial resources of patients with multimorbidity vary [35], the creation of supportive financial structures also accounts for community support that may be inaccessible to patients with fewer resources.The healthcare professionals reported that they struggled with how to provide physical comfort in their GP practices. A systematic review revealed differences in preferences regarding essential aspects of physical comfort provided in healthcare organizations among departments and occupants [58]. Additional research is needed to identify specific aspects of physical comfort preferred by patients with multimorbidity. The study participants reported several barriers in this dimension. They acknowledged that they had difficulty involving patients’ relatives in care delivery because they are simply not used to doing so, and not all healthcare professionals were aware of the benefits of doing so. Patients with chronic diseases have been found to involve their family members and friends more often when their care needs become too complex to self-manage and when worse health outcomes become more likely [59,60]. The study participants also reported that their consultation time is too limited to incorporate all aspects of PCC. As patients with multimorbidity often have physical complaints, most of the professionals’ attention is devoted to these problems, leaving limited time to address relatives’ needs and questions [40,61]. Finally, the healthcare professionals experienced difficulties when they faced contradicting needs of patients and their family members. In another study, patient–family disagreements also were identified as a barrier to family involvement in primary care [62]. This study revealed that patients with multimorbidity do not think their GPs’ tasks include the discussion of emotional aspects of their conditions, as has previous research (38). GPs likely feel the same, although a 2014 mental healthcare reform in the Netherlands designated emotional support as a GP task [63]. The aforementioned barrier that consultation time is often spent fully on the addressing of the physical aspects of patients’ conditions also applies to this dimension. However, as patients with multimorbidity often experience high emotional burdens related to their conditions, emotional support of these patients should receive more attention [27,28].Healthcare professionals participating in this study emphasized the importance of patients’ possession of health literacy and communication skills, which allows them to participate in PCC delivery. The alignment of information provided with multimorbid patients’ needs and backgrounds has been shown to be important to increase patient-centeredness [34]. This study revealed wide variation in such literacy and skills among patients with multimorbidity. This is in accordance with the previous identification of subgroups of patients with multimorbidity based on personal resources such as communication and health literacy skills [35]. Moreover, health literacy skills are often considered to be fundamental for patients who want to be in charge of their care [64]. Previous research provides insight in how PCC delivery can be aligned to the (differences in) care needs of patients with multimorbidity [38]. Furthermore, this study revealed a barrier related to the provision of information to patients with multimorbidity, as most available information is disease specific. The same barrier was identified in a systematic review describing the challenges that GPs face in managing patients with multimorbidity [40].According to the study participants, optimal PCC delivery requires that all healthcare professionals in an organization are motivated to achieve change and improvement, and that the environment is supportive. When not all such professionals are motivated or able to change, improvement may be difficult. Consequently, larger teams may add complexity to the achievement of improvement. According to Fleuren et al. [56], organizational size, colleagues’ support, and the extent to which the task orientation beliefs of healthcare professionals fit the innovation goals are important determinants for healthcare innovation. The study participants reported three barriers in the continuity and transition dimension. They reported that adequate information sharing is difficult to achieve when working with large teams of healthcare professionals across multiple settings. A study investigating how GP practices should organize their care for patients with multimorbidity to increase patient-centeredness showed that multidisciplinary work is very important and can be strengthened by the organization of multidisciplinary meetings [34]. A systematic review showed that fragmentation between primary and secondary care poses a major challenge to the provision of care to patients with multimorbidity [40]. Second, the study participants reported that data protection laws restrict information sharing among healthcare professionals from multiple disciplines involved in individual patients’ care. Third, they emphasized that data information systems within organizations and for entire care chains are not optimally designed for concurrent functioning. Previous studies have revealed similar challenges to the continuity of care [65,66]. The inadequacy of information and communications technology systems may endanger the continuity of care, which is especially important for patients with multimorbidity, many of whom require multidisciplinary healthcare teams. Optimal technology and supportive laws are often described as prerequisites for the effective and sustainable implementation of healthcare delivery [56,57].The barriers identified in this study pose true challenges in the effort to effectively and sustainably implement PCC at the patient, organizational, and national levels. At the patient level, most identified barriers were related to the variation in patients’ care needs and health literacy skills. These differences should be considered when developing care plans according to the PCC framework. At the organizational level, this study showed that not all healthcare professionals are aware of and/or trained in all elements of PCC delivery. Training and education of healthcare professionals should be initiated to increase their awareness and skills related to patient-centered communication, the involvement of patients’ family members and friends, and the discussion of patients’ emotional status, thereby improving care delivery to patients with multimorbidity. At the national level, challenges are related to data protection laws that restrict information sharing among healthcare settings, and to the lack of financial structures supporting PCC implementation; both of these factors are considered to be prerequisites for the effective and sustainable implementation of healthcare delivery [56,57]. Future research and policies should focus on meeting organizational preconditions to enable investment in preventive care across the lifespan and to make PCC the best way forward.Several limitations of this study should be considered when interpreting its results. First, the generalizability of the results may be limited, as this study was conducted with primary healthcare professionals in the Noord-Brabant region of the Netherlands. Future research should investigate the experiences of healthcare professionals with regard to barriers to PCC implementation in other regions, countries, and healthcare settings. Second, the sample of nine healthcare professionals may be considered to be small. However, this sample size is similar to those used in other qualitative health and well-being studies [67,68,69,70]. We selected it carefully, inviting 50% of all healthcare professionals from the GP practices participating in the PCC improvement program. Furthermore, the data are rich and were discussed during a meeting with all PCC program participants for validation; all healthcare professionals agreed with the findings, and no new theme was raised. PCC has the potential to entail the tailored delivery of primary care according to the needs of patients with multimorbidity. PCC implementation in practice, however, is often difficult due to the existence of barriers. At the patient, organizational, and national levels, barriers were identified in all eight dimensions of PCC (patient preferences, information and education, access to care, physical comfort, emotional support, family and friends, continuity and transition, and coordination of care) in this study. They include difficulties with the achievement of mutual understanding between patients and healthcare professionals, the lack of healthcare professionals’ training and education in new skills, data protection laws that impede adequate documentation and information sharing, time pressure, and conflicting financial incentives. These barriers pose true challenges to effective and sustainable PCC implementation for patients with multimorbidity. J.M.C., S.J.K., and A.P.N. drafted the design for data collection. J.M.C., S.J.K., and A.P.N. were involved in subject recruitment; S.J.K. conducted the interviews and performed the qualitative analysis. J.M.C., S.J.K., and A.P.N. interpreted the data. S.J.K. drafted the manuscript and J.M.C. and A.P.N. contributed equally to its refinement. All authors have read and approved the final version. All authors have read and agreed to the published version of the manuscript.Funding for this research was provided by the Dutch healthcare insurance companies CZ and VGZ. The reported research findings are based solely on the analytical results.The medical ethics committee of Erasmus Medical Centre, Rotterdam, the Netherlands, determined that the rules stipulated in the Medical Research Involving Human Subjects Act did not apply to this study (protocol no. MEC-2018-021). Verbal informed consent was obtained from all participants in this study and tape recorded.The data are available upon (reasonable) request.The authors thank all of the healthcare professionals who took the time to participate in this study. The authors declare that they have no competing interest.Framework of the eight dimensions of patient-centered care, as defined by the Picker Institute [24].Six steps of thematic analysis according to Braun and Clarke [47].Descriptive statistics.GP: General practitioner, NP: Nurse practitioner.Overview of barriers to patient-centered care (PCC) for patients with multimorbidity.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06058.txt ADDED Viewed

	@@ -0,0 +1 @@

+ The production of pharmaceutical ingredients, intermediates and final products strongly depends on the utilization of water. Water is also required for the purification and preparation of reagents. Each specific application determines the respective water quality. In the European Union, the European Pharmacopeia (Ph. Eur.) contains the official standards that assure quality control of pharmaceutical products during their life cycle. According to this, the production of water for pharmaceutical use is mainly based on multi-stage distillation and membrane processes, especially, reverse osmosis. Membrane distillation (MD) could be an alternative process to these classical methods. It offers advantages in terms of energy demand and a compact apparatus design. In the following study, the preparation of pharmaceutical-grade water from tap water in a one-step process using MD is presented. Special emphasis is placed on the performance of two different module designs and on the selection of optimum process parameters.The forecast of the European pharmaceutical market predicts a 3.9% growth between 2019 and 2024, while the global market is expected to rise by about 4.2% in the same period [1]. For the production of pharmaceutical ingredients, intermediates, and final products, as well as for the purification and preparation of reagents, water is crucial. However, water used for pharmaceutical applications is subject to strict regulations. In the European Union, the European Pharmacopeia (Ph. Eur) determines the quality of water for pharmaceutical use as well as the processing measures. Within the current pharmacopoeias, a distinction is made between purified water (PW) and water for injection (WFI) [2]. PW is used for the production of medical products that are neither pyrogen-free nor sterile. Pyrogens are substances that can cause fever in humans during parenteral intake (bypassing the intestine, e.g., intravenously). WFI is water for the production of medical products, solutions and dilutions for parenteral use. Drinking water with a quality according to the respective national regulations is the raw material for both PW and WFI production [2]. The Ph. Eur. sets limits and requirements for the quality of PW and WFI; some quality parameters are depicted in Table 1.PW is produced by distillation, reverse osmosis (RO) in combination with electrodeionisation (EDI), ion exchange or other suitable methods [3]. In contrast, the Ph.Eur. significantly limits the processes that may be applied for the production of WFI. The most commonly used method for WFI generation is multi-stage pressure column distillation [3].Membrane distillation (MD) may be an alternative process for the generation of pharma-grade water. MD is a thermal separation process based on vapor pressure differences between the feed and the distillate sides of porous, hydrophobic membranes [4]. When drinking water is used as a feed solution, these membranes allow the passage of water vapor only and retain suspended and dissolved matter on the retentate side; thus, the condensed product obtained is theoretically 100% pure water [5,6].MD is investigated worldwide as a low-cost, energy-saving alternative to conventional separation processes such as distillation and reverse osmosis (RO). MD can be operated at a relatively low feed temperature between 50 °C and 80 °C. Some studies [7,8,9] employed newly developed single-layer and multilayer graphene-based membranes for application in membrane distillation. Xu et al. [7] reported about graphene oxide (GO) nanolayers that were deposited on the permeate side of PVDF membranes. Generally, these membranes show an advantage for membrane distillation because of their unique water channels. Application in direct-contact membrane distillation enhanced the desalination performance due to a decrease of the vapor pressure at the permeate side. Salt rejection was improved to 99.9%. Deterioration of the permeate quality caused by membrane wetting was avoided by the properties of the GO layer. Grasso et al. [8] prepared porous composite membranes from functionalized PVDF membranes that were coated with graphene. The generated membrane was applied in direct-contact membrane distillation (DCMD) and showed long-lasting salt rejections >99.9%. Huang et al. [9] reported about photothermal membrane distillation (PMD) using a PTFE membrane coated with an ultrathin graphene-based film. Compared to the unmodified membrane, the transmembrane water flux of the modified PTFE membranes showed an enhancement of about 78% under solar illumination.For the operation of MD, it is important that the dry pores are not wetted by the liquid feed, which is directly in contact with the membrane. In contrast to RO, MD is not limited by the osmotic pressure that is generated; furthermore, equipment costs are lower [10]. As MD is operated at low temperature levels, the use of waste heat [11] as well as solar thermal energy and geothermal energy is possible [12,13]. On the other hand, a disadvantage of MD is the low permeate flow compared to RO [7]. This restriction may be overcome by the further development of graphene-based membranes. MD processes can be categorized into four basic module configurations, which play a fundamental role in separation efficiency and processing costs: direct-contact membrane distillation (DCMD), air-gap membrane distillation (AGMD), sweeping-gas membrane distillation (SGMD), and vacuum membrane distillation (VMD), as shown in Table 2. Furthermore, alternative configurations with low energy consumption and improved flux have been developed [14].There is hardly any information available on the application of MD for the generation of pharma-grade water. A study [16] reported about the application of membrane distillation for the generation of WFI without intermediate steps. Data on the performance were not provided.The investigations were carried out using a test plant that was especially designed by Wilhelm Werner GmbH (Figure 1). It allows the operation of both AGMD and VMD configurations, which were chosen for the experimental investigations due to their advantages, i.e., little conductive heat losses and relatively high permeate flux.Generally, the plant was designed for testing the suitability of MD for the generation of pharma-grade water from tap water. The design criteria considered neither hygienic design nor energetic optimization. Hauser [17] stated that conventional detachable connections need to be replaced by solid connections or by hygienically designed detachable connections and sealing points to prevent the growth of microbiological impurities.Figure 2 depicts the piping and instrumentation diagram of the two chosen configurations.Deukum GmbH, Germany, provided a plate-and-frame membrane module Type MDX100 with membrane packages for AGMD and VMD, as shown in Figure 3. The module was operated in a counterflow mode, i.e., the feed and distillate flew along the membrane in different directions. The velocity of the respective solutions determines the thickness of the fluid dynamic boundary layer, which creates a resistance to material and heat transport. The membrane package used for AGMD consisted of the flat membranes (polyethylene), spacers (silicone) and condenser sheets (PP). The membrane package used for VMD consisted of the flat membranes (polyethylene) and spacers (silicone). Each membrane package was sealed with silicone, and the individual components were bonded together by heat treatment.In total, 30 microporous polyethylene membranes type 14P02F from Lydall Performance Materials, each with an effective membrane area of 0.01 m2, were stacked in each module, leading to a total membrane area of 0.3 m2. The membrane characteristics are reported in Table 3.Untreated tap water provided by the city of Leverkusen, Germany, served as feed solution. Tap water served as a cooling agent for the vapor condensation as well.The electrical conductivity is an important quality parameter for the separation efficiency in the MD process. Conductivity was determined with a conductivity meter type JUMO Blackline CR-EC built into the pipeline.The central factor for the efficiency assessment of MD is the transmembrane distillate flow, represented as mass flow or volume flow [18]. To compare the performance of different module or plant concepts, the distillate flow can be related to the efficient membrane surface AM. The resulting value is referred to as flux.The degree of desalination is described by the salt retention R, which can be calculated using the conductivity of distillate σD and feed σF [19]:(1)R=(1−σDσF)·100%The rejection of impurities of all kinds can be described by the decontamination factor DF [20]:(2)DF=cDcF·100%In batchwise investigations, the yield Y describes the ratio of the generated distillate volume to the feed volume:(3)Y=VDVF·100%The investigations included factor variation of influencing parameters to determine optimum operating conditions and long-term experiments. Furthermore, investigations concerning hygienic safety of the process were conducted.Before the examinations were carried out, the statistical variation of the distillate flux was tested at different working points. Statistical variation was so low that each test presented in our paper was carried out only once.Full factorial parameter variations of feed temperature TF, feed volume flow FF and cooling water volume flow FK, each on three factor levels, were performed. For VMD, also the applied negative pressure was varied. The cooling water temperature in all tests was kept constant at approximately 17.5 °C. The designations 1, 2 and 3 in Section 3 refer to the factor levels Low (1), Medium (2) and High (3). Furthermore, a reference value (0) is given for VMD, indicating operation without negative pressure.Table 4 summarizes the experimental conditions.Long-term investigations were conducted with both AGMD and VMD configurations of the test plant. The generated distillate was drained constantly, while the concentrate was fed back to the feed tank. Thus, the concentrate concentration increased over time. Table 5 depicts the parameters that were chosen for these investigations. A higher cooling water volume flow could not be applied, as previous experiments showed that in this case, the feed temperature could not be kept stable for a sufficiently long time. Furthermore, it turned out that in the VMD configuration of the test plant, the feed tank contracted at a pressure below −0.15 bars. Since the employed vacuum pump could not be monitored automatically, the duration of the VMD investigation had to be restricted to 10 h for safety reasons.For the determination of the total bacteria count in the distillate, the plant was first sterilized for 10 min using a 2 wt% hydrogen peroxide solution. Afterwards, the peroxide solution was rinsed out, and the plant was operated for 24 h at constant operating conditions prior to sampling of the feed solution and distillate. Table 6 indicates the operating conditions.The microbiological investigations took place in an external lab. The total bacteria count of the feed water was measured according to the Germen Trinkwasserverordnung TVO based on DIN EN ISO 6222. The total bacteria count of the distillate was measured according to TVO and Ph. Eur. The examination based on the Ph. Eur. served to check whether the distillate met the requirements for PW or WFI according to Ph. Eur. The decontamination factor DF was calculated using the data of distillate and feed water determined according to TVO.The influence of the peroxide solution on spacer, condenser sheet and hydrophobicity of the applied membrane was examined. All materials were inserted for 24 h in a 2 wt% H2O2 solution, rinsed and then dried. Spacer and condensation film were visually inspected. A change of the contact angle of the membrane indicates whether the sterilization solution negatively affects the necessary hydrophobicity of the membrane. Thus, a drop test with distilled water was performed on untreated and treated membranes. For this, distilled water was dripped to the membrane sample with a pipette. The contact angle of the drop with the respective membrane sample was determined graphically.Figure 4, Figure 5 and Figure 6 show the results of the full factorial parameter variation.Figure 4 depicts the influence of the feed temperature on the transmembrane distillate flux. Generally, the distillate flux increases linearly as the temperature rises.The influence of the feed volume flow on the transmembrane distillate flux is depicted in Figure 5.It is apparent that generally, the distillate flux increased with increasing feed volume flow, but the influence of crossflow conditions in the feed chamber on the distillate flux became weaker, as indicated by the slope of the curves, which become flatter with increased feed volume flow.Figure 6 shows the influence of the cooling water flow on the transmembrane distillate flux.With increasing volume flow of the cooling water, the distillate flux increased for almost all factor level combinations. It is noticeable that for many factor stage combinations, the increase in distillate flow became less evident with the increase of the cooling water volume flow.Figure 7, Figure 8, Figure 9 and Figure 10 show the results of the full factorial parameter variation.The influence of the applied vacuum ∆p on the transmembrane distillate flux is shown in Figure 7.Generally, the transmembrane distillate flux increased with increasing negative pressure. An exception was provided by the measuring series TF1, FF1, FK1, during which the distillate flux seemed to be unaffected by the applied pressure. The increase of the distillate flux was particularly pronounced across almost all factor levels at the lowest applied pressure of −0.15 bar.Figure 8 depicts the distillate flux as a function of the feed temperature TF.For all factor level combinations, the distillate flux increased linearly with increasing feed temperature, as the driving force—vapor pressure—increased.Figure 9 depicts the transmembrane distillate flux as a function of the feed volume flow FF.The increasing feed volume flow led to an increase in the transmembrane distillate flux, as the crossflow reduced the hydrodynamic boundary layer on the feed side of the membrane.Figure 10 depicts the distillate flux as a function of the volume flow of the cooling water FK.For all factor stage combinations, an increase in transmembrane distillate flux with increasing cooling water volume flow can be seen.The temperatures of all solutions could be kept constant over time. During the investigation period, 63.1 kg of distillate were generated. This corresponds to a yield of ca. 49%. Based on the amount of distillate, an average distillate flux of 4.4 L/m2h was calculated. This is slightly lower than the value of 4.7 L/m2h achieved with the same parameters during the factor variation investigations.Figure 11 depicts the resulting conductivity of feed/concentrate and distillate as a function of time.The electrical conductivity of the feed increased linearly during the long-term investigation. Within the investigation period of 48 h, the conductivity almost doubled from 410.7 µS/cm2 to 806.3 µS/cm. The conductivity of the distillate increased over time and followed the conductivity of the feed. The initially low conductivity in the first hour did not reflect the performance of the membrane. Due to the low flow of both configurations, the conductivity sensor was mounted in a filling pipe. At the start of the long-term investigations, the pipe was filled with air. Contact with air at the beginning of the test runs influenced the sensor until it was fully covered with water. The rise and fall of electrical conductivity can be explained by the start-up performance of the MD plant and remaining solution in the membrane modules. The electrical conductivity of the distillate increased from, initially, 0.9 µS/cm to 2.3 µS/cm. Thus, during the entire test period, the distillate had an electrical conductivity, which corresponded to the quality of PW.The calculated salt retention R during the AGMD long-term investigation is shown in Figure 12.Basically, the salt retention R was permanently above 99.7%.The temperatures of all solutions could be kept constant over time. During the investigation period, 9.14 kg of distillate was generated. This corresponds to a yield of ca. 17.8%. Based on the amount of distillate, an average distillate flux of 3.05 L/m2h was calculated. This is slightly lower than the value of 3.16 L/m2h achieved with the same parameters during factor variation investigations.Figure 13 depicts the resulting conductivity of feed/concentrate and distillate as a function of time.The electrical conductivity of the feed increased linearly during the long-term investigation. Within the investigation period of 10 h, the conductivity increased from 435 µS/cm2 to 530 µS/cm. The conductivity of the distillate increased over time and followed the conductivity of the feed. The initially low conductivity can be explained by start-up procedures until the pipe that contained the sensor was filled with liquid. The electrical conductivity of the distillate increased shortly within the first hour, from, initially, 0.7 µS/cm to 1.6 µS/cm. Then, the conductivity decreased again, reached a value of 0.9 µS/cm and remained constant until the end of the investigation. Thus, during the entire test period, the distillate had an electrical conductivity, which corresponded to the quality of PW and WFI.The salt retention R of the VMD as a function of the time is depicted in Figure 14.The initially high salt rejection in the beginning of the long-term investigation is attributed to the insufficient construction details of the applied membrane plant, as described before. Salt retention was permanently above 99.6%. Decrease and increase in salt retention within the first five hours is justified by the course of the distillate conductivity, (Figure 12).The results of the microbiological studies are presented in Table 7. In each case, two samples were taken from the feed water and from the distillate.To determine the decontamination factor with regard to the total bacteria count, the respective results depicted in Table 5 were averaged. From these arithmetic average values, the decontamination factor for the two incubation temperatures was calculated. From this, the mean decontamination factor ∅DF was calculated. The decontamination factor for the total bacteria count is shown in Table 8.About 4% of the total bacteria in the feed could be found in the distillate. Therefore, the MD provided a rejection of the microbiological impurities of approximately 96%. It is noticeable that the decontamination factor for the incubation at 22 °C is significantly lower than that for the incubation at 36 °C.Table 9 depicts the total bacteria count estimated according to Ph. Eur.The total bacteria count in the distillate met the requirements of PW but did not meet the requirements of WFI.The drop test prior to and after the insertion of the membrane material showed that sterilization with an aqueous 2% H2O2 solution did not negatively affect the hydrophobicity of the membrane, as the measured contact angle was >90% in both conditions. The spacer and film material were also suitable for sterilization with the aforementioned solution.Table 10 depicts the effects of the feed temperature TF, feed volume flow FF and cooling water flow FK on the transmembrane distillate flux FD during AGMD.It is obvious that the feed temperature TF is the decisive factor to increase the transmembrane distillate flow. The increase in the feed volume flow led to a reduction in the boundary layer resistance. However, this effect was limited. The effect of the cooling water flow FK on the distillate flux was negligible.Table 11 depicts the effects of the feed temperature TF, feed volume flow FF, cooling water flow FK and applied negative pressure on the transmembrane distillate flux FD during VMD.Similar to what observed for AGMD, the feed temperature TF had the greatest influence on the transmembrane distillate flux during VMD, followed by the feed volume flow FF and the cooling water volume flow FK. The applied negative pressure ∆p had the smallest effect on the distillate volume flow. However, ∆p had to be kept above −0.2 bar, as the feed tank design did not allow lower values. Accordingly, the effect of negative pressure on the performance of VMD could not be predicted.The long-term investigations showed that both AGMD and VMD were able to produce distillates with an electrical conductivity within the limiting values of PW.Noticeable is the dependence of the electrical conductivity of the distillate on the conductivity of the feed during AGMD. It is assumed that an intrusion of the feed solution into the distillate cycle occurred. The intrusion of the feed was caused either by minimal defects of the sealings or by a too high pressure on the feed side of the membranes that caused mass transfer through the pores of the actually hydrophobic membranes (exceedance of the liquid entry pressure LEP). In addition, the investigations showed that a certain maximum concentration of the concentrate limited the process, as the quality of the distillate depended on it. The achievable yield of AGMD thus seems limited.The long-term investigation of the VMD did not show a comparable behavior. The conductivity of the distillate did not increase when the concentration of the feed rose. However, VMD investigations were significantly shorter compared to AGMD ones, so the concentrate concentration did not reach a hypothetical critical value.The achievable distillate fluxes were smaller compared to the respective fluxes measured during parameter variation. It is assumed that the reason is vapor pressure reduction due to the increasing salt concentration of the concentrate. The driving force of membrane distillation during the long-term investigations was reduced accordingly and led to the decrease of the distillate flux.The total bacteria count in the distillate met the requirements of PW but not those of WFI. The measured bacteria counts were 30–40 times higher than the limit value of WFI and 250–330 times below the limit value of PW. However, neither the test plant nor the module fulfilled the demands of hygienic design, as the detachable connections in the product and distillate-contacting area were realized with screw connections.The investigations carried out showed that single-stage MD is suitable for the production of pharmaceutical-grade water. The generated distillate reached the claims of Ph. Eur. for PW regarding the electric conductivity and the total bacteria count.The feed temperature was the decisive factor for the performance of both AGMD and VMD configurations. All other factors played a minor role. VMD showed no significant increase in transmembrane distillate flow with an applied vacuum. The main reason for this behavior is the limited negative pressure on the distillate side that could be achieved in the test facility. Future studies of VMD should therefore be carried out below −0.2 bar.Future studies must also provide information on whether MD is comparable to distillation and membrane-assisted processes in terms of yield and specific energy demands. The decisive factor for this is the decoupling of the quality of the distillate from that of the feed.In addition to the necessary technical investigations, the approval of the legislator is crucial for the establishment of MD as a possible alternative to distillation and cold production by RO in a strictly regulated market. Then, the decision of pharmaceutical manufacturers to use MD will depend on the strict adherence to operational safety. This also includes the approval of FDA or other institutions for any kind of material used in the plant.Further applications of membrane distillation in the field of ultrapure water generation seem to be promising for the treatment of water for hydrogen production. The electrolysis used for water splitting requires a fully de-salted and softened feed water. Hygienic requirements similar to those of the pharmaceutical industry do not exist. Excess heat from the electrolysis plant can be used to heat the feed water.Conceptualization, T.K., H.