ijerph6

Med-MDPI/ijerph_6/ijerph-18-03-01015.txt

@@ -0,0 +1 @@
+We aimed to identify persistent asthma phenotypes among adolescents and to evaluate longitudinally asthma-related outcomes across phenotypes. Adolescents (13–17 years) from the prospective, observational, and multicenter INSPIRERS studies, conducted in Portugal and Spain, were included (n = 162). Latent class analysis was applied to demographic, environmental, and clinical variables, collected at a baseline medical visit. Longitudinal differences in clinical variables were assessed at a 4-month follow-up telephone contact (n = 128). Three classes/phenotypes of persistent asthma were identified. Adolescents in class 1 (n = 87) were highly symptomatic at baseline and presented the highest number of unscheduled healthcare visits per month and exacerbations per month, both at baseline and follow-up. Class 2 (n = 32) was characterized by female predominance, more frequent obesity, and uncontrolled upper/lower airways symptoms at baseline. At follow-up, there was a significant increase in the proportion of controlled lower airway symptoms (p < 0.001). Class 3 (n = 43) included mostly males with controlled lower airways symptoms; at follow-up, while keeping symptom control, there was a significant increase in exacerbations/month (p = 0.015). We have identified distinct phenotypes of persistent asthma in adolescents with different patterns in longitudinal asthma-related outcomes, supporting the importance of profiling asthma phenotypes in predicting disease outcomes that might inform targeted interventions and reduce future risk.Asthma is one of the most common chronic diseases in children worldwide [1]. Despite advances and changes in guidelines, there is no known treatment for asthma and the main goal, which is to achieve disease control, remains challenging [2].Children with uncontrolled asthma need to use asthma medication more frequently and are more likely to use healthcare services due to their asthma, with an increase in unscheduled medical visits, and hospital admissions [3,4,5]. Likewise, adolescents have poorer outcomes and worse adherence due to a lack of self-management skills and insufficient health literacy knowledge [6].Classification and understanding of heterogeneous asthma phenotypes are the starting point to establish individualized management plans [7] and might lead to improvements in asthma control. Recent studies using data-driven methods provided novel insights into meaningful and accurate asthma phenotypes based on real-life data [8,9,10].However, phenotypic characterizations of adolescents with asthma are very limited, and longitudinal studies that potentially predict long-term outcomes and personalizing treatments are scarce. Moreover, variables collected longitudinally are of extreme importance to evaluate the stability of any derived phenotypes and to further validate them, particularly in unselected subjects with asthma from the general population. The INSPIRERS studies assessed adherence to inhaled medication among adolescents with persistent asthma, collecting real-life data over time, with the potential to support and enable patient-centered care and research [11,12].Therefore, this study aims to identify persistent asthma phenotypes among adolescents and to evaluate longitudinally asthma-related outcomes across phenotypes.We performed a secondary analysis of data from adolescents (13–17 years) enrolled in the prospective, observational, and multicenter INSPIRERS studies, which have been described previously [11,13]. Briefly, the INSPIRERS studies assessed adherence to asthma inhalers among adolescents and adults with persistent asthma. During a face-to-face baseline medical visit, patients were invited by their physicians to participate in the study, and a convenience sample was obtained, between March 2018 and January 2020, at 30 primary and secondary care centers from Portugal and Spain.The study protocol was approved by the Ethics Committee of all participating centers. Before enrolment in the study, adolescents signed an assent form and a written consent form from the parent(s) or legal guardian(s) was also obtained. In this study, we followed the STROBE statement for reporting of observational studies [14].Participants were eligible for this secondary analysis if they: (1) had a previous medical diagnosis of persistent asthma, (2) were between 13 and 17 years old, and (3) had an active prescription for a daily inhaled controller medication for asthma. Exclusion criteria included a diagnosis of a chronic lung disease other than asthma or other chronic condition with possible interference with the study aims.Participants completed a face-to-face baseline visit (T0), where physicians reported patients’ asthma treatment; an assessment of asthma control according to the Global Initiative for Asthma (GINA) [2]; last reported value of percent predicted Forced Expiratory Volume in the first second (FEV1); number of exacerbations in the past year (defined as episodes of progressive increase in shortness of breath, cough, wheezing, and/or chest tightness, requiring a change in maintenance therapy [15]) and of unscheduled medical visits in the past year (consultations at primary care, specialist’s office, hospitalizations, or emergency department).A sociodemographic and clinical questionnaire was administered to the participants, including height and weight, an assessment of asthma control during the previous 4 weeks by Control of Allergic Rhinitis and Asthma Test (CARAT) questionnaire [16]. CARAT total score (CARAT-T) is calculated by summing up the scores of all 10 questions, resulting in a range of 0–30 points. CARAT has two domains: upper airways (CARAT-UA, range: 0–12) and lower airways (CARAT-LA, range: 0–18). A score > 24 on CARAT-T, >8 on CARAT-UA, and ≥16 on CARAT-LA were used to define control regarding total, upper, and lower airways symptoms, respectively [17].Participants were interviewed by phone at 1 week (T1), 1 month (T2), and 4 months (T3) after the face-to-face baseline visit, and their asthma control was again assessed using CARAT. Also, exacerbations and unscheduled medical visits in the past 4 months were recorded at T3.Latent class analysis (LCA) was applied to 10 variables easily collected in medical visits both primary and secondary care, collected at T0: sex (Male/Female), asthma symptom’s onset before the age of 6 years (Yes/No), presence of comorbidities (Yes/No, yes if at least one of the following physician-reported: atopic dermatitis, rhinitis, and rhinosinusitis), body mass index (BMI) ≥ 85th percentile [18] (Yes/No), exposure to environmental tobacco smoke (ETS) (Yes/No), Pre-BD FEV1 < 80% (Yes/No), CARAT-UA and CARAT-LA (controlled/uncontrolled), ≥1 exacerbation (Yes/No), and ≥1 unscheduled healthcare visits (Yes/No). The optimal number of classes resulting from the variables was determined by evaluating k classes versus k-1 classes sequentially, until adding a class no longer significantly improved the model, measured by Lo-Mendell-Rubin-adjusted likelihood ratio test. The best model was determined by the largest entropy and the lowest Bayesian information criteria (BIC) values [19]. Longitudinal differences in clinical variables (CARAT, exacerbations, and unscheduled medical visits) were assessed by comparison between baseline, T1, T2, and T3 data. The ratio of the number of exacerbations/unscheduled healthcare visits per month was also calculated.Categorical variables were presented as absolute frequencies and proportions. Continuous variables were presented according to their distributions: mean and standard deviation (sd) or median (percentile 25–percentile 75: P25–P75). Group comparisons were performed by independent t-test, Mann–Whitney, and Kruskal–Wallis tests for continuous variables or Chi-square test for categorical variables. Multiple testing was conducted using the Bonferroni correction when needed. Longitudinal changes in each variable were assessed using a generalized linear model with pairwise comparisons of means, to analyze the differences between the classes over time.Statistical analyses were performed using IBM SPSS Statistics V.26.0 (IBM Corporation, Armonk, NY, USA), and MPlus 6.12 (Muthén & Muthén, Los Angeles, CA, USA) was used to conduct LCA analysis. Plots were created using GraphPad Prism V.6.0 (GraphPad Software, La Jolla, CA, USA). The level of significance was set at 0.05.The sample consisted of 162 adolescents with a median (P25–P75) of 15 (14–16) years, 128 of which (78%) completed the three follow-up interviews. Baseline characteristics are shown in Table 1.There were no significant differences between sex and baseline characteristics, except in: GINA asthma control, with females having significantly more proportion of uncontrolled asthma than males (p = 0.033); and CARAT-T and CARAT-LA, with females presenting lower scores, indicating more symptoms and poorer control (Table 1). Moreover, most of the adolescents used a single inhaler (64%) and ICS/LABA was commonly used as a controller asthma medication (Table 1).A three-class model was selected as the best solution for these data (Table 2), with a significantly better fitting than a two-class model (p = 0.001), and a non-significantly different fit from a four-class model (p = 0.195). Furthermore, the entropy of the three-class model was 0.851, a very good overall certainty in classification, and this was the best fit for phenotype identification as it had the lowest BIC value with minimal loss of entropy.Adolescents in class 1 (n = 87) were equally distributed regarding sex, with most of them having at least one comorbidity (72%). Although only 17% had FEV1 < 80%, they were highly symptomatic at baseline, with a high proportion having unscheduled healthcare visits and exacerbations in the past year (57% and 97%, respectively).Class 2 (n = 32) was characterized by a predominance of females (80%), a higher proportion of overweight/obese adolescents, and all had FEV1 above 80%. However, this class had the highest proportion of adolescents with uncontrolled upper/lower airways symptoms, with a low proportion of exacerbations (11%) and without unscheduled healthcare visits.Class 3 (n = 43) included mostly males, half of whom had uncontrolled upper airway symptoms and the majority having controlled lower airways symptoms. Similar to class 2, in class 3 we observed a very low proportion of exacerbations and unscheduled healthcare visits.Of the 162 adolescents included in the baseline visit (T0), 139, 136, and 128 adolescents respectively completed the T1, T2, and T3 follow-up phone interview. Figure 1 and Figure 2 show the longitudinal changes in CARAT scores (across T0, T1, T2, and T3) and in the number of exacerbations/healthcare unscheduled per month (between T0 and T3).Adolescents in class 1 presented an increase in CARAT scores across all time points (Figure 1) and had the highest mean number of exacerbations and unscheduled healthcare visits/month, both at baseline (0.25 and 0.10, respectively) and follow-up (0.44 and 0.14, respectively) (Figure 2).Class 2 at T3 follow-up presented a significant increase in the proportion of adolescents with controlled lower airways symptoms (p < 0.001), but also a significant increase in the number of exacerbations/month (mean difference = 0.22, p = 0.009). The number of unscheduled healthcare visits also increased, although non-significantly (mean difference = 0.14, p = 0.10).In class 3, although most participants kept a high proportion of lower airway symptom control at follow-up, there was a significant increase in the mean number of exacerbations/month (mean difference = 0.27, p = 0.015).In our study, we identified three distinct phenotypes (classes) of persistent asthma in adolescents that presented different patterns in longitudinal asthma-related outcomes. These classes differed significantly concerning airways symptoms, exacerbations, and the need for unscheduled healthcare visits, particularly in their longitudinal changes at follow-up.To identify the phenotypic heterogeneity over time in children and adolescents with asthma, clustering techniques have been applied to broad cohorts, aiming to describe and monitor asthma phenotypes [8,9,10,20]. However, the present study focused on a specific population, adolescents with persistent asthma, and used the obtained phenotypes as a starting point for the follow-up of these asthma phenotypes, enabling the assessment of the trajectories of asthma control.In our study, both at cross-sectional and longitudinal levels, the obtained classes of adolescents with persistent asthma are clinically reasonable. Class 1 was a troublesome exacerbation-prone asthma phenotype with a high proportion of uncontrolled disease and more unscheduled healthcare visits, while class 3 was the mildest phenotype of adolescents with persistent asthma, being similar to other phenotypes found in the literature among this population [8,21,22,23]. Class 2, compared to the other classes, included predominantly female adolescents, who were more frequently obese and had later-onset asthma symptoms (>6 years-old), with a high proportion of uncontrolled disease, but with few self-reported exacerbations/unscheduled healthcare visits. Class 2 had the lowest CARAT scores (T, UA, and LA) at baseline, and significantly improved after four months, even with higher CARAT LA scores than class 1. Moreover, among class 2, there was a meaningful change in CARAT T score (improvement of 5 points), above the minimal clinically important difference reported for CARAT questionnaire (3.5 points) [24], although it was not further assessed in confirmatory studies, namely in adolescents. To our knowledge, this is the first time that this phenotype is described in adolescents with persistent asthma; however, additional studies that include more participants and more comprehensive variables are needed to validate it.All classes reported an increase in the number of exacerbations per month in the follow-up, and this was significant in classes 2 and 3. A possible effect of the seasonality of the exacerbations might be a cause of this increase; however, when interpreting these results, one should bear in mind that a possible memory bias could explain this increase, as participants were likely better able to remember exacerbations during the past 4 months, compared to the first assessment at baseline (focused on the past 12 months). Also, exacerbations at T0 were physician-reported and at T3 were self-reported, without clinical validation. Specifically, class 3 was composed mainly of male adolescents that, although keeping symptom control, reported a significant increase in exacerbations. This suggests an underestimation of their airway symptoms that, together with the fact that adolescents often wish to take charge of their health and/or stop the medication due to stigma [6,25], might lead to poorer outcomes. Moreover, these findings support not only the importance of continuously monitoring childhood asthma to reduce the impact of this disease [1] but also that, in older children, asthma monitoring should be based both on symptomatic patterns and in objective variable expiratory airflow limitation.Indeed, evaluation of symptoms plays a key role in asthma diagnosis and management [2]; however, diagnostic tests, such as lung function tests, are also important for the diagnosis and assessment of disease. In our study, FEV1 was predominantly normal (>80%) in classes 1 and 2, which was in line with the findings of Lee et al. [8] that described that in some clusters/phenotypes of children and adolescents with persistent asthma, lung function was normal, particularly in those with a lower proportion of atopy. However, because we did not assess atopy/sensitization, a direct comparison is limited, and this should be further studied.Similar to our findings, it has already been described that children with uncontrolled asthma need to use more asthma medication and are more likely to use healthcare services for asthma, especially unscheduled medical visits and hospital admissions [4,26]. This points to the importance of improving knowledge on asthma control in the long-term, particularly among adolescents. Innovative mobile health (mHealth) applications could be a means to approach this issue, as smartphones are now widespread [27] and apps are very appealing to inspire behavior changes, through gamification and social support, particularly in this age range [11,28].There are many opportunities regarding longitudinal real-life data combined with mHealth applications, such as InspirerMundi [29] and MASK [30], which are becoming increasingly popular among physicians, patients, and the general public. However, future studies that combine hypothesis-independent clustering and real-life data extracted from mHealth applications applied to asthma diagnosis and management are needed.There are limitations in this study that should be acknowledged. First, the INSPIRERS studies cohort is not a random sample of individuals with persistent asthma, and inclusion was based on the presence of physician-diagnosed asthma without the need for objective measurements, such as lung function tests, to support the diagnosis. However, our findings must be further validated to be generalized to the greater population of adolescents with persistent asthma. Second, the lack of data regarding atopy and viral infections should be considered in the interpretation of our findings, as they play a major role in asthma control and management in later childhood and adulthood [31]. Third, as with any data-driven clustering, there are limitations in the interpretation of derived classes as being a true set of clinically meaningful subgroups [32]; however, no clustering/group effect on the asthma-related outcomes was observed (data not shown). Also, the choice of the variables included in the LCA model was based on parameters being easily reported at a medical consultation, both in primary and secondary care, and that could potentially be useful in disease management; however, the inclusion/replacement of other variables representing different disease domains, such as adherence to treatment, medication and inflammatory biomarkers, should be explored. Finally, the cluster stability could not be assessed because of the limited number of adolescents in T3; however, we aimed to monitor the baseline latent classes concerning longitudinal asthma-related outcomes.We have identified three distinct phenotypes of persistent asthma in adolescents that presented different patterns in longitudinal asthma-related outcomes, supporting the importance of profiling asthma phenotypes in predicting disease outcomes, which might inform targeted interventions and reduce future risk.Conceptualization, R.A. (Rita Amaral) and C.J.; Data curation, R.A. (Rita Amaral), C.J., R.A. (Rute Almeida), A.M.P. and J.A.F.; Methodology, R.A. (Rita Amaral), C.J., A.M.P. and J.A.F.; Writing—original draft, R.A. (Rita Amaral); Writing—review and editing, R.A. (Rita Amaral), C.J., R.A. (Rita Amaral), A.M.P., M.A.-C., S.M., J.C.C.R., J.C., L.A., A.C., A.S., M.F.T., M.F.-M., R.R.A., A.S.M., R.M.F., R.F., P.L.P., N.N., D.B., A.T.B., M.J.C., T.F., J.G., C.V., A.M., M.J.V., P.M.S., J.F., A.M., C.S.P., N.S., C.C.L., A.A., M.L.M., C.L. (Carlos Lozoya), C.L. (Cristina Lopes), F.C., C.C.L., R.C., I.V., S.d.S., E.S., N.R. and J.A.F. All authors have read and agreed to the published version of the manuscript.This research was funded by ERDF (European Regional Development Fund) through the operations: POCI-01-0145-36 FEDER-029130 (“mINSPIRE—mHealth to measure and improve adherence to medication in chronic respiratory diseases—generalisation and evaluation of gamification, peer support and advanced image processing technologies”) cofunded by the COMPETE2020 (Programa Operacional Competitividade e Internacionalização), Portugal 2020 and by Portuguese Funds through FCT (Fundação para a Ciência e a Tecnologia).The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Centro Hospitalar de S. João—EPE (protocol code 258-17 and date of approval: 5th of January 2018). 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 ethical restrictions.The authors want to acknowledge the collaboration and participation of the participants, parents, and physicians involved in the data collection.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.Longitudinal changes in the three CARAT scores (T, UA, and LA), in classes 1 (n = 89), class 2 (n = 31) and class 3 (n = 43). Error bars indicating standard error of the mean. CARAT: Control of Allergic Rhinitis and Asthma Test; T0: baseline assessment; T: total; UA: Upper airways; LA: Lower airways. * significant overall longitudinal changes (p < 0.05). T0: baseline assessment; T1: 1-week; T2: 1-month; T3: 4-month follow-up.Longitudinal changes in the rate of exacerbations and unscheduled healthcare visits per month, among the three classes. Error bars indicating standard error of the mean. Number of subjects with exacerbations: class 1 (T0: n = 81; T3: n = 33), class 2 (T0: n = 3; T3: n = 10) and class 3 (T0: n = 3; T3: n = 16). Number of subjects with unscheduled healthcare visits: class 1 (T0: n = 49; T3: n = 12), class 2 (T0: n = 0; T3: n = 2) and class 3 (T0: n = 0; T3: n = 3). * significant for p < 0.01. T0: baseline assessment; T3: 4-month follow-up.Characteristics of adolescents at the time of enrolment (T0), according to sex (n = 162).1 Comorbidities included rhinitis, rhinosinusitis, or atopic dermatitis. 2 Independent t-test, Mann–Whitney, or Chi-square tests. In bold are statistically significant p value at <0.05. BMI: body mass index; CARAT: Control of Allergic Rhinitis and Asthma Test; T: Total; UA: Upper airways; LA: Lower airways; GINA: Global Initiative for Asthma, Pre-BD: Pre-bronchodilator; FEV1: Forced expiratory volume in one second; ICS: inhaled corticosteroid; LABA: Long-acting beta-agonists; SABA: Short-acting beta agonists; LAMA: Long-acting muscarinic antagonists.Results of latent class analysis at the time of enrollment (T0).Numbers are presented as n (%). Three participants were excluded due to missing data on the main variables in the model. 1 Chi-square test. In bold are statistically significant p-value at <0.05. BMI: body mass index; ETS: environmental tobacco smoke; Pre-BD: pre-bronchodilator; FEV1: Forced expiratory volume in one second; CARAT: Control of Allergic Rhinitis and Asthma Test; UA: upper airways; LA: Lower airways.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01016.txt

@@ -0,0 +1,23 @@
+Owing to the convenience, reliability and contact-free feature of Mobile payment (M-payment), it has been diffusely adopted in China during the COVID-19 pandemic to reduce the direct and indirect contacts in transactions, allowing social distancing to be maintained and facilitating stabilization of the social economy. This paper aims to comprehensively investigate the technological and mental factors affecting users’ adoption intentions of M-payment under the COVID-19 pandemic, to expand the domain of technology adoption under the emergency situation. This study integrated Unified Theory of Acceptance and Use of Technology (UTAUT) with perceived benefits from Mental Accounting Theory (MAT), and two additional variables (perceived security and trust) to investigate 739 smartphone users’ adoption intentions of M-payment during the COVID-19 pandemic in China. The empirical results showed that users’ technological and mental perceptions conjointly influence their adoption intentions of M-payment during the COVID-19 pandemic, wherein perceived benefits are significantly determined by social influence and trust, corresponding with the situation of pandemic. This study initially integrated UTAUT with MAT to develop the theoretical framework for investigating users’ adoption intentions. Meanwhile, this study originally investigated the antecedents of M-payment adoption under the pandemic situation and indicated that users’ perceptions will be positively influenced when technology’s specific characteristics can benefit a particular situation.With the increasingly widespread popularity of mobile devices, our daily lives have significantly changed, especially in terms of financial transactions. Mobile payment (M-payment) has been dramatically adopted in various industries in recent years. According to a WorldPay report, M-payments accounted for 22% of the global points of sale spending in 2019, and this percentage will increase to 29.6% in 2023 [1]. Moreover, China’s overwhelming adoption of M-payments (Alipay and Wechat Pay) at the point of sale by using Quick Response (QR) codes drove nearly half (48%) of the point-of-sale payments in 2019 [1]. Various previous studies have facilitated the understanding of adoption intentions of M-payment in different contexts [2,3,4]. However, there are still deficiencies of determinant variation and theoretical evidence of different perspectives in emergency conditions [5].The 2019 novel coronavirus (COVID-2019) broke out in December of 2019 and has dramatically expanded globally. As of 7 December 2020, there were 66,243,918 confirmed cases of COVID-19 and 1,528,984 deaths worldwide, reported by the World Health Organization [6]. Due to the high risk of COVID-19 transmission, reducing contact among people and maintaining social distancing was highly recommended by the WHO (2020b) and Tang et al. (2020) [7,8]. In this sense, the contactless characteristic of M-payments can potentially contribute to users’ mental and physical expectations to support their transaction processes and protect their safety. Accordingly, adoption of M-payment in China has significantly increased during the COVID-19 pandemic. According to a report from China banking and insurance news (2020), during the COVID-19 pandemic, the number of transactions made by M-payment was 22.4 million in the first quarter of 2020 in China, up 187% from the previous year (2019) [9]. Meanwhile, based on the CNNIC (2020) report comparing the smartphone users who used M-payment from 2019 to 2020, this percentage increased from 73.5% in June 2019 to 85.3% in March 2020 and reached 86.0% in June 2020 in China, which indicates that M-payment contributes to maintaining individual and organizational transactions during emergency situations [10]. Furthermore, users’ payment habits and business models have changed from traditional face-to-face transactions to contactless M-payment transactions during the pandemic, which in turn efficiently supports the survival of various business and maintains the development of the social economy under an emergency situation. Therefore, what factors influence users’ intentions to adopt M-payment during the pandemic? It becomes dramatically valuable to understand customers’ behaviors under the pandemic for relevant researchers and stakeholders to comprehensively investigate information technology adoption under an emergency situation to develop business strategies correspondingly.Traditional adoption models (e.g., Technology Acceptance Model (TAM) and the Unified Theories of Acceptance and Use of Technology (UTAUT)) evaluate users’ intentions determined by technological perceptions with an obvious limitation of influence from users’ mental perceptions [11,12]. Notably, based on the recommendations of governments and the WHO (2020b) regarding restrictions of direct and indirect contacts among people under the pandemic situation [7], the contactless feature of M-payment potentially influenced users’ attitudes regarding the benefits of using M-payment for daily transaction, which indicates that environmental conditions affect users’ mental process with regard to adopting M-payment [13]. Thus, this paper involved mental accounting theory (MAT) to explain customers’ psychological cognitions of the benefits of using M-payment under a pandemic situation. In order to fill the gap of limited integration of technological and mental perceptions on technology adoption, this study incorporates MAT with UTAUT to comprehensively investigate the antecedents of M-payment adoption on users’ perspectives. Specifically, perceived benefits are considered as an important factor in terms of users’ expectations and will help determine their decisions [14]. Meanwhile, due to the influence of the pandemic, perceived security and trust are also considered as additional antecedents of users’ adoption intentions of M-payment [15]. Perceived security is the most significant determinant of trust, positively affecting users’ intentions of using M-payment [16]. Therefore, this study proposes a new adoption model, including perceived benefits, performance expectancy, effort expectancy, social influence, perceived security, trust and behavioral intention, to investigate users adopting M-payment during the COVID-19 pandemic in the following sections: Section 2: theoretical backgrounds of the utilization of M-payments during the COVID-19 pandemic, MAT and UTAUT; Section 3: development of hypotheses and research model; Section 4: research methodology and data demonstration; Section 5: data analysis; Section 6: discussion; Section 7: theoretical and practical implications; Section 8: limitations and future research recommendations; Section 9: conclusions.M-payment, as an information interaction electronic financial transaction method for paying goods, services and bills by mobile devices [5], consists of three leading contactless technologies, including Short Message Service (SMS), Near Field Communication (NFC) and Quick Response (QR) codes [2]. Due to the convenient, open and secure features of M-payment, a new business climate has been formulated by the wide adoption of M-payment, as financial transactions, are able to take place anywhere, anytime and by anyone, which has established colossal market potential in various contexts, especially under pandemic situations [17]. Many researchers have investigated various factors affecting M-payment adoption by reviewing theoretical frameworks and variables, supporting that relevant knowledge and understanding of users’ adoption intentions of M-payment is determined by technological and mental perceptions, as shown in Table 1. However, few studies have analyzed the adoption intentions determined by mental and technological factors conjointly under an emergency situation.COVID-19, as a global pandemic, has dramatically influenced people’s daily lives and the world economy. According to relevant studies [8] and a report from the WHO (2020b), COVID-19 has significant transmission risk by direct contact with infected people and indirect contact with surfaces in the immediate environment or with objects used on an infected person. In this sense, the contact rate can significantly contribute to the infection risk of COVID; thus, the contactless feature, as a typical characteristic of M-payment, provides mental and physical support to protect and maintain users’ experience in transactions [20]. Moreover, due to the restrictions imposed by the Chinese Government to avoid direct contact and maintain social distancing during the COVID-19 pandemic, M-payment had been widely adopted for its contactless feature and trustworthy performance. Users’ positive cognitions and feelings of safety when using M-payment as the main payment method have been formulated, which reduces the virus transmission risk, protects personal safety and supports the social economy [9].Mental accounting theory (MAT), proposed by Thaler (1985), is defined as the set of individuals’ cognitive operations to categorize, organize and evaluate the consequences of their decision-making in financial activities [21]. Specifically, MAT explains that personal desires influence the cognitive processes of individuals, and their psychological processes for valuing a specific technology should be taken into consideration in the environment of voluntary usage [22]. Accordingly, based on the normative principle of fungibility at the point of purchase, mental accounting is engaged, and decision-making is based on the evaluation of perceived benefits of the purchase activity [23]. Concretely, in the technology adoption aspect, a consumer’s decision of adoption is based on the perceived benefits of utilization of technology [13]. Moreover, MAT can also be incorporated into an adoption model to complementarily explain customers’ intentions of technology adoption [24]. Cheng and Huang (2013) incorporated MAT into TAM to investigate the mental factors affecting customers’ intentions of adopting high-speed railway mobile ticketing services [25]. Park et al. (2018) proposed that the multidimensional perceived benefits of M-payment services are influenced by social influence and technology anxiety, which indicates that users’ willingness of using M-payment is significant determined by the external environment and internal technological perception [14]. Furthermore, MAT provides a theoretical basis to explain consumers’ decisions under conditions of risk and uncertainty [13]. Combined with the disaster of COVID-19, customers’ psychological processes of adopting M-payment are significantly influenced by the contactless feature of M-payment, which is appropriately adapted to the environmental situation, public restriction and users’ requirements. Therefore, MAT is appropriate to apply for explaining users’ mental cognitions of using M-payment under the COVID-19 pandemic.UTAUT was developed by Venkatesh et al. (2003). It consists of performance expectancy, effort expectancy, social influence and facilitating conditions as determinants of behavioral intentions to use a new technology system [26]. UTAUT has been applied in various contexts of technology adoption. It has been revised with additional variables to explain users’ behavioral intentions [4]. For example, Khalilzadeh et al. (2017) integrated security-related factors with the UTAUT model and validated that security and trust have a strong effect on customers’ adoption intentions of NFC M-payments in the restaurant industry [15]. Marinković et al. (2020) modified the UTAUT model with extra variables (perceived trust and satisfaction) to evaluate customers’ usage intentions of M-commerce [27]. Moreover, UTAUT has also been integrated with other models to evaluate users’ behavioral intentions [2,28]. Di Pietro et al. (2015) integrated TAM, DOI and UTAUT to verify M-payment adoption intentions [2]. Oliveira et al. (2014) integrated UTAUT with the initial trust model and task–technology fit model to investigate users’ behavioral intentions of adopting mobile banking in Portugal [28]. However, UTAUT focuses on technological expectations rather than mental expectations, which weakly explains users’ expectations determining their intention of technology [12]. Thus, it is necessary to integrate UTAUT with MAT to explain users’ technological and mental perceptions complementarily on usage intention of M-payment during the COVID-19 pandemic. The development of hypotheses and research models is illustrated in the following section.According to MAT, when consumers perform a particular behavior, they tend to evaluate a possible beneficial outcome [21]. Perceived benefits represent users’ perceptions of the functional benefits of M-payment services, which determine their decisions of adoption [14]. Perceived benefits support a better understanding of users’ mental perceptions of adoption intentions in various technologies, such as online shopping [29], and mobile banking [30]. Meanwhile, perceived benefits have been identified as multidimensional benefits, including utilitarian, hedonic and social values, which are determined by social influence and technology uncertainty [14,24]. However, few studies focus on the perceived benefits of technology characteristics corresponding to a particular condition. Specifically, in a pandemic situation, social distancing is an efficient way to decrease COVID-19 transmission risk among people [7,31]. Compared with traditional payments, the contactless characteristic of M-payments supports users in maintaining social distancing to avoid direct and indirect contacts from cash or point of sale terminals during a transaction process. This aspect allows users to formulate their opinions on the perceived mental and physical benefits of personal safety and provides convenience and utility when using M-payment technology as a financial transaction method in the COVID-2019 pandemic. Thus, perceived benefits are considered as a mental factor to influence the users’ adoption intentions of M-payment during the COVID-19 pandemic, expressed as the following hypothesis.
+              Perceived benefits have a positive effect on the behavioral intentions to adopt M-payments during the COVID-19 pandemic.
+            Performance expectancy is defined as an individual’s perception in terms of the use of an information system facilitating the completion of a task and work performance [26]. Performance has been conceptualized by using attributes related to the system’s efficiency, speed and accuracy in task completion [11]. Especially during the COVID-19 pandemic, users show more concern toward payment efficiency and accuracy. Concretely, in the M-payment adoption aspect, performance expectancy has significantly positive effects on users’ adoption intentions in various contexts [2,3,32,33]. Therefore, when users perceive M-payment as a useful way to accomplish their transactions during the pandemic, they will choose M-payment instead of traditional payment. Accordingly, this paper proposes the following hypothesis.
+              Performance expectancy has a positive effect on the behavioral intention to adopt M-payments during the COVID-19 pandemic.
+            According to UTAUT, effort expectancy is referred to as “the degree of ease associated with the use of the system” [26]. Effort expectancy influences users’ attitudes toward adopting M-payment [17], revealing an even higher influence than performance expectancy [34]. Specifically, Liébana-Cabanillas et al. (2018) found that effort expectancy is the most significant factor affecting users’ intentions of using NFC M-payment systems in public transportation [3]. Moreover, effort expectancy has also been verified to have a positive impact on performance expectancy in various technology adoption contexts [2,17,35]. Therefore, the following hypotheses are proposed.
+              Effort expectancy has a positive effect on the behavioral intention to adopt M-payments during the COVID-19 pandemic.
+            
+              Effort expectancy has a positive effect on the performance expectancy to adopt M-payments during the COVID-19 pandemic.
+            In terms of UTAUT, the definition of social influence is “the degree to which an individual perceives that significant others believe he or she should use the new system” [26]. Slade et al. (2015) explained that it is an underlying assumption that users prefer to consult their social network to reduce any anxiety arising from uncertainty [36]. Especially during the COVID-19 pandemic, recommendations and suggestions from important, relevant people are more important for individuals’ decisions and actions. From previous studies, social influence has been widely tested in the different contexts of its impact on usage intention of mobile technologies [15,24,33,36]. Morosan and DeFranco (2016) presented that social influence has a significant effect on the intention of using M-payment [37]; Kerviler et al. (2016) illustrated that social influence plays a considerable role in explaining users’ intentions of using M-payment [24]. Moreover, social influence, as a determinant for formulating users’ attitudes, significantly affects the perceived multibenefits of users with regard to using M-payment services [14]. Thus, relevant hypotheses are proposed as follows.