-J.R. and F.R.; Investigation, C.N. and T.K.; Methodology, C.N., T.K., H.-J.R. and F.R.; Supervision, T.K. and H.-J.R.; Visualization, C.N. and F.R.; Writing—original draft, C.N., T.K., H.-J.R. and F.R. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.Data are contained within the article.In October 2020, Cornelius Nellessen received the DME’s commitment for a scholarship. DME German Seawater Desalination e.V. is a company and university-neutral, non-profit association and the central platform for seawater desalination in Germany.The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.MD test plant.P + I schemes of the test plant. (a) AGMD configuration; (b) VMD configuration.Membrane stack provided by Deukum. (a) General set-up; (b) single membrane spacer; (c) membrane.Distillate flux as a function of the feed temperature during AGMD at a constant cooling water temperature of 17.5 °C. The numbers refer to the different levels, low (1), medium (2) and high (3), described in Table 4.Distillate flux as a function of the feed volume flow (crossflow) during AGMD. The numbers refer to the different levels, low (1), medium (2), and high (3), described in Table 4.Distillate flux as a function of the cooling water flow. The numbers refer to the different levels, low (1), medium (2), and high (3), described in Table 4.Distillate flux as a function of the applied negative pressure during VMD. TF = const., FF = const., FK = const. during the respective measurement series. The numbers refer to the different levels, low (1) and medium (2), described in Table 4.Distillate flux as a function of the feed temperature during VMD. Δp = const., FF = const., FK = const. during the respective measurement series. The numbers refer to the different reference levels, without pressure (0), low (1), medium (2), and high (3), described in Table 4.Distillate flux as a function of the feed volume flow (crossflow) during VMD. Δp = const., TF = const., FK = const. during the respective measurement series. The numbers refer to the different reference levels, without pressure (0), low (1), medium (2) and high (3), described in Table 4.Distillate flux as a function of the volume flow of the cooling water FK during VMD. Δp = const., FF = const., TK = const. during the respective measurement series. The numbers refer to the different reference levels, without pressure (0), low (1), medium (2) and high (3), reported in Table 4.Feed and distillate conductivity during long-term investigations of AGMD. TF = 64.6 °C, FF = 60 L/h, FK = 130 L/h, TK,in = 16.9 °C.Salt rejection during long-term investigations of AGMD. TF = 64.6 °C, FF = 60 L/h, FK = 130 L/h, TK,in = 16.9 °C.Feed and distillate conductivity during long-term investigations of VMD. TF = 56.5 °C, FF = 45 L/h, FK = 130 L/h, TK,in = 13.1 °C, Δp = −0.1 bar.Salt rejection during long-term investigations of VMD. TF = 56.5 °C, FF = 45 L/h, FK = 130 L/h, TK,iin = 13.1 °C, Δp = −0.1 bar.Limiting values for PW and WFI [2,3].1 CFU colony-forming unit. 2 IU international unit.Characteristics of common MD module configurations [14,15].Both sides of the membrane are in contact with a liquid.The extraction of the condensation enthalpy of the vapor takes place on the permeate side in cooling waterSimple operationSimple plant setupSimple module setupHigh fluxHighest heat losses caused by conduction through the membranesNot suitable for the removal of non-volatile organicsThe feed side of the membrane is in contact with a liquid, the permeate side is in contact with a stagnant gas layerThe condensation of the distillate takes place on a separate dense surface (condensation film) adjacent to the air gap, which is cooled by cooling waterLow conductive heat lossesLow fouling tendencyNo wetting on the permeate side of the membraneAdditional resistance to mass transferComplicated module designLow fluxThe feed side of the membrane is in contact with a liquid, the permeate side is in contact with a sweep gas stream (e.g., air)Water vapor condensation takes place outside the moduleNo wetting on the permeate side of the membraneLow thermal polarisationLarge condenser required (small amount of permeate in a large volume of sweep gas)Low fluxThe feed side of the membrane is in contact with a liquid, the permeate side is under negative pressureWater vapor condensation takes place outside the moduleHigh fluxLittle conductive heat lossesHigher risk of pore wettingVacuum pump and external condenser requiredMembrane characteristics. Product Datasheet Lydall Performance Materials, “14P02F: Microporous Polyethylene Film”.Conditions of factor variation investigations.Operating conditions of long-term investigations.Operating conditions prior to microbiological sampling.Total bacteria count according to TVO, (CFU/mL).Decontamination factor.Total bacteria count according to Ph. Eur.Effects of the parameters on the transmembrane distillate flux FD during AGMD.Effects of the parameters on the transmembrane distillate flux FD during VMD.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06059.txt ADDED Viewed

	@@ -0,0 +1,9 @@

+Just-in-time adaptive intervention (JITAI) has gained attention recently and previous studies have indicated that it is an effective strategy in the field of mobile healthcare intervention. Identifying the right moment for the intervention is a crucial component. In this paper the reinforcement learning (RL) technique has been used in a smartphone exercise application to promote physical activity. This RL model determines the ‘right’ time to deliver a restricted number of notifications adaptively, with respect to users’ temporary context information (i.e., time and calendar). A four-week trial study was conducted to examine the feasibility of our model with real target users. JITAI reminders were sent by the RL model in the fourth week of the intervention, while the participants could only access the app’s other functionalities during the first 3 weeks. Eleven target users registered for this study, and the data from 7 participants using the application for 4 weeks and receiving the intervening reminders were analyzed. Not only were the reaction behaviors of users after receiving the reminders analyzed from the application data, but the user experience with the reminders was also explored in a questionnaire and exit interviews. The results show that 83.3% reminders sent at adaptive moments were able to elicit user reaction within 50 min, and 66.7% of physical activities in the intervention week were performed within 5 h of the delivery of a reminder. Our findings indicated the usability of the RL model, while the timing of the moments to deliver reminders can be further improved based on lessons learned.Inactive lifestyle and a lack of physical activity can lead to serious physical and mental health issues [1]. People are therefore advised to engage in regular physical activity (i.e., at least 150 min of moderate-intensity activity every week for adults [2]). However, many individuals struggle to maintain a healthy activity level. Consequently, the promotion of healthy physical activity behavior is still an open challenge [3]. Interventions using mobile exercise applications are considered promising for supporting physical activity behaviors, as mobile devices are well integrated into people’s daily lives [4] and can continuously deliver interventions to users [5].Recently, just-in-time adaptive intervention (JITAI) has gained attention and been considered to be an effective strategy in the field of mobile healthcare intervention [6]. JITAI is an intervention design aiming to provide the right type and amount of support, at the right time, by adapting to an individual’s changing internal and external state [7]. In the JITAI framework, identifying the right moment for intervention is a crucial component. As argued by Fogg, in daily life, the personal state that the user is involved in when an intervention is delivered can play an important role in determining its effectiveness [8]. For instance, a reminder about physical activity sent during a meeting will likely be ineffective, no matter how persuasive its content might be.JITAIs in mobile health applications are mainly used to prevent certain health threats, including addictive behaviors such as overeating [9,10], smoking [11] and prolonged sedentary behaviors [12]. In other words, that research mainly concentrates on finding ‘vulnerable states’ in which a user is urged not to perform a certain behavior. A few recent systems have started to research the right moments of ‘opportunity’ to promote a certain behavior, although they are still at the early stage. Ding et al. [13] described the potential of such a system and demonstrated how sending reminders at the right moments could improve the persuasiveness of mobile interventions. In their study, researchers set predetermined opportunistic moments in which the user is invited to walk (e.g., when the individual is overusing a smart phone or when he or she is sedentary). Expanding on this previous research, we proposed the incorporation of a data-driven ‘moment of opportunity’ in a smartphone application to promote user physical activity.However, identifying the ‘opportunity’ moments from data to deliver interventions in mobile applications is demanding, as the right moments for an individual can change according to personal momentary needs. One possible solution is to use the reinforcement learning (RL) technique, as it can continuously adapt the intervention strategy based on both the momentary state and the feedback of users [14]. RL has been used in mobile applications to support a certain physical activity. Most researchers have continuously adapted the content or type of interventions by tracking the momentary behaviors of people. For instance, Rabbi et al. [5] recommended different types of physical activity adapting to users’ changing needs using a bandit algorithm. Yom-Tov et al. [15] personalized the contents of messages for patients with diabetes to encourage their physical activity. Zhou et al. [16] delivered interventions to users suggesting adaptive and personalized daily step goals. In an intervention focused on weight loss, Forman et al. [17] optimized the combination of text messages and human coaching intervention based on a cumulative reward score. Conversely, several researchers have concentrated on using RL to adapt to the moments of intervention delivery with respect to users’ on-going status and behavior. In one recent work, Liao et al. [18] developed the HeartSteps application to actively decide whether to provide interventions at five fixed moments per day. However, too many interactions over a short period (e.g., 5 times per day) could add burden for the user and adversely impact engagement [19]. Driven by this practical issue, we allow a limited number of reminders within each week in our smartphone application, and use RL to identify the ‘right’ moments for sending a restricted number of reminders. As far as we know, this has never been studied in mobile health intervention settings.In this paper, we concentrate on motivating users to perform a physical activity by delivering reminders at personalized moments. The moments when reminders are sent depend on the personal momentary context of users (i.e., time and calendar). To examine the usability of such an RL-based smartphone exercise application, a feasibility study with real target users was conducted and presented. We aim to explore both the user behavior (i.e., interactions with the reminders) and the user experience after receiving reminders delivered at the personalized moments. Thus, the following three hypotheses were addressed and examined:Reminders sent at personalized moments can trigger active reactions within a short period of time.Reminders sent at personalized moments can trigger users to start a physical activity.Reminders sent at personalized moments are perceived as useful by users, based on user feedback from questionnaires and interviews.Reminders sent at personalized moments can trigger active reactions within a short period of time.Reminders sent at personalized moments can trigger users to start a physical activity.Reminders sent at personalized moments are perceived as useful by users, based on user feedback from questionnaires and interviews.To examine the proposed hypotheses, we first designed and integrated an RL model in a smartphone exercise application to deliver reminders based on users’ momentary temporal context. Then, we conducted a feasibility study, in which 11 target users participated. At the end, we collected and analyzed data from 7 users who used the smartphone exercise application for 4 weeks and received reminders sent at moments determined by the RL model. The 7 users were also asked to fill out a questionnaire and have an interview to share their user experience.In our study, a smartphone exercise application (the PAUL app) was developed to send reminders adaptively at moments suggested by a reinforcement learning (RL) model. The development of the PAUL app was conducted based on both theoretical foundations of behavior change [20] and a focus-group study with our target users [21]. The design process and practical implementation of this smartphone exercise application can be found in [22]. In this subsection, we mainly describe the integration of a reinforcement learning model in the PAUL app.In Figure 1, we present an overview of how the PAUL app iteratively interacts with the user to decide the right moments to deliver reminders, where the RL model is marked in green. In this interaction procedure, the RL model is a decision-maker in the smartphone exercise app. On the one hand, the RL-based decision-maker determines the momentary status of a user based on personal context information collected by the app at each time step, then estimates whether this is the right moment to send a reminder to this user. Furthermore, the RL-based decision-maker records the physical activity behavior of the users after receiving the reminders. In this manner, the RL model learns the preference of this user from his or her historical data, and keeps the state-of-the-art knowledge to make the next decision.We decided that the PAUL app should decide on whether to send a reminder or not at every hour from 8:00 to 20:00 in each day. Additionally, since too-frequent interactions with the user are not desirable in our practical task, we set the maximum number of reminders sent. In general, the PAUL app sends a maximum of 14 reminders each week. The content of the messages were drafted based on previous literature [23] and evaluated in a pilot test. Based on the results of questionnaires, a message library of 141 messages was built. The messages are positively framed, focus on effectiveness and immediate outcomes, and are tailored to the activity type (i.e., running, walking or both).We modeled our practical task (i.e., how to learn the optimal strategy for delivering reminders with respect to the momentary context of this user) as a reinforcement learning problem. For each user, we formalized the problem as a Markov decision process (MDP). The MDP framework is a mathematical abstraction of sequential decision-making problems [14]. Figure 2 depicts the interaction between an agent and an environment in our MDP. Here, the agent represents a mobile health system that learns the optimal strategy to interact with a target user, which is our environment. Our agent and environment interact in a sequence of discrete and finite time-steps {1,2,…,t}, which can be naturally broken into episodes. At each time step, the agent observes the contextual representation of the environment, and on that basis selects an action. Afterwards, the environment passes a numerical reward (inferring its feedback on the given action) back to the agent, and updates itself in a new state. Based on this trial-and-error mechanism, the agent adapts its policy for maximizing an expected long-term reward. We follow the standard definition in [14] and define the key elements of our MDP as follows:S: A set of observed states, each state st∈S is a vector of contextual variables of the environment (i.e., user), including ‘the number of reminders left in this week’, ‘timestep from the last run’, ‘timestep from the last reminder’ and ‘hour of the day’, ‘weekday’, ‘calendar availability’.A: A set of actions taken by the agent (i.e., app), two actions, ‘sending reminder’ and ‘not sending reminder’, are denoted as at=1 and at=0 respectively.T: A probabilistic transition function, defining the probability of moving from st−1 to st given a specific action at−1.R: A reward function, representing the probability that a reward rt was received by the agent. If the target user performs a physical activity at time t, the reward rt=1; if not, rt=0.S: A set of observed states, each state st∈S is a vector of contextual variables of the environment (i.e., user), including ‘the number of reminders left in this week’, ‘timestep from the last run’, ‘timestep from the last reminder’ and ‘hour of the day’, ‘weekday’, ‘calendar availability’.A: A set of actions taken by the agent (i.e., app), two actions, ‘sending reminder’ and ‘not sending reminder’, are denoted as at=1 and at=0 respectively.T: A probabilistic transition function, defining the probability of moving from st−1 to st given a specific action at−1.R: A reward function, representing the probability that a reward rt was received by the agent. If the target user performs a physical activity at time t, the reward rt=1; if not, rt=0.In our case, we aim to optimize the frequency of physical activities in a week with the PAUL app. Following the setup of sending reminders in the PAUL app, the RL model decides on which action (sending reminder or not) to take at every hour from 8:00 to 20:00 in each day. In total, there are 84 decision time-steps in each week. If a physical activity is performed by the user before the next decision time step (within one hour), we gave a reward of 1.0 to our app (otherwise zero reward is assigned). Since we allow a constrained number of reminders, our optimization goal becomes how to wisely deliver the maximum number of reminders at the ‘right moments’ in an episode to maximize the user’s physical activity frequency in that period.In this section, we explain one important property of our RL model. A problem that many RL-based mobile intervention systems suffer from is the cold-start problem [24]. As very few (or even no) experiences with the user are available at the beginning, RL models often require the application to interact many times with the users prior to performing well (learning how to deliver the reminders at the right moments in our case). To remedy this problem, we pre-learned a generalized initial delivery strategy in our RL model from the large-scale empirical running data. An overall framework of our RL smartphone exercise application is shown in Figure 3.Following this procedure, we first clustered users from a large-scale historical running data and identified the target users [25]. The used running data contains around 440,000 runs over 4 years performed by more than 10,000 users. In the data analysis, we characterized the target users as people who struggled with maintaining running behavior (it is consistent with our prioritized participants in the feasibility study). Moreover, as each data point includes the timestamp and weather information at the start moment for every run, we also captured when and under which weather conditions our target users prefer to start runs. Based on those preferences, we then developed an RL algorithm to learn an initial strategy using this empirical data in simulations [26]. In this manner, we aim to obtain a generalized strategy that delivers a limited number of reminders to our target users at reasonable moments (better than a random initialized one in code-start RL models).The feasibility study consisted of three steps, namely enrollment, intervention and feedback. First, in the enrollment step, we recruited the participants, visited all eligible participants and asked them to fill out a questionnaire to assess their socio-cognitive characteristics. Second, our participants used the PAUL smartphone application in their own smartphone from 18 November to 15 December 2019. They were asked to turn on the reminder function of their smartphones (i.e., to receive pop-up notifications) and to give access to their digital calendar (i.e., to provide calendar availability data, but they did not have to). We separated this 4-week study period into 2 phases: baseline (3 weeks) and intervention (1 week). In the first 3 weeks, participants had access to the PAUL app, but no JITAI reminders were sent by the app. These initial three weeks without the JITAI reminders allowed us to ensure the usability of basic functions in the app and to collect the baseline behaviors of participants. Then, in the final intervention week, the PAUL app delivered a maximum of 14 reminders to each participant, asking them to perform a physical activity (either a walk or a run). The moments of delivery were decided by the RL model based on the user’s personal momentary information (i.e., time preference and calendar availability). During the whole period, all participants were encouraged (but not obliged) to use the PAUL app as much as possible, and several data on participants’ behaviors (e.g., interactions with the app) were monitored and collected automatically by the PAUL app. Third, at the end of the study, all participants answered a questionnaire on the usability of the reminder function in the PAUL app. In this feedback step, we also conducted a one-to-one interview (20–40 min) with each participant, to understand their experience and feelings on using the PAUL app in the manner we designed.After being granted approval from the local ethics review board (No. ERB Review Geo S-19253), we started our study by recruiting participants. To maintain a healthy lifestyle, people are advised to engage in a sufficient amount of physical activity on a regular basis. Therefore, our prioritized participants are people who are struggling with maintaining a healthy activity level and would like to participate in more physical activities. During the recruitment, we also accepted participants who perform physical activity regularly but still need to move more to reach the suggested amount.To control the influence of living location on user behavior, we selected three parks in city centers of the Netherlands (i.e., Transwijk at Utrecht, Oosterpark at Amsterdam and Sloterplas at Amsterdam) and only recruited participants who live close (less than 5 km and 20 min bike ride) to either park. We built up a website to introduce the relevant information of our study for interested individuals and to include an online eligibility questionnaire, which is used for assessing their stage of healthy behavior change (as defined by the Transtheoretical Model (TTM) [27]). Our stakeholders in the project also aided with the recruitment by contacting neighborhood organizations (e.g., promotion materials such as flyers were placed in public areas around the parks). We also recruited participants via Facebook advertisements. Based on the participant applications, we enrolled 11 participants who were aged between 18 and 55 years and met the following criteria: not meeting the physical activity guidelines of 150 min per week, with no medical condition preventing them from performing PA (defined by PAR-Q [28]), owning an Android smartphone, not currently participating in another PA or health-related intervention and have proficient knowledge of the Dutch language (as the app is in Dutch). In total, 7 of 11 participants completed the whole study and received reminders sent at personalized moments by the RL model, whose demographic information is presented in Table 1.During the study, our PAUL smartphone app automatically tracked the physical activities performed by the participants and recorded their reactions after receiving reminders. On the one hand, the data collection of physical activity starts when the user clicks the ‘start exercise’ button of the app and then continuously tracks all the data involved in a run in the background, until it is terminated by the user in a comparable manner. For each physical activity, a dataset is collected summarizing total distance and run time, and marking the timestamp and GPS location at the start point of this physical activity. On the other hand, the data collection of users’ active reactions to reminders starts when they give positive feedback. The positive feedback can be given by either tapping on the pop-up notification corresponding to the reminder or opening the application in a short period of time. Otherwise, we consider users ignore the reminders because they were not sent at the right moments. Based on the collected data, we address the following two outcome measurements.The number of reminders actively reacted by a user: a user was considered to ‘actively react to a reminder’ if he or she taped on the pop-up notification after this reminder was delivered.The number of physical activities triggered by a reminder: a physical activity was defined as ‘triggered by a reminder’ if the user performed a physical activity after he or she actively reacted to a delivered reminder.Using those measurements, we aim to quantify user behavior after receiving our designed reminders and examine the first two hypotheses in Section 1.In addition, to explore the experience of users on receiving the reminders and to answer the third hypothesis, we designed several questions for a questionnaire and an individual interview based on previous studies [13,29,30]. These questions also concentrate on the receptivity of reminders, which is a concept that anticipates a subjective overall reaction of the user to a notification [30]. We particularly measure two underlying factors of receptivity, which are the willingness to be interrupted (known as reachability) [29] and the likelihood of influencing the recipient’s future actions (known as actionability) [30]. All interviews were audio-recorded and transcribed. In Table 2, we present the questions asked in the questionnaire and the interview respectively (the original ones are all in Dutch). Among them, all questionnaire questions are multiple choices with a 4- or 5-point Likert scale (see the scale details in Section 3.2).In this study, we mainly used descriptive analysis and analyzed the data collected from three resources (including the PAUL application, a questionnaire, and interviews). We not only statistically described user objective behaviors from the PAUL application data, but also conducted the descriptive analysis to understand user experience from the questionnaire and interview data.Regarding the PAUL application data, two datasets were collected and analyzed. On the one hand, a log dataset collected the app usage events of every participant and particularly marked the timestamp of those events. Table 3 gives examples of our collected data in the log dataset. Following two defined outcome measurements in Section 2.2.3 (i.e., reminders actively reacted to by a user and physical activity triggered by a reminder), we tracked the objective behaviors of users after receiving reminders from the PAUL application. On the other hand, a session dataset contained detailed information about all physical activity sessions (i.e., user ID, starting timestamp, duration, length, etc.).During this process, to make sure the data analysis is valid, we confirmed the consistency of collected data. For instance, a walking or running session data entity was included in the analysis only if its starting timestamp was marked as the same in both datasets from the PAUL app. Additionally, to determine the reliability of the user experience data, we conducted a cross-check between answers in the questionnaire and interviews of same participants.In this section, we present the results of the usability of our RL-based application, including user reaction times and physical activity behaviors, as well as their perceived feedback.In this section, we present the quantitative results based on the quantified outcome measurements defined in Section 2.2.3, which are mainly related to users’ objective behaviors.First, we summarized the physical activities performed by our participants, where the records with less than 60 s were excluded (i.e., the duration of run or walk is less than one minute). We realized that 6 out of 20 physical activities performed by all 7 users were in the intervention week (i.e., accounting for 30%), which is higher than the average in all four weeks. Meanwhile, in total, 79 reminders were sent to our participants in the intervention week. On average, each user received 11.3 reminders per week.Meanwhile, we present the results of two kinds of reactions performed by all users in Figure 4. We observed that participants tend to behave in a diverse manner after receiving the designed reminders. For instance, user No. 3 actively clicked on most of the pop-up reminders sent to him or her, while none of them transferred into an actual physical activity performance. Conversely, although user Nos. 2, 4 and 7 only reacted to a few reminders (by clicking them), they did seem to be triggered by those reminders and performed physical activities afterwards. To demonstrate the detailed behavior of participants, we present the behavior of user No. 2 during the intervention week in Figure 5 (who was the most active participant in our study).In addition, by summarizing the data of our participants, we noticed several common behaviors. First, we observed that 4 out of 6 physical activities in this week were performed within 5 h of the delivery of a reminder (accounting for about 66.7%). Those 4 physical activities were contributed by 3 different participants (i.e., user Nos. 2, 4 and 7). Interestingly, user No. 7 only ran one time during the whole feasibility study, which was triggered by our reminder. Second, we found that 18 out of 79 reminders received a reaction from our participants by clicking on the pop-up notification (accounting for 22.8%). Importantly, every participant reacted to at least one reminder. The distribution of the interval time between the reaction (timestamp to click a reminder) and delivery of a reminder is calculated and given in Figure 6. From the figure, we can observe that most reminders (83.3%) were actively reacted to by our participants within 50 min.In this section, we present the qualitative results based on a questionnaire and an interview. We separate the detailed results of the questionnaire in Figure 7 (Q1–Q6 related to the timing of the reminders) and Table 4 (Q7–Q9 related to other subjective experience of the reminders), while we jointly discuss the findings from the questionnaire and the interview.In summary, all participants indicated that it was very important to receive reminders to initiate a running or walking session. Some explained that reminders ‘confront’ them with not doing the exercise and reaching their goal (user No. 5) or that they would otherwise forget to go for a run or walk (user No. 7) or to use the application (user No. 3). One of the participants also claimed that although the reminders of applications sometimes annoyed her, she did think they are useful, because you would not download such an application if you do not need to be pushed:
+            “...no signal at all is not a kick in the ass, and apparently you need it, that’s why you need an app, yeah, and you install it”
+          Furthermore, our participants also highlighted that the right ‘timing’ of reminders is significant to make sure they are effective. If the messages are continuously sent at times that people cannot engage in physical activity, they can end up feeling annoyed, or even feel discouraged and disappointed in themselves for not being able to do more activity, while they do want to do it:
+            “And every time you think, “Now I can’t, now I can’t.” [sighs] And that feels like giving up. […] That’s just demotivating for, um, using the app very much. And not necessarily for moving. That you think after a while: “Oh I remove the app, because this is not nice anymore”. Like that. I didn’t have that yet, but it can happen.”