+              Social influence has a positive effect on the behavioral intention to adopt M-payments during the COVID-19 pandemic.
+            
+              Social influence has a positive effect on the perceived benefits to adopt M-payments during the COVID-19 pandemic.
+            Trust is defined as users’ willingness to expect a positive outcome of technology’s future performance and a subjective belief that the service provider will fulfil their obligations [38]. Meanwhile, the COVID-19 pandemic has brought uncertainty and social pressure to individuals’ daily transaction processes. Trust of M-payment platforms can increase the likelihood of users using them to make contactless M-payments rather than traditional payments [27,39]. Zhu et al. (2017) validated that trust has the most significant effect on the behavioral intention to use M-payment [39]. Meanwhile, many studies have also verified the effect of trust significantly determining users’ usage intentions of M-payments [16,18,39]. Zhou (2013) modified a trust-based adoption model and found that trust has significant direct and indirect impacts on the behavioral intention to use M-payment [20]. Moreover, trust has also been validated as an additional variable of UTAUT, which positively influences performance expectancy, consequently affecting user behavioral intentions to use M-payment [15]. Similar results have been found in other studies [35], including trust against perceived risk and uncertainty when adopting new technology [15,16]. Moreover, perceived risk combines uncertainty with the seriousness of the potential outcome [24], which negatively influences the multidimensional perceived benefits [14]. Thus, it can be summarized that trust has a positive impact on perceived benefits, which has also been supported by Khalilzadeh et al. (2017). Therefore, this study proposes the following hypotheses.
+              Trust has a positive effect on the behavioral intention to adopt M-payments during the COVID-19 pandemic.
+            
+              Trust has a positive effect on performance expectancy to adopt M-payments during the COVID-19 pandemic.
+            
+              Trust has a positive effect on perceived benefits to adopt M-payments during the COVID-19 pandemic.
+            Perceived security is defined as “the degree to which a customer believes that using a particular M-payment procedure will be secure” [40]. In terms of conducting a financial transaction, lack of security—perception of security against the risk associated with mobile transactions—is one of the most frequent reasons of users refusing to adopt M-payments [41]. Previous studies have proved that perceived security is an important factor determining whether users will adopt M-payments [2,3,42]. Johnson et al. (2018) found that perceived security has the most significant positive impact on a user’s intention to adopt M-payment [43]. Moreover, perceived security significantly increases users’ trust by protecting users from transactional uncertainties and risks [15,44]. Shao et al. (2018) verified that security is the most significant antecedent of customers’ trust towards affecting usage of M-payment in both male and female groups [16]. Therefore, perception of perceived security of M-payment, considered as an extra variable of UTAUT, is a crucial guarantee for establishing users’ trust in using M-payment under a pandemic. Accordingly, this study proposes the following hypotheses.
+              Perceived security has a positive effect on the behavioral intention to adopt M-payments during the COVID-19 pandemic.
+            
+              Perceived security has a positive effect on trust to adopt M-payments during the COVID-19 pandemic.
+            Based on the above hypotheses, all measurement items were adapted from previous studies [4,8,11,14,15,16,19,24] and have been reasonably modified to correspond to the research purposes to explain the mental and technological factors affecting users’ behavioral intentions with regard to adopting M-payments under the COVID-19 pandemic. Specifically, users’ adoption intentions of M-payment under COVID-19 pandemic is conjointly determined by the variables from the revised UTAUT model (for explaining users’ technological perceptions) and perceived benefits, (as the variable of MAT, representing users’ mental cognitions and psychological acceptance of using M-payment under pandemic conditions). The questionnaire is presented in the Appendix A. Moreover, this study revises the UTAUT model, integrating performance expectancy, effort expectancy and social influence with additional variables, perceived security, trust and perceived benefits from MAT to establish a research model, depicted in Figure 1, with the proposed hypotheses relations.In order to validate the proposed conceptual model and examine the research hypotheses, the online questionnaire survey was designed and applied to data collection. Specifically, the questionnaire consisted of two parts. The first part contained respondents’ demographic data with close-ended questions, consisting of gender, age, education, occupation and M-payment experience. The second part was developed by implementing constructs and items from previous hypotheses, consisting of 27 measurement items as indicators to explain perceived benefits, performance expectancy, effort expectancy, social influence, trust, perceived security and behavioral intention. In order to reduce confusion and save time for the participants [45,46], a five-point Likert scale (from 1 to 5, representing “strongly disagree” to “strongly agree”) was applied for representing the items of each construct.The main survey target of this research was smartphone users who used or intend to use M-payment serviced in China during the COVID-19 pandemic. In order to avoid the impact of culture and language differences, the questionnaire was translated into the Chinese language by a professional translator, and then reversely translated into English, followed by confirmation of the translation equivalence. The questionnaire data were collected from a Chinese social media platform, named Wechat, for a three-week period during the height of the COVID-19 pandemic in China, from 11 March 2020 to 31 March 2020.According to the N: q rule proposed by Jackson (2003), an ideal sample size-to-parameters ratio would be higher than 20:1 [47]; therefore, the sample size of this study should be higher than 140. This study dispatched a total 1000 online questionnaires via Wechat, 864 data were collected on 1 April. After removing the answers with missing values, a total of 739 valid questionnaires were accepted, achieving a final response rate of 73.9%. According to the guideline from Ryans (1974) [48], the Kolmogorov–Smirnov test was applied for verifying the nonresponse bias of the sample by comparing the groups between males and females. The demographic distribution of the sample was 45.74% male and 54.26% female; 53.86% of participants were in the age bracket between 21 and 30; 61.71% of participants held bachelor’s or college degrees (this group is more active on social media and so more likely to respond to the questionnaire) [49]; employees and students were the two main groups of participants, with percentages of 43.03% and 23.68%, respectively; 56.16% of total responses used M-payments at least one time per day and 93.78% at least one time per one week during the COVID-19 pandemic, which is in accordance with a report from Ipsos (2020) expressing that the penetration rate of M-payment among mobile Internet users in China (those who have used M-payment in the last three months) is 96.9% [50]. The reason of this high rate of adoption of M-payment during the pandemic can be summarized as follows. Firstly, based on the restrictions from the Chinese Government [10], due to daily transactions using contact being restricted during the COVID-19 pandemic, people tended to complete the transactions in a contactless way. Secondly, according to the suggestions and recommendations from to government and WHO [7], avoiding contacts among peoples is an efficient way to reduce the transmission risk of COVID-19. Thus, M-payment had been widely adopted by customers and retailers for general transactions. Thirdly, M-payment apps were applied to track users’ health statuses during the pandemic, such as Alipay Health Code being assigned a color code (green, yellow or red) to indicate users’ health statuses. Therefore, M-payments are dramatically adopted by smartphone users in China not only to support daily transactions, but also to confirm their health statuses during the COVID-19 pandemic. Specific sample demographics are listed in Table 2.The covariance-based structural equation modelling (CBSEM) technique was conducted for quantitative data analysis. SPSS 17 and AMOS 22 were applied in this study, through the two-step approach suggested by Anderson and Gerbing (1988), including validating the measurement model and testing structural model. The maximum likelihood estimation was conducted in the model assessment [51].A measurement model aims to assess fitness between indicators and latent variables. Exploratory factor analysis (EFA) was applied to examine the construct reliability, and a standard method factor analysis, confirmatory factor analysis (CFA), was applied to assess the convergent and discriminant validity of the measurement model. All seven hypothesized latent constructs in the CFA model were allowed to covary and were determined by related measurement items as reflective indicators.Construct reliability was tested by Cronbach’s alpha. As presented in Table 3, all Cronbach’s alpha values of latent variables are in the range of 0.807 to 0.897, all exceeding the 0.70 suggested by Nunnally and Bernstein (1994) [52], which means that construct reliability has been demonstrated.Convergent validity was assessed by standardized factor loading of all sample items. Table 3 shows that all items loadings are in the range of 0.807 to 0. All loadings are ideally greater than 0.70 [53], which demonstrates eligible convergent validity of the measurement model. Moreover, completed convergent validity was assessed by Composite Reliability (CR) and the average variance extracted (AVE) criteria. As shown in Table 4, the constructs have CRs in the range 0.811 to 0.898, all above 0.7 [54]. Meanwhile, all constructs have AVEs in a range of 0.589 to 0.688, which all meet the suggestions by Fornell and Larcker (1981) (that AVE should be higher than 0.5) [55], which means the latent variables explain more than half of the variance of the indicators. Therefore, the consistency of measurements among the indicators and latent variables has been proved.Discriminant validity reflects whether two factors are statistically different. It is evaluated by using two criteria. Firstly, according to Fornell and Larcker (1981), the square root of the AVE should be greater than all correlations between each pair of constructs [55]. Table 4 shows that for each factor the square root of AVEs is larger than its correlation coefficients between all latent constructs, which proves that each construct shared more variance with its associated indicators than with any other construct [55]. Secondly, all AVEs are greater than the maximum shared squared variance (MSV) [56]. Thus, the scales satisfy the criterion of discriminant validity suggested.Meanwhile, the following criteria (the ratio of chi-square to degrees of freedom (X2/df) < 3, comparative fit index (CFI) > 0.9, goodness of fit index (GFI) > 0.9, adjusted goodness of fit index (AGFI) > 0.9, normalized fit index (NFI) > 0.9, Tucker–Lewis index (TLI) > 0.9, and root mean square error of approximation (RMSEA) < 0.05) were applied to evaluate the fitness of the model. Table 5 shows that all the model-fit indices of the measurement model (X2/df = 1.832, CFI = 0.979, GFI = 0.948, AGFI = 0.935 NFI = 0.959, TLI = 0.979, RMSEA = 0.034), respectively, exceeded the common acceptance levels, which demonstrates a qualified fitness of the measurement model.Further, this study examined the potential common method bias by Harman’s one-factor test (Podsakoff et al. (2003)) in SPSS [57]. The results show that the largest variance explained by an individual factor is 40.99% (<50%). The result confirms that none of the factors can individually explain the majority of the variance. Moreover, a CFA was applied to assess the fitness of a single-factor model (all items as the indicators of one factor) [58]. The results of the model fit present a poor fitness, which include v2/df = 15.999(>3), CFI = 0.603 (<0.9), GFI = 0.566 (<0.9), AGFI = 0.485 (<0.9), NFI = 0.588 (<0.9), TLI = 0.569 (<0.9) and RMSEA = 0.143 (>0.08). Therefore, both tests confirm that no common method bias appeared in this study.The assessment results of the measurement model validate the construct reliability and convergent and discriminant validity of constructs satisfactorily. The constructs can be used to test the structural model.The maximum likelihood estimation method and bootstrapping technique (500 samples and 95% significance level) were applied for assessing the structural model. Firstly, the model fit of the structural model was evaluated similarly to the measurement model. The results were presented in Table 5, which confirms a qualified goodness of fit of structural model.Secondly, the variance (R2) of endogenous variables was assessed to evaluate the explanatory power of the structural model. As shown in Table 6 and Figure 2, the explained variances of performance expectancy (R2 = 0.48), perceived benefits (R2 = 0.28), trust (R2 = 0.44) and behavioral intention (R2 = 0.71) all exceed the recommended minimum value, 0.1 [59]. Thus, the structural model substantially explains the dependent variable.Moreover, the results of hypotheses testing show that behavioral intention of adopting M-payments during the COVID-19 pandemic is the most significantly determined by performance expectancy (ß = 0.426; p < 0.001), followed by, perceived benefits (ß = 0.283; p < 0.001), social influence (ß = 0.277; p < 0.001), trust (ß = 0.234; p < 0.001) and perceived security (ß = 0.221; p < 0.001). Thus, Hypotheses H1, H2, H5, H7 and H10 are validated, respectively. However, the results show that effort expectancy (ß = −0.209; p < 0.001) has a negative effect on behavioral intention; therefore, H3 was rejected. Moreover, both effort expectancy (ß = 0.470; p < 0.001) and trust (ß = 0.401; p < 0.001) are statistically significant in terms of explaining the performance expectancy. Thus, Hypotheses H4 and H8 are confirmed. Meanwhile, the results illustrate that the perceived benefits are significantly influenced by social influence (ß = 0.272, p < 0.001) and trust (ß = 0.233, p < 0.001), respectively. Therefore, Hypotheses H6 and H9 are confirmed. In addition, H11 is also accepted by the result of perceived security (ß = 0.586; p < 0.001), significantly determining trust.Based on the data analysis results, ten of the eleven hypotheses were confirmed in this study, which demonstrates that the current study exhibits an appropriate adoption model to explain antecedents of users’ adoption intentions of M-payment under the pandemic.Specifically, performance expectancy had the most significant positive impact on users’ adoption intentions of M-payments during the COVID-19 pandemic (Hypothesis 2), which corresponds to the vast majority of previous studies [37,60]. It can be confirmed that the utility and practicability of M-payment technology can improve users’ payment efficiency under emergency situations. Especially, M-payment provided a fast payment process without any direct or indirect contacts among people, significantly influencing users’ adoption intentions during the pandemic. Users will feel M-payment is a more useful and more reliable method than traditional payments to support their transactions under the pandemic.Meanwhile, performance expectancy is significantly determined by effort expectancy (Hypothesis 4) and trust (Hypothesis 8), which is in accordance with findings from previous studies [2,35]. This study initially validated the effects of effort expectancy and trust on performance expectancy under the pandemic, which explains the absence of a confirmation of the simplicity and trustworthiness influencing the perceived functional utility when using M-payment under an emergency situation. Accordingly, the results support the accessibility and operability of the technology’s interface and function are positively formulate users’ performance expectancy; meanwhile, the reliability and trustworthiness of the technology’s services are essential to shape the high utilization of the technology under an emergency situation.Moreover, the second largest significant effect on users’ behavioral intentions to adopt M-payments during the COVID-19 pandemic is caused by perceived benefit (Hypothesis 1). This result illustrates that perceived benefits correspond with individuals’ mental expectations related to contributions of M-payment under the pandemic. Specifically, perceived benefits, such as M-payment’s efficiency, not only influence users’ perceived technological perceptions, convenience and utility [14,24], but also increase perceived safety benefits by M-payment’s contactless characteristic. Concretely, users’ mental expectations are satisfied by perceiving more reliability and safety of using contactless payment to reduces contacts among people and maintains social distancing to decrease the COVID-19 transmission risk [7,31]. Thus, perceived benefits reflect users’ mental cognitions of technology’s features which can overcome a particular environmental issue, which in turn significantly influences users’ adoption intentions.Meanwhile, under the condition of COVID-19 pandemic, perceived benefits as mental expectations are significantly influenced by social influence (Hypothesis 6) and trust (Hypothesis 9). The effects of social pressure and opinions of important, relevant people play an important role in influencing an individual’s mental expectations, affecting his/her behavioral intention [14]. When users receive recommendations from their close friends or families indicating that M-payment is beneficial for protecting their personal safety by avoiding contact with people during a transaction process to reduce the infection risk of COVID-19, they tend to consider M-payment as a helpful and valuable payment method. Moreover, trust was analyzed and found to have a significant effect on perceived benefits in this study. The reputation and trustworthiness of M-payments are potentially determined by the contactless advantage of M-payment in optimizing users’ experiences and supporting their safety during the COVID-19 pandemic, which emphasizes users’ perceived benefits towards adopting M-payments during the emergency situation.Furthermore, social influence as the third important factor has a statistically significant impact on behavioral intention (Hypothesis 5), which means the opinions, recommendations and support from close relationships of users are essential in the formulation of users’ behavioral intentions to adopt M-payments during the COVID-19 pandemic. This result is supported by previous studies in normal situations [33,36]. Especially under the pandemic, people are relying more on the support and recommendations of important people in their lives—their family and close friends more easily influence their behaviors. Accordingly, the reputation of M-payment and word-of-mouth effect are considered crucial for attracting users’ adoption intentions of M-payment to formulate a new payment habit by the influence of the pandemic.In addition, this study confirms Hypothesis 7 and Hypothesis 10—trust and perceived security have statistically significant effects on explaining users’ behavioral intentions of using M-payments during the COVID-19 pandemic. Specifically, consumers have developed trust in M-payment platforms through their reliable performance and mature legal framework protection, and so they worry less about financial risks to reap more benefits from the service [39]. Thereby, users’ adoption intentions are influenced by technological and privacy security and users’ trust from technological and mental perspectives [16,18,43]. Moreover, Hypothesis 11 also proved that perceived security significantly associates with trust. In this sense, perceptions of users’ perceived security could reduce uncertainty as well as crucially guaranteeing the M-payment performance to improve users’ trust of M-payment platforms [15]. It demonstrates that trust and perceived security have a significant association, and both factors conjointly determine users’ adoption intentions of M-payments under the pandemic. Furthermore, M-payments involve sensitive and personal data; therefore, it is necessary to ensure the reliability and credibility of M-payment platforms for securing transactions and protecting personal information [61]. Moreover, based on the security, trustworthiness and reliability of M-payment platforms, users can accept the records of their transaction times and locations during the pandemic to be utilized by governments and health institutions to track contacts among payment processes, for monitoring, updating and reporting the pandemic transmission status. Accordingly, users can clearly and opportunely be made aware of the virus infection situation among them, which positively influences their intentions to use M-payment during the COVID-19 pandemic to reduce the infection risk.However, Hypothesis 3 was rejected in this study, which means easiness of understanding and handling M-payment systems does not have a direct impact on a user’s behavioral intentions to adopt M-payments during the COVID-19 pandemic. Similar results are supported by previous M-payment studies [3,62]. The main reason for this result is because users have become accustomed to smartphone functions and become more skillful through their previous utilization of various applications on smartphones [60]. Meanwhile, under the COVID-19 pandemic, user behavior is determined more by other perceptions related to personal safety, such as reliability, utility, security, trustworthiness and benefits, which can provide multidimensional supports for protecting transaction processes during a pandemic. Thus, the easiness of using M-payment is a less critical or surmountable factor determining users’ adoption intentions during the pandemic.This study contributes three main theoretical implications. First, this study was empirical and examined the factors affecting users’ adoption intentions of M-payments under the pandemic situation, which is absent from evaluations of previous studies. Consequently, the study dramatically enriched the literature of technology adoption during a pandemic. Specifically, this study illustrates a worthwhile direction to understand users’ adoption intentions by not only examining users’ perceptions from technological perspectives, but also assessing users’ mental expectations. Moreover, users’ technological and mental perceptions of technology are significantly influenced by the emergency situations. Therefore, this study provides a future sight for relevant research to analyze new technology adoption from technological and mental perspectives conjointly and corresponding with the specific situation, especially for emergency situations.Second, this study integrated the UTAUT model with perceived benefits from MAT and two extra variables, perceived security and trust, which significantly contributes to the theoretical development and framework coordination of the emerging literature on information technology adoption. Simultaneously, this study demonstrates a substantial contribution to the theoretical expansion of UTAUT and MAT by initially proposing and verifying new causal paths (PB → BI, SI → PB, TR → PB, TR → PE, TR → BI and PS → BI) and rejecting the path EE → BI for investigating the interactions of variables in the new comprehensive model. Therefore, the integrative research approach presented in this study can serve as a beneficial and valuable reference to modify and evaluate new adoption models for investigating novel technology adoption.Third, this study initially focused on technology characteristics corresponding to the pandemic situation as a potential antecedent determining users’ mental and technological perceptions. Specifically, the contactless feature of M-payment avoids contacts during transaction processes and maintains social distancing, which improves the perceived multidimensional benefits of the users and optimizes their experience of using M-payments under the pandemic situation. Meanwhile, based on the disaster status of the COVID-19 pandemic, the effort expectancy became less important than other variables for determining whether users would adopt M-payment. Thus, it is important to consider whether a particular technology’s features can influence users’ interpretations of the perceived mental and technological benefits corresponding to particular situations or conditions to comprehensively explain technology adoption in an emergency situation.Moreover, four main practical implications are demonstrated in this study. First, the current research enhances the existing knowledge of adoption intention of M-payments in an emergency situation and enriches the understanding of how a pandemic changes users’ payment habits. It suggests that a pandemic might bring suffering to people or society. Furthermore, it can also facilitate the development of new technology that can bring benefits to individuals, organizations and society to survive in the emergency situation, which is valuable for relevant stakeholders to consider the pandemic to establish appropriate business strategies.Second, this study could be valuable to start-up companies, policymakers, government bodies and private service providers who are interested in M-payment services. M-payment has become increasingly popular and provides useful services for efficient transaction processes, particularly in emergency conditions. In the context of a pandemic, M-payment can increase personal safety perception and maintain the stable development of business. Based on the finding from this study, as well as providing an easy-to-use operating application, relevant stakeholders should initially recognize the importance of M-payment in formulating users’ perceived benefits and design system attributes accordingly under the pandemic situation. Meanwhile, M-payment service providers should guarantee the compatibility, efficiency and security of transactions to meet customers’ requirements and match their lifestyles. In addition, enhancing the public impression of M-payment and stimulating a positive word-of-mouth social effect would improve the technology providers’ reputation in different situations.Third, this study supports new technology providers with a comprehensive understanding of customers’ adoption intentions, determined by technological and psychological perceptions conjointly. Consequently, relevant stakeholders should focus on taking advantage of the features of technology (such as the contactless characteristic of M-payment) corresponding to its benefit to a particular situation (such as avoiding direct or indirect contacts to decrease COVID-19 transmission risk) in terms of maintaining service quality, reliability and efficiency to meet consumers’ physical and mental concerns and optimize their experience, thereby increasing acceptance among the target population.Finally, the findings and results of this study could be applied as references for other online-to-offline (O2O) service industries in a pandemic situation. Relevant businesses could utilize the results to develop appropriate strategies that combine the benefits of technology characteristics with users’ technological perceptions and mental expectations to expand markets to adapt to different emergency situations and build better customer bases.There are several limitations inherent in this study which need to be acknowledged. Firstly, the data collection was restricted to China during a particular period of the COVID-19 pandemic; the results may not be generalized to different countries and various situations. Future studies should replicate this model and collect data from different nationalities and consider specific benefits corresponding to particular situations. Furthermore, the research model can be examined through cross-cultural studies for better understanding of the variations in different cultural backgrounds.Secondly, there were limited variables and interactions of the variables analyzed in this study—e.g., the variables selected in this study were mainly from a technology adoption aspect. Future research can put more effort into integrating the relations between variables, such as social influence affecting perceived security [15] and use technological indicators with the variables from a health and risk aspect. Meanwhile, in order to gain a deeper understanding of the mental and technological factors affecting adoption intentions with regard to novel technology, future research can incorporate research models with other variables, such as a cultural moderator, satisfaction, etc., which are also recommended in previous studies [42,63,64].Thirdly, as the data collection period was limited, and that data were homogeneously distributed and collected through Wechat (a mobile social media application in China), in this study, the data collection process is recommended to chronically and integrally cover the users from different areas (urban and rural areas) over a different period of using M-payment in various patterns (online and offline surveys).Finally, there was no distinction between the types of M-payment patterns (such as SMS, NFC and QR), platforms of M-payment (such as Apple pay, Samsung pay, Wechat pay, and Alipay) and patterns of electronic transaction (such as, electronic transaction via computer, electronic transaction via mobile device). Therefore, a future study can focus on distinguishing the different payment methods or payment platforms of M-payment techniques in accordance with specific research objectives.In conclusion, we proposed a theoretical adoption model integrating UTAUT with perceived benefits from MAT and two additional variables, trust and perceived security, to appropriately explain the mental and technological factors affecting users’ behavioral intentions of adopting M-payment during the COVID-19 pandemic in China. This research model provided extensive explanatory power when explaining that users’ payment habits had changed due to the influence of pandemic, and that adoption intentions of M-payment were determined by technology perceptions and mental expectations conjointly. Performance expectancy, perceived benefits, social influence, trust and perceived security are significant in facilitating users’ adoption intentions of M-payments during the COVID-19 pandemic. Specifically, the contactless characteristic of the M-payment technique is beneficial in maintaining social distancing and protecting personal safety under a pandemic. This study also explored new causal relationships and found that perceived benefits are significantly determined by social influence and trust. Moreover, performance expectancy is influenced by effort expectancy and trust, towards explaining users’ behavioral intentions of using M-payments during the COVID-19 pandemic.Furthermore, this study provides several significant theoretical and practical contributions on on investigating novel technology adoption in a particular situation, which contributes to the knowledge and understanding of the extension of the UTAUT application, explaining that users’ payment habits have changed because of the pandemic and adoption intention of M-payment is determined by users’ technological perceptions and mental expectations. In addition, this study recommends that researchers and relevant stakeholders focus on a particular characteristic of M-payments that corresponds with the pandemic, which can influence the perceived mental and technological benefits of the user. Understanding users’ behaviors is an efficient way to analyze new technology adoption and develop an appropriate strategy for optimizing users’ experiences.Conceptualization, Y.Z. and F.B.; methodology, Y.Z.; software, Y.Z. and F.B.; validation, Y.Z.; formal analysis, Y.Z.; investigation, Y.Z.; resources, Y.Z..; data curation, Y.Z.; writing—original draft preparation, Y.Z.; writing—review and editing, Y.Z. and F.B.; visualization, Y.Z. and F.B.; supervision, F.B.; project administration, Y.Z. and F.B.; funding acquisition, None. All authors have read and agreed to the published version of the manuscript.This research received no external funding Not applicable.Informed consent was obtained from all subjects involved in the study.The data that support the findings of this study are available from the corresponding author upon reasonable request. I would like to express my gratitude to all those who helped me during the writing this research paper. I would like to express my heartfelt gratitude to my supervisor and the online survey participants. The authors declare no conflict of interest.Questionnaire with constructs, items and references.Research model with proposed hypotheses relations.Hypotheses testing results.Literature reviews related to Mobile payment (M-payment).UTAUT: Unified Theory of Acceptance and Use of Technology; TAM: Mental Accounting Theory.Demographic distribution of the sample.Item loadings and Cronbach’s alpha of structures.Descriptive statistics and correlation among constructs.CR = Composite reliability; AVE = average variance extracted; MSV = maximum shared squared variance.Model-fit indices of the measurement model and structural model.Hypotheses testing results.PB = Perceived benefit; PE = performance expectancy; EE = effort expectancy; SI = social influence; PS = perceived security; TR = trust; BI = behavioral intention; ***: p < 0.001.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01017.txt

@@ -0,0 +1 @@
+According to the World Health Organization, cancer is the second leading cause of death in the world. In Spain, about a quarter of a million cases were diagnosed in 2017, and 81% of the Spanish population has used, at least once, some kind of complementary therapy. Said therapies are increasingly being used by cancer patients. The purpose of the study is to analyse the effectiveness of complementary therapies among cancer patients. A systematic peer review was conducted following the PRISMA-ScR guide in four databases (PubMed, CINAHL, Scopus and WOS). The inclusion criteria were Randomised Clinical Trials, published between 2013 and 2018, with a value of 3 or more on the Jadad Scale. The protocol was registered in PROSPERO (CRD42019127593). The study sample amounted to 1845 patients (64.55% women), the most common being breast cancer patients (794), followed by lung cancer patients (341). Fifteen complementary therapies were identified. We found two studies for each of the following: electroacupuncture, phytotherapy, hypnotherapy, guided imagery and progressive muscle relaxation. From the remaining ones, we identified a study on each therapy. The findings reveal some effective complementary therapies: auriculotherapy and acupuncture, laser moxibustion, hypnosis, Ayurveda, electroacupuncture, progressive muscle relaxation and guided imagery, yoga, phytotherapy, music therapy and traditional Chinese medicine. On the other hand, electroacupuncture, laser moxibustion and traditional Chinese medicine presented adverse effects, and kinesiology did not show effectiveness.Cancer is the second leading cause of death in the world, especially lung cancer (1.69 million deaths) [1]. Nearly a quarter of a million new cases of cancer were diagnosed in Spain, amounting to 149,000 in men. The most common cancers in men were prostate (22.4%), colorectal (16.6%), lung (15.1%) and urinary bladder (11.7%) cancer; and in women, breast (28.0%), colorectal (16.9%), uterus (6.2%) and lung (6.0%) cancer [2] were the most common.Traditional or complementary medicine is not usually included in the public health systems of developed countries, although its use is increasing. In this regard, the population who has used complementary therapies (CTs) at least once amounts to 70% in Canada, 42% in the United States, 81% in Spain, 49% in France and 31% in Belgium. Some of the countries that include them in their national health system are Canada, the United Kingdom, Germany and Switzerland. The prevalence of their use is 80% in African countries and 40% in China, where there is an integration between traditional Chinese medicine and acupuncture within the public health system [3,4,5,6,7].The use of CTs is increasing in the West because of the growing accessibility to health information, as well as an increase in concern about side effects caused by treatments or drugs; however, in Africa and China, this increase is due to the accessibility and affordability of CTs, as well as the strong relationship that traditional medicine has with their belief system and the increasing number of health professionals who practice it [3,4,5,6,7,8,9,10,11].CTs are a set of knowledge, practices and skills based on indigenous experiences, beliefs and theories from different cultures—whether or not they can be explained—used both for health maintenance and for the prevention, diagnosis, improvement or treatment of physical or mental illnesses [12]. In a study by Borm et al. [10], CTs are defined as “approaches and practices that are typically not part of conventional medical care”. They are considered complementary therapy when these treatments are used in combination with conventional medicine. However, if they are used as a substitute for conventional medicine, they are called alternative therapy. CTs’ classification is wide due to their high diversity and different combinations among them. The United States National Center for Complementary and Integrative Health (NCCIH) classifies them into three large groups. First, we find the group of natural products, which are easily found in pharmacies and herb shops and are usually sold as diet supplements. This group includes products composed of herbs, minerals, vitamins and probiotics. Around 17.7% of adults in the United States used at least one natural product in 2012. Next, we find the mind and body practices group. These practices include yoga, acupuncture, relaxation techniques, therapeutic touch, movement therapies, tai chi, qi gong and hypnotherapy. Among them, yoga, chiropractic treatment, osteopathy, massage therapy and meditation are most frequently used by adults. Lastly, the final group incorporates other complementary methods like traditional Chinese medicine, Ayurveda, naturopathy and homeopathy [10,13,14,15].In cancer patients, CTs are different from standard medical oncology treatments and include special diets, vitamins, herbs or methods such as acupuncture or massage, among others [16,17]; CTs are mainly oriented to the treatment or prevention of various symptoms caused by cancer treatments or by the disease itself [10]. In addition, the use of CTs causes many patients to feel that they are taking an active role in their health and in the treatment of their illness [18,19,20], as we can also read in the study by Alimujiang et al. [21].The research question is based on the above. How effective are CTs used to treat cancer patients? The participants were the cancer patients who undergo CTs; the interventions were the various CTs. Comparisons were made between the effectiveness of the various CTs used, and the results were related to the benefits and risks of the various CTs used in cancer patients and the design of the selected randomised clinical trials (RCTs) studied.The general objective of this study was to update the knowledge regarding the effectiveness of complementary therapies used in cancer patients. To this end, the characteristics of the studies were analysed, the different complementary therapies were described and their effectiveness in cancer patients was compared.A qualitative systematic review was carried out, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) guide [22] and different authors [23,24,25]. The review protocol was recorded in the international prospective register of systematic reviews PROSPERO under the number CRD42019127593 (National Institute Health Research, 2019) [26]. This review was carried out using a peer-review process (M.D.G.-M. and M.S.T.