+          According to Q4 in Figure 7, more than half of the participants (n = 4) noticed most reminders immediately after their delivery (including 2 participants almost always noticed them). This result is consistent with the quantified results of number of reminders actively reacted by users shown in Figure 6, which indicates that our reminders were sent at the right timing of ‘capturing the immediate attention of a user’.In addition, we found that three participants did not perceive the reminders to be sent at the right times (see Q1 in Figure 7). Two of them also stated that there were too few messages (one participant even claimed that he or she never received any reminders), which might explain their negative feedback about the delivery timing. In other words, there is a possibility that the reminders were not sent at times they needed them (e.g., in the first 3 weeks or as frequent as they wanted), instead of at bad times. We also noticed that our reminders did not always motivate people to perform a physical activity. Only about 57% of participants (n = 4) perceived the reminders can motivate them to engage in physical activity (see Q2 in Figure 7), which is similar to the number of participants who actually performed a physical activity during the intervention week (n = 3) (see Figure 4). Interestingly, some participants discussed the reasons that they thought the reminders were not sent in good times in the interview. One important factor is the weather. Since the Dutch weather tended not to be attractive during our feasibility study (e.g., windy, rainy and short daytime), a few participants claimed that they were not very motivated to perform in such weather conditions.Most participants did not feel annoyed (n = 4) or interrupted (n = 4) by the reminders (see Q5 and Q6 in Figure 7), while they perceived the reminder function (e.g., only one participant turned off the reminders for a couple of days in Table 4). Thus, although the reminders were not always sent at optimal times, they also did not frustrate people to a degree that would result in stopping using the application or perform physical activity.During the interviews, different topics arose regarding the frequency of the reminders. In general, it seems the preferred number of reminders is related to the goal of the participants. Some participants want to plan their activities and receive a certain number of reminders based on their own settings. However, other people, who do not like to plan their activities, prefer to have reminders daily, so they can decide at that point of time whether they feel like running or walking. See the following answers from the participants:
+            “Well, if I set a goal of three times a week, I want a reminder three times a week. And that I can choose on which days, so I can schedule it in advance.”
+            “Um, I think one every day, because otherwise a day will soon go by. Say with one of those apps, I’ll miss it quickly. So if there is such a reminder, I thought “oh yeah, I should do that too”. And if you don’t, after all, chances are I won’t do anything about it that day. And if I got that reminder, I’m more likely to think ’oh yes, I had this as my goal, I have to do this for a while’.”
+          In this paper, we follow just-in-time adaptive intervention (JITAI) and propose to send reminders at adaptive times based on users’ momentary temporal and calendar information. We first integrated the reinforcement learning approach in a mobile exercise application for determining the right moments to deliver a reminder. Then, we conducted a 4-week feasibility study and evaluated the usability of such a reminder delivery method. This study demonstrates that sending reminders at personalized moments through a RL-based mobile application is feasible. These JITAI reminders were sent at times that the participants appear to be receptive, but not always at times that the participants had the opportunity to exercise.First, the results shown in Section 3 provided positive evidence for the hypotheses. We found that reminders sent in our personalized moments were able to attract user reactions within a short period of time. For instance, all participants actively reacted to the reminders and over 83.3% reactions were done within 50 min of receiving the reminders. This is an important property to increase the reachability of reminders [31] (introduced in Section 2.2.3), as it can respect the user’s willingness of being interrupted. For instance, if a participant is busy with something, it is hard to interrupt him and let him receive the content of this reminder. Meanwhile, reminders sent in our personalized moments could trigger users to start a physical activity. For instance, an inactive participant only performed one physical activity during the entire feasibility study, which was triggered by our reminder. In the interviews, our participants also highlighted that the right ‘timing’ of reminders are significant to make sure they are effective. Those findings are in line with earlier research that indicated the effectiveness of delivery ‘timing’ in mobile interventions [7,8].Several interesting scientific contributions have also been made in our study. First, it is among the first studies to examine the relationship between the reminder time, the reaction time and the action time (actual performance time) for physical activity interventions. For instance, according to the questionnaire, most participants would like to have about 1 h between the reminder and the actual exercise. In addition, shown in the behavioral results, the reaction time and action time were about 1 h and 5 h respectively. As the reaction and action behaviors could reflect the reachability and actionability of interventions respectively [29,30], we think those two interval times (action time and reaction time) shall be further explored in determining the effectiveness of mobile health interventions. Second, as far as we know, our RL method incorporated a constraint on interaction frequency (i.e., the maximum number of reminders per week) for the first time in RL-based mobile intervention applications. According to the questionnaires and interviews, only 2 out of 7 users expressed that there were too many reminders. Additionally, most of participants rarely felt annoyed or interrupted by the reminders. As the frequency of interactions could add burden for the user engagement [19], we therefore recommend that future applications also restrict the frequency of reminders when developing RL methods for mobile health interventions. Furthermore, we generated results from both quantified outcome measurements and qualitative studies (i.e., questionnaire and interview). By comparing those results, we noticed an interesting phenomenon. A participant claimed that he or she did not receive any reminder in the interview (in Section 3.2); however, based on our data collection, all participants have clicked on the notification of a reminder at least once (see Figure 4). We think this phenomenon was caused by the deviation between physical behavior and awareness of the participant. As discussed in a recent review, the automatic detection from sensing data could reduce the burden of self-reporting [32]. This finding indicated the importance of combining results from the non-invasive methods with the qualitative methods in feasibility studies.Besides the interesting findings, there are still several limitations to our research. First, a general methodological challenge concerns the fact that this is only a small-scale feasibility study, where the number of participants was limited, and the duration of intervention was relatively short. These setups could influence the accuracy of our results. Additionally, due to the short duration, our RL model only used a pre-learned delivery for sending the reminders and was not able to adjust the strategy for each individual participant. Those reasons drive us to conduct a large-scale study with more participants and a longer intervention period. As examining the feasibility prior to large-scale effectiveness testing is an important preparation step in the development phase of a digital intervention [33], we think the study protocol and lessons learned in this feasibility study will be useful in the future study. Additionally, to examine the effectiveness of our reminder delivery function, we would like to cover a comparison between the adaptive timing and the non-adaptive timing in the large-scale study. Second, although we did not set up any gender bias during the participant recruitment, we are aware that most of our participants were female (accounting for 71.4%). Furthermore, due to the technical restriction, only people who have an Android smartphone could participate in this study. For those two reasons, it may be the case that our results are not generalized enough. We therefore would like to overcome this problem in the future by better balancing the gender of participants and expanding the possible users.The lessons learned from this study also motivate us to improve our intervention design in future research. For instance, as shown from the user experience results, we still need to improve the delivery timing of our reminders for influencing the user’s future actions, which corresponds to the actionability of reminders [30]. As mentioned by some participants, factors such as the weather can be very effective. Previous JITAI studies also indicated that several factors (e.g., weather and personal mood) can play an important role in determining well personalized timing of reminders [5]. We therefore would like to take more momentary context information into account to improve the strategy of our reinforcement learning approach. Second, similar to the preferred timing, the preferred frequency of reminders is dramatically different among participants based on the questionnaire and interview results. For instance, some participants prefer 2 or 3 messages a week, while others would like to have two a day. We think this phenomenon will be taken into account in the future study. One possible solution is to give a different initialized strategy to the different user clusters based on their frequency preference [34].Conceptualization, K.S. and S.W.; methodology, S.W. and H.v.H.; smartphone app development, R.D.D.d.B. and S.W.; feasibility study, K.S., S.W., M.S. and N.N.; data analysis, S.W.; interview, K.S.; writing—original draft preparation, S.W., K.S. and B.K.; writing—review and editing, all authors; supervision, M.S., D.E. and B.K.; funding acquisition, M.D. and B.K. All authors have read and agreed to the published version of the manuscript.This work is funded by Playful Data-driven Active Urban Living project under NWO and SIA grant 629.004.013.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Science-Geo Ethics Review Board at Utrecht University (Geo S-19253, 26 September 2019).Informed consent was obtained from all subjects involved in the study.The datasets for this article are not publicly available because the collected data (e.g., application usage data, interviews) are privacy sensitive. No consent has been given to publicly share this data, but the data can be made available on request for verification purposes. Request to access the datasets should be directed to the corresponding author (s.wang2@uu.nl).The authors especially thank the cooperator MYLAPS for providing the smartphone application dataset. We also appreciate the precious comments from our colleagues Victor Dourado, Paula Castro and Chao Zhang.The authors declare no conflict of interest.Overview of the reminder delivery procedure in the PAUL smartphone exercise application, regarding integration of a reinforcement learning model.The interaction flow of our RL model, where the agent representing the smartphone health application and the environment is a user (Please note that the term reward is a specific entity in the MDP model. In our case, it indicates whether a physical activity is performed after the user receives a reminder sent by the RL model).The overall framework for our pre-learned RL exercise application.The reaction behaviors of all 7 users after receiving reminders from the PAUL app.The reminder delivery and corresponding behaviors of user No. 2 in the intervention week, where a bell represents one reminder, a tick represents one short-term reaction (clicking the reminder pop-up), and a smile represents one physical activity.Distribution of the interval time between reminder delivery and reaction (when participants clicked a reminder).The results of the top 6 questions in the questionnaire, showing perceptions of the timing of reminders. Each unit on the X coordinate represents one participant.The demographical information of 7 participants who completed the feasibility study.Questions in the questionnaire and the interview about user experience on received reminders.Example data records indicate the behaviors of a user after receiving a reminder.The results of other questions in the questionnaire about reminder-related properties.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06060.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Studies from India and several eastern African countries found that the impact of dairy animal ownership on household nutrition varied greatly, depending on the socio-geographic context. The purpose of this study was to examine the association between livestock ownership and household dietary quality in rural Kolar district, India. We collected data from a household survey in four study villages (n = all 195 households of the four villages) of Kolar district, applying a cross-sectional design. Kendall’s rank correlation coefficient was employed to determine the correlation between milk consumption and other dietary variables. Multivariable logistic regression was used to describe the relationship between dairy animal ownership and household milk consumption. Households owning dairy animals more often had access to irrigation (58.3% vs. 25.2%) and were less often woman-headed (2.4% vs. 22.5%). Household milk consumption was significantly correlated with consumption of vegetable variety, egg, and meat (all p-values < 0.05). After adjusting for multiple confounders, the odds ratio of milk consumption between dairy animal-owning households as compared to other households was 2.11 (95% confidence interval 0.85, 5.45). While dairy animal ownership was found to be associated with improved dietary quality, larger households were in a better position to adopt dairy animals, which, in turn, might contribute to better household nutrition.The period 2016–2025 has been declared the Decade of Action on Nutrition by the United Nations [1]. The underlying reason is that child and maternal malnutrition continues to be a major global challenge as a top risk factor for morbidity and mortality worldwide [2], including India [3]. Although considerable progress has been made over the past several years, in 2016 a high prevalence of stunting (38%), wasting (21%), and undernutrition (36%) among children under 5 years of age were recorded in India [4]. Dietary diversity is an important determinant of the nutritional status of children [5,6]. Milk is an important source of animal protein, especially for under 5-year-old children [7,8]. However, analysis of nationally representative data revealed that 5–15% of children in rural areas were at risk of quality protein deficiency, worsening to 26–42% among adults in poor households [9]. Livestock ownership, especially dairy animals, has been promoted in rural India to support livelihood and nutrition [10].Literature on the impact of dairy animal ownership on diet and nutrition is available from India and several eastern African countries. Livestock ownership was estimated at 59.7% for households across rural settings in India in 2016 [4], and was deemed an important factor in determining consumption of animal-sourced foods [5]. The linkage between dairy animal ownership and milk consumption was specifically emphasized [11]. This was corroborated by studies from Ethiopia [8] and Uganda [12]. However, several factors such as globalization, urbanization, changing access to education, livelihood and progressive reduction in farm sizes are impacting agriculture as a viable livelihood option, complicating its linkages with food security and dietary quality in rural areas, as was found in a study in northern India [7]. Moreover, little is known regarding the association between dairy animal ownership and household milk consumption in villages in southern India.In a recently conducted qualitative study in Kolar district, a rural semi-arid area in southern India, we found that households that adopted dairy animals for livelihood through the support of watershed development (WSD) projects perceived positive impacts on milk consumption and dietary quality [13]. This pointed to a need for additional quantitative research to deepen the understanding of context-relevant factors, especially keeping in mind the planned promotion of livestock rearing as part of WSD projects in India [14]. Hence, the basis for the current paper was set, and the opportunity for this analysis presented itself when we conducted a cross-sectional baseline survey, as part of a health impact assessment (HIA) of a planned WSD project in four villages of the Kolar district [15]. The specific objectives of this paper were to: (i) describe the status of socio-demographic factors based on dairy animal ownership; (ii) assess the status of health determinants in the study villages based on dairy animal ownership; (iii) determine the correlation of milk consumption with other dietary variables; and (iv) examine the association between livestock ownership and milk consumption in the designated project area.A cross-sectional survey was conducted in the study area during April to July 2019 to characterize socio-demographic factors and health determinants. As the number of people living in each village was relatively small, all households were invited to participate. Data from the cross-sectional survey were used for multivariable logistic regression analysis.The study was conducted in four neighboring villages in the southern part of Malur sub-district of Kolar district in Karnataka state, India. The total population of these villages was officially enumerated at 1340 individuals in 2011 [16]. The main economic activity is agriculture, with finger millet being the most important crop [16]. The region is drought-prone and vulnerable to climate change [17,18]. The sub-district is in close proximity to Bengaluru, but is largely rural, and has a literacy rate of 62.6%. One out of 10 people (9.4%) belong to scheduled tribes (ST), and one quarter (25.4%) belong to scheduled castes (SC) [19]. In the study villages, 30.8% were of ST and 7.7% were of SC. These villages had access to water, electricity, and sanitation, whereas one village did not have an anganwadi (creche). Only 2.1% of households in the project villages reported never consuming meat [15,20]. There was a perception among some households that milk of cattle of high-yielding varieties was unhealthy to consume [13]. Often it was the women of the households who were involved in managing the cattle and milking them. The richer households grew fodder crops, and poorer households used finger millet fodder and also bought fodder to feed the cattle [13], while some pastures were also reportedly available in these villages (6.3% of total land) [16]. Almost all the collected milk was sold to the local dairy, with a little bit for household consumption, or sold to neighbors if they requested [13]. The road connectivity to the nearby town was good, though some of these villages were heavily dependent on private vehicles to access the town [15]. The prevalence of underweight among children was found to be 26.5% in these villages [15,20], which is comparable to the rest of rural Kolar (28.5%, in 2015) [21]. Among 15- to 49-year-old women in rural Kolar, the prevalence of overweight or obesity was 15.2%, and that of anemia was 45.9% [21]. Often one or more members of the household, especially from SC background, traveled outside the village for livelihood and better remuneration [13,15].These four villages were chosen for the study because a WSD project was planned for the villages by a local non-governmental organization (NGO). These specific villages are geographically part of a micro-watershed, which had been identified for future WSD interventions by local governmental departments. As part of the baseline studies prior to implementing the WSD project, a survey was carried out by our team to deepen the understanding of the social and health situation in the villages, and the relevant data were used for the current research. As several similar villages in this semi-arid region would be subjected to WSD projects, studies of this nature can provide context-specific insights. In addition, the potential of WSD projects to incorporate health aspects such as nutrition was identified by a technical committee in 2006 [22], and hence, such studies can guide nutrition-related activities of WSD projects.WSD projects have been carried out in semi-arid areas of India by the government, often in partnership with NGOs, aiming to enhance soil and water conservation, improve agricultural productivity, and support livelihoods through approaches such as livestock rearing, especially among the landless households [14]. There are other context relevant interventions conducted as part of these projects by the local project-implementing NGOs, for instance, supporting self-help groups [23]. It was opined by project managers that beneficiaries often opted for high-yielding milch cattle as they were perceived to be a sustainable livelihood source [13].A structured and pilot-tested questionnaire covering topics of household demography, occupation, agriculture, diet, sanitation, and access to healthcare was used. The questionnaire was administered by trained field enumerators, and data were directly entered into electronic tablets on the Open Data Kit (ODK) platform [24]. Any woman of the household (aged ≥ 15 years) was requested to be the respondent (in some cases this was not possible, so a man became the respondent). Further details of the baseline survey methodology are available elsewhere (see reference [20]).Data were downloaded from the ODK Aggregate platform and read into R statistical software version 3.5.1 (on RStudio Version 1.1.456) [25]. Variables were summarized to find missing and inappropriate values. Categorical variables were checked for small cell sizes. We examined the associations of socio-demographic factors and other health determinants with ownership of dairy animals—which is the main explanatory variable (binary) for the regression analysis. Dairy animals were defined as high-yielding variety cows, local variety cows, and/or buffaloes. Because all households in the study area were included in the study, confidence intervals (CIs) were not calculated for the descriptive statistics. Kendall’s rank correlation coefficient (tau) was used to calculate the correlation between milk consumption (binary) and other dietary variables (i.e., vegetable consumption, fruit consumption, egg consumption, and meat consumption; each with four ordered categories).Crude odds ratios (ORs) and 95% CIs were calculated for describing the relationship between household milk consumption (the main outcome; binary variable) and a set of pre-determined covariates (i.e., ownership of dairy animal, household size, sex of household head, existence of children in household, caste (scheduled caste or SC, scheduled tribe or ST and other; with SC and ST considered as marginalized groups in most regions in India), land ownership, access to irrigation, self-help group (SHG) membership and a dummy variable for wage labor as main income source. Household income (self-reported) data were available but not used as a covariate in our analysis because of high potential for measurement errors, and literature from India suggesting that income was an unsuitable predictor of nutrition in rural areas [11]. Several included covariates are proxies for income, for instance, land ownership, access to irrigation, and ownership of motorized vehicles. The covariates were based on a hypothesized causal model of household milk consumption in rural areas in southern India, as ascertained from literature [5,7,8,11,12,26] and knowledge of the local context (Figure 1).A multivariable logistic regression model adjusting for various covariates was fitted to the data to better understand the causal relationship between dairy animal ownership and household milk consumption. Variables (ownership of other livestock) showing high correlation with the main explanatory variable were eliminated from the final model (Phi coefficient = 0.69). The ORs and 95% CIs were reported and interpreted. Results were considered significant at p < 0.05. CIs were reported only for the ORs despite having conducted a census, to provide an impression of the uncertainty around the estimate, especially keeping in mind the small population size, to aid in generalizability of study results outside of the study population.Sensitivity analysis was performed by using propensity score matching (PSM) through nearest neighbor matching (NNM) with replacement, using the MatchIt package in R statistical software, to increase the comparability of the two groups based on the following covariates: household size, sex of household head, child in household, caste, land ownership, access to irrigation, SHG membership, and wage labor. The quality of matching was assessed through quantile-quantile (QQ) plots and two sample t-tests for each covariate. As the matching was deemed to have improved comparability (though having expectedly reduced the sample size), multivariate logistic regression was carried out on this data subset to better understand the association between dairy animal ownership and household milk consumption.A total of 195 households were included in the four project villages (response rate: 100%; this includes all households of the project villages that were available in the village during the survey period). Over 93% of the respondents were adult women. The median age of all respondents was 35 years (25th–75th percentile: 27–45 years). Most of the respondents were illiterate (56.9%). Characteristics of the respondents are summarized in Table 1. Less than half of the households (43.1%; n = 84) owned at least one dairy animal. Households owning dairy animal(s) were more often larger than households not owning a dairy animal (median of 5 persons vs. 4), had greater land-holdings (2.75 acres vs. 2.09 acres), increased access to irrigation (58.3% vs. 25.2%), more frequent ownership of a motorized vehicle (92.9% vs. 81.1%) and higher ownership of other livestock (95.2% vs. 26.1%). Households owning dairy animal(s) were also less likely to be woman-headed (2.4% vs. 22.5%), SC (2.4% vs. 11.7%), or have recent history of seasonal migration (2.4% against 9.0%).Details on ownership of livestock in the study population are shown in Figure 2. Those owning dairy animals (local variety cows, buffaloes, and/or high-yielding variety cows) more often had greater variety of livestock (including chicken, goats, sheep, oxen, and/or calves).The status of selected health determinants is summarized in Table 2. One in five households (20.5%) have experienced food insecurity during the past two years. A higher proportion of households owning dairy animal(s) consumed milk (88.6% vs. 62.2%). Latrine ownership was high across both groups (92.8%). Dairy animal-owning households more commonly opted for private healthcare services in case a household member had fever (38.1% vs. 15.3%). Health insurance coverage was only reported by 44.1% of households. Milk consumption was significantly correlated with vegetable consumption (Kendall’s tau = 0.16, p = 0.017), egg consumption (Kendall’s tau = 0.27, p < 0.001), and meat consumption (Kendall’s tau = 0.14, p = 0.049). The correlation statistics are summarized in Table 3.Of the 195 households, 52 (26.7%) reported that they did not consume milk. Covariates that showed strong crude associations with milk consumption include owning dairy animal(s) (OR: 4.50, 95% CI: 2.17, 10.14), household size (OR: 2.00, 95% CI: 1.56, 2.66), woman-headed households (OR: 0.19, 95% CI: 0.08, 0.43), land ownership (OR: 1.67, 95% CI: 1.23, 2.35), access to irrigation (OR: 3.19, 95% CI: 1.57, 6.99), and owning a motorized vehicle (OR: 9.72, 95% CI: 4.04, 25.4) (Table 4).The full model for the relationship between dairy cow ownership (primary explanatory variable) and milk consumption (main outcome) was adjusted by household size (count), woman-headed household (binary), whether general caste (binary), child in household (binary), wage labor as main income source (binary), land owned (continuous), access to irrigation (binary), membership in SHG (binary), and ownership of motorized vehicle (binary). The multivariate logistic regression model output indicated that the adjusted OR for household milk consumption was 2.11 (95% CI: 0.87, 5.45) between households owning and not owning dairy animals. Evidence of association was found for household size (adjusted OR: 1.88, 95% CI: 1.34, 2.77), ownership of motorized vehicle (adjusted OR: 4.08, 95% CI: 1.23, 14.31) and wage labor as primary income source for family (adjusted OR: 2.89, 95% CI: 1.04, 9.03).The sensitivity analysis included 128 households (84 households owning dairy animals and 44 not owning dairy animals). Previously observed imbalances between households owning and not owning dairy animals were minimized through the matching for all included covariates, except for household size (difference persisted after matching at p = 0.03). The adjusted OR for milk consumption was 2.20 (95% CI: 0.77, 6.45) in this subsample.Milk consumption was found to be both significantly correlated with other markers of a diverse and high-quality diet (i.e., vegetable, egg, and meat