-B.), and when there was any discrepancy, a consensus was reached [27].The following electronic databases were consulted: PubMed, CINAHL, Scopus and Web of Science. Medical Subject Headings descriptors were used. In order for the search strategy to meet the requirements, if reproducible and delimited [25], the same strategy was performed on the different databases: treatment outcome AND neoplasms AND complementary therapies.The inclusion criteria were as follows: RCTs published in scientific journals addressing the effectiveness of CTs in cancer patients, published in the last 5 years; in English, Spanish or Portuguese; with RCTs scoring 3 or higher on the Jadad Scale [28]. There were four stages followed in the selection process of the studies. First, the studies were identified following the established search strategy. Second, an analytical reading of the abstracts, titles and keywords was carried out, and unrelated studies and duplicates were discarded. Third, a thorough and critical reading of the full-text studies was carried out, and those that did not meet the inclusion criteria were discarded. Fourth, studies that did not meet the requirements for methodological assessment according to the Jadad Scale were discarded.The following data were extracted from the characteristics of each selected study: objectives, study design/methodology, sample/study period/country, complementary therapy/interventions and findings. In addition, the limitations or biases exposed in the various selected studies were collected.On the one hand, we analysed the sociodemographic variables in terms of the characteristics of the selected CT studies performed on cancer patients, and on the other hand, we analysed the different CTs and their effectiveness. For the analysis of the data, a description of the different variables (frequencies and percentages) that appeared in the different clinical trials of the selected studies was made.There were 182 studies identified through the search strategy and, after eliminating duplicates and reading the abstracts and full text, there were 19 studies left. Four of these were excluded after performing the methodological quality assessment using the Jadad Scale, as they obtained a score of less than 3. Fifteen studies were finally selected (Figure 1), of which two studies scored 5 points [29,30], seven scored 4 points [31,32,33,34,35,36,37], and six scored 3 points [38,39,40,41,42,43].The study sample included 1845 patients (64.55% women), of whom 794 had breast cancer [31,32,33,34,36,39,40,43], 50 had head and neck cancer [35], 220 had pancreatic cancer [38], 104 had prostate cancer and 104 had breast cancer [36], 341 had lung cancer [41], 142 had liver cancer [37] and 194 had an unspecified type of cancer [29,30,42]. Supplementary Material Table S1 presents the objectives, study design/methodology and sample/data collection/country.Regarding the types of CT, 15 therapies were identified. We found a study of each of the following therapies: kinesiology [31], Ayurveda [32], music therapy [39], yoga [40], laser moxibustion [29] and traditional Chinese medicine [30]. We found one study addressing two therapies (acupuncture and auriculotherapy) [41]. We also found two studies of each of the following: electroacupuncture [32,37], phytotherapy [33,38] and hypnotherapy [34,42]. In addition, we found two studies addressing progressive muscle relaxation (PMR) and guided imagery [36,43], and one of them also covered hypnotherapy, acupressure and meditation [43].As for the effectiveness of the various CTs, two studies fail to demonstrate any effectiveness for these therapies: kinesiology [31] and phytotherapy [33]. On the other hand, we found some effective therapies for all the factors studied: hypnosis [34,42], Ayurveda [32], laser moxibustion [29] and auriculotherapy together with acupuncture [41]. We also identified some therapies that are effective on almost all factors: electroacupuncture [42], yoga [40], PMR and guided imagery [36,43], phytotherapy [38] and traditional Chinese medicine [30]. However, there are two therapies that have been shown to be effective on only one factor in the study: electroacupuncture [37] and music therapy [39]. Moreover, some authors identify some safe CTs with no associated adverse events: kinesiology [31], auriculotherapy together with acupuncture [41], PMR and guided imagery [43], hypnosis [34,42], phytotherapy [38], Ayurveda [35] and music therapy [39]. Supplementary Material Table S2 presents CTs, interventions and findings.Table 1 presents the limitations and/or biases of the selected studies.In terms of the effectiveness of CTs in cancer patients, there are therapies that show no effect compared to the control group or the therapy to be compared, such as kinesiology [31] or phytotherapy [33]. In contrast, there are therapies, such as Ayurveda [35], hypnosis [34,42], laser moxibustion [29] and auriculotherapy together with acupuncture [41], which present significant results, in terms of effectiveness throughout the follow-up of the RCT, as a solution to resolve symptoms derived from some treatments or from the disease itself. These would be, for acupuncture and auriculotherapy, a decrease in the occurrence of postsurgical constipation [41]; for laser moxibustion, a decrease in cancer-related fatigue [29]; for hypnosis, an improvement in fatigue [34,42], sleep disturbance and pain [42], coinciding with the results of other authors concerning the pain-decrease effect [44]; and for Ayurveda, a decrease in the occurrence of radiodermatitis and an improvement of its symptoms [35].Regarding electroacupuncture, Xie et al. [37] reported that of the three elements studied (nausea–vomiting, anorexia and the MD Anderson Symptom Inventory (MDASI) scale), only a significant decrease in anorexia was observed. In contrast, Mao et al. [32] obtained better results in their study because they observed improvements in fatigue, anxiety and depression. However, we cannot compare both RCTs as they do not analyse the same elements.Regarding studies on phytotherapy, the study by McCann et al. [33] mentioned above argues that the intervention carried out did not produce effective results. Although in another study it is shown that there is no clear scientific evidence about mistletoe effectiveness [10], Tröger et al. [38] suggest that phytotherapy with mistletoe therapy has benefits in 13 of the 15 items obtained from the European Organisation for Research and Treatment of Cancer (EORTC) scale, highlighting improvements in pain, fatigue, anorexia and depression. This indicates that not all phytotherapy interventions are effective, but that mistletoe therapy applied in the RCT by Tröger et al. [38], for example, has great benefits for the patient. The effectiveness of these therapies is supported by other authors, such as in the study by Codini et al., where the effectiveness of vitamin C is studied and it is found that this treatment provides broad benefits for cancer patients [45].Regarding PMR and guided imagery therapy, Charalambous et al. [36] and Stoerkel et al. [43] agree that its use improves fatigue and pain but contradict each other in terms of anxiety, which is found to decrease in the RCT by Stoerkel et al. [43] and has nonsignificant results in the RCT by Charalambous et al. [36]. This result is due to the fact that PMR and guided imagery were not the only therapies used in the study by Stoerkel et al. [43], as other therapies such as self-hypnosis, acupressure and meditation also played a role, making the set of different therapies more effective, as we can observe in the studies by Araujo et al. [46] and Tacón et al. [47], where it is highlighted that meditation yields positive effects like lower anxiety levels in patients with cancer, in addition to other factors, as the study by Zhang et al. also considers [48].Studies of traditional Chinese medicine [30], music therapy [39] and yoga [40] also fail to show significant effects on all study factors. In the first study [30], there is significant improvement in chest discomfort, fullness and distension and shortness of breath, but not for palpitations and pain. Moreover, other authors consider the use of traditional Chinese medicine in the decrease of side effects caused by cancer treatments such as chemotherapy or radiotherapy [49]. In the second [39], there is a greater reduction in perioperative anxiety in both intervention groups, but not in the other factors to be studied. However, in the yoga RCT [40], they are not present in all of them, but only one of the components studied does not give a significant positive result. On the other hand, in the systematic review by Behzadmehr et al. [44], yoga and music therapy are effective in the reduction of pain, a factor that was not highlighted in the studies about music therapy [39] and yoga [40] of this review. Similar results were found in the study by Gosain et al. [50], where they affirm that, in addition to yoga being effective, this therapy also increases tolerance to side effects of chemotherapy.Thus, it is observed that therapies such as electroacupuncture [32,37], phytotherapy [38], PMR and guided imagery [36,43], traditional Chinese medicine [30], music therapy [39] and yoga [40] are effective in some of the items studied, and they are mostly beneficial.Regarding the risks or adverse effects of CTs in cancer patients, the RCTs on electroacupuncture [37], laser moxibustion [29] and traditional Chinese medicine [30] described side effects, although they are listed as mild and rapidly resolving symptoms. Depending on the therapy, these are mild erythema, with 3 cases in 39 patients using laser moxibustion [29]; mild tingling and redness, with 2 cases in 79 patients using electroacupuncture [37]; and skin irritation and allergic reaction, with 4 cases in 36 patients using traditional Chinese medicine [30]. However, in relation to traditional Chinese medicine, the study of Wang et al. affirms that there is little evidence to validate the therapy’s safety [51].On the other hand, we identified nine therapies in different studies that are considered to be safe, i.e., that do not pose any adverse effects. These are kinesiology [25]; auriculotherapy together with acupuncture [35]; hypnosis [28,36]; phytotherapy [32]; Ayurveda [29]; music therapy [33]; and PMR and guided imagery together with self-hypnosis, acupressure and meditation [37], coinciding with the results of other authors such as Araujo et al. [46] and Gosain et al. [50], who show that meditation is a safe intervention. Despite this fact, as it is observed in the studies by Thomford et al. [52] and Thomford et al. [53], in relation to phytotherapy, CTs can cause interactions with conventional medicine, some with fatal clinical outcomes. However, we found that some studies did not mention the risks or adverse effects in the use of certain therapies, such as electroacupuncture [32], PMR and guided imagery [36], yoga [40] and phytotherapy [33].As for limitations, the first is related to the impossibility of retrieving all existing information on the topic of study, since searches have been restricted to the last 5 years and by language. We tried to alleviate this language limitation by including English, Spanish and Portuguese, since most studies are published in English. The second limitation is the heterogeneity of the studies in terms of samples, study periods, interventions and results, although some authors believe that if the data are treated rigorously and methodically, as has been done in this review, they are a beneficial contributing source to the research [54].The findings reveal some effective CTs: auriculotherapy and acupuncture, laser moxibustion, hypnosis, Ayurveda, electroacupuncture, progressive muscle relaxation and guided imagery, yoga, phytotherapy, music therapy and traditional Chinese medicine. However, other CTs—such as electroacupuncture, laser moxibustion and traditional Chinese medicine—presented adverse effects, and kinesiology did not show effectiveness. These results may be useful when making decisions in clinical management, although it would be advisable to continue research on this subject, carrying out more RCTs that can provide greater certainty about the influence of sample size, follow-up periods and/or other variables on the results. All of this should be performed in order to evaluate the scientific evidence and to corroborate or contrast the results obtained in this review, so as to obtain greater insight into the effectiveness of CTs for cancer patients.The following are available online at https://www.mdpi.com/1660-4601/18/3/1017/s1: Table S1: Characteristics of the selected studies. Table S2: Complementary therapy/interventions and findings.Conceptualization and methodology, M.D.G.-M. and M.S.T.-B.; writing—original draft preparation, M.D.G.-M.; writing—review and editing, M.D.G.-M. and M.C.-C.; visualization, M.C.-C. and M.S.T.-B.; supervision, M.D.G.-M. All authors have read and agreed to the published version of the manuscript.This research received no external funding. The review protocol was recorded in the international prospective register of systematic reviews PROSPERO under the number CRD42019127593. Available online: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=127593.The authors declare no conflict of interest.Study selection flowchart.Limitations and/or biases of the selected studies.RCT = Randomised clinical trial.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01018.txt

@@ -0,0 +1 @@
+Medication-induced jaw osteonecrosis (MRONJ) is a rare and serious disease with a negative impact on patients’ quality of life, whose exact cause remains unclear and which may have a multifactorial origin. Although there are different therapeutic protocols, there is still no consensus. This case series evaluated three patients diagnosed with staged 2 MRONJ treated at the University of Murcia dental clinic according to the protocols described by the Spanish Society of Oral and Maxillofacial Surgery and the American Association of Oral and Maxillofacial Surgeons. Within 12 months of the application of therapeutic protocols, the lesions were completely healed in all cases. Radiography showed slow but progressive healing with normal bone structure. Conservative treatment with antibiotics, chlorhexidine rinses and minimally invasive surgical intervention with necrotic bone resection is effective in treating stage 2 of MRONJ. In cases of refractory osteonecrosis, the application of platelet and leukocyte-rich fibrin (PRF-L) in the surgical approach improves the outcome in soft tissue healing and bone regeneration but further research is needed to confirm its effectiveness.Medication-related osteonecrosis of the jaw (MRONJ) is an uncommon but serious debilitating disease, whose cause remains unclear, although it may have a multifactorial origin [1]. MRONJ is characterized by exposed bone that does not heal in patients with a history or continued use of anti-resorptive or antiangiogenic agents and without a history of exposure to radiation in the head and neck [2,3].The first cases of MRONJ were described by Marx during the 2000s in patients receiving bisphosphonates (BPS) [4], and was named bisphosphonate-related osteonecrosis of the jaw (BRONJ) [5]. However, from 2010 onwards, an increase in the prevalence of osteonecrosis was observed in patients treated with anti-resorptive and antiangiogenic drugs other than bisphosphonates, such as denosumab, and this was named denosumab-related osteonecrosis of the jaw (DRONJ) [6,7]. In 2014, the American Association of Oral and Maxillofacial Surgeons (AAOMS) recommended replacing the name BRONJ with MRONJ [8].Anti-resorptive drugs such as BPS modulate bone metabolism by inhibiting bone reabsorption, and limiting the activity of osteoclasts [9], although they also have an antiangiogenic effect [10]. They are effective in the treatment of diseases such as osteoporosis, bone metastases, multiple myeloma and Paget’s disease, among others [11,12,13]. In addition to BPS, other anti-resorptive drugs are used to treat these diseases, including denosumab, a RANKL inhibitor that is a monoclonal antibody which inhibits osteoclast function and, therefore, bone reabsorption. It is administered subcutaneously every 6 months to decrease the risk of fractures in patients with osteoporosis [14] and monthly in bone metastases of solid tumors [8]. Unlike BPS, RANKL inhibitors do not bond with the bone and their effects on bone remodeling are mostly lessened 6 months after treatment discontinuation [8,15].Antiangiogenic drugs interfere with the formation of new blood vessels by binding to various signaling molecules and interrupting the angiogenesis signaling cascade. These new drugs have proven effective in treating gastrointestinal tumors, renal cell carcinomas, neuroendocrine tumors and other malignancies, but favor the onset of MRONJ [8].According to the AAOMS, a patient is considered to have MRONJ if all the following conditions are met:-Current or prior treatment with anti-resorptive or antiangiogenic agents.-Maxillofacial intra- or extraoral bone fistula persisting for >8 weeks.-No history of maxillary radiation therapy or bone metastases [8].Current or prior treatment with anti-resorptive or antiangiogenic agents.Maxillofacial intra- or extraoral bone fistula persisting for >8 weeks.No history of maxillary radiation therapy or bone metastases [8].The AAOMS also defines the following grades or stages (Table 1):Various treatments are recommended for the different grades or stages of MRONJ (Table 2) [8]:Regardless of the disease stage, mobile bone fragments should be removed without exposing unaffected bone. The extraction of symptomatic teeth within the exposed necrotic bone should be considered because extraction is unlikely to exacerbate the necrotic process [8].The Spanish Society of Oral and Maxillofacial Surgery (SECOM) recommends the following protocol for the treatment of MRONJ (Table 3) [16]:Risk factors for MRONJ include not only medication and dental interventions, but also demographic, systemic and genetic factors [17]. The highest prevalence is in females in the peri- or postmenopausal period or where there is a decrease in estrogen levels, leading to alterations in bone homeostasis, although this may also be a reflection of the underlying disease for which pharmacological agents (osteoporosis, breast cancer) have been prescribed. Patients with cancer, anemia and diabetes have a higher risk of MRONJ [18]. Genetically, studies have reported that single nucleotide polymorphisms are associated with an increased risk of MRONJ. A 2012 gene study showed that patients with a single nucleotide polymorphism in the RBMS3 gene, which is associated with bone density and collagen formation, had a 5.8-fold risk of jaw osteonecrosis [19]. Another study analyzed polymorphisms related to the activity of farnesyl diphosphate synthase, an enzyme specifically inhibited by bisphosphonates, and found a positive correlation between carrier status and jaw osteonecrosis [8,20].Smoking has been recognized in the literature as a risk factor for MRONJ. In the oral cavity, carcinogens in smoking delay wound healing and worsen periodontal disease, as well as promote soft tissue epithelial changes and cancer. In addition to affecting healing, nicotine in smoking may cause increased vasoconstriction in bone, leading to ischemic states that may underlie the pathophysiology of osteonecrosis [21]. However, there is no sufficient scientific evidence to corroborate it. For example, in a case-control study, smoking was assessed as a risk factor for MRONJ in cancer patients, but was not statistically significant [22]. In another more recent case-control study, no direct association of smoking with MRONJ was found in a sample of cancer patients exposed to zolendronate [18]. Vahtsevanos et al. did not observe an association between smoking and MRONJ [23]. More studies with statistically significant results are needed to determine whether tobacco should be considered a risk factor or not.Studies comparing the effects of denosumab and bisphosphonates found a similar incidence of MRONJ [24]. Denosumab (DMAb) is a monoclonal antibody directed against RANKL and exerts its activity by forming a complex with RANKL and thus inhibiting the activation of its descending pathway [25]. Based on this premise, it is possible to consider that the mechanism of action of denosumab mimics the activity of osteoprotegerin (OPG), a soluble anti-RANKL decoy receptor that is physiologically produced by osteoblasts and inhibits osteoclasts. Therefore, an important difference compared to bisphosphonates is represented by the fact that denosumab is specific for the RANKL pathway. Furthermore, due to the different mechanisms of action, it is assumed to have a time-limited effect. On the other hand, bisphosphonates (BP) have a chemical similarity with pyrophosphate sharing the presence of two phosphate groups within the molecule. BPs can be divided into two classes: BP that do not contain nitrogen and that contain nitrogen [26].Non-nitrogen containing BPs are represented by clodronate and etidronate. These compounds were discovered earlier, but they are less active and for this reason their use is substantially restricted to bone metabolic disorders. They exert an antiresorbent effect that is obtained by competing with the recruitment of intracellular ATP in osteoclasts, thus blocking the energy reserve of these cells, inducing their apoptosis [27].Nitrogen-containing BPs such as zoledronate (ZA) and pamidronate were developed later, exhibit stronger activity and are predominantly used in neoplastic bone diseases, but lower doses are also prescribed in benign conditions. The mechanisms of action of these drugs have not been fully determined. However, those that present nitrogen has been shown to be more potent and bind and inhibit key enzymes of the intracellular mevalonate pathway, thus preventing the prenylation and activation of small GTPases that are essential for bone resorption activity and survival. osteoclasts [28]. The final effect is similar to that observed with non-nitrogenous family members and is represented by apoptosis of osteoclasts [29] Denosumab and bisphosphonates have differing pharmacokinetics, resulting in the treatment of osteonecrosis caused by denosumab being more effective. The main factor is that the half-life of bisphosphonates is approximately 10 years, while denosumab is eliminated from the body 6 months after treatment cessation, allowing a faster resolution of osteonecrosis and decreasing the risk after drug discontinuation [24]. We believe that more studies are necessary to assess in short periods of time (less than 3 months) whether the effects of denosumab are comparable to those of BP. In longer periods of 6 months or more, it does seem to be associated with an increased risk of MRONJ compared to bisphosphonates. However, it has been seen that cessation of denosumab therapy may be associated with a faster resolution rate of MRONJ than that of BP; although this requires more prospective studies [30].In a systematic analysis conducted with a total of 8963 patients, it was determined that the use of denosumab was associated with a significantly higher risk of MRONJ compared to treatment with bisphosphonates (BP) [31].On the other hand, Limones et al. In a systematic review and meta-analysis published in 2020, they conclude that the use of denosumab is associated with a higher risk of developing maxillary osteonecrosis when compared to zolendronic acid (bisphosphonate), obtaining statistically significant differences between denosumab and BP to develop MRONJ after 1 year (p = 0.030), 2 years (p = 0.006) and 3 years of exposure (p = 0.007) [32].The objective of this study was to describe the results obtained in three patients diagnosed with MRONJ treated by the University of Murcia dental clinic according to the SECOM [16] and AAOMS [8] protocols.A 60-year-old woman smoker diagnosed with breast cancer, had received oral alendronate (Fosamax weekly, MSD, Madrid, Spain) for 10 months to combat osteoporosis caused by ovarian suppression, chemotherapy treatment, and the postmenopausal state itself. She presented discomfort and pain in the left hemimandible 10 weeks after removal of the first left lower molar. The examination showed an area with exposed necrotic bone measuring 8 × 6 mm2 with exudate (Figure 1); the patient was diagnosed with stage 2 MRONJ.Immediate conservative treatment included amoxicillin/clavulanic acid 2000/125 mg (Augmentin Plus, GSK, Madrid, Spain) every 12 h for 15 days, oral hygiene education and 0.12% chlorhexidine rinses every 12 h (Perio-Aid, Dentaid, Barcelona, Spain). The oncologist was consulted about discontinuation of bisphosphonates. At 15 days, partial closure of bone exposure without associated discomfort was observed; initial conservative treatment was maintained for another 15 days. At 30 days, soft tissue closure was complete, and the patient remained asymptomatic. Antibiotic treatment was discontinued, but 0.12% chlorhexidine rinses were maintained for an additional 15 days (Perio-Aid, Dentaid, Barcelona, Spain). At 12 months, there was normal soft tissue and no symptoms or signs of infection (Figure 2).An 80-year-old female non-smoker with type II diabetes, osteoarthritis and osteoporosis received oral ibandronate (Bonviva, Roche GSK, Madrid, Spain) for 12 months. She presented with discomfort in the left mandible 3 months after extraction of the second lower left molar.Oral examination showed a fistula with inflammation and erythema of adjacent soft tissue without bone exposure and X-ray showed a lesion compatible jawmaxil osteonecrosis. The patient was diagnosed with MRONJ stage 2 (Figure 3). Immediate conservative treatment included amoxicillin/clavulanic acid (2000 mg + 125 mg every 12 h for 15 days) (Augmentin Plus, GSK, Madrid, Spain) and 0.12% chlorhexidine rinses (Perio-Aid, Dentaid, Barcelona, Spain) every 12 h for 30 days. The rheumatologist was consulted to discontinue bisphosphonates.At 15 days, there was no discomfort and a significant reduction in fistula suppuration (Figure 4a), so treatment was maintained for another 15 days. At 30 days, there were no symptoms and an almost complete reduction in fistula suppuration (Figure 4b). Two months later, the patient reported discomfort and increased fistular suppuration, but without bone exposure. After re-establishing conservative treatment with amoxicillin/clavulanic acid (2000 mg + 125 mg every 12 h for 15 days) (Augmentin Plus, GSK, Madrid, Spain) and 0.12% chlorhexidine rinses (Perio-Aid, Dentaid, Barcelona, Spain) every 12 h for 30 days, minimal surgery for the removal of necrotic bone tissue was decided upon. After local anesthesia with articaine 4% + epinephrine (Inibsa, Barcelona, Spain), a flap was made at full thickness without discharges, necrotic bone tissue was removed, and smooth curettage of the bone walls with a dental spoon was made until bleeding was produced. The surgical bed was irrigated with serum and 0.12% chlorhexidine (Perio-Aid, Dentaid, Barcelona, Spain). The wound was stitched with 7/0 non-resorbable monofilament (Seralene, SERAG Wiessner Iberia, Madrid, Spain) (Figure 5).Fifteen days post-surgery there was no suppuration or discomfort, but the initial treatment was maintained for 7 more days (Figure 6a). At 30 days, there was closure of the soft tissue without signs of infection (Figure 6b) and 3 months later X-ray showed normal bone healing (Figure 6c). At 12 months, there was normal soft tissue with no symptoms or signs of infection.A 72-year-old female non-smoker with osteoarthritis and osteoporosis received denosumab 60 mg solution for injection (Prolia®, AmgenEurope B.V. Breda, Netherlands) every six months for two years. She developed discomfort and bone exposure after surgery for dental implants in the upper left jaw two weeks earlier (Figure 7). The patient was treated with amoxicillin 875 mg/clavulanic acid 125 mg every 8 h (Augmentin 875/125 mg, GSK, Madrid, Spain) and 0.12% chlorhexidine rinses every 12 h.Oral examination showed a bone exposure of 8 × 4 mm2 without suppuration and associated with discomfort, with the surrounding soft tissue inflamed with exudate and bleeding; the implant was stable with no pain on percussion. The diagnosis was MRONJ stage 2. Conservative treatment was instituted including amoxicillin 875 mg/clavulanic acid 125 mg every 8 h (Augmentin 875/125 mg, GSK, Madrid, Spain) and 0.12% chlorhexidine rinses every 12 h (Perio-Aid, Dentaid, Barcelona, Spain) for 15 more days. The family physician was consulted to cease denosumab: the last 60 mg dose had been injected two months earlier. At 15 days, the patient reported no discomfort in the area; scans showed persisting bone exposure of the same size as at the beginning of conservative treatment, but with normal surrounding soft tissue, although with slight edema; the implant remained stable and painless on percussion. cone beam computed tomography (CBCT) was used to assess the extent of bone necrosis (Figure 8) and, as it affected only the exposed part, we resected the exposed necrotic bone without an access flap by with minimum curettage, screwing in of the healing plugs and continuing with antibiotics and chlorhexidine rinses for 15 days.At two weeks, the patient reported discomfort again and although there was less bone exposure (Figure 9) exudate was observed. Surgery for the elimination of all necrotic bone, curettage and regeneration by platelet and leukocyte-rich fibrin (PRF-L) was carried out. After local anesthesia with articaine 4% + epinephrine (Inibsa, Barcelona, Spain), a flap was made at full thickness without discharge, extending towards the distal. Necrotic bone tissue was eliminated, and smooth curettage of the bone walls was made with a dental spoon to generate bleeding from the bone walls of the defect. The surgical bed was irrigated with serum and 0.12% chlorhexidine (Perio-Aid, Dentaid, Barcelona, Spain) (Figure 10).Before suturing the wound, several clots and PRF-L membranes obtained by centrifuge at 2700 rpm for 12 min of 40 mL of the patient’s blood were inserted (Figure 11).The wound was sutured with non-resorbable polytetrafluoroethylene 6/0 monofilament (Seramon®, SERAG Wiessner Iberia, Madrid, Spain) and an intraoral control X-ray taken (Figure 12). The same antibiotic therapy and 0.12% chlorhexidine rinses were maintained for a week.After 7 days, sutures were removed, and an incidence-free healing and lack of symptoms were observed; 0.12 % chlorhexidine rinses were maintained for 15 days. Reviews at 1, 6 and 12 months showed normal healing of soft tissues and bone tissue (Figure 13). All patients gave informed written consent after a full description of the surgery in accordance with the guidelines of the Helsinki World Medical Association Declaration and the revision of the 2013 Guidelines for Good Clinical Practice.Twelve months after carrying out the protocols described by SECOM and AAOMS for the handling of stage 2 MRONJ, the three cases presented completely-resolved lesions. Clinical examinations at 1, 6 and 12 months showed a fully healed gum with a healthy appearance without pain or discomfort. Unlike case 1, case 2 did not initially improve with antibiotics and 0.12% chlorhexidine rinses but, after minimally-invasive resection of exposed necrotic bone there was complete remission of the lesion. Radiographically, slow but progressive healing of the affected bone area, with a normal bone structure, was observed. In case 3, in which the involvement was associated with the implant, which worsened the prognosis by communication between the oral medium and the necrotic bone tissue, a first attempt to remove the fragment was not successful and complete resolution was not achieved and a second surgical approach had to be used to remove more necrotic bone tissue and graft PRF-L, with antibiotics and 0.12% chlorhexidine rinses, resulting in normal tissue healing and favorable bone regeneration.The pathogenesis of MRONJ is not fully defined. Exaggerated osteoclastogenesis inhibition, and reduced bone replacement and angiogenesis may trigger osteonecrosis [7,17]. The process of bone turnover begins with the action of osteoclasts through bone resorption, followed by the formation of new bone tissue from osteoblasts. Bisphosphonates produce a significant increase in osteoclast apoptosis, while other antiresorptive drugs nullify their function and differentiation. Osteoclasts are activated, among other proteins, by RANK-L, the inhibitor of which is an antibody that prevents the binding of RANK-L with its nuclear receptor and therefore, preventing the function of osteoclasts. All of this results in a considerable decrease in bone resorption, regeneration and remodeling, increasing bone density and therefore reducing the effects of osteoporosis [8,25].Although the optimal treatment of osteonecrosis is unclear, all scientific societies—AAOMS [7,8], SECOM [16], the Italian Society of Maxillofacial Surgery (SICMF) and the Italian Society of Pathology and Oral Medicine (SIPMO) [33], agree that minimally-invasive treatment, with the use of broad-spectrum oral antiseptics, is the best. Chlorhexidine gluconate, from the biguanide group, is the most widely used agent as it has a wide antimicrobial spectrum and substantivity, allowing chemical control of the biofilm. They are used in MRONJ stage 0, 1, 2 and 3 [34,35]. In the three clinical cases reported here, 0.12% chlorhexidine every 12 h, combined with antibiotics and painkillers, were important parts of treatment. In fact, in case 1, surgery was not necessary even though it was stage 2 MRONJ. In cases 2 and 3, surgery was necessary, together with chlorhexidine rinses and antibiotics, to achieve complete healing of soft tissues and bone, since no improvement was observed only with chlorhexidine. Bagan et al. [34] recommended, in addition to 0.12 chlorhexidine rinses twice daily, irrigation of the area affected every 72 h for 4 weeks and, if there was improvement, continuation of the rinses for one month.The best antibiotic schedule for the treatment of osteonecrosis is not well defined. In fact, their use in both intermittent and continuous cycles to prevent osteomyelitis and superinfection of the soft parts has been proposed [36,37]. Some authors have recommended clindamycin due to its efficacy against Gram-positive flora and its affinity for brain tissues [38], although Marx et al. [36] discouraged it as single therapy, and recommended penicillin derivatives as first choice drugs. The antibiotic administered should cover Actinomyces, Fusobacterium, Eikenella, Bacillus, Staphylococcus, Streptococcus and Treponemas [39]. Of all these, the bacterium that stands out is Actinomyces, as it has been observed in histological sections of osteonecrosis, although it is unknown whether it plays a primary (co-producer of MRONJ) or secondary role (colonizes the necrotic bone once in place), due to the difficulty of microbiological isolation, since no biopsy is usually performed in patients receiving bisphosphonates, for obvious reasons [40].Ideally, an antibiogram would be performed but, if not possible, the most appropriate antibiotics in the acute infection phases are beta lactams (amoxicillin or amoxicillin + clavulanic acid) [41]. According to SECOM [16], the guideline should be amoxicillin 875 mg/125 mg clavulanic acid or amoxicillin 2000 mg/125 mg and, in case of allergy to penicillin, clindamycin 300 mg every 8 h for 7 days, or levofloxacin 500 mg every 24 h, which have shown good results [37]. Other reports used clindamycin for two weeks, followed by amoxicillin-clavulanic acid for two weeks, and then penicillin G depending on the culture, although without specifying for how long [41]. In cases of beta lactam allergy, macrolides (azithromycin) or quinolone (ciprofloxacin or levofloxacin) are also used alone or in combination. If these are not effective, clindamycin or metronidazole [8,42] is added. In our cases, good results were obtained following the SECOM protocol, and administering 2000 mg/125 mg of amoxicillin- clavulanic acid in cases 1 and 2, and amoxicillin-clavulanic acid 875 mg/125 mg in case 3 for 15 days, and prolonging treatment for 15 days if no improvement was observed.The non-surgical treatment in stage 3 patients, consisting of systemic antibiotics, pentoxifylline 400 mg (which promotes blood circulation), tocopherol (vitamin E) IU twice daily and 0.12% chlorhexidine rinses four times daily, significantly improves pain, symptomatology and bone exposure [43].If mycosis associated with osteonecrosis is suspected, nystatin or fluconazole should be used. Some studies also prescribed penicillin G + IV metronidazole [44], levofloxacin + metronidazole [45], piperacillin + tazobactam or imipenem + cilastatin [46,47], with favorable results. Mycosis was not suspected in any of our three cases, so no antifungal treatment was necessary.Reports suggest antibiotics should be given for one week [48], ten days [49], fifteen days [50], three or four weeks, or until healing is complete [47]. Most studies agree that antibiotic treatment should be long-term until clinical remission of osteonecrosis-related signs and symptoms is achieved [47].Antibiotics fail in some cases, possibly due to the difficulty of reaching a target in an environment with little vascularization. In these cases, if symptoms are mild, some authors suggest bone debridement and closing with mucous flaps, always with antibiotic treatment [51]; this involves the removal of the exposed necrotic bone by surgical curettes, and it may be difficult to obtain margins with viable healthy bone: bisphosphonates or denosumab have little influence [45,52]. Antibiotic-impregnated membranes that may act as local anti-infectious drug release systems, which not only fill non-vital space after surgical debridement, but also provide high concentrations of antibiotics at the potential infection site, without increasing serum antibiotic levels, are currently being studied [53]. For example, using a collagen matrix impregnated with gentamicin as a topical supplement for perioperative antibiotic prophylaxis. The gentamicin-collagen sponge used by Chia et al. consists of a type I purified bovine collagen matrix impregnated with gentamicin sulfate 2.0 mg/cm, and the authors suggest it reduces the risk of postoperative infection after being applied in 92 patients [54].When surgery is used, a signal that may indicate the “healthy” bone margin is normal bleeding in the bone, which indicates there is sufficient potential for healing, although this is not always the case, and in many cases it cannot serve as a golden rule [55]. In addition, the use of ultraviolet light after prescribing tetracycline (250 mg 4 times daily for 3 days) or doxycycline (100 mg 2 times daily for 10 days) has been suggested as a means of detecting the extent of osteonecrosis and delimiting the margin of the resection successfully, even for surgical debridement [56]. Most common minimally invasive surgery technique is removal of symptomatic bone sequestration with minimal injury to the soft tissues [41]. However, the removal of necrotic bone may, in some situations, worsen the clinical picture. The use of surgical ultrasound in bone removal has shown good results as it allows careful cavitation due to the vibration around osteonecrosis, causing degradation of bacterial complexes and favoring revascularization and wound healing, although more experimental studies are needed [57,58].In addition to performing careful surgery technique, recent studies have also used growth factors such as platelet-rich plasma (PRP) or platelet and leukocyte-rich fibrin (PRF-L) in order to stimulate angiogenesis and the repair of local bone tissue [59]. In this procedure, all affected alveolar bone is removed, leaving only the basal component, grafting the PRP or PRF-L, and attempting primary closure of the mucosa [60,61]. Dinca et al. (2014) used PRF-L in patients with jaw osteonecrosis stage 2 following intravenous bisphosphonate therapy in the alveoli post-extraction [62]. The sample used was small and the study had limitations, but there were no postoperative complications in any of the 10 cases studied and after 30 days there was no evidence of bone exposure [62]. In our study, PRF-L membranes were applied in the second surgical approach in patient 3, achieving better results than in the first approach where PRF-L was not applied, and resulting in complete soft tissue healing and bone regeneration. The use of platelet- and leukocyte-rich fibrin membranes in patients with MRONJ appears to be a promising treatment due to the association between MRONJ and bone remodeling suppression, antiangiogenic effects, reduced immune response and soft tissue toxicity; however, further research is needed to confirm these effects [63].Another therapeutic option for the treatment of MRONJ is the use of stem cells. In an experimental study in rabbits under treatment with bisphosphonates and immunosuppressants, stem cells prevented osteonecrosis and significantly improved healing after tooth extraction and local transplant of stem cells derived from adipose tissue [64]. There are few human studies of this therapy: there is a report of a 71-year-old patient with multiple myeloma in the jaw, who was treated with autologous bone marrow stem cells, platelet-rich plasma, beta tricalcium phosphate and demineralized bone matrix. At 6 months, healing and bone formation had improved. The authors concluded that cell therapy could be applied in patients with refractory osteonecrosis [65].This case series suggests the effectiveness of the various protocols published for the treatment of stage 2 MRONJ, although the small number of patients included is a limitation. Therefore, randomized clinical trials with a control group and more patients included, using the same protocol and treatment parameters, are recommended.Conservative treatment with antibiotic therapy, chlorhexidine rinses, minimally invasive surgery with necrotic bone resection and the application of PRF-L membranes has been shown to be an effective treatment for stage 2 MRONJ.Studies of denosumab-related jaw osteonecrosis have shown it can be resolved more quickly with the suspension of the drug for 6 months compared with bisphosphonates, but during treatment, denosumab resulted in a higher incidence of MRONJ with less trauma, and sometimes spontaneously; further studies are needed.The novel therapeutic options for the treatment of MRONJ, such as platelet and leukocyte-rich fibrin (PRF-L) represent improvements in soft tissue healing and bone regeneration, although more research is needed to confirm their effectiveness.G.P.-Z.; Y.M. performed the clinical examination, diagnosis, treatments, and follow-up. G.P.-Z.; Y.M.; A.J.O.-R. wrote the manuscript. Y.M. and J.A.M. performed the literature review. All authors read and agreed to the published version of the manuscript. This research received no external funding.All patients gave informed written consent after a full description of the treatment in accordance with the guidelines of the Helsinki World Medical Association Declaration and revision of the 2013 Guidelines for Good Clinical Practice.Informed consent was obtained from all subjects involved in the study.The authors declare no conflict of interest. (a) Clinical aspect of the injury. (b) X-ray of the affected area.(a) Clinical aspect and (b) X-ray of the affected area 12 months after treatment.(a) Fistula in the jaw and (b) X-ray of the affected area.