consumption) and elevated among households owning dairy animal(s) (OR: 2.11), even after controlling for multiple confounders. The association between dairy animal ownership and milk consumption was not statistically significant after adjusting for all covariates (see the column “Adjusted OR” in Table 4), which might be explained by the relatively small sample size (n = 195). Of note, we included all households in the four villages that will be affected by the project. It is conceivable that families owning dairy animals would consume milk when the dairy animal is producing milk. However, families owning dairy animals were, in general, different from those not owning dairy animals. For example, they often owned greater assets (land, access to irrigation, and motorized vehicle(s)) than those that did not. This was also consistent with the finding that woman-headed and SC households owned dairy animals far less frequently. This suggests ownership of dairy animals was associated with an overall higher socioeconomic status in the study area. The investment of purchasing and managing a high-yielding dairy animal may be prohibitive to those without assets [7,12]. However, even after matching for these socioeconomic covariates through PSM, a positive association was found between dairy cow ownership and milk consumption, suggesting that this finding is not entirely dependent on socioeconomic status.The advantage of owning land and irrigation access for meeting fodder and water needs of dairy animals was reported by other studies [7,12]. Indeed, some local farmers from nearby villages revealed cultivating only fodder crops in their irrigated fields, focusing solely on dairy for livelihood [13]. This indicates that while dairy animals have the potential to contribute dietary quality and diversity, the impact may be disproportionately higher for richer households, as has been shown in an earlier study [12].The main finding of an association between dairy animal ownership and household milk consumption was corroborated by a large study from India [11], and also smaller studies from Ethiopia (23% increased frequency) [8], Uganda [12,27], and Kenya [28]. This was found to be especially important in areas without access to markets [8], which was not the case in our study area where dairies have been established.While milk was significantly correlated with other markers of a diverse and high quality diet, it was not the only source of protein and micronutrients in the study area, as has also been reported in literature [9]. There is also consumption of finger millet, pulses (a regular feature in meals), eggs and meat, the latter two being more frequent among households owning dairy animals. These findings are in contrast to what was observed in some villages in northern India where consumption of milk and milk products were found to be more critical to dietary quality [7]. The importance of understanding local context in the contribution towards household nutrition is emphasized [7].The proportion of households that reported having experienced food insecurity during the last two years was similar for households with or without dairy animal(s) (17.9% vs. 22.5%). These percentages do not indicate the frequency and severity of the experienced food insecurity. In addition, it is difficult to draw causal inferences in the context of dairy animal ownership as this is a cross-sectional study.Food consumption at the household level cannot be extrapolated to nutritional status of individuals within the household, as shown before [6,11]. A study from Ethiopia indicated positive impact on reducing stunting [8]. Studies from Uganda found significant positive impact [12], no impact [27], or even negative impact [26] of dairy animal ownership on child nutrition, and hence, there must be other contextual factors, such as availability and use of sanitation and intra-household competition for resources. Small ruminants (e.g., goats and sheep) were found to contribute to better nutrition outcomes in Uganda [26] and the poorest households in Kenya [29]. Several other factors complicating this relationship have been elucidated in the literature, including wealth, resource constraints, and experience of financial shocks [8].WSD projects locally have helped overcome the obstacle of high initial investment by providing grants and loans to procure livestock, preferentially to poor woman-headed households through SHGs [13]. Currently SHG membership was somewhat lower among households without dairy animals (33.3% against 41.7%), and this can be expected to improve through the planned WSD project [15]. Beneficiaries in earlier local WSD projects perceived financial and nutritional benefits following the adoption of dairy animal(s) [13]. On similar lines, an intervention study in Rwanda on donation of livestock to households was able to demonstrate impact on child nutrition [30]. However, keeping in mind that managing dairy animals is labor-intensive and harbors various costs [31,32]—including accessing water and feed [7], all households may not be able to adopt it. The varying success of dairy programs in villages in northern India due to the role of availability of land and labor in the household has also been reported [33].Interventions encouraging dairy animal ownership as part of the WSD project should take into account whether it is feasible for low-income households to maintain a dairy animal long-term. Additionally, challenges of water and feed are worsened during droughts [7], which occur regularly in Kolar district. Local anecdotal evidence (assimilated during a recent study [13]) reported that several households sold their dairy animals a few years ago following a period of intense drought. Financial returns from dairy animals were also reportedly lower in areas with high groundwater exploitation [32], such as in the study area. In Tamil Nadu, an “economically transforming” state, it was also observed that smallholder farmers had downsized dairy farming in the 10 years prior to the 2018 study for various economic and cultural reasons [34], and this may have bearing for Kolar, which is a neighboring district of Tamil Nadu. In addition, promotion of dairy animals comes with health and ethical challenges such as antimicrobial resistance, especially for high-yielding varieties [35]. Therefore, this strategy could be reviewed accordingly.We found a strong association of milk consumption and household size (OR: 1.88, 95% CI: 1.34, 2.77) (Table 4), which is in contrast to findings from a large Indian dataset [11]. Two factors might explain this observation. First, wealthy households in the study area lived as joint families, as they have the financial and human resources to buy and manage dairy animals. Second, the poorest households were those of elderly women living alone. The strong association between wage labor and milk consumption (OR: 2.89, 95% CI 1.04, 9.03) may also be related to few households consisting only of elderly poor women living alone unable to engage in wage labor. Reportedly, regular wage labor in construction industry and domestic work in nearby cities was providing adequate returns to young people from this area [13].It is not possible to draw conclusions on causal relationships from cross-sectional data. Reverse causality between household milk consumption and dairy animal ownership is plausible if milk consumption is considered a proxy for wealth/income. Keeping the literature and context in mind, this is unlikely. However, as ownership of cattle was strongly associated with wealth indicators, the association with household milk consumption should be interpreted with caution. Another limitation of the analysis was the lack of data on other milk products. In our preceding work in the same region, we found that part of the milk was consumed in fermented form (curd) [13]. As this curd was made from fresh milk within the household, we assumed it was represented within the data on milk consumption. Finally, the findings of the present study mainly apply to the study area, but may also provide insights on what can be expected in the drought-prone rural regions in southern India.Further research could adopt a prospective mixed-method design, focus on differential benefits experienced by various types of adopting households, and also study challenges being faced by each in taking up and managing dairy animals. Adding outcome measures (e.g., nutritional status among children within the household, hemoglobin levels among adults) as part of the survey would be good to indicate the size and distribution of direct health impacts of these interventions. This kind of evidence is currently lacking [5]. The experience from Gujarat also indicates the need to consider larger economic aspects and cultural dynamics in dairy promotion [36], and such studies with social science perspectives need to be conducted in Kolar’s context.Keeping in mind that the data for the study came from a baseline survey, the additional benefits of conducting comprehensive health impact assessments for agricultural projects was revealed—fostering empirical research in neglected settings [37]. Indeed, baseline survey data can be leveraged to better understand agriculture and nutrition linkages, besides other locally relevant health outcomes.Our study revealed that dairy animal ownership was quite common in the study area (55.9% of the households, as compared to 59.7% for rural India on average [4]). We found evidence suggestive of causal relationship between dairy animal ownership and household milk consumption in the four villages in the southern part of Kolar district. Households consuming milk were found to have a better dietary quality in terms of vegetable variety, frequency of meat consumption and frequency of egg consumption. In terms of the factors associated with adoption of dairy animals, we found that wealth, household size, land ownership and access to irrigation were important. Our findings also illustrated how context plays a role in determining effects of interventions in rural areas, for instance, the effect of household size.Health- and equity-sensitive rural development schemes are needed for achieving Sustainable Development Goals 3 (good health and well-being) and 10 (reduce inequalities). More specifically, a call has been made for development policies to be nutrition-sensitive [38,39]. Both WSD projects and the livestock mission have been recognized for the potential to also address nutritional challenges. Based on our findings and the literature, we recommend that, while there is merit for continued support for livestock programs, there is a need for incorporating contextual insights into program design to ensure that it is relevant and to have realistic expectations of returns in the form of better nutrition and/or improved livelihood. This emphasizes the role of district offices, local organizations and research institutions in the process. The type of livestock would also be an important factor. As considerable resources are put into these initiatives, careful monitoring and evaluation of these interventions and schemes is essential, with periodic revision of interventions based on key findings. One-size-fits-all approaches cannot be expected for diverse contexts.We used the opportunity of a baseline survey of a planned project to contribute to literature and local planning. Further social science-oriented studies can provide further insights on the utility and appropriateness of livestock interventions for improving nutrition. Similar studies from other parts of the country could also further enhance our understanding about agriculture–nutrition interlinkages towards addressing undernutrition.Conceptualization, A.P. and A.F.; methodology, A.P. and A.F.; formal analysis, A.P.; investigation, A.P. and A.F.; resources, A.P., M.S.W., and J.U.; data curation, A.P.; writing—original draft preparation, A.P.; writing—review and editing, A.F., M.S.W., and J.U.; visualization, A.P.; supervision, A.F., M.S.W., and J.U.; project administration, M.S.W.; funding acquisition, A.P., M.S.W., and J.U. All authors have read and agreed to the published version of the manuscript.The first author was a Ph.D. student from India sponsored by the Swiss Government Excellence Scholarships (ESKAS; 2017.0562). This paper was part of his doctoral thesis. No other funding was obtained for this study. The Federal Commission for Scholarships for Foreign Students (Bern, Switzerland) did not have any role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.The study was conducted according to the guidelines of the Declaration of Helsinki, and obtained ethical approval from the Padmashree Institute of Clinical Research in Bengaluru, India (reference no. IEC-BIO-004; approval date: 10 August 2018) and the Ethics Commission of Northwestern and Central Switzerland (EKNZ, reference no. BASEC Nr Req-2018-00839, approval date: 19 October 2018).Informed consent was obtained from all subjects involved in the study.Data is available upon request to the author.We would like to thank Shiva Shankar and his team at the MYRADA Kolar Project for the field support during this study. We also thank the peer reviewers for their constructive and encouraging inputs on the manuscript.The authors declare no conflict of interest.Factors potentially influencing milk consumption at household level in the study area (boxed variables have been included in the analysis); SHG, self-help group.Ownership of livestock in the study population based on a survey conducted between April and July 2019 in four villages in Kolar district, India.Socio-demographic characteristics of the study population from a household survey conducted between April and July 2019 in four villages in Kolar district, India.a 25th and 75th percentile; b one acre = 4046.86 m2.Select health determinants in the study population based on a survey conducted between April and July 2019 in four villages in Kolar district, India.a 25th and 75th percentile.Correlation of milk consumption with other dietary variables based on data from survey conducted between April and July 2019 in four villages in Kolar district, India.* Significant at p-value < 0.05.Crude and adjusted odds ratios (ORs) and a sensitivity analysis (SA) comparing household milk consumption with the explanatory variables based on data collected between April and July 2019 from four villages in Kolar district, India.*** p-value < 0.001; ** p-value < 0.01; * p-value < 0.05; a due to small cell sizes, only a dummy variable for general caste was used in the final model; b highly correlated with ownership of dairy animal, hence not included in final model; NA, not available; SA, sensitivity analysis with data subset determined by propensity score matching; SC, scheduled caste; SHG, self-help group; ST, scheduled tribe.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06061.txt ADDED Viewed

	@@ -0,0 +1 @@

+ The aim of this study was to investigate the characteristics of social health and its association with resilience among older adults living alone excluded from the public care service due to their relatively good health. For this cross-sectional study, we surveyed older adults aged between 65 and 80 years using questionnaires to measure the social health status and levels of resilience of the participants. We conducted a hierarchical regression analysis to confirm the association between resilience and social network. Finally, data from 266 community-dwelling older adults were analyzed. We discovered that participants had social networks with a mean score on the Lubben Social Network Scale 18.13 ± 7.98, which means they were socially isolated. The network size (standardized β = −0.149, p < 0.05) and contact frequency (standardized β = 0.136, p < 0.05) correlated positively with higher levels of resilience. A hierarchical model accounted for 48.0% of the variance in resilience. The results suggested that interventions by the public health service to protect social health are needed for older adults living alone even when they are physically, emotionally, and cognitively healthy. In addition, smaller network size and higher frequency of contacts may be considered to strengthen resilience, which is a protective factor in social health.The burden of caring for elderly parents has been transferred from offspring to the public sector in the Republic of Korea [1,2,3]. This transition began with industrialization, urbanization, and nuclearization of family, and finally affected the traditional family values of supporting older parents. In addition, a rapid increase of aging population due to longer life span and continuous decline in birthrate triggered social issues, such as increasing socioeconomic care burdens caused by the increase in the older adults living alone or with dementia. Social isolation, characterized by living alone, is related to the risk of dementia, and the need to promote social networks is also suggested to prevent dementia [4]. Community care has been suggested for older adults living alone due to their higher rate of unsatisfied needs in social relationships or social participation as well as economic status and health conditions, compared with older adults living with family or housemates [3]. In 2007, the Ministry of Health and Welfare in South Korea established the Comprehensive Support Center for the Elderly Living Alone to administer community organizations that provide care and support services for older adults living alone [5,6]. Since then, various public care projects have been implemented by community organizations nationwide, such as the home helper dispatch service, the direct care worker dispatch service, and the emergency and safety care service [7,8].As of 2019, there were 1.47 million senior citizens living alone in South Korea, which include 350,000 receiving the aforementioned public care services [9,10]. The services comprise mainly of safety support, daily life education, such as health education or nutrition education, and domestic help, focusing on physical assistance [7,8]. These services address the lower levels in the Maslow’s hierarchy of needs but do not include support for social participation or emotional support that correspond to an upper level of needs [11]. As the present public care service for senior citizens is biased towards physical health, the service partially addresses the concept of health, which encompasses physical, mental, and social aspects [12]. Therefore, with the present public care service, there could be unmet demands for mental health services [13]. The social health service also needs to be addressed as a public care service considering the fact that older adults perceive being connected with others as a component of health [14].Resilience is also important for older adults living alone to protect their health status, as enhanced resilience helps older adults to overcome negative life events, which may cause health problems [15,16]. Resilience is defined as the personal ability to adjust to and overcome impacts of adverse events, which restores the stability of life [17,18]. Resilience in older adults is not likely to decline with ageing, but it is associated with various forms of social health, such as family network [15,19], social participation [20], or social support [21]. Given that older adults living alone are exposed to a higher risk of decrease in social participation, poor social support, or loneliness compared with older cohabiting individuals [22,23], studies investigating and protecting social health of those who live alone are needed. However, practical implications for protecting the social health of older adults living alone reported in previous studies are insufficient because the samples inadequately reflected the characteristics of older adults living alone. In addition, the result may be moderated by health issues, because the social network is connected to health issues, such as impaired cognitive function [24,25], physical disabilities [25], and depression [26,27].In this study, we investigated the characteristics of social health in a healthy sample of older adults living alone to understand the status of social health. In addition, we analyzed the relationship between resilience and both network size and contact frequency to assess the importance of social network protection in healthy older adults living alone and to make suggestions for maintaining their social health.We collaborated with the Comprehensive Support Center for the Elderly Living Alone to design a sampling method. We received a list of community-dwelling older adults who participated in the 2018 National survey of the elderly living alone and were excluded from the public care service because they were assessed as relatively wealthy or healthy people. The list included contact information of those who agreed to the use of their information for nonprofit services.We conveniently selected two regions in the Gangwon province, one of the 16 administrative districts of South Korea, for our sample frame. Two regions, located close to the research team office, were selected to allow research team members to visit the participants easily if it was necessary. We selected a medium-sized city and a small town to include residents of both a city and a rural area. The sample frame consisted of 90.80% (n = 4756) of city-dwellers and 9.20% (n = 482) of town-dwellers. This rate reflected the rate of older adults living alone in cities and rural areas of South Korea. Considering the rate of refusal or drop-out, the contact information of 1013 older adults was extracted for this study by using the purposive sampling method.We made calls to the people on the list and introduced the research. When the recipient was interested in participating in the study, after the phone call, we sent them a letter by post to explain the study. In a few days, we made a follow-up call to check if they had read the study explanation and to confirm their intent to participate in the study.We recruited cognitively, physically, and emotionally healthy participants. The exclusion criteria were:Aged older than 80Scores less than −1.5 standard deviation in the Mini-Mental State Examination for Dementia Screening (MMSE-DS) [28]Extremes in any dimension of mobility, self-care, usual activities, pain/discomfort, anxiety/depression in the Euro Quality of Life Questionnaire 5-Dimensional Classification, three-level version (EQ-5D-3L) [29]Scores higher than 9 in the Korean version of Short form Geriatric Depression Scale (SGDS-K) [30]Aged older than 80Scores less than −1.5 standard deviation in the Mini-Mental State Examination for Dementia Screening (MMSE-DS) [28]Extremes in any dimension of mobility, self-care, usual activities, pain/discomfort, anxiety/depression in the Euro Quality of Life Questionnaire 5-Dimensional Classification, three-level version (EQ-5D-3L) [29]Scores higher than 9 in the Korean version of Short form Geriatric Depression Scale (SGDS-K) [30]After confirming the participation intention over the phone, a trained investigator visited a participant for a screening test to confirm that the participant was without unhealthy symptoms in cognition, subjective health, or emotion. The screening tools for this study comprised the MMSE-DS [28], the EQ-5D-3L [29], and the SGDS-K [30]. This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board, and we received written consents from all participants before the survey. The same investigator who conducted the screening test collected the data through a one-on-one interview within a day. Visiting participants’ home was recommended to have an interview in a silent and settled environment. For a few participants who did not want to invite the investigator to their home, we prepared a meeting room located close to the research team office for interview.We used the Korean version of the Lubben Social Network Scale (LSNS) to assess participants’ social health status in this study [31]. Participants’ network size and contact frequency were measured with 6 of the 10 questions in the LSNS [31]. Social relationships are a part of social health. Measuring network size and contact frequency can be used to quantitatively evaluate the social health characteristics [32]. The number of close people or frequency of contacts have been commonly used to understand participants’ social network [24,33,34].Generally, the result of the LSNS evaluates scores of 10 questions or subtotal scores according to social resources. The maximum score of the 10 questions is 50. A score higher than 31 indicates low risk of isolation, whereas scores between 26 and 30 suggest moderate risk for isolation, and a value between 21 and 25 indicates a high risk for isolation. Lower scores (<21) indicate isolation [35]. In this study, we used subtotal scores of network size and contact frequency in addition to a total score of the LSNS. Cronbach’s alpha for 10 questions in the present study was 0.724.We measured the network size with four questions:“How many relatives do you see or hear from at least once a month?”“How many relatives do you feel close to; i.e., how many of them do you feel at ease with when talking about private matters or calling for help?”“Do you have any close friends; i.e., do you have any friends with whom you feel at ease or talk to about private matters or can call for help? If so, how many?”“How many of these friends do you see or hear from at least once a month?”“How many relatives do you see or hear from at least once a month?”“How many relatives do you feel close to; i.e., how many of them do you feel at ease with when talking about private matters or calling for help?”“Do you have any close friends; i.e., do you have any friends with whom you feel at ease or talk to about private matters or can call for help? If so, how many?”“How many of these friends do you see or hear from at least once a month?”The response options included zero, one, two, three or four, five to eight, and nine or more. The maximum total score of the four questions was 20. Higher score indicated larger size of social network.For contact frequency, we used two questions:“Tell me about the relative with whom you have the most contact. How often do you see or hear from that person?”“Tell me about the friend with whom you have the most contact. How often do you see or hear from that person?”“Tell me about the relative with whom you have the most contact. How often do you see or hear from that person?”“Tell me about the friend with whom you have the most contact. How often do you see or hear from that person?”The answer options comprised less than monthly, monthly, a few times a month, weekly, a few times a week, daily or almost daily. The maximum score of the two questions was ten. A higher score suggested a higher frequency of social contact.Multidimensional Individual and Interpersonal Resilience Measure (MIIRM) [36] was used to measure the resilience. The MIIRM was developed to measure older adults’ resilience and the tool contains 22 questions in 8 subcategories of self-efficacy, emotional regulation, emotional expression and communication, optimism, perceived economic and social resources, access to support network, relational accord, and spirituality and religiosity. The 22 questions consist of 9 questions with a 4-point scale, 11 questions with a 5-point scale, and 2 questions with a 10-point scale. The total score for the 22 questions ranges from 22 to 111. A higher score means a higher level of resilience. Concurrent validity of the MIIRM was confirmed with a correlation (r = 0.648, p < 0.001) between the MIIRM and the 10-item Connor–Davidson Resilience Scale (CD–RISC) [37]. In the development study, the internal consistency in the total score of MIIRM was represented by Cronbach’s α = 0.72 [36].As we do not have a Korean version of MIIRM yet, all items in the MIIRM were translated into Korean by the research team to be applied in this study. We used 18 questions for analysis, excluding 4 questions related to access to support network to avoid improper correlation with independent variables. Cronbach’s alpha for the 18 questions was 0.781.We surveyed participant’s age, sex, years of living alone, years of education, existence of living children (0 = none, 1 = one or more than one), existence of living siblings (0 = none, 1 = one or more than one), having religion (e.g., Christianity, Catholicism, Buddhism, etc.) (0 = no, 1 = yes), and perceived socio-economic status (SES) using a questionnaire on a 5-point Likert scale (5 = very satisfied, 4 = satisfied, 3= moderate, 2 = unsatisfied, 1 = very unsatisfied), which we developed for this study.We investigated the perceived health issue in usual activities, the number of chronic diseases, cognitive function, and subjective health as health-related variables.We used one question from the EQ-5D-3L [29] to assess the perceived health issue in usual activities. Because we used the EQ-5D-3L as a screening assessment and excluded those who answered “extreme problem” on any dimension, the participants’ answers were 1 (no problem) or 2 (some problems). Participants were asked how many chronic diseases they had from the list of stroke, heart disease, hypertension, diabetes, hyperlipidemia, pulmonary tuberculosis, and insomnia. Participants additionally described if they had other diseases, not in the list. The score of the MMSE-DS in the screening test was used to indicate cognitive health status: a higher level of cognitive function was associated with higher scores. Subjective health was measured using a 5-point Likert scale (5 = very healthy, 4 = healthy, 3 = moderate, 2 = unhealthy, and 1 = very unhealthy).Loneliness was measured using the Revised UCLA Loneliness Scale (RULS) [38]. Higher scores indicated higher levels of loneliness. Cronbach’s alpha for the 20 questions of the RULS in the present study was 0.873. The SGDS-K [30] score in the screening test was used. A higher score suggested higher level of depression.We used the SPSS Statistics software (SPSS Inc., Chicago, IL, USA), version 23, for all statistical analysis. We analyzed the data (Supplementary File S1) using a two-tailed test at a significance level of 0.05.Descriptive analysis was used for general and social network characteristics of the sample. Before conducting the descriptive analysis to characterize the social network, we averaged the scores of network size (four questions) and contact frequency (two questions) separately to determine the number of people who the participants were in contact with and their contact frequency generally. Mean scores of network size and contact frequency were used as indices. We developed a conversion table to interpret indices based on the scales in the LSNS. Chi-square tests and independent t-tests were used to assess the difference in social network based on gender.Analysis of variance was used to examine whether or not the level of resilience differed along with the difference in the network size or contact frequency. Scheffe’s post hoc test was used for comparison between groups.We conducted hierarchical multiple regression to analyze the association between resilience and both network size and contact frequency. The variables validating their significant relationship with resilience were entered into a hierarchical regression model. We used Pearson correlation analysis for continuous variables and Spearman correlation analysis for ordinal variables. Independent t-tests were used to confirm the association between resilience and nominal variables. Socio-demographic variables were entered into model 1; health-related variables were entered into model 2; emotion-related variables were entered into model 3; and social network size and contact frequency were entered into model 4.In this study of 300 cases, 266 cases without any missing values were analyzed. The participants’ general characteristics and descriptive statistics of study variables are presented in Table 1. The participants’ average age was 71.85 years. The proportion of females was 59.0%.Characteristics of older adults’ social network are presented in Table 2. Five male participants (1.9%) responded “none” to all questions about network size. Seventeen participants (female: 5, male: 12) (6.4%) answered “less than monthly” to all questions about contact frequency. There were four participants (1.5%) who answered “none” or “less than monthly” to all questions about network size and contact frequency. No statistical difference was found between male and females in chi-square tests (network size: χ2 = 4.84, df = 3, p = 0.184; contact frequency: χ2 = 6.85, df = 3, p = 0.077).The average score of the LSNS (total score) was 18.13 (±7.98) and the score was lower than the cut-off score of 21, which suggests isolation [35]. The total score of females was significantly higher than that of males (t = 2.85, p = 0.005). While the score of network size was not significantly different between males and females (t = 0.36, p = 0.722), the score of contact frequency among females was significantly higher than that of males (t = 3.82, p = 0.001).Mean scores of resilience by group according to network size and contact frequency are presented in Table 3. Participants with a larger network size tended to show higher resilience. Participants who reported their network size as three or more than three persons showed a significantly higher score of resilience than those who had less than one person (p = 0.021) and one to less than two persons (p = 0.037). Participants with more frequent contacts tended to exhibit higher scores of resilience. Participants who reported their contact frequency as weekly to less than a few times a week showed a significantly higher score of resilience than those who reported their contact frequency as less than a few times a month (p = 0.004). Participants who reported their contact frequency as a few times a week or more often showed a significantly higher score of resilience than those who reported contact frequency less than few times a month (p < 0.001) and a few times a month to less than weekly (p < 0.001).The correlation coefficients of continuous and ordinal variables are presented in Table 4. Nominal variables, which were associated with resilience via t-tests, included existence of living children (t = −2.023, p = 0.044), religion (t = −4.502, p < 0.001), and perceived health problems in usual activities (t = 2.340, p = 0.020).The results of hierarchical regression analysis are presented in Table 5. Religion, years of education, and subjective economic status were entered into model 1 including socio-demographical variables, excluding the variables of age (p = 0.344), gender (p = 0.326), years of living alone (p = 0.659), and existence of living siblings (p = 0.162), which were not significantly related to resilience. Subjective health and cognitive function were entered into model 2 among health-related variables, excluding the number of chronic diseases (p = 0.134). Loneliness and depression were entered into model 3 as emotion-related variables. Network size and contact frequency were entered into model 4.The final model explained 48.0% of the variance in resilience (Table 5). The variance inflation factor (VIF) ranged from 1.088 to 1.794 in the final model and there was no multicollinearity.Religiousness (β = 0.243, p < 0.01) and higher level of perceived SES (β = 185, p < 0.01) in model 1, higher level of subjective health (β = 0.173, p < 0.01) in model 2, and lower level of loneliness (β = −0.319, p < 0.01) and depression (β = −0.240, p < 0.01) in model 3 significantly correlated with higher levels of resilience. After controlling for socio-demographic characteristics, the health-related variables, and emotion-related variables, the network size (β = −0.149, p < 0.05) and contact frequency (β = 0.136, p < 0.05) showed statistically significant predictive impact on resilience in model 4. While network size was negatively correlated with resilience, contact frequency was positively correlated with resilience. The results showed that the participants with a smaller network size were likely to have higher levels of resilience, and the participants with more frequent social contacts were likely to have higher levels of resilience.In this study, we found that the LSNS score of our sample of community-dwelling older adults living alone excluded from the public care service due to their relatively good health conditions indicated that they were socially isolated. Their size of social network and frequency of social contacts were associated with resilience.Given that the study sample comprised of community-dwelling older adults who were excluded from the public care service due to their relatively better health condition and socioeconomic status, their poor social health status was noteworthy. This result suggests the need for interventions to reduce social isolation, which is one of the serious health problems among older adults [39,40].In this study, a positive association between resilience and contact frequency was confirmed by hierarchical regression analysis. Considering that resilience is linked to depression or mental health, it is likely to be in line with previous studies suggesting a positive aspect of the number of contacts. The positive effect of frequent contact with social network members for older adults may be supported by previous literature, such that frequent contacts with social network members had a positive effect on mental health and frequent contacts with friends or relatives indirectly ameliorated depressive symptoms [26,41].However, the higher resilience associated with a larger network size based on a simple comparison of mean scores was not validated by the linear relationship. The association between smaller network size and higher levels of resilience in hierarchical regression modeling analysis could mean that community-dwelling older adults regarded emotional intensity of relationship as more important than larger size per se. Social support based on high-quality relationship is a preventative factor to overcome adversity [21]. Hence, resilience might be positively correlated with frequent contact with small networks of members who guaranteed higher quality in relationship. According to the socioemotional selectivity theory [42], as people age, older adults tend to be more motivated for participation which gives them higher emotional satisfaction. In the same vein, participants may prefer frequent contact with higher quality in a small group. These results may suggest that the two dimensions of network size and contact frequency should be considered together when we explore the characteristics of older adults’ social health.As an alternative to helping older adults to reinitiate or maintain important social ties, joining a mutual support group could be feasible. Participating in a self-help group has a positive effect on psychological well-being by enhancing resilience through regular gathering and sharing of information or experience related to common interests or difficulties [43]. In this study, we found that the level of resilience was different depending on social network conditions. Considering the fact that improving resilience has a positive effect on wellbeing in later life [44,45], it is needed to protect older adults’ resilience via the public care service for their social health. To use or demonstrate the effect of self-help groups among community-dwelling older adults as a preventative approach, further studies are needed to determine the optimal size of group members or frequency of gathering and develop the specific framework such as online or in-person sessions.Considerations for developing interventions for older adults to improve their resilience are not only the components of social network (i.e., size or contact frequency) but the purpose of the networking as well. Decreasing loneliness or depression, which showed relatively higher coefficients among the variables of the final model in our results, could be considered as the purpose of the self-help group. In addition, it is recommended to explore through further studies what type of network (e.g., conversation, task execution, or recreational activities) is suitable for older adults who need to develop new social ties. Also, we need to take into consideration the COVID-19 pandemic-accelerated exposure of older adults to the use of information and communication technology. Considering the spread of telehealth and the social atmosphere where people consider a none-contact service safer than an in-person service [46], it might be necessary to develop online interventions.Our results have some limitations. First, the conversion table and average scores we used to determine the participants’ network size and contact frequency may be insufficient to identify the properties of the variables. We adopted this strategy to measure network size and contact frequency because of the difficulty in finding a validated tool focused on network size and contact frequency per se. Second, we could not control the unique characteristics of resilience in this cross-sectional study. The levels of individual resilience could be temporarily affected by individual adversities at the time of assessment. Although we recruited a homogeneous set of participants with respect to health condition and used validated tools to measure resilience, various level of individual resilience may limit the interpretation of the study results. Third, the convenience sampling in this study limits us to generalize the result for a population of older adults living alone in relatively good health conditions.Healthy community-dwelling older adults living alone experienced social isolation, and their network size and contact frequency were related to their resilience. This study suggests that interventions addressing social health by the public health service are needed for older adults living alone even when they are physically and mentally healthy. In addition, protecting the present important social ties and enriching the network through frequent contact may be considered to strengthen resilience, which is a protective factor in social health.The following are available online at http://doi.org/10.5281/zenodo.4723552, File S1: The minimal dataset for analysis.Conceptualization, S.P., T.H.K., T.R.E.; Formal analysis, S.P.; Investigation, T.R.E., S.P.; Methodology, S.P., T.H.K.; Project administration, S.P., T.H.K., T.R.E.; Writing—Original Draft Preparation, S.P.; Writing—Review and Editing, S.P., T.H.K.; Funding Acquisition, T.H.K. All authors have read and agreed to the published version of the manuscript.This research was supported by a grant of the Korea Health Technology research and development project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea (Grant number: HI18C1207).This study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Wonju Severance Christian Hospital (Approval number: CR 318026, date of approval: 10 July 2018).Informed consent in written form was obtained from all participants involved in the study.The data presented in this study are available as supplementary material.The authors declare no conflict of interest. The sponsors had no role in the design, execution, interpretation, or writing of the study.General characteristics of participants and descriptive statistics (n = 266).SD = standard deviation, SES = socioeconomic status, MMSE-DS = Mini-Mental State Examination for Dementia Screening, RULS = Revised UCLA Loneliness Scale, SGDS-K = Korean version of Short-form Geriatric Depression Scale, MIIRM = Multidimensional Individual and Interpersonal Resilience Measure. †: Subtotal score of 18 items in MIIRM was used for analysis in this study. Mean and SD for the total score of the 22 items was 71.26 (11.26).Characteristics of participants’ social network (n = 266).n = 266 (Female: 157, Male: 109), LSNS = Lubben Social Network Scale; Percentage (%) indicates the proportion in the column. ** p < 0.01.Difference in mean scores of resilience according to social network (n = 266).* p < 0.05, ** p < 0.01. a—Less than 1 person; b—Less than 2 persons; c—Less than 3 persons; d—More than 3 persons; e—Less than a few times a month; f—A few times a month–Less than weekly; g—Weekly–Less than a few times a week; h—A few times a week or more often.Correlation analysis.Edu = years of education, SES = perceived socioeconomic status, SH = subjective health, Cog = cognitive function, Lon = loneliness, Dep = depression, Size = size of social network, Freq = frequency of social contact, Res = resilience. * p < 0.05; ** p < 0.01; †—r with Pearson correlation test; ‡—r with Spearman correlation test.Hierarchical regression analysis (Standardized beta coefficient).Dependent variable: resilience, SES = socioeconomic status, * p < 0.05, ** p < 0.01.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06062.txt ADDED Viewed

	@@ -0,0 +1,22 @@

+Background: Some swimmers reach high performance level at a relatively young age. The purpose of this study is to determine the relationship between adolescent female swimmers’ 100 m front crawl race (Vtotal100) and several anthropometry, body composition, and physiological and specific strength indices. Methods: Nineteen adolescent female swimmers were examined for biological age (BA) and body composition. Oxygen uptake was measured during water-flume stage-test front crawl swimming with ventilatory thresholds examination. Specific strength indices were assessed during 30 s of tethered swimming. Stroke rate (SR), stroke length (SL), and stroke index (SI) were also examined. Results: BA was strongly correlated with anthropometrics and tethered swimming strength indices, and showed moderate to strong correlation with ventilatory thresholds. Speed of swimming in the race was moderately to largely correlated with speed at V˙O2 max−VV˙O2max (r = 0.47–0.55; p < 0.05)—ventilatory thresholds (VAT, VRCP) (r = 0.50–0.85; p < 0.05), SL (r = 0.58–0.62; p < 0.05), and SI (r = 0.79–0.81; p < 0.01). Conclusion: Results confirmed a significant role of biological maturation mediation on body composition and body size, ventilatory indices, and specific strength indices. BA was not a significant mediation factor influencing the swimming kinematics (SL, SI) and speeds of VAT, VRCP or VV˙O2 max, which were strong predictors of the 100 m race.Female competitive swimming is a sport in which systematic training sessions and high training loads are implemented at an early age [1,2]. Training more and more extensively with gradually and periodically increasing intensity is undertaken along with the phenomenon of progressive biological maturation. The onset of puberty and mental, morphological, and physiological maturation interacts with the development of determinant factors, which affect swimming performance at the age-group level [3,4]. In considering competitive achievements, individual variations in intervals between earlier or later occurrences of growth spurts in one competitive age group cause variations in motor abilities as well as performance. Young swimmers at higher maturity levels are more likely to perform better than their less mature peers due to greater aerobic and anaerobic abilities [5,6,7,8]. Indeed, physical parameters such as body height, free fat mass or more athletic constitution, or bioenergetical indices such as swimming oxygen consumption have been shown in previous studies [2,9] to have an influence on the swimming performance of young women swimmers. Nevertheless, considering the swimming elite in the long term, Timakova and Klyuchnikova [9] pointed out that female swimmers with relatively slower maturation prevailed over the others, reaching high performance levels in adulthood. Monitoring growth as well as somatic and physiological traits in relation to biological maturation is therefore crucial to young athletes’ training optimization and expectations of adequate sports performance [8]. Anthropometric measurements from earlier studies [10] revealed a strong relationship between morphological indices and swimming performance. Cochrane et al. [11] stated that morphological characteristics of young swimmers influence swimming performance and vary by events. Vorontsov et al. [7] claim that the strongest effect of maturity on physical conditioning and strength was indicated in the group of girls aged 13–14.The number of studies of young swimmers in which oxygen uptake is measured directly in swimming is still very limited [12]. Malina et al. [13] noted that aerobic power of more mature female and male adolescent trained children was higher than in their less mature peers (differences ranging from 0.2 up to 1.0 L·min−1). Anaerobic metabolism is more developed in adults than in children, and it indicates that the maturation process influences the level of anaerobic energy production [4,14]. Tethered swimming is the most specific in-water test for strength measurement [15]. Nevertheless, according to Moran et al. [6], anaerobic power, peak force of arms, and free fat mass values are mediated by maturity, and swimmers who are already at peak high velocity are more likely to respond strongly to strength training.When training age-group young women swimmers, technique should be the main concern, because it is one of the most important factors influencing present and future performance. Stroke length and stroke rate or stroke index are essential for an efficient swimming technique. Therefore, here the authors aim to analyze in adolescent female swimmers the influence of a set of morphological, specific strength, and physiological indices on the 100 m front crawl swimming race.These studies are conducted while considering the impact (correlation or mediation effect) of biological age (BA) on swimming determinant factors: (a) body composition, (b) physiological and specific strength, and (c) kinematic indices of 100 m front crawl swimming. The subsequent aim of this study is to identify a set of variables which influence 100 m front crawl performance in female swimmers, but which are not directly related to BA.The authors expect that BA will differentially influence the particular set of indices which could be higher for indicators related to body dimensions and strength, and lower for those related to specific swimming abilities: stroke kinematics and speed on metabolic thresholds (VAT, VRCP, VV˙O2 max).Nineteen female swimmers (age 13.4 ± 0.26, min: 12.71, max: 13.73 years; height 1.66 ± 0.07 m; body mass 55.5 ± 9.3 kg) participated in this study. They were recruited from the most successful swimmers in their age category from the Cracow, Poland region. All of them were healthy and had licenses from the Polish Swimming Federation. All swimmers went through 4–5 years of systematic swimming, trained in at least 10 training units weekly and took part in national level competitions and national swimming championships for their age group. Despite swimming style specialization, all the participants performed in freestyle events regularly. The study was approved by the Regional Medical Chamber in Cracow on 5 June 2020 (No. 94/KBL/OIL/2020). All participants and their parents provided informed consent for their participation in intensive physical effort during this study (parents of all participants became acquainted with the study program and with a short description of the tests).The body composition analyzer Tanita BC-418 (Tokyo, Japan) was used to assess segmental body composition. In addition to body mass (BM, kg) measurement, this device uses bioelectrical impedance analysis (BIA), a method of analyzing electrical responses to a weak electrical current introduced into the body. It is a research method that allows one to assess the human body composition with regard to extracellular and intracellular water, fat and lean mass, and cell mass, based on the differentiation of tissue resistance [16]. The body fat estimated by BIA has almost perfect reproducibility, making it an applicable research tool in studies that investigate body composition changes. FFM estimated by BIA correlates almost perfectly with reference methods, regardless of sex. Moreover, regarding quality, BIA has shown high reproducibility (correlation coefficient between 0.95 and 0.99 [17]. BIA is a reliable method of assessing the tissue composition of the body; its reliability and validity have been recognized in many independent studies: Jackson et al. [18]; Aandstad et al. [19]; Dave et al. [20]; Cortesi et al. [21]; Vasold et al. [22]. This method is successfully used by both untrained people and by athletes of all disciplines [23,24]. The participants dressed in underwear, stood with electrodes on their bare feet, and gripped handheld electrodes. This procedure provided data on fat free mass (FFM, kg), total body water (TBW, kg), and predicted muscle mass of body segments: arms (mm arms, kg), trunk (mm trunk, kg), and legs (mm legs, kg). FFM and TBW values were also converted to percentages of BM. Biological age (BA) examinations of participants were conducted by an experienced anthropologist, who used the following calculation: BA = (BHage + BMage)/2, where BHage (height age) = age obtained from percentile charts (growth charts by The Children’s Memorial Health Institute; 50th percentile was used to align height and mass with age) on the basis of the participant’s body height and BMage (mass age) = age obtained from percentile charts on the basis of the participant’s body mass (growth charts by The Children’s Memorial Health Institute, standardized and validated for the Polish population; 50th percentile was used to align height and mass with age). Additionally, pubertal development was assessed, by an experienced, formally trained anthropologist. Namely, Tanner stages based on pubic hair scale were estimated [25] and the date of menarche was obtained retrospectively (year, month, and, if possible, exact day). The recall method has previously been widely used in research, as well as justified as a reliable way to obtain the age of menarche [26,27]. It is a validated method, used for several dozen years all over the world [28,29]. It was also used many times in the youth population from Cracow [30].Participants took part in two test trials. One contained tethered swimming, 100 m front crawl race and anthropometric measurements. During the second one, separated by 48 h, stage test in water flume was implemented. Before each test, the swimmers completed a 1000 m in-water warm up with low to moderate intensity.The stage in a water flume (Figure 1) was conducted in a laboratory-controlled environment. All swimmers were informed about the testing procedure and performed a 1000 m in-water warm up, as before a competition. Participants wore a nose clip and were attached to a respiratory valve system with an expired air analyzer (Start 2000 MES, Poland). One minute of slow-paced swimming ensured their adjustment to the testing conditions. After an initial speed of 0.93 m s−1 providing moderate intensity of 30–40% V˙O2max, every two minutes a whistle signaled for them to increase speed by 0.06 m s−1. Breath by breath, exhaled air was continuously sampled and saved (Ergo2000M software MES, Poland). V˙O2max, aerobic threshold (AT), and respiratory compensation point (RCP) were estimated [31]. The test was terminated after complete exhaustion and inability to maintain required swimming pace, reaching criteria of V˙O2max examination [32]. Speed of the water flow and oxygen uptake (VO2AT and VAT, VO2RCP and VRCP, V˙O2max and VV˙O2max) were assessed. In our study, all participants met the mentioned criteria, with RER (1.18 ± 0.17).In the 30 s tethered swimming test, participants wore a waist belt and were connected to a steel pole (fixing point 0.49 m above the surface) by a 5.65 m steel cable and attached dynamometer (with 100 Hz recording frequency). The following indices were collected:
+		Maximum value of force (Fmax, N);Average value of force (Fave, N);Force decline (Fdecline, N), calculated from decrease in average force production in 0–10 and 20–30 s of the recording duration;Average impulse per single cycle (Iave, N·s), defined as the integral of force over a period of time containing all full cycles divided by a number of completed cycles:
+		Iave=∫t0t1Fdtn
+		where: t0 is the beginning of the first full cycle and t1 is the ending of the last full cycle in the 30 s period.Maximum value of force (Fmax, N);Average value of force (Fave, N);Force decline (Fdecline, N), calculated from decrease in average force production in 0–10 and 20–30 s of the recording duration;Average impulse per single cycle (Iave, N·s), defined as the integral of force over a period of time containing all full cycles divided by a number of completed cycles:
+		Iave=∫t0t1Fdtn
+		where: t0 is the beginning of the first full cycle and t1 is the ending of the last full cycle in the 30 s period.The 100 m race was carried out in a 25 m swimming pool that meets International Swimming Federation (FINA) requirements. The final results and split times of the trials were measured with an automatic timing device (Omega, Switzerland). Each one of the trial series was performed by three to four swimmers in order to imitate competition conditions. All trials were recorded with a (JVC GC-PX100BE, Japan) camera with 50 Hz framing. The camera was placed on a tripod at the bleachers, seven meters above the water surface on an extension in the middle point of the pool. The swimmers started from the blocks at the sound signal. Markers were placed at the side of the pool to indicate the line of 7 m from each of the walls and 10 m from the starting block. The pool was divided into three zones: (a) I turn zone (7 m), (b) surface swimming zone (11 m), and (c) II turn zone (7 m). Including the first 10 m start zone, it resulted in a) 59 m for VSTF (start, turn, finish velocity) calculation and b) 41 m for Vsurface (surface swimming velocity) examination. Times for separate sectors were measured when swimmers’ heads crossed the imaginary line linking markers at the sides of the pool (Kinovea ver. 0.8.15 software). The 100 m front crawl speed (Vtotal100) was calculated from the final time taken to complete the 100 m distance. The average SR (cycle min−1) (ICC = 0.99, 95%, CI = 0.960–0.997) was calculated from 12 cycles (3 cycles from each of the 4 laps, measured in the surface swimming zone). SL was calculated from the 11 m surface swimming zones, during 4 laps. Stroke index SI (m2·cycle s −1) was calculated as: SI=Vsurface·SL.Individual, mean, and standard deviations (SD) computations for descriptive analysis were obtained for all studied variables. Measures of skewness, kurtosis and the Shapiro–Wilk test were used to assess the normality and homogeneity of the data.One-way ANOVA with repeated measures and Tukey’s HSD post hoc test were carried out to detect and present differences between: (Vtotal100, Vsurface, VSTF). To identify the relationship between the variables, the Pearson correlations were computed between:
+		(a)Anthropometric, body composition indices and all the indices, tethered swimming test (Fmax , Fave, Fdecline, Iave) and swimming speed (Vtotal100, Vsurface, VSTF),(b)Stage test and swimming speeds or tethered swimming variables, and(c)SR, SL, SI and Vtotal100, Vsurface, VSTF.Anthropometric, body composition indices and all the indices, tethered swimming test (Fmax , Fave, Fdecline, Iave) and swimming speed (Vtotal100, Vsurface, VSTF),Stage test and swimming speeds or tethered swimming variables, andSR, SL, SI and Vtotal100, Vsurface, VSTF.To examine the possible mediation effect of BA on variables of VRCP, VV˙O2max, VAT, SI and SL, which correlated the most with Vtotal100, were tested by mediation analysis with the Sobel test. Mediation analysis was made on the basis of three regression models [33]. The tests were conducted with STATISTICA 13.1 software (TIBCO Software Inc, Palo Alto, CA, USA). A significance level of p ≤ 0.05 was established. Mediation analysis was prepared using R software ver. 4.5.0 with mediation package.The magnitude of the correlations was determined using the modified scale by Hopkins (Hopkins, WG. Measures of reliability in sports medicine and science. Sports Med 30: 1–15, 2000.): trivial: r < 0.1; low: 0.1–0.3; moderate: 0.3–0.5; high: 0.5–0.7; very high: 0.7–0.9; nearly perfect > 0.9; and perfect: 1. To get significant results (p < 0.05) with sufficient power (80%) to detect at least a correlation coefficient of 0.6, the minimum required sample size for this study is 19. The formula for calculation is based on two-tailed test (Guenther, W C. 1977 Desk Calculation of Probabilities for the Distribution of the Sample Correlation Coefficient. The American Statistician).There was a significant difference between measured average speed values of: Vtotal100, Vsurface, and VSTF (F = 127.0, p ≤ 0.001). Post hoc Tukey’s (HSD) test confirmed significant differences among all of the measured averages (p ≤ 0.001). Figure 2 presents differences between all of the three analyzed averages.Biological age (BA) presents high correlation relationship with body composition indices of FFM and TBW (r = −0.56, p ≤ 0.05). The highest correlations (r = 0.88 to 0.92, p ≤ 0.001) were found between biological age and h, BM, FFM, TBW, and mm total (Table 1).The anthropometric indices of height, body mass, and muscle mass of particular body parts all showed significant moderate to very high relationship with tethered swimming indices. All tethered swimming indices correlated moderately to very highly with mm total, mm arms (3.43 ± 0.54), mm trunk (24.03 ± 3.14), and mm legs (6.43 ± 0.93). BA was correlated at a very high level with maximal propulsion force and average impulse (Table 2).There was no significant correlation between body composition, tethered swimming indices, and free swimming speeds: Vtotal100, Vsurface, VSTF.The anthropometric, body composition indices correlated moderately to very highly with: V˙O2AT, V˙O2RCP, and V˙O2max values. Significant correlations were observed between the tethered and stage tests Fmax and ventilatory indices V˙O2AT and V˙O2RCP (r = 0.53 and r = 0.50, p ≤ 0.05, respectively) (Table 3).Significant correlations were observed between VAT, VRCP, VV˙O2max and all swimming speeds: Vtotal100, Vsurface, and VSTF. The level of ventilatory indices expressed in (L · min−1) did not significantly correlate with swimming results. The best predictor of swimming results was VRCP (Table 4).High to very high correlations were observed between SL and SI with Vtotal100, but SR did not correlate with this speed (Table 5). The kinematic indices were not correlated with body height.For BA mediation analysis we selected variables (VAT, VRCP, VV˙O2max, SL, and SI) which were significant predictors of the swimming race in this study (Vtotal100). This was carried out to examine the effect of maturation (BA), which could influence the relation of selected predictors on swimming performance.None of the tested variables—VRCP, VV˙O2max, VAT, SI, or SL—were identified as mediated by biological age. The mediation analysis showed moderate or strong correlations between variables and Vtotal100 (Figure 3a–e).This study showed that 100 m front crawl race results of adolescent female swimmers were significantly related to swimming endurance: VAT, VRCP, VV˙O2 max and stroke kinematics SL and SI. The main finding is that those particular relationships were observed without significant mediation effect of BA. As far as we know, VV˙O2 max was not presented when evaluating adolescent female swimming performance. In this study, VV˙O2 max showed a significant relationship (r = 0.47, p ≤ 0.05) with the 100 m front crawl race, but also our choice of VAT and ,VRCP as predictors of swimming performance is rare in young swimmers.In young swimmers, even in short race distances with duration over one minute, aerobic power development is crucial [1]. Malina et al. [13] noted that aerobic power of swimmers more advanced in maturation was higher than in their less mature peers. In longitudinal studies of the development of cardiorespiratory capacity, which have rarely been conducted with young swimmers, there has been observed an ability to perform increased volume and intensity of training load [8]. Those changes linked with growth and development of aerobic power translate into an increase in swimmers′ achievements [1]. V˙O2 max (L · min−1) results presented in our study show no significant correlation (r = 0.19, p ≤ 0.05) with 100 m race results, as in the Unnithan et al. [34] study of female adolescent swimmers (age 15.3 ± 1.5 years). Lack of significant correlations between V˙O2 max and performance in our study might be caused by not fully utilized cardiovascular and respiratory range and by use of anaerobic energy source in the 100 m race. The noted results also significantly show higher speeds of VAT and VRCP in relation to the 100 m race, which must be the advantages of better developed swimming economy and endurance of the best swimmers. In our study, the results for absolute maximal oxygen uptake (3.01 ± 0.42 L · min−1) were higher than in the study of Plyley, Wells and Schneiderman-Walker [35] (2.7 ± 0.3 L · min−1) assessed during tethered swimming. The reasons for that could be that females in our group, despite similar chronological age, were taller with greater body mass and reached greater exhaustion in the flume stage test (higher values of RER). The VRCP relationship with the 100 m front crawl race presented in our study shows a similar strength as that of critical velocity of 30 min aerobic endurance test of young females (age 11.5 ± 0.6 years) and swimming performance in their personal best events (r = 0.55) [36]. We found that a 100 m front crawl race of young female swimmers is highly correlated with an ability to maintain higher swimming intensity and endurance—VRCP.In this study, anaerobic tethered swimming indices were not significantly related to 100 m race but remained influenced by BA. Taylor S, MacLaren D, Stratton G [5] observed a substantial increase in mean force production in tethered swimming in 13-year-old swimmers, which is explained by the development of the glycolytic energy system caused by maturation. Geladas, Nassis, and Pavlicevic [37] have not found a significant relationship (r = −0.18, p ≤ 0.05) between grip strength and a 100 m race of young female swimmers (age 12.68 ± 0.06 years), but in male participants this correlation was strong (r = −0.73; p ≤ 0.01). In a study of Silva et al. [38], tethered swimming indices Fmax and Fave were not significant predictors of young female swimmers’ sprint performance. A study of Oliveira et al. [39] revealed the significance of controlling for the maturation effect of specific strength evaluation in adolescent swimmers, showing biological maturation mediated positively between anthropometric or body composition and the propulsive force of arms. As mentioned by Vorontsov [8], early maturers demonstrate greater physical abilities and performance level than their peers who are normal or late maturers. According to Moran et al. [6], swimmers who have already reached the peak height velocity are more likely to respond more to strength training. In our study, anthropometrics showed no significant relationship with the 100 m race, so we can state the advantage of better efficiency of swimming technique (SI, SL) of leaders instead of strength. However, undoubtedly, also in female swimming the appropriate level of strength must be reached. This study did not find in female swimmers significant influence of height on the 100 m race, SL, or SI, but Lätt et al. [2] observed significant correlation (r = 0.41, p ≤ 0.05) between SL and height of adolescent swimmers. Silva et al. [40] showed longer SL and higher height in age 11–12 female swimmers, affected by advanced calendar age and technique development of earlier maturers. Geladas et al. [37] also did not find any relationship between anthropometrics and the 100 m race in 13-year-old female swimmers, but they revealed body height, hand length, and horizontal jump association with BA, which explained only 17% of the variance of the 100 m race. Zuozienė and Drevinskaitė [41] reported in young female swimmers (11.8 ± 0.4) a lack of significant correlation between anthropometrics and the 200 m race. They concluded that, in girls versus boys, anthropometrics predicts swimming performance less. On the other hand, physical characteristics of young swimmers might influence swimming performance [2,11]. Toussaint and Beek [42] pointed out that young swimmers’ ability to increase maximal swimming velocity is associated with better force-generating capacity caused by age-related growth in muscle size. Vorontsov et al. [7] concluded that the strongest effect of maturity on physical development and strength could be observed in girls aged 13–14.The relationship between stroke kinematics and the 100 m race showed that SI and SL are good swimming speed predictors, especially here for young females, because they are free of BA mediation. This indicates that their size depends mostly on the type of technique dedicated to swimming training. Mezzaroba and Machado [43] pointed out that SR, SL, and SI included in multiple linear regression of swimmers aged 10–17 could explain almost 100% (R2 = 0.99) of 100 m race results of young swimmers. Jürimäe et al. [4] stated that SI may be an important indicator of swimming economy in adolescent swimmers. Morais et al. [44] revealed a strong relation between adolescent girls’ (12.31 ± 1.09 years) 100 m race time result and SI (r = −0.82, p ≤ 0.01) and SL (r = −0.61, p ≤ 0.05), which is very similar to our speed performance and kinematics (r = 0.81, p ≤ 0.01 and r = 0.61, p ≤ 0.01, respectively). Lätt et al. [12] also presented high partial correlation between time of 100 m front crawl performance in young male swimmers (age 15.2 ± 1.9 years) and SI (r = −0.643; p ≤ 0.05).Despite objective strengths of the presented study, some limitations should also be noted. For example, the method used to assess biological age is not validated, or the small sample of participants examined may limit the application of the conclusions in regards to the wide swimming community.This study analyzed significant predictors of the 100 m front crawl race in adolescent female swimmers: front crawl swimming endurance (VAT,VRCP,VV˙O2max) and stroke kinematics (SL, SI). The noted predictors were not mediated by BA. These results showed that young female swimmers rely on trained physiological capacity and efficient front crawl stroke technique and less on somatic traits or strength. The identified predictors are certainly susceptible to the influence of well-thought-out, planned swimming training.
+              -
+              Biological age must be taken into consideration when evaluating young female swimmers’ abilities in regards to training and performance;
+              -
+              Efficiency of sprint swimming technique reflected by SL and SI, crucial in young female swimmers may be more dependent on the training used and less dependent on biological age;
+              -
+              Swimming speed at ventilatory thresholds and maximal oxygen uptake is valuable in terms of assessment of the physiological build-up in relation to performance in adolescent female sprint swimming.
+        Biological age must be taken into consideration when evaluating young female swimmers’ abilities in regards to training and performance;Efficiency of sprint swimming technique reflected by SL and SI, crucial in young female swimmers may be more dependent on the training used and less dependent on biological age;Swimming speed at ventilatory thresholds and maximal oxygen uptake is valuable in terms of assessment of the physiological build-up in relation to performance in adolescent female sprint swimming.Conceptualization, K.S.; methodology, K.S., M.S.; validation, K.S.; formal analysis, K.S., M.S., A.S.; resources, K.S., M.S.; data curation, K.S., M.S., A.S.; writing—original draft preparation, K.S., M.S.; writing—review and editing, K.S., M.S., A.S., Ł.K., A.R.-P., P.K., T.R., B.K.; supervision, K.S., M.S.; project administration, K.S.; funding acquisition, K.S., M.S., A.S. All authors have read and agreed to the published version of the manuscript.This research received no external funding.The research was approved by the Bioethics Committee at the Regional Medical Chamber (No. 94/KBL/OIL/2020).All participants and their parents provided informed consent for their participation.The data presented in this study are available on request from the corresponding author.The authors declare no conflict of interest.One of the swimmers going through a stage test procedure in a water flume.Comparison between average speed values of all of the distance (Vtotal100), surface swimming zones (Vsurface), and start, turn, and finish zones (VSTF) measured during 100 m crawl stroke race. * Significant difference from the other speeds; p ≤ 0.001.Mediation models illustrating the level of mediation effects of the relation between the independent variables of: (a) VRCP, (b) VV˙O2max, (c) VAT, (d) SI, (e) SL and dependent variable of Vtotal100. β and corresponding p-values are presented. β of total (c), direct (c’), and indirect (a,b) effects are presented with 95% confidence intervals, p-value.Correlations between BA and anthropometric, body composition indices: h, BM, FFM, FFM, mm total. In the top row, there are mean values and standard deviations (m ± SD) of anthropometric, body composition indices with corresponding ranges (min–max) presented. In the lower line there are values of Pearson correlations with corresponding p values.Correlations of anthropometric, body composition indices: BA, h, BM, mm total, mm arms , mm trunk, and mm legs with tethered swimming indices: Fmax, Fave, Iave, and Fdecline.Correlations of stage test physiological and kinematic indices: V˙O2AT, V˙O2RCP, and V˙O2max with anthropometric, body composition indices.Correlations of stage test physiological and kinematics indices VAT, VO2AT, VRCP, VO2RCP, VV˙O2 max, and V˙O2max with Vtotal100, Vsurface, VSTF.Correlations of kinematic indices of SR, SL, and SI with: Vtotal100, Vsurface, VSTF.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06063.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Children on farms have been identified as a population vulnerable to injury. This review seeks to identify child farm-related injury rates in Australia and to determine the key hazards and contributing risk factors. This critical review utilised the PRISMA guidelines for database searching. Research from the year 2000 onward was included as well as earlier seminal texts. Reference lists were searched, and the relevant research material was explored. Our primary focus was on Australian peer-reviewed literature with international and grey literature examples included. Evidence suggests that there is limited Australian research focusing on child farm-related injuries. Child representation in farm-related injuries in Australia has remained consistent over time, and the key hazards causing these injuries have remained the same for over 20 years. The factors contributing to child rates of farm injury described in the literature include child development and exposure to dangerous environments, the risk-taking culture, multi-generational farming families, lack of supervision, child labour and lack of regulation, limited targeted farm safety programs, underuse of safe play areas, financial priorities and poor understanding and operationalisation of the hierarchy of control. It is well known that children experience injury on farms, and the key hazards that cause this have been clearly identified. However, the level of exposure to hazards and the typical attitudes, behaviours and actions of children and their parents around the farm that contribute to chid injury remain unexplored.Globally, injuries are a leading cause of death and disability among children [1]. Unintentional injuries have replaced infectious diseases as the most serious child health issue in the industrialised world [2,3,4] —accounting for approximately 40% of deaths in developed countries [1] and contributing to disability, future health care needs, intellectual development delays and reduced economic and productivity prospects [5]. In Australia, injury is the leading cause of death and hospitalisation of children aged one to fifteen years [6,7]. Both fatality and hospitalisation rates rise with increasing remoteness. Between 2013 and 2017, the rate of fatal injuries among Australian children in outer regional, remote and very remote areas was three times higher than in major cities [6]. Farming families are defined as families where at least one adult is a farmer or farm manager [8]. According to the Australian Bureau of Statistics, in 2016, there were 87,325 farming families in Australia, including 48% consisting of parents with children living with them [8]. Children and adolescents on farms have been identified as vulnerable; suffering premature death, morbidity and disability from injury [1,9]. The blurred distinction between the farm as a home and a workplace means children are exposed to hazards typically not present in most homes [10,11,12,13,14,15,16,17,18]. Additionally, most farming parents want to involve their children in the unique farming lifestyle, often giving them on-farm work responsibilities at a young age [17,19]. This review aims to: (1) identify child mortality and morbidity rates from farm-related injuries in Australia; (2) explore the key hazards causing these injuries; and (3) understand factors that may contribute to the risk of child farm-related injury. This critical review identifies significant work in the field, and analyses and evaluates the quality of the existing literature on child farm injury and safety [20]. A broad and deep coverage of the topic was undertaken. We focused on an area of research that has not previously been comprehensively explored in Australia. The PRISMA guidelines were followed for the database search process [21] (Figure 1). Articles were removed during the screening and eligibility phases due to inability to find full texts, date of text, wrong population focus, wrong outcomes and wrong settings. Studies were included if they encompassed children aged under 18 years. A snowballing and saturation approach was then adopted, with reference lists searched for further relevant material. The key focus was on Australian peer-reviewed literature. However, given the limited recent research, grey literature and international research—particularly North American—were included to provide examples relevant to the Australian context. Initial search terms used were: child*, adolescen*, teenag*, paed*, farm*, agricultur*, injur*, morbidit*, mortalit*, safe*, risk* and behav*. The Boolean operators ‘AND’ and ‘OR’ were utilised to increase the search scope. The databases searched were Medline, Embase, Eric, Informit, Global Health and SafetyLit. As limited recent research has been conducted on child farm safety in Australia, literature from 2000 to 2020 and earlier seminal texts were included. Article titles and abstracts were assessed to ascertain relevance and duplicates removed [22]. Full-texts were then read to determine compatibility with other inclusion criteria:Research reported on human subjectsEnglish languagePeer-reviewed or grey literatureFocus on the farming contextFocus (primary or secondary) on children on farmsResearch reported on human subjectsEnglish languagePeer-reviewed or grey literatureFocus on the farming contextFocus (primary or secondary) on children on farmsMandryk and Harrison [23] first explored work-related deaths of Australian children for the period 1982–1984. They acknowledged existing workplace controls as effective in child injury prevention, with farms as the exception. Since this time, the majority of research has largely focused on specific regions or farming hazards [11,24,25,26,27,28,29]. Peachey and colleagues [30] provided the most recent literature on child farm-related fatalities across Australia. Between 2001 and 2019, 222 children under 15 years of age died from farm-related injuries. This represented 15% of all farm-related deaths, which is consistent with earlier research by Lower and colleagues [31]. It is widely acknowledged that boys make up the majority of farming fatalities, accounting for 73% of child farm deaths between 2001 and 2019 [30]. Peachey and colleagues [30] emphasised the persistent rate of child farm fatalities over time, with only a 0.009 reduction in deaths per year over the 18-year period. This minor reduction aligns with the trend identified by Lower and colleagues [31]; a reduction in overall farm-related fatalities in the early 1990s and then a stabilisation of rates from 2005. It has been hypothesised that the small reduction in deaths is due to the decline in farmers and consequently a decrease in children on farms [30]. This could also be due to improved treatments and emergency transport services in rural areas (i.e. injuries occur but do not result in as many fatalities). With no comprehensive research investigating non-fatal child farm-related injuries, it is challenging to determine and interpret trends over time. While previous literature has largely adopted a focus on the broader farming population, consistent child representation in farm fatality statistics has been reported [14,32,33,34]. Lower and Herde [14] identified 326 farm-related deaths on the National Coroners Information System between 2003 and 2006; 17% were children. Safework Australia [35] attributed 51 workplace deaths to children between 2003 and 2016. The suggestion from this data—that children only make up 7% of the 744 agriculture workplace deaths—is counter to the more targeted research previously conducted by Lower and colleagues [14] over the same period. This highlights the limitations in farm workplace death reporting and indicates that not all non-work-related fatalities that occur on farm are captured. Lower and Herde [14] reported that 48% of on-farm deaths between 2003 and 2006 were non-work related (bystanders or leisure activities). This even spread of work-related and non-work-related deaths highlights the blurred division between the farm as a workplace and home—with nearly all of the 55 child farm deaths between 2003 and 2006 identified as non-work-related. While recent work conducted by Peachey and Lower [30] has helped to fill the gap in knowledge on child farm fatalities in Australia, a clear understanding of the incidence of farm-related injuries among children is still lacking. Most existing literature focuses on overall farm or workplace deaths, often identifying children as a secondary population [24,36,37]. This lack of Australia-wide literature provides challenges in quantifying, comparing and drawing conclusions. Since the 1980s, children on farms have been recognised in academic literature as a population that is particularly vulnerable to farm-related injury [38]. Despite research mentioning child farm-related non-fatal injuries [39,40,41], there is yet to be comprehensive research examining this in Australia. To date, most non-fatal injury research has concentrated on the state of New South Wales, limiting its validity (and the validity of proposed prevention measures) across the wider Australian context. An Australian Institute of Health and Welfare report provided the most recent Australia-wide farm-related injury hospitalisation data [39]. Approximately 22,000 people were hospitalised from 2010–2011 to 2014–2015 due to farm-related injuries, including over 2000 (9%) children under 15 years. Although focused solely on New South Wales, between 2010 and 2014, Lower and Mitchell [41] attributed 6270 hospital admissions to farm-related injuries, of which 16.3% were children. As hospital record procedures vary, and cause and location of injury is not always reported, and an accurate picture of non-fatal farm-related injuries may not be captured. Farms present a diverse range of hazards to children through exposure to machinery, grain silos, farm buildings, water sources, chemicals, noise, dust, animals, heights, sun and firearms [42,43]. However, the key hazards—accounting for 75% of child fatalities—have changed little over time. These seven hazards include water bodies, quad bikes, tractors, utility vehicles (trucks), cars, motorbikes and horses [30]. In Australia, drowning is the leading cause of death in children under 15 years [44]; accounting for 31% of child farm-related deaths between 2001 and 2019 [30]. This was most commonly represented by the drowning of children under four years of age in farm dams [30,45]. Quad bikes (sometimes referred to as All-Terrain Vehicles (ATVs)) have emerged as a leading cause of farm-related morbidity and mortality [10,14,30,39,40,46]. Approximately 80% of Australian and New Zealand farms have at least one quad bike that is utilised daily [47]. The use of quadbikes by children on farms (residents and visitors), both as drivers and passengers, is a contentious issue due to the traumatic injuries sustained, largely due to their heavy weight and large size compared to children [48]. Peachey and colleagues [30] attributed over 20% of all child (under 15) fatalities on farms to quadbikes and motorbikes between 2001 and 2019. Henley and Harrison [39] described 42% of child farm-related injury hospitalisations between 2010–2011 and 2014–2015 were due to quadbike and motorcycle incidents. Specifically, young males, aged five to fourteen, are overrepresented in motorbike/quadbike injury statistics [39,46,49]. On farms, motorbikes and quadbikes are used for both work and recreation—this is reflected by the even split of work and recreation fatalities in Australia [50]. Children are most often represented in recreational motorbike/quadbike injuries [30,50]. Although not focused solely on farms, research conducted by both McIntosh and colleagues [50] and Amey and Christey [51] unsurprisingly identified at least half of the children injured on quadbikes were riding adult-sized bikes. It is widely acknowledged and included in the manufacturer’s guidelines that children under 16 years should not drive or be passengers on quadbikes [51]. Children on farms are exposed to heavy vehicles and machinery that they would otherwise not come in contact with outside the farm environment [11,14,29,30,46]. Standard vehicle safety procedures are often not adhered to when on the farm. For example, seatbelts—whilst compulsory when riding in a vehicle on Australian roads—have been identified as the least utilised child safety practice on farms [29]. Vehicle incidents on farms typically involve children riding in, or on the utility tray, riding unrestrained in the cabin, and as casualties of vehicle rollovers or low speed run-overs [30]. Although health promotion campaigns have discouraged children riding on tractors, they remain a leading cause of injury [18,42,52]. With farm machinery being expensive and the ever-present economic difficulties in the rural sector, farmers may not have the latest or safest equipment [53]. While boys are more likely to be injured by motorbikes, girls are more likely to experience injury from horses [39]. Approximately 80% of horse-related injuries occur to girls, with the most vulnerable aged 10 to 14 years [29,39,46]. The prevention of child farm injuries is a complex issue for both parents and safety professionals [54]. Children are vulnerable due to their physical, cognitive and behavioural characteristics [55]. Haddon’s Injury Model explains that injuries occur due to an uncontrolled interaction between a host, agent and environment. These interactions are unique in regard to child injuries as a parent/adult is usually responsible for controlling or supervising the interactions [56]. The consistent rate of farm injuries indicates that children are regularly engaging with farming hazards [54]. Typically, injury type varies according to age and developmental stage. Other potential influencing factors include farm type, season and periods of increased farm activity [57,58]. Understanding the context of injury is important for developing prevention strategies. Visitors to the farm are also considered vulnerable to farm-related injuries, although little research has explored this [36,43]. There is a challenge in understanding the rate of child farm visitor injuries as the number of visitors to farms is unknown [30,39]. Typically, farm visitors have less knowledge of farm hazards and risks. Peachey and colleagues [30] identified every third case of child farm-related fatality as a farm visitor; representing 21% of child farm drownings, along with 43% of quad bike, 50% of motorbike and the majority of utility fatalities [30]. Australian farmers characteristically view safety as common sense, considering incidents as ‘unlucky’ rather than avoidable [16,59]. Risk-taking traits and attitudes to risk are commonly passed on through generations and attributable to geographic isolation, frequent exposure to occupational hazards, personal behaviours and the influence of cultural norms [16,60,61]. Greater risk-taking, combined with increased exposure to hazards, can result in a heightened likelihood of injury. A recent report by Farmsafe Australia [62] highlighted four aspects of farming culture that create barriers to improvement in farm safety; money and time pressures, belief that incidents ‘will not happen to me’, reliance on common sense and the generational transfer of farming methods and attitudes. There has been a view that little can be done to increase the health and safety of farmers, families and workers [63]. Fatalistic attitudes reflecting the inevitability of injury are thought to provide parents with perspective on the impossible task of preventing all possible harm—releasing them from overprotection and responsibility [54]. Nilsson [17] reported that parents worry everyday about what could happen to their children on the farm and that parents acknowledged they cannot protect their children at all times on the farm because they have their own farm work to complete. Conversely, Dixey [64] explored the importance of considering health promotion solutions in different contexts. What appears ‘rational’ to one culture may not to another, and people’s lack of knowledge on causes of injury can incite the belief of injury as an ‘act of God’ and fatalism. Families play a pivotal role in child development, particularly in modelling and enforcing safety practices. Children imitate behaviours and actions around them, assuming what they see is right [1,17]. Farming families often give children responsibilities and tasks at a young age, and parents may overestimate their child’s physical, social and cognitive capacity to complete tasks [17,19,30]. Traditionally, farming methods are passed down from generation to generation with on-the-job trial and error. Limited formal farming and safety training can mean unsafe operations are perpetuated [16,63]. The blurred distinction between work, home, personal identity and occupation in farming can result in behaviours that are typically not acceptable within other workplaces [16]. Elliot and colleagues [54] determined that farming parents are aware of the occupational hazards for their children. However, farmers under 55 years—those most likely to be parents of dependent children—have the lowest safety scores [65]. As such, they are the least likely to demonstrate a positive example to their children. However, excluding children from participating in farming practices can be perceived as threatening to the continuation of traditions and values [54,66]. Rather, parents believe that exposure, education and knowledge about risks should equip children to be safe [54,66]. Internationally, migrant/seasonal workers have been identified as a population vulnerable to injury [67]. In the United States, these workers’ children often work in the fields beside them from around 12 years of age [68]. Migrant/seasonal farm workers are a key workforce in Australian agriculture, particularly in horticulture [69]. However, they are not identified in injury data, and it is unknown if their children are exposed to the farm workplace or even provide assistance to their parents. Children live in an environment built for adults; with little power or control and a reduced capacity to predict or respond to danger [1,42]. There is a stereotypical view that farms are healthy places for children to grow and develop and can provide an enriching life. However, the farm is also a work environment with hazards not typically present at most homes [66]. Children’s size, limited physical strength, inexperience, increasing independence, impulsivity, curiosity and immature coordination increase their injury risk. Additionally, they have poorly developed danger identification and are influenced by peer pressure [70]. Children are unable to extricate themselves from potentially deadly situations where this may be possible for adults [71]. Potential hazards and injury risks differ with a child’s stage of development:Toddlers—falls when playing and exploring the worksite and through interaction with animals [7].Young children (five to nine years)—distraction and impulsivity [18].Older children/adolescents (10–14 years)—performing farm tasks or during recreation; able to identify hazards and follow basic rules/procedures but are easily distracted, challenge authority, are overly confident and influenced by peers [42,72].Toddlers—falls when playing and exploring the worksite and through interaction with animals [7].Young children (five to nine years)—distraction and impulsivity [18].Older children/adolescents (10–14 years)—performing farm tasks or during recreation; able to identify hazards and follow basic rules/procedures but are easily distracted, challenge authority, are overly confident and influenced by peers [42,72].Child safety in the farming environment is complex—parents must consider the home, workplace, geographical isolation and their child’s cognitive development and physical ability. They also need to match child safety and ability with the needs and demands of farming life. A delicate and difficult balance is required between encouraging children to grow and develop through stimulation and inclusion in the family farm life and providing a safe environment [66,73]. According to Elliot and colleagues [54], farming parents weigh up the perceived benefits and hazards of involving their children in the farm workplace, and that when these are balanced, the benefits outweigh the risks. However, other research indicated that parents overstate the cognitive ability of children [17,19,30]. Recent work by Summers and colleagues [74] reported that children break rules and take risks, but also recognise that they are modelling the unsafe behaviour of adult relatives. Farming is one of the most dangerous industries for adults, and when children engage in agriculture activities, the risks of injury are increased [75]. Evidence shows the importance of parents selecting developmentally appropriate tasks for their children when engaged in farm work [2,76]. However, this can be difficult to achieve as Neufield and colleagues [77] noted that parents have a clear belief system that justifies their child farm work. They suggested any effort to reduce risks to children would need to consider the benefits parents perceive farm work gives their children and how the need for labour determines the task over their child’s maturity and development levels [77]. Nilsson [17] suggested these behaviours could be a reflection of an industry culture that values productivity over human conditions. Agriculture is a priority industry in the Australian Work Health and Safety Strategy 2012–2022; however, little progress is evident [78,79]. Farm safety has traditionally not been a high priority of government and industry bodies [79]. While regulation has proven successful in other dangerous industries (e.g., mining and construction) [80], enforcement in agriculture has proven challenging. Within the farming industry, differences in regulations exist between commodities [81]. For example, the horticulture, cotton and dairy sectors have higher health and safety performance when compared to other agricultural sectors (e.g., sheep and beef) as they must follow more stringent quality assurance measures [81]. Regardless of the presence of similar dangers, small/medium sized farms are less likely to engage in safety measures to protect children than larger farms [37]. Socio-economic status may contribute to this lower engagement. As most farms are small businesses—with the workplace also the home/private property—farmers may view regulations as optional and an infringement on their personal home and private property. Voluntary safety standards and common sense are relied on to prevent injury [16,82]. While international literature has described children’s involvement in farming tasks [83,84,85], the roles performed by Australian children are less understood. Kennedy and colleagues [86] identified Australian children as actively involved in farming; commonly engaging with machinery, vehicles, firearms and livestock. These tasks were normalized, considered fun, and identified as activities that families can do together. The presence of family members and children as workers is a distinguishing feature of the agricultural industry [56]. Agriculture accounts for 71% (108 million) of the world’s child labour [9]. The Food and Agriculture Organization of the United Nations [9] defines child labour as “work that is inappropriate for a child’s age, affects children’s education, or is likely to harm their health, safety or morals”. Child labour is a multifaceted issue, influencing schooling, healthcare, labour market conditions, social protection, basic services, social norms and cultural practices [87]. Most child labour is non-paid informal work for the family, and varies around the world from deliberate exploitation to more subtle methods commonly seen in Australia—such as children staying home from school to assist during shearing. While parents may not consider this child labour, it has the potential for negative health and education consequences [87]. Research is yet to explore the full economic contribution of Australian children to the agricultural industry. Despite the known presence of children within the farm workplace, Occupational Health and Safety (OHS) standards do not currently protect them. While a range of laws apply across Australia stipulating the age at which children can commence employment [88,89], most farm work performed by children is unregulated due to the exemption for working in a family business [82,90]. Worksafe Victoria [90] stipulates that children on farms should only complete ‘light work’ that is not likely to harm their health, safety and welfare, and does not restrict their ability to attend school. However, the term ‘light work’ remains vague. The Child Employment Act 2004 (Victoria) provides more specific guidance on work that could cause harm to children, including moving vehicles and uncontrolled animals—both of which are prevalent on farms. It is currently unknown whether new workplace manslaughter laws will increase the protection for children assisting on farms; however, the Victorian Farmers Federation were denied an exemption to family farms—therefore, they should be enforced [91]. While legislation may provide some protection to children, there appears to be no examples of prosecution following a child farm-related fatality. Ultimately, parents and guardians are left responsible for judging their child’s involvement in farm tasks [85]. Leaving decision-making regarding child safety at farm workplaces to parents/guardians may not be sufficient protection [52,66]. The perceived costs and concerns of reduced productivity may limit the adoption of OHS solutions [16]. Durey and Lower [92] suggested that farmers only spent money on safety if it improved the business’ overall financial position. Investments in safety often compete with investments to increase productivity and general family needs. Typically, farmers choose lower-order, cheaper safety options [93]. Saluja and colleagues [2] explained that parents acknowledge there are changes they can implement to enhance their child’s safety but are faced with limited economic and practical resources. To address this, financial incentives, such as tractor rollover protection rebates, have been effective in the past in increasing investments in safety 37]. The government of the Australian state of Victoria recently (2021) announced the provision of $5000 rebates on capital works and equipment to improve farm safety, inclusive of fencing for safe play areas on farms [94]. Pollock [95] estimated the cost of farm-related deaths between 2001 and 2004 to the Australian economy was just over $650 million (based on 2008 dollars). However, this did not take into consideration the costs of non-fatal injuries (hospital costs and time lost from work). Safework Australia [96] identified that agricultural workers required longer periods off work following an injury compared to all other industries. In Australia, employed workers who are injured on farms are eligible for workers’ compensation. However, according to the 2013 Safework Australia report [12], only half of the workers in the agricultural sector were eligible for workers compensation, and further, many that are eligible do not put in a claim. Many farmers are self-employed and, therefore, not eligible for workers’ compensation as they are not defined as a worker. These individuals are able to purchase loss of income insurance policies for injury or disability; however, these policies usually do not cover other family members working on the farm, such as children and partners (who are also often not defined as workers). Importantly, the universal healthcare system (Medicare) in Australia does give residents access to health and hospital care at little to no cost [97]. Therefore, investment in farm safety has the potential to result in improved economic outcomes for farmers.Child farm safety in Australia came under focus in the 1990s following reports of high rates of injury [11]. Previously, the focus of agricultural OHS approaches was on employees of larger enterprises [57]. While Farmsafe Australia promoted interventions targeting child safety on farms (e.g., safe play areas, seatbelts and helmet use), funding ceased in 2006 30]. It has been estimated that 82% of child deaths and 58% of child hospitalisations could have been prevented if the Farmsafe strategies were in place [46]. The Rural Injury Prevention Primary Education Resource (RIPPER) was developed in 2005 to align with the school curriculum and educate children on farm safety. However, this program is yet to be evaluated. The farm safety guidelines developed in the late 1990s may still be relevant today, as the major farm hazards have remain unchanged. However, the effectiveness of these guidelines has not been assessed [30]. Education—together with culture and behaviour changes—is required for effective prevention of child farm injuries [30,52]. Typically, education programs are considered less effective injury prevention interventions, and their long-term influence is unknown. They should, therefore, be implemented alongside other interventions, such as engineering controls and regulation [30,82]. As farming communities are heterogeneous, tailored prevention programs are required to address specific needs, empower communities to make better health and safety decisions and to create universal action [60]. Children are active agents in their own risk taking behaviour and, likewise, their own injury prevention. Targeting children in prevention efforts has been shown to—in the short term—influence their safety behaviour and that of their parents [70,72]. The hierarchy of control aims to create safer workplace environments by providing a step-by-step ranking approach to eliminate or reduce risks of injury [98,99,100] (Figure 2). Hazard elimination is viewed as the highest level of control in the hierarchy (e.g., filling in old sheep dips or removing keys from vehicles). Further stages are the reduction of risk through substitution (e.g., substitute horse/motorbike with transport appropriate for the child’s size) and engineering controls (e.g., altering motorbike throttle speed). The use of administrative controls (e.g., family rules that everyone must follow) and personal protective equipment (PPE) (e.g., helmets and seatbelts) are the final stages of the hierarchy [98,99]. Dosman and colleagues [101] suggested that these steps could be used as a series of options—selected depending on the agricultural task or process involved. Adherence to a variety of the controls is more reliable in reducing injury risk [99,101]. While, the hierarchy has not been evaluated specifically in relation to children on farms, it has been well documented in workplace injury prevention. This may be important, as the elimination of children from the farming workplace is often not possible, and involving the family is strongly engrained within the farming lifestyle. Lower and Temperley [79] estimated that 70% of quad incidents could be averted using engineering controls (such as crush protection devices), PPE (such as helmets), and hazard elimination (ensuring children do not ride). Safety campaigns have promoted the need for children to wear safety protection, such as a helmet, when riding horses and motorbikes; the need for seatbelts when driving in a utility around the farm; and features such as a safety switch in the workshop or on machinery seats when a driver is not seated [102]. However, the actual implementation of these is largely unknown, as is the removal rates of these safety switches after purchase. While the majority of engineering controls target farming adults, some concentrate on children, for example, throttle adjustment to limit motorbike speed [18]. Limited regulation of the agricultural industry means most engineering controls must be voluntarily adapted [17,82] or legislated—such as the recent Australian Competition and Consumer Commission announcement regarding improved safety requirements for quad bikes manufactured or supplied in Australia [103]. If farmers do not employ workers, they may be exempt/think they are exempt from many workplace rules as they are performing work on private property. Given that the children assisting may not be considered employees, the adoption of engineering controls may not be compulsory. PPE can protect the body from injury. However, as seen in the hierarchy of control, this is the lowest level of control. There is also limited understanding of the utilisation of the wide range of PPE available to children on farms. National coronial evidence highlighted that only 20–25% of children who died in a quad bike fatality were wearing helmets [47,50]. A higher rate of helmet use was identified among non-fatal injuries, with 36% of hospitalisations noting the use of helmets [48]. While this suggests the protective influence of helmets against more serious injury, the evidence was based on data from only one hospital, and may not be generalisable. Self-report surveys of farmers in New South Wales identified higher rates of helmet use—38% of quad and 46% of motorcycle riders, and 73% of horse riders [37]. However, self-report responses are vulnerable to participants providing socially desirable answers. Supervision is the most common method of protecting children from danger. Young children require constant supervision with a safe play area and rules, while observation is imperative for older children performing developmentally appropriate tasks [46,90]. However, inadequate supervision is a leading cause of child farm injury [54]. Over half of Australian child farm deaths between 2001 and 2019 had no active supervision [30]. Active supervision is described as specific and overt behaviours by parents/supervisors with children, such as scanning, escorting and interacting, to prevent problem behaviour [104]. Simultaneously performing farm work and providing effective supervision is challenging [105]. Stiller and colleagues [46] identified that 58% of farmers with children living on, or regularly visiting their farm, provided care while in the farming workplace. Only 30% of these farmers acknowledged that this was due to limited access to childcare services, suggesting that many parents consciously choose to take children into the farm workplace. This may be due to the long work hours, days of work, cost, isolation from family networks, potential travel required to utilize services, or the desire to engage children in the farming lifestyle. Safe play areas are effective safety measures for the prevention of child farm-related injury by providing secure boundaries to contain the movement of young children (under five) into unsafe work areas [42,52,53,105,106]. Safe areas are recommended ahead of other active interventions, such as supervision, as there is less chance of error. However, combining a safe play area with close active supervision has the potential to reduce the risk of child injury across all aspects of the farm [106]. Although broadly promoted, research suggests that the uptake of safe play areas is limited [53,81] and varies between commodity groups [25]. Surveys of farmers in the Australian state of New South Wales suggested that only 44% of farms had safe play areas with a fence that would be difficult for a small child to infringe [106]. Further work is required to understand how to engage farmers and identify barriers to normalising and adopting this relatively simple and cost-effective safety measure [53]. A range of limitations exist in the data used to quantify child farm-related injuries, potentially resulting in conservative statistics. These include: The National Farm Injury Data Centre consistently monitors Australian print media for farm-related injuries, providing annual reports on both farm-related fatal and non-fatal injuries [107]. However, farm-related injuries are not always published in the media, especially in the case of non-fatal injuries.Child farm-related injuries/fatalities are only captured from hospital and coronial records—less severe injuries are missed (e.g., attendance at General Practitioner clinics or bush nursing centres) [30].Hospital and coronial records are collected differently between jurisdictions and so are likely to capture varied data types. Additionally, these records are not developed for the purpose of injury research.The process of data linking is vulnerable to errors [41].Injury data does not consistently capture occupation or place of injury, resulting in records not including farm identifiers.Reliance on self-reporting farm residence can be subjective.The number of children living on—or visiting—farms remains unreported, resulting in challenges determining the rate of children injured on farms.Child farm injuries are not always accurately reported, as children helping out in the family business are not classified as employees and, consequently, not captured in workers compensation data. Data accuracy is reliant on parents reporting their child’s activity at the time of injury. Parents may ‘hide’ injuries or describe them as ‘recreational incidents’, not as workplace accidents.The National Farm Injury Data Centre consistently monitors Australian print media for farm-related injuries, providing annual reports on both farm-related fatal and non-fatal injuries [107]. However, farm-related injuries are not always published in the media, especially in the case of non-fatal injuries.Child farm-related injuries/fatalities are only captured from hospital and coronial records—less severe injuries are missed (e.g., attendance at General Practitioner clinics or bush nursing centres) [30].Hospital and coronial records are collected differently between jurisdictions and so are likely to capture varied data types. Additionally, these records are not developed for the purpose of injury research.The process of data linking is vulnerable to errors [41].Injury data does not consistently capture occupation or place of injury, resulting in records not including farm identifiers.Reliance on self-reporting farm residence can be subjective.The number of children living on—or visiting—farms remains unreported, resulting in challenges determining the rate of children injured on farms.Child farm injuries are not always accurately reported, as children helping out in the family business are not classified as employees and, consequently, not captured in workers compensation data. Data accuracy is reliant on parents reporting their child’s activity at the time of injury. Parents may ‘hide’ injuries or describe them as ‘recreational incidents’, not as workplace accidents.The reviewed literature provides a foundation for the understanding of child farm-related injuries in Australia. It is apparent the representation of children in farm-related injury statistics has remained consistent over the last decade [30,39]. Furthermore, the literature is in agreement on the key farming hazards that invariably cause these injuries. Although these hazards (water bodies, quad bikes, motorbikes, utility vehicles/cars, tractors and horses) have been well-identified, they remain the key sources of child farm-related injury [30]. This review also identified factors noted in the literature as contributing to the increased vulnerability of children on farms. It is important to understand the limitations in the current literature and key gaps in knowledge. To date, Australian research on child farm-related injury has largely been quantitative in design, resulting in little understanding of typical family farming behaviours. As this review highlights; it is known that children experience injury on farms, and the key agents responsible for these injuries are well-identified. However, research is yet to explore children’s level of exposure to these farming hazards. Furthermore, the typical attitudes, behaviours and actions of both children and parents on farm safety remain unknown. In the early 2000s, Marlenga and colleagues [84] highlighted this, acknowledging that patterns of childhood injury were frequently interpreted without an understanding of their level of exposure to occupational risk. In order to prevent these injuries from happening, an understanding of the social context of child farm-related injuries is required. Research suggests that farmers are often willing to share their stories to minimise the risk for others [108]. Absent from the literature is the exploration into the lived experiences of child farm-related injury. First-hand experiences of injury—if conveyed appropriately—can provide an invaluable source of knowledge. While an emotive topic, lived experience can be used to increase the understanding of context and long-term consequences of injury and inspire positive action to reduce risks of farm-related injury. Typically, farmers view their peers as relatable and trustworthy, with narrative-based stories from fellow farmers being more influential than statistics [108]. Farm-related injury can have lasting effects on the mental wellbeing of victims as well as family members [109]. Farmers can hold attitudinal barriers that undermine changes in beliefs regarding the adoption of safer farming practices [2]. In order to achieve a behaviour change, firstly, a desire for change is required as well as ensuring farmers have the capacity and information required to take action [110]. As identified in the review, evaluation is yet to be conducted on the previously identified intervention strategies [11,30]. Evaluating these may provide insight into the social and behavioural determinants influencing parents’ decision making in regard to their child’s farm safety. Lack of adaption may potentially indicate a lack of concern or knowledge. Furthermore—and given that injury rates remain steady—understanding the rates and patterns of implementation of farm safety measures would assist in identifying whether these previous interventions were inappropriate, ineffective, or not widely implemented. This is imperative in guiding the development of future prevention strategies. Summers and colleagues [74] described that, in order to develop effective farm safety programs, both parents and children need to be involved. Ehrlich and colleagues [111] further explained this need as they matched parent and child surveys on their knowledge, habits and attitudes around safety behaviours and identified parental reports on their children’s behaviours as often inaccurate. Aside from three studies in the early 2000s, engagement with children in farm safety research in Australia has been minimal [70,72,112]. The behaviours and attitudes of children and adolescents influence not only their immediate health, but also their health and wellbeing into adulthood. Resnick and colleagues [113] reported “at least 70% of premature adult deaths reflect behaviours started or reinforced during adolescence” (p. 1565). There is a need to understand child and adolescent health as it can mean improved future health and ultimately economic outcomes, due to enhanced productivity [113]. Injury prevention needs to be viewed as a whole family issue—engaging both parents and children is important in creating long-term change to safety cultures and behaviours on Australian farms. Over the last 20 years, the Australian agriculture industry has increased production, and the predicted future returns remain positive [114]. The majority of agriculture production investment and research has focused on the safety and demands of the consumer, such as the traceability of products for food safety, quality, provenance, authenticity and the animal welfare practices associated with production [115]. A greater focus should be provided on the health and safety of the farming families producing the food and fibre [116]. In many ways, the need for improved child farm safety has never been greater [30,117]. Farmsafe Australia [62] recently highlighted farming culture as a key contributing factor to the consistent injury statistics over the last decade, suggesting the need to target the next farming generation. Pickett and colleagues [118] wrote a powerful editorial piece about their work in Canadian coronial courts, describing the desperate need to address the issue of child farm-related injuries: “I find that I am angry, but am not sure at whom – at farm parents who expose their children to risks, at a rural society that appears to accept these tragedies as part of their fate, at the coroners and health and safety professionals that have yet to challenge the status quo. This is no longer an academic exercise. Something must change.” (p. 259) As Barbara Lee, the Director of the United States National Children’s Center for Rural and Agricultural Health and Safety, highlighted, “children depend on us—as parents, as farmers, as medical providers and as safety professionals to keep them safe” [[119], p. 1]. This call to action is equally relevant to child farm safety in Australia. An evolution of farming culture and behaviours/practices to create a desire for change is required for an improvement in farm safety to occur [30]. A greater understanding of the context in which child farm safety occurs is a necessary precursor to such evolution. J.A., A.K., J.C. and S.B. conceived the study; J.A. designed and wrote the draft paper; A.K., J.C. and S.B. supervised the study. All authors contributed to reviewing the draft. All authors have read and agreed to the published version of the manuscript. J.A. is supported through an Australian Government Research Training Program Scholarship.Not applicable/not required for this study.Not applicable.The authors declare no conflict of interest.PRISMA diagram of database research publication selection [21].Hierarchy of Control (modified from the National Institute for Occupational Safety and Health) [100].Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_7/ijerph-18-11-06064.txt ADDED Viewed

	@@ -0,0 +1 @@

+ Healthcare professionals have a key role in promoting physical activity, particularly among populations at greatest risk of poor health due to physical inactivity. This research aimed to develop our understanding of healthcare professionals knowledge, decision making and routine practice of physical activity promotion with older adults. A cross-sectional survey was conducted with practicing healthcare professionals in general practice, physiotherapy, occupational therapy and nursing in Ireland and Northern Ireland. We received 347 eligible responses, with 70.3% of all respondents agreeing that discussing physical activity is their job and 30.0% agreeing that they have received suitable training to initiate conversations with patients about physical activity. Awareness of the content and objectives of national guidelines for physical activity varied considerably across the health professions surveyed. Less than a third of respondents had a clear plan on how to initiate discussions about physical activity in routine practice with older adults. Assessment of physical activity was not routine, neither was signposting to physical activity supports. Considering the COVID-19 pandemic and its implications, 81.6% of all respondents agreed that healthcare professionals can play an increased role in promoting physical activity to older adults as part of routine practice. Appropriate education, training and access to resources are essential for supporting healthcare professionals promotion of physical activity in routine practice. Effective physical activity promotion in healthcare settings has the potential for health benefits at a population level, particularly in older adult populations.The health benefits of physical activity for older adults are well established [1]. There is strong evidence that physical activity contributes to increased physical function, reduced impairment, independent living, and improved quality of life in both healthy and frail older adults [2]. International guidelines recommend that all older adults (65+ years) should aim to do at least 150–300 min of moderate intensity or 75–150 min of vigorous intensity aerobic activity throughout the week, with muscle strengthening and multicomponent balance training on 2 or more days per week [3]. In addition to recommending that all older adults should undertake regular physical activity, these guidelines also emphasize the benefits to older adults of ‘moving more’ and sitting less throughout the day, as doing some physical activity is better than none [3]. However, for many older adults, ageing is defined by rapid declines in levels of physical activity, loss of mobility and functional independence, and premature morbidity [4]. Therefore, this stage of life represents an important period for promoting physical activity to improve functions of daily living and slow progression of disease and disability [5].Effective national action to reverse trends in inactivity across the life course requires a ‘systems-based’ approach, with action across different sectors [6]. As part of this approach, healthcare services can play an important role by implementing systems for patient assessment and counselling, and by strengthening the provision of and access to opportunities and programmes that enable older adults to increase their levels of physical activity [7]. Healthcare professionals play a central role in the promotion of a variety of health behaviours and are ideally positioned to promote physical activity [8]. Evidence suggests that they can positively impact patient behaviour by routinely assessing physical activity levels and by using brief practical interventions (advice or counselling on how to initiate and maintain healthy behaviours) with links to community-based support for behaviour change [9]. These are key actions in promoting health enhancing physical activity to reduce the incidence of chronic disease and/or to manage a range of chronic conditions and weight maintenance [10]. Indeed, recent healthcare policy developments on the island of Ireland (Ireland and Northern Ireland) recognize that preventing and reducing chronic disease across the life course by addressing inactivity (amongst other modifiable risk factors) requires a cultural shift on improving health and wellbeing through strategies that focus on health promotion and disease prevention [11].Enhancing our understanding of current practice in relation to physical activity promotion in the health services is crucial to inform the development of evidence-based strategies for improving uptake of policy into practice. Previous research suggests that healthcare professionals instigate brief physical activity interventions opportunistically in a quarter of appropriate instances [9] and there is a disparity between the development of guidelines for physical activity and their dissemination and integration into routine clinical practice for many healthcare professionals [12,13,14]. Barriers to physical activity promotion have been reported by healthcare professionals, including lack of awareness, expertise, and lack of time and incentive [8]. It is unclear what level of knowledge around physical activity exists across a broad range of healthcare professionals on the island of Ireland, and to what extent physical activity promotion with older adults is involved in decision making in routine practice. This research is therefore focused on developing our understanding of healthcare professionals approaches to promoting physical activity to older adults in Ireland and Northern Ireland by evaluating their knowledge of current physical activity guidelines and exploring the factors which may influence their clinical judgements and decision making in promoting physical activity to older adults in routine practice. As this research took place during the COVID-19 pandemic, healthcare professionals perceptions of the implications of the public health and social measures for older adults’ levels of physical activity, and their resultant behaviour(s) related to physical activity promotion with older adults in routine practice were also explored.This was a cross-sectional study of practicing healthcare professionals in general practice, physiotherapy, occupational therapy and nursing in Ireland and Northern Ireland in 2020. Healthcare professionals who were not registered to practice in one of these four professional groups, those who were retired, working outside of Ireland or Northern Ireland, or who did not have clinical contact with older adults (defined as ≥65 years) as part of routine care were excluded from the study.A 43-item (3-section) survey was developed using the theoretical domains framework (TDF). The TDF is an integrative framework of theories of behaviour change developed to identify influences on healthcare professional behaviour in the implementation of evidence-based recommendations [15]. It has been used extensively to identify barriers and facilitators for individual uptake of evidence-based practices and for implementation design and research [16].Section 1 of the survey captured demographic and employment data as well as healthcare professionals self-reported levels of physical activity, knowledge of physical activity guidelines and awareness of resources to facilitate knowledge and practice development. Section 2 assessed TDF domains of healthcare professional’s behaviour in assessment, discussion, and prescription of physical activity in routine practice. Healthcare professionals perceptions of the implications of the COVID-19 pandemic on older adults’ levels of physical activity and views on the role of physical activity promotion to older adults in light of the pandemic were also examined. Section 3 of the survey used clinical vignettes (3 for each healthcare profession) to enable healthcare professionals to self-report routine practice in relation to physical activity promotion with older adults. Analysis and findings for Section 3 of the survey are reported in a separate publication.A research project advisory group (N = 10) participated in pilot testing, refinement, and approval of the survey. This group included appointed representatives of the Royal College of General Practitioners in Northern Ireland and the Irish College of General Practitioners in Ireland; the Chartered Society of Physiotherapy in Northern Ireland and the Irish Society of Chartered Physiotherapists in Ireland; the Royal College of Nursing in Northern Ireland and the Department of Public Health Nursing in Ireland; and the Royal College of Occupational Therapists in Northern Ireland and the Association of Occupational Therapists Ireland. The survey was live for a 2-month period from mid-August to mid-October 2020, during which time the link to the survey was promoted through focused email distribution and promoted widely on social media.All eligible returned surveys were included in the analysis regardless of missing data; consequently, the number of total responses for each survey item is varied. Descriptive and explorative analysis of the data were performed using IBM SPSS V.24 (IBM Inc., Armonk, NY, USA). Descriptive statistics were used to describe data from the survey reporting frequencies of responses. Pearson’s chi squared tests were used to compare knowledge of guidelines for physical activity with assessment, discussion, prescription, and signposting of physical activity in routine practice. Knowledge of the guidelines was classified by dividing the participants into two groups: (1) those who correctly recalled three specific components of the physical activity guidelines for older adults (number of minutes per week of moderate intensity physical activity, vigorous intensity physical activity, and the number of days of strength, balance, and flexibility activities recommended for optimal health benefits); and (2) those who indicated that they did not know the guidelines, or who incorrectly recalled one or more specific component of physical activity guidelines for older adults. Statistical significance was set at p value <0.05.In total, 573 responses were received. Of these, 143 did not meet the inclusion criteria, and a further 83 did not complete components of the survey required to be included in subsequent analysis. A total of 347 respondents met the inclusion criteria and their responses were included in subsequent analysis. Participant characteristics are presented in Table 1. Of the respondents, 44 (12.7%) were male and 299 (86.2%) were female. Three quarters of respondents (74.4%) were healthcare professionals practicing in Ireland. Nearly half of all respondents (49.0%) were physiotherapists. The proportion of all respondents who reported that they had 26+ years of practice experience was 27.4% (n = 95). Most respondents worked in the public sector (80.7%, n = 280). The proportion of respondents who achieved the recommended level of moderate intensity physical activity over a week was 40.6% (n = 141).Responders were asked how aware they were of the content and objectives of national guidelines in their jurisdiction. Of all respondents, 42.7% (n = 148) (62.9% of physiotherapists, 34.2% of nurses, 22.2% of general practitioners and 19.4% of occupational therapists) agreed that they were aware of the content and objectives of national guidelines for physical activity in their jurisdiction (Table 2). Of all respondents, 35.4% (n = 123) agreed that they were aware of the content and objectives of national guidelines for physical activity for older adults in their jurisdiction. The percentage of those that ‘agreed’ varied considerably across the healthcare professions surveyed (Table 2).Responders were also asked about their knowledge of three specific components of physical activity recommendations for older adults (number of minutes per week of moderate intensity physical activity, vigorous intensity physical activity, and the number of days of strength, balance, and flexibility activities recommended for optimal health benefits). Of all respondents, 61.1% (n = 212) reported that they knew how many weekly minutes of moderate intensity physical activity were recommended for older adults, and 38.6% (n = 134) of all respondents reported that they knew how many minutes of weekly vigorous intensity physical activity were recommended for older adults. However, when prompted, 49% (n = 170) correctly recalled the number of minutes of moderate intensity activity, and 27.4% (n = 95) correctly recalled the number of minutes of vigorous activity. 54.2% (n = 188) of all respondents reported that they knew how many days per week older adults were recommended to perform strength, balance, and flexibility training. The proportion who correctly recalled the number of days when prompted was 51.6% (n = 179). Of all respondents, 24.5% (n = 85) correctly recalled all three specific components of physical activity guidelines for older adults (Table 2).Most respondents (64.6%) agreed that physical activity guidelines have a place in routine practice. However, only 26.5% of physiotherapists, 15.8% of nurses, 14.0% of occupational therapists, and 5.6% of general practitioners agreed that there is sufficient time allocated to implement physical activity guidelines for adults/older adults in day-to-day work (Table S1).Of all respondents, 47.6% (n = 165) (64.1% of physiotherapists, 47.4% of nurses, 33.0% of occupational therapists and 11.1% of general practitioners) reported that they were aware of resources to facilitate their knowledge development and practice of assessment/discussion/prescription of physical activity with patients as a part of routine care. The most frequently cited resources included government health department websites, professional body websites and health profession specific websites, and training programmes and research projects with online physical activity resources and toolkits (Figure 1).Table S1 shows the theoretical domains of healthcare professional’s behaviour in assessment, discussion, and prescription of physical activity in routine practice.Responders were asked about their use of screening tools to measure patients’ physical activity levels. Nearly half of all respondents (48.7%, n = 169) reported that they ‘never’ formally assess whether a patient is active or inactive as part of routine practice. The proportion who reported that they ‘sometimes’, ‘usually’, or ‘always’ assess whether a patient is active or inactive was 40.0% (n = 139). A variety a functional assessment tools (e.g., the Timed Up and Go (TUG) test, 6-min walk test), questionnaire based (e.g., General Practice Physical Activity Questionnaire (GPPAQ)) and device-based measures of physical activity (e.g., pedometers) were reported. Having knowledge of three specific components of physical activity guidelines for older adults was significantly associated with formally assessing whether a patient is active or inactive as part of routine practice (Table S1).Responders were asked if they considered it a part of their professional role to promote physical activity to patients. Most respondents (70.3%, n = 244) agreed that discussing physical activity with patients was part of their work as a healthcare professional. Many (74.1%, n = 257) ‘agreed’, or ‘somewhat agreed’ that it was easy to remember to discuss physical activity with patients (74.1%, n = 257) and that they were confident that they could discuss physical activity as part of routine practice even when the patient was not motivated (72%, n = 250) or when there was little time (63.8%, n = 221). Overall, 30.0% of all respondents (n = 104) agreed that they had received suitable training to initiate conversations with patients about physical activity. This percentage was higher for physiotherapists (44.1%) (Table S1). We found that 36.1% of general practitioners, 21.4% of occupational therapists, and 23.7% of nurses did not agree that they have received suitable training to initiate conversations with patients about physical activity.Most respondents (68.0%, n = 236) agreed that discussing physical activity with older adults was part of their work as a health professional, with the majority ‘agreeing’, or ‘somewhat agreeing’ that it is something they do automatically (67.6%, n = 233) and is useful (82.4%, n = 286). Many ‘agreed’, or ‘somewhat agreed’ that they were aware of how to initiate conversations about physical activity with older adults (83.5%, n = 290), that they had the skills to initiate conversations with older adults about physical activity (80.1%, n = 278) and that it was something that they typically did within their organization (72.1%, n = 250). There was a significant association between having knowledge of three specific components of physical activity guidelines for older adults and initiating conversations with patients about physical activity as part of routine practice (Table S1). However, even though many respondents agreed that that they intended to discuss physical activity in their next consultation/appointment with an older adult as part of routine care (55.0%, n = 191), fewer agreed that they had a clear plan on how to initiate discussions about physical activity in routine practice with older adults (30.5%, n = 106).Responders were asked how often they refer/signpost patients through referral programmes or community-based schemes. We found that 12.1% of respondents ‘always’ signposted patients to other physical activity services (i.e., exercise referral programmes/community-based physical activity initiatives). Having knowledge of three specific components of physical activity guidelines for older adults was significantly associated with signposting patients to other physical activity services as part of routine practice (Table S1). Most physiotherapists (64.1%), occupational therapists (63.1%), general practitioners (61.2%), and 50.0% of nurses reported that they ‘sometimes’ or ‘usually’ signposted patients to other physical activity services. The physical activity services that they ‘sometimes’, ‘usually’, or ‘always’ signposted to patients included exercise referral programmes; community-based active retirement groups; community-based healthcare professional follow-up (e.g., fall prevention groups, community physiotherapy/occupational therapy services); online self-management groups/virtual classes and resources. Most respondents (68.6%) agreed that assessing/discussing/prescribing physical activity with an older adult as part of routine practice would benefit the public health agenda. 23.3% of occupational therapists, 21.1% of nurses and 2.8% of general practitioners agreed that they were supported to use physical activity assessment/discussions/prescription in everyday practice. This number was higher for physiotherapists (38.8%).To explore healthcare professionals understanding of the potential impact of public health restrictions on older adults’ levels of physical activity, and their resultant behaviour(s) related to physical activity promotion with older adults in routine practice, the following questions were asked (see Table 3). The proportion of all respondents who indicated that the public health and social measures introduced to prevent the spread of COVID-19 had reduced older adults’ levels of physical activity was 71.2%, and considering this, 81.6% of all respondents agreed that healthcare professionals can play an increased role in promoting physical activity to older adults. Overall, 84.7% stated that they were ‘more likely’ (47.8%), or the ‘same as usual’ (36.9%) to discuss physical activity with older adults as part of routine practice considering the COVID-19 pandemic and its implications.Healthcare professionals play a pivotal role in educating patients about the benefits of being more active and motivating their patients to engage in a more active lifestyle [18,19]. In this study most respondents agreed that discussing physical activity is their job (TDF domain: Social/Professional Role and Identity), and that it is easy to remember to do (TDF domain: Memory, Attention & Decision processes). However, the majority had not received suitable training in initiating discussions about physical activity with patients (TDF domain: Skills). Our survey also shows that many healthcare professionals are unaware of current guidelines for physical activity in older adults (TDF domain: Knowledge)—one in four correctly answered questions about the content of these guidelines, and less than a third of respondents had a clear plan on how to initiate discussions about physical activity in routine practice with older adults (TDF domain: Behavioural Regulation).The results of this survey have identified a range of theoretical domains that can be targeted to support healthcare professionals in their role of promoting active lifestyle change with patients. In particular, the domains of Knowledge, Skills and Behavioural Regulation were identified. These domains map directly onto the ‘Capability’ component of the COM-B model of behaviour change [20]. Building healthcare professionals’ ‘Capability’ to promote physical activity in routine practice through ‘Knowledge’ development: appropriate education on guidelines for physical activity in prevention and treatment of disease is essential. In this study, having a detailed knowledge and recall of physical activity guidelines for older adults was associated with formal assessment, initiating discussion, and referral/signposting to physical activity services as part of routine practice. Building capability through ‘Skill’ development: relevant training on initiating discussions/brief interventions to support patients with behavioural change (perhaps through motivational interviewing) is equally important. Evidence suggests that brief practical interventions by clinicians may improve short and long-term engagement with active lifestyles [21] and that components of motivational interviewing are central to this approach as a theory consistent and evidence-based technique to strengthen an individual’s motivation for change [22].The recently introduced Making Every Contact Count (MECC) strategy places the responsibility for providing brief (and opportunistic) interventions on all healthcare professionals who may have patient contact in Ireland and Northern Ireland. By integrating MECC at the initial/undergraduate level it is planned that brief interventions will become central to many consultations [23]. In this survey most respondents were fully qualified with regular patient contact for more than 6 years, which suggests that continuing education/professional development on physical activity promotion will be crucial to maximise their impact on and support for sustained behaviour change with their patients. This finding is consistent with other studies that have highlighted the need for postgraduate training for healthcare professionals to address health behaviour change [23,24], but also the need to integrate physical activity training and its relationship with health at an undergraduate level. Several reports indicate that this issue is now receiving greater attention in undergraduate curricula [25,26].Many healthcare professionals surveyed never formally assess whether a patient is active or inactive as a part of routine practice. In part, this may reflect the level of training and support that healthcare professionals surveyed have received on physical activity promotion broadly, and on physical assessment more specifically. The routine assessment of physical activity (and sedentary behaviour) in the health system is the basis for the surveillance of physical inactivity as a risk factor [27]. Assessing a patient’s level of physical activity can provide a valuable insight into health status and is an essential first step that can lead to important intervention opportunities, if appropriate. Those who reported that they assess physical activity and/or functional status in older adults used a variety of tools and resources. Whereas the use of tools to support the systematic screening and delivery of brief interventions is recommended, some have advocated the use of a standardised physical activity tool as a ‘vital sign’ in patients’ consultations, that should be a standard of care for all patient visits with the potential to highlight inactivity and prompt a brief intervention (counselling or referral) [28].Few respondents ‘always’ signposted patients to other physical activity services (i.e., exercise referral programmes/community-based physical activity initiatives). The majority reported that they did this ‘usually’ or ‘sometimes’. Raising awareness of community resources (passive or active signposting) or prescribing activity as a means of referral is seen as a ‘formal’ acknowledgement of the problem (inactivity) and provides both a legitimacy to the issue and an opportunity to do something about it. Previous research has suggested that patients ‘like’ the option of being connected to resources on specific physical activity opportunities by a health professional, to consider and potentially follow up on [29].How the topic is raised and linked to a patient’s specific health conditions is central to patient acceptance to the topic [29], highlighting again the need for specific and ongoing education and training on initiating discussions and delivering brief interventions as part of routine care. Research suggests that the ‘motivation’ provided by a healthcare professional is key to whether a patient accepts offers of being signposted to community physical activity opportunities [30]. The potential for patients to meet with people in similar circumstances from the wider community to engage in physical activity can provide an opportunity to build relationships and a supportive solution to increase activity levels [31]. Social support is important for the adoption and maintenance of physical activity, particularly for older adults [32].The role that healthcare professionals play in promoting physical activity as part of routine care, and the support that they provide for older adults will be increasingly important in addressing the fall-out from COVID-19 on older adult’s health. It is likely that the proportion of the older adult population inactive and at risk from disease and disorders related to inactivity will have increased [33]. Those who are socio-economically disadvantaged, frail, living with multimorbidity or disability or living in residential care, may have been disproportionately affected [34].This study captured views from a diverse range of healthcare professions. To the authors’ knowledge, this is the first study to have completed this type of analysis across the island of Ireland. The TDF was utilised as an evidence-based method for identifying individual and organisational determinants of health professionals’ behaviour and decision-making. In this context, questions used within the questionnaire had established content validity and extensive piloting and pre-development work, which with the input of an expert advisory group, improved the overall validity of the final survey. The questionnaire was distributed through professional body networks across the island of Ireland and the number of valid responses (conceivably affected by the COVID-19 pandemic, during which healthcare professionals were facing unprecedented demands on their time) compares favourably to other studies conducted with healthcare professionals. However, consideration should still be given to the generalisability of the survey findings. Selection bias is an issue that needs to be considered in this context, as it is possible that healthcare professionals who are interested in and utilise physical activity in routine practice were more motivated to participate. The smaller number of respondents from general practice and nursing (relative to physiotherapy and occupational therapy) is also a potential limitation of the research.Healthcare professionals have a key role in the promotion of physical activity as part of a whole-systems approach. This is highlighted in the strategic objectives of National and International Policy.This study has shown that healthcare professionals consider it a part of their role to discuss physical activity, and many reported that it was feasible to initiate discussions about physical activity even in the face of commonly reported barriers (little time available/patient not motivated).Successful implementation of physical activity promotion in routine practice will have substantial health benefits at a population level, particularly for older adults who stand to benefit the most from increasing levels of activity. However, continuing education and training is essential to support healthcare professionals’ knowledge and skill development if they are to be successful in this role.The following are available online at https://www.mdpi.com/article/10.3390/ijerph18116064/s1, Table S1: Theoretical domains of healthcare professionals’ behaviour in assessment, discussion, and prescription of physical activity in routine practice.C.C. and R.O. were involved in the conception, design and methodology, administration, analysis, and writing—original draft preparation, review, and editing. Both authors have read and agreed to the published version of the manuscript.This research received no external funding.The study was conducted according to the guidelines of the Declaration of Helsinki and approved by an Independent Peer Review panel (Ref: 2020-03-HCP).Informed consent was obtained from all subjects involved in the study.Data is contained within the article or supplementary material (Table S1).The authors would like to acknowledge the support from members of the Research Project Advisory Group in pilot testing, refinement, approval, and promotion of the survey.The authors declare no conflict of interest.Resources accessed by participants for knowledge development of physical activity promotion.Characteristics of survey participants.* Measured by Single Item Metric [17].Participant’s awareness of physical activity guidelines and resources.* Questions linked to ‘Knowledge’ domain of TDF ** Of those who replied ‘Yes’ and correctly recalled how many minutes/days when prompted.The COVID-19 pandemic: implications for older adults’ physical activity.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.