(a) Fistula at 15 days and (b) fistula at 1 month.(a) Full thickness flap, (b) fragment extracted from necrotic bone, (c) surgical bed after curettage and irrigation and (d) monofilament suture 7/0.(a) Affected area 15 days post-surgery, (b) 30 days post-surgery and (c) X-ray 90 days post-surgery.Bone exposure in the distal implant at 7 (a) and 15 (b) days post implant-surgery.(a) Bone exposure at 30 days post implant-surgery. (b) White arrows: extension of bone necrosis at cross-sections at the crestal level and two millimeters towards apical on CBCT.(a,b) Clinical situation at 45 days after removal of necrotic bone tissue and antibiotic therapy.(a) Surgical access by a flap without discharge; (b) 7 × 4 mm bone fragment.(a) PRF-L membranes; (b) Adaptation of PRF-L in bone defect and affected implant.(a) Stress-free suture; (b) Resulting bone defect in post-surgical intraoral X-ray.(a) Soft tissues with normal appearance at 12 months; (b) Radiographically stable bone tissue at 12 months.Grades or stages of MRONJ according to the AAOMS [8].Recommended treatments according to the grade or stage of MRONJ.No treatment indicated.Preventive patient education.Systemic management, including painkillers and antibiotics.Clinical follow-up with quarterly antibacterial mouthwashes. Patient education and review of indications for continuous therapy with bisphosphonates.Symptomatic treatment with antibiotics.Antibacterial mouthwash.Pain management.Debridement to relieve soft tissue irritation and infection control.Antibacterial mouthwash.Antibiotic therapy and pain management.Surgical debridement or resection to alleviate long-term infection and pain.Treatment protocol recommended by SECOM for MRONJ.Quantification in millimeters of the exposed area.If possible, stop treatment with anti-resorptive or antiangiogenic drugs.Rinses with 0.12% or 0.2% chlorhexidine every 12 h for 15 days.Control at 15 days:-Equal or smaller size: maintain treatment for another 15 days.-Increased exposure: apply stage 2 treatment.Equal or smaller size: maintain treatment for another 15 days.Increased exposure: apply stage 2 treatment.Control at one month.Points 1–3 as in stage 1.Empirical antibiotic therapy (when no culture or antibiogram is available):-Amoxicillin/clavulanic acid 2000/125 mg, every 12 h for 15 days.-Allergic patients: Levofloxacin 500 mg every 24 h for 15 days. -Alternative: Azithromycin.Amoxicillin/clavulanic acid 2000/125 mg, every 12 h for 15 days.Allergic patients: Levofloxacin 500 mg every 24 h for 15 days. Alternative: Azithromycin.Administer oral nonsteroidal anti-inflammatory drugs (NSAIDs).Control at 15 days:-Improvement: move to stage 1 treatment.-Equal or worse: maintain treatment for another 15 days and request computed tomography (CT). Improvement: move to stage 1 treatment.Equal or worse: maintain treatment for another 15 days and request computed tomography (CT). Control at one month:-Improvement: move to stage 1 treatment.-Equal or worse: move to stage 3 treatment.Improvement: move to stage 1 treatment.Equal or worse: move to stage 3 treatment.Point 1–3 as in stage 1.Point 2 and 3 as in stage 2.Under local anesthesia, eliminate bone sequestration and extraction of teeth involved if necessary, irrigation of the surgical bed with 0.12% chlorhexidine and suture with resorbable material.Control at 15 days:-Improvement: maintain antibiotics, anti-inflammatories and rinses for another 15 days.-Equal or worse: maintain antibiotics, anti-inflammatories, and rinses for another 15 days.-Control at one month-Improvement: implement preventive measures and reinstate treatment with anti-resorptive or antiangiogenic drugs.-Equal or worse: new conservative surgery in serious circumstances: Pathological fracture: curettage of necrotic bone tissue and reconstruction plate (avoid grafts). Involvement of inferior border: block resection and reconstruction plate. Extraoral fistula: debridement to eliminate necrotic bone causing mucous irritation.Improvement: maintain antibiotics, anti-inflammatories and rinses for another 15 days.Equal or worse: maintain antibiotics, anti-inflammatories, and rinses for another 15 days.Control at one monthImprovement: implement preventive measures and reinstate treatment with anti-resorptive or antiangiogenic drugs.Equal or worse: new conservative surgery in serious circumstances: Pathological fracture: curettage of necrotic bone tissue and reconstruction plate (avoid grafts). Involvement of inferior border: block resection and reconstruction plate. Extraoral fistula: debridement to eliminate necrotic bone causing mucous irritation.Pathological fracture: curettage of necrotic bone tissue and reconstruction plate (avoid grafts). Involvement of inferior border: block resection and reconstruction plate. Extraoral fistula: debridement to eliminate necrotic bone causing mucous irritation.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01019.txt

@@ -0,0 +1,17 @@
+In the landscape of Covid-19 pandemic, people’s well-being, to some extent, can be affected through virtual reality tourism because it has the opportunity to enhance their level of well-being and destination recovery. To verify this empirically an investigation was conducted among people who used Quanjingke, the largest tourism-related virtual reality website in China, during the pandemic. The specific the aim of this paper is to prove the effectiveness of virtual reality tourism in promoting people’s subjective well-being. Hence, an integrated model with the constructs of peripheral attribute, core attribute and pivotal attribute, presence, perceived value, satisfaction, and subjective well-being was proposed and tested. The results indicate that attributes of virtual reality tourism have a positive effect on presence during virtual reality experiences, which positively influences perceived value. The values of virtual reality tourism perceived by tourists result in their satisfaction. Eventually, it was found that tourists’ subjective well-being is improved due to their satisfaction with virtual reality tourism. Practical suggestions are also provided based on the findings.From the beginning of 2020 up to now, most of the world has been struggling with the COVID-19 pandemic. Rigorous restrictions, like entry bans and quarantines, and preventive measures are imposed throughout the world to halt the spread of the coronavirus, causing a downturn in economic activity and sapping the economic momentum of countries. There is no exception for the tourism industry with regard to such a recession. The COVID-19 pandemic has caused a 70% fall in international tourist arrivals (overnight visitors) during first eight months of 2020 compared to the same period of the previous year. Particularly, international arrivals declined 81% in July and 79% in August, which would usually be the peak season of the year, the latest data from the World Tourism Organization (UNWTO) indicates [1].Considering the quality of life and residents’ well-being, our lives have been dramatically changed by the pandemic. Waves of strict lockdowns are “inevitable” and social distancing is required by governments and organizations to prevent virus transmission, which, to some extent, induce negative effects on the mental health and well-being of residents and the effects will extend beyond those who have been directly affected by the virus [2]. Ma and Yang found that the onset of the coronavirus epidemic led to a 74% drop in overall emotional well-being which is an important constituent of subjective well-being (SWB) [3]. Given the importance of SWB to residents’ lives, it is time to take reasonable precautions to help us bring health and well-being to the forefront. Recommendations that could contribute to an alteration in residents’ well-being, including the promotion of physical activity and sleep hygiene etc. are contained in the relevant literature [4]. The improvement of the well-being of humankind has been an object of many public policies [5]. On the individual level, experiences during travel and satisfaction with travel contribute to improving well-being [5]. Given the risk of increasing the opportunities for spreading the virus and getting infected, it would be better to find an alternative way to travel without physical movement when travels are postponed on a large scale and staying at home is advocated.In the current world, which has been filled with artificial intelligence, we are becoming increasingly reliant on technology. For example, virtual reality (VR) is expected to be one of the significant technology products in the tourism industry. By providing accessible tourism for all and enhancing tourist experience, immersion, and visualization, VR may be an invaluable resource for transforming traditional tourism with intangible and experiential essence into a product [6,7]. Therefore, it seems that VR tourism has the potential to provide residents with the access to experience tourist sites in VR and can play a vital role in improving residents’ well-being [8]. The subject of this study, Quanjingke (QJK) provides 360-degree and ultra-high-definition panoramic images and guided tours and its language is Chinese only. According to the latest report from one of the most authoritative websites in China, QJK is the largest tourism-related VR website in China and it has around 1.5 billion active users and provides over 60,000 panoramic pictures and VR videos all over the whole country [9]. The large amounts of users of QJK and rich VR resources make it reliable for collecting meaningful data for the current study. At the early stage of the pandemic, it is noteworthy that an article page published on QJK’s official account of WeChat induced the number of 67 million page views, leading to over 40 million unique visitors and 150 million page views for its website and its popular app, “Beautiful China” [9]. Many Chinese accepted the new idea of “take it home”, which means to encourage potential tourists to “carry” tourist resorts to their home by using laptops and smart phones via which the interesting photos or videos and intelligent tour guides can be attained during the pandemic. With rich VR tourism resources and large amounts users, QJK enables us to access high quality data for empirical research related to VR tourism in the context of the pandemic.The subjective evaluation of well-being is frequently referred to as SWB, and the subjective appraisal of well-being is the measure of well-being achieved when answering questions, which means that SWB can be measured by questionnaire in a self-reported way [10]. The adoption of VR associated with enhancing subject well-being has been examined in various contexts. For example, Li et al. investigated the effectiveness of using virtual reality computer games in promoting the subjective well-being of children with cancer [11]. In spite of this, exploring the role of VR tourism on enhancing residents’ subjective well-being remains in its infancy, as an integrated model of VR tourism has not been established [12]. To be concrete, researchers emphasize certain facets of VR tourist behavior (e.g., perception of authenticity and attitude) for the purpose of tourism marketing [12]. In addition, a handful of the literature employing theoretically integrated model remains on too broad scope rather than concentrating on a specific VR tourism product, leading to potentially diverse conclusions. Focusing on the VR tourists of QJK, we strive to fill the aforementioned research gaps by establishing an integrated model. In this study, structural equation modeling (SEM) was employed to test the proposed model and hence the relationship between VR tourism and residents’ subjective well-being was explicated clearly on both theoretical and empirical level.Incorporating the constructs of PCP attributes, presence, perceived values, and satisfaction into an integrated model, the current study aims to explore how the VR tourism promote residents’ subjective well-being by delving into its mechanisms. The model incorporates constructs frequently used in the consumer behavior domain, encompassing PCP attributes (i.e., peripheral attribute, core attribute and pivotal attribute), perceived value (i.e., functional value and emotional value), and satisfaction, with the construct unique to VR tourism research (i.e., presence). Consequently, the theoretical and practical implications are summarized based on the results of empirical analysis. Thus, the policy makers, residents, tourist destinations and VR tourism operators will all benefit from the research findings. Through the era of PC, realistic online (including mobile) content that stimulates the five senses through VR (virtual reality) and AR (augmented reality) is growing. Virtual reality refers to a system that creates a three-dimensional visual and auditory experience in real time and expresses an object in a simulated form [13]. The concept of virtual reality began to be used as a theoretical approach in the field of HCI (Hyper Converged Infrastructure) in the mid-1970s, but it began to be actively used in the 1980s. Various studies and attempts have been made in the field of virtual reality [14,15]. In addition, as related contents increase, various distributions are being made in connection with culture, art, sports, and tourism. In the future, more fields using virtual reality technology such as games, education, medical care, manufacturing, and e-commerce are expected to increase. Burdea and Coiffet defined virtual reality as an interface between humans and computers that enables users to immerse themselves and interact in real time [15]. As one of the fields of virtual reality application, VR tourism (with the synonyms “virtual tour”, “panoramic tour” etc.) provides VR tourists with the online service to experience traveling in virtual environments by creating multimedia elements and simulating real tourist sites and unreal situations [8].Just as e-service enterprises in the travel industry offer various online services (travel planning, hotel reservations, and rental car services), VR tourism operators provide various types of virtual services related to destination experiences [12]. Therefore, it is crucially important to grasp the general properties of VR tourism and to measure VR tourists’ evaluation of the service performance of VR tourism in order to explain the mechanism by which VR tourists improve subjective well-being through VR travel.Scholars have proposed the various attributes of e-service quality based on their context. Among various e-service quality-related studies, several studies that are related to VR tourism are as follows. Argyriou et al. proposed five primary attributes of VR tourism (“narrative”, “virtual scenes”, “actor role”, “navigation”, “gamified”) that are important to VR tourism quality [16]. Chiao et al. identified a virtual reality tour-guiding platform consisting of “itinerary planning”, “virtual game-based design”, “cultural tourism features” and “tourism English��� [17]. Hahn et al. initiated a user-centered design of a virtual reality heritage tourism system composed of three basic attributes: “VR environment”, “optimization” and “player interaction” [18]. Although various attributes related to VR tourism have been proposed based on their context, there is no consensus on the attributes of VR tourism. Moreover, it seems that prior research did not examine specific attributes from a holistic and systematic perspective because VR tourism attributes remain largely fragmented. In terms of more advanced approaches to VR tourism attributes, Philip and Hazlett proposed the hierarchical structure model called the PCP (pivotal, core, peripheral) attribute model, which can offer support in this area. In the model, pivotal attributes that focus mainly on output are considered attributes of the most intrinsic central level of quality of service. Pivotal attributes refer to the most decisive and core attributes in providing any product or service to consumers regardless of personal preference. In the case of QJK, fulfilment of travel needs and experience could be a pivotal attribute. Core attributes that users need to interact with to achieve the pivotal attribute are considered as the process and middle level. The core attribute, which encompasses the pivotal attribute, acts as a mediator to help realize the ultimate goal, a satisfying VR tourism experience. For this reason, ease of use, personal information protection, and security can be included in the category of core attribute. Peripheral attributes representing “completeness to the entire service encounter” or “roundness” are considered input and lowest level [19]. The peripheral attribute, in connection with the core and peripheral attributes, plays a role of making the product distinct from other types of products. In the case of VR tourism, tourists are expected to recognize the difference between VR travel and other previous tourism products mainly through the peripheral attribute. Specifically, in terms of VR tourism, interface design and operating system quality are expected to be included in the surrounding properties. Considering the characteristics of these three attributes, the PCP attribute model seems to contribute to developing a framework of VR attributes as a scientific body of knowledge [20,21,22].The common underpinning paradigm of the PCP attribute model has received much attention from researchers in the field of marketing. The three-rank attribute model is always trimmed to a two-level attribute model when it is employed in practice, with peripheral (input) being the low level and core attribute (output) being the intrinsic ranking. For example, Skard et al. investigated consumers’ inferences about sustainable products with green core attributes and green peripheral attributes [23]. Wang et al. examined antecedents of brand experience in a historical and cultural theme park with the core and peripheral attributes [24]. However, limited research has investigated the PCP attribute of VR tourism. Built on the skeletal framework of the PCP attribute model and primary streams of literature about VR tourism, the current study develops a specific PCP attribute model for VR tourism to fill the gap. In the current study, VR tourism consists of pivotal attributes which indicate the output of VR tourism (e.g., VR tourism fits well with tourist’s travel needs), core attributes which refer to the process of VR tourism (e.g., VR tourism is easy to use) and peripheral attributes which represent the input of VR tourism (e.g., the user interface design is fascinating). The configuration of the PCP attributes is shown in Figure 1.Presence is the subjective experience of the VR environment, whilst users are physically in real world [6]. The term of presence, also known as telepresence, is widely accepted as a sense of “being there”, a psychological effect, in non-physical space [7,25]. Social psychology researchers and practitioners have noticed the significance of understanding presence and the relationship between presence and VR attributes. Presence is crucial for evaluating VR effectiveness [6,7,25,26]. To put it another way, when the level of presence experienced by a participant is low, the side effects may be produced. Nichols et al. addressed the important role of presence in VR and identified attributes (i.e., content and design, scene registration or update lags, head-mounted display optics and design, display and interaction) that may produce “sickness” [26]. Orth et al. postulated four informational attributes (i.e., mystery, complexity, legibility, and coherence) with construal level theory and examined how to achieve presence in virtual service environments [27].The main discussion dominating the literature is that presence is characterized as transportation, a sort of subjective experience or sensation of “it is here” and “being there”. This sense of transportation is usually labeled using a two-dimension metaphor, arrival (being present in VR) and departure (not being present in VR) [7,28]. The discussion notes the critical dynamic process in which VR tourists continuously suppress input information that is incompatible with his or her desired VR experience and construct the mental model needed to experience presence [29]. The PCP attribute model proposed in the current paper is designed to understand this transportation process from input to output. In addition, the PCP model includes various determinants of presence which can be generally divided into external stimuli (VR environment delivery) and internal tendencies (user features) [6,30,31,32]. For example, richness, one of VR environment delivery, is reflected in an item of pivotal attribute (“VR tourism fits well with my travel needs”). Even if specific determinants are not included in the current item scale, they can be categorized as one of the PCP attributes. Consequently, the following hypotheses are proposed:
+            The peripheral attribute has a positive effect on presence in VR tourism.
+          
+            The core attribute has a positive effect on presence in VR tourism.
+          
+            The pivotal attribute has a positive effect on presence in VR tourism.
+          In general, value is an abstract and polysemic concept. The mainstream of the academic literature has been focusing on perceived value instead. Perceived value is a useful and critical construct for identifying tourist behavior in tourism industry. In many cases, perceived value has been regarded as a multidimensional concept, although sometimes as unidimensional one, with regards to overall value [33]. It is commonly understood from the consumers’ standpoint in the literature. In the early 1988, Zeithaml captured a widely accepted definition from four prior definitions. Perceived value is defined as “the consumer’s overall assessment of the utility of a product or service based on perceptions of what is received and what is given” [34]. It is built on the dual conception, “get-give” tradeoff [35]. Accordingly, VR tourists may trade off perceived benefits (e.g., convenience, utilitarian features, positive emotions) and perceived sacrifices (e.g., time, money, effort) [33]. The tradeoff conception conceives perceived value as a temporally dynamic process: pre-use, at the time of use and after use [36]. Although ubiquitous dimensions of perceived value are proposed in the literature, they echo the two underlying ones, functional value and emotional value [36,37,38,39,40]. Functional value refers to the rational and utilitarian value perceived by individuals. Emotional value is the feelings or affective states generated by a product or service [36,37].The research on the consequences of presence in VR has converged on emotional response. For example, Yung et al. established a conceptual model comprising the consequences of presence in VR on emotional response by a critical review of presence research [6]. Gorini et al. discovered the similar findings when they evaluated the emotional response produced by VR [41]. Furthermore, presence is found to be crucial for improving perceived effectiveness and usability [42,43]. Brade et al. evaluated impact of presence on perceived usability using a mobile navigation task [43]. Likewise, Sun et al. demonstrated that presence is positively related to functional value in virtual environments [42]. Additionally, the extant literature has confirmed that tourist experience influences their perceptions of functional and emotional values. For example, Song et al. examined the impact of tourist experience on perceived value with temple stays [44]. According to the theory of presence, presence is a sort of subjective experience [6]. Above all, the hypotheses are suggested as follows:
+            Presence in VR tourism positively influences functional value.
+          
+            Presence in VR tourism positively influences emotional value.
+          Customer satisfaction is defined as the measurement of how the actual experience generated by the service or product fulfills the customers’ expectations [45]. It is the central concept of marketing from which the term “tourist satisfaction” derived [46]. In the literature about tourism, the tourists’ overall satisfaction is usually in line with their levels of return visits to the destination, loyalty, and the retention of tourists [47]. Hence, managing tourist satisfaction is crucial for the successful development of the tourism industry. Accordingly, VR tourist satisfaction is of substantial importance for understanding the effectiveness and performance of VR tourism. In recent studies, research has been exploring the topic of satisfaction in VR. Hudson et al. investigated the moderating effect of immersion, interaction and social interaction in VR on users’ satisfaction [48]. Kim and Ko found that the effect of VR on flow experience, which will improve media user satisfaction, decreases as sport involvement increases [49]. Thus, satisfaction in VR has received much attention.Besides, a range of tourism research has validated the relationship between satisfaction and perceived value, with an increasing number of studies reporting that tourist satisfaction is positively affected by perceived value. For example, Song et al. confirmed the clear relationship between tourist satisfaction and perceived value, showing that functional and emotional values influence tourist satisfaction [49]. Similarly, Wang et al. examined the positive impact of functional value and emotional value on consumers’ satisfaction level at a theme park [24]. However, few studies explore the effect of perceived value on satisfaction in VR tourism. To fill in the gap, we propose the following hypotheses:
+            Functional value has a positive effect on satisfaction in VR tourism.
+          
+            Emotional value has a positive effect on satisfaction in VR tourism.
+          The psychological study of subjective well-being has developed since Warner Wilson’s critical review [50]. Defined as people’s evaluation of their well-being, subjective well-being (SWB) is an essential element for improving positive physical and mental health and quality of life. [10,12,50]. In the context of tourism, SWB is, on one hand, the social outcome of tourism development: on the other hand, SWB is beneficial for tourism operators, policy makers, and tourists to promote understanding of the impacts of the tourism industry [51]. The fact that tourism contributes to tourist SWB has been confirmed in the literature. Meng et al. investigated the SWB of Chinese rural–urban migrants in the context of rural tourism, revealing that returning to rural destinations improves tourists’ SWB as they achieve an important lifetime goal via such experiences [52]. Through exploring the nature of tourists’ experiences, Knobloch et al. suggested understanding tourist consumption experiences beyond their momentary effects and considering a broader scope of well-being [53]. Thus, this empirical research rests on cognitive bases (e.g., the accomplishment of goals) and effects (hedonic balance) [51].In the literature concerning the study of tourism, the research has recently started to focus on the link between tourist satisfaction and SWB. Saayman et al. investigated the impact of travel experience on tourists’ experience which further influence their SWB [54]. Similarly, Su et al. reported that overall customer satisfaction has a positive influence on SWB [51]. Nonetheless, such attempts are not observed in VR and VR tourism settings. Based on prior findings, the following hypotheses are formulated:
+            Satisfaction in VR tourism has a positive influence on SWB.
+          A thorough review of the literature concerning the related constructs and topics was undertaken before the original measurement items were developed. To ensure an appropriate questionnaire with good readability and effectiveness, two experts, and Mr. Ma, the chief executive of QJK, were asked to assess the content validity of the questionnaire and some obscure expressions in it were removed or modified. A 5-point Likert-type scale ranging from 1= “strongly disagree “to 5= “strongly agree” was applied to measure questionnaire items (see Appendix A). Built on the implication of PCP model and variables of VR tourism attributes in prior research, multiple items used to measure peripheral, core, and pivotal attributes were adopted [21,24,55]. In specific, the peripheral attribute was assessed using a 5-item scale including “operating system compatibility and applicability” and “interface design”. The core attribute was determined by a 4-item scale comprising “ease of use”, and “privacy and security”. The pivotal attribute was measured by a 5-item scale involving “fulfillment and advantage of VR tourism to users”. The 4-item scale of presence (e.g., “In the VR tourist world, I had a sense of being there”) was adapted from Bogicevic et al. and Schuemie et al. [25,29]. Based on suggestions from prior studies, two dimensions of perceived value (i.e., functional value and emotional value) were measured with four items respectively for each one (e.g., “The VR tour on QJK has a consistent level of quality” for functional value; “Using QJK for VR travel gives me a feeling of happiness” for emotional value) [33,36,44]. Satisfaction was operationalized with three items which were recommended by Lee et al. and Song et al. [40,44]. Finally, subjective well-being was assessed with four items, as suggested by Kim and Hall [8].With the assistance of QJK, researchers contacted with VR tourists from the top four metropolises in China (i.e., Beijing, Shanghai, Guangzhou, and Shenzhen) via a Group Chat created by QJK. WeChat and QQ are the most prevalent social media platforms whose users are active. The Group Chats on WeChat and QQ were established as channels for QJK users’ to communicate with and give feedback to us. It was much easier to conduct the survey in the four cities which have a significant number of QJK’s users from various areas of China. Compared with an offline survey, it was more suitable to perform the research online in this study because the most of the users of QJK are active online and they registered with their real names. Based on this, an online anonymous survey was conducted among residents who have used QJK for VR travel during the COVID-19 pandemic (from February to November in 2020). The data were collected online from 19 November to 11 December 2020 by employing convenience sampling. The researchers sent friend requests to the potential respondents via WeChat or QQ in the first instance. Next, we described the purpose of survey, the time when responses were due, and compensation. After accepting our invitation on WeChat or QQ, each of respondents was asked to fill in the self-administered questionnaire online. 589 questionnaires were distributed and a total of 542 respondents completed the questionnaire. After excluding invalid questionnaires that were completed hastily or in repetitive response patterns, the remaining 490 completed questionnaires were finally used for the empirical analysis.The data were analyzed using R and descriptive statistics was performed at first. Based on Anderson and Gerbing’s suggestions, the current study conducted structural equation modelling (SEM) with a two-step approach [56]. In order to ensure internal consistency together with construct validity and reliability, confirmatory factor analysis was firstly implemented to examine the measurement model for all variables. Moreover, SEM was performed to examine the proposed research model and hypotheses. Figure 2 is the proposed conceptual model.The complete respondent demographic characteristics are provided in Table 1. As shown in Table 1, the number of females (51.6%) was slightly higher than that of males (48.4%). Among the 490 respondents, 31.2% were single and 59.2% were married. A wide range of occupations were present, including technicians and professionals (20.4%), businessmen and self-employed (23%), service workers (5.7%), office workers (8.4%), officials (10.4%), students (10%), freelancers (13.5%), and retired people (4.9), with the level of education ranging from less than high school (11.8%), to a postgraduate degree (18.2%). The majority of reported monthly incomes were more than CNY 5000. In terms of residence, 21.2% of respondents were in Beijing, 24.9% were in Shanghai, 27.4% were in Guangzhou, and 26.5% were in Shenzhen. The age group of 20–29 years old was dominant, representing 37.1%, followed by age groups of 30–39 years old (29.2%) and 40–49 years old (21.6%).Generally, two approaches are used to assess structural equations, maximum likelihood (ML) and robust methods. The commonly used ML estimation is used when the data follow the assumption of a multivariate normal distribution. If the data do not meet the criteria for a multivariate normal distribution, the study results provided through ML are considered unreliable [57]. In this case, another approach like robust estimation should be performed. To test the multivariate normal distribution assumption, Mardia’s standardization coefficient is used. If the value exceeds 5, the collected data are considered not to satisfy the assumption of multivariate normal distribution. In this study, the MLM (maximum likelihood estimation with robust standard errors and a Satorra–Bentler scaled test statistic) estimator, which is one of the powerful methods, was used because Mardia’s standardization coefficient (66.032) in this study exceeded the cutoff value of 5 [58].Hair et al. suggested that normed S-B χ2 below 3 is associated with a good model fit if sample size is less than 750. Values of 0.9 or greater show good model fit for the Comparative Fit Index (CFI), Normed Fit Index (NFI) and Non-Normed Fit Index (NNFI). For the Root Mean Square Error of Approximation (RMSEA), a cut-off criterion is needed [59]. As presented in Table 2, the overall fit of the measurement model is satisfactory: S-B χ2 (df) = 739.268 (467), Normed S-B χ2 = 1.583, CFI (Comparative Fit Index) = 0.975, NFI (Normed Fit Index) = 0.936, NNFI (Non-Normed Fit Index) = 0.972, RMSEA = 0.034.The Cronbach’s alpha of the latent variables varied from 0.863 to 0.939, indicating the acceptable reliability of the measurement model. The standardized factor loadings of the items ranged from 0.772 to 0.907, which were all statically significant (p < 0.001) and exceeded the cut-off point of 0.5. The values of average variance extracted (AVE) were all greater than the recommended value of 0.5, varying from 0.675 to 0.782. Composite reliability (CR) for all variables ranged from 0.862 to 0.940, which exceeded the critical value of 0.7 [60]. In addition, all AVE values were greater than the values of squared correlations among latent constructs [61]. Therefore, the convergent and discriminant validity was confirmed [59].As illustrated in Figure 3, the overall fit of the structural model is satisfactory: S-B χ2 (df) = 886.124, Normed S-B χ2 = 1.831, CFI = 0.964, NFI = 0.923, NNFI = 0.960, RMSEA = 0.041. Based on the cut-off values suggested in the prior discussion, the results demonstrate that the structural model’s fit is satisfactory.As for Hypothesis 1, which predicted a positive relationship between peripheral attribute and presence was supported (β PEA→PRE = 0.199, t = 4.597, p < 0.001). The hypothesized positive relationship between core attribute and presence (H2) was accepted (β CA→PRE = 0.315, t = 7.330, p < 0.001). As presumed by Hypothesis 3, the pivotal attribute had a positive effect on presence in VR tourism (β PIA→PRE = 0.439, t = 11.201, p < 0.001). It presented that presence positively influenced functional value (β PRE→FV = 0.696, t = 22.997, p < 0.001) and emotional value (β PRE→FV = 0.684, t = 21.012, p < 0.001). Function value (β FV→SAT = 0.354, t = 11.054, p < 0.001) and emotional value (β EV→SAT = 0.576, t = 20.201, p < 0.001) were each found to have a positive impact on satisfaction. Finally, satisfaction positively influenced subjective well-being (β SAT→SWB = 0.783, t = 28.622, p < 0.001). The predicted relationships, coefficients, t-values and results of hypothesis test are shown in Table 3.This study established and tested an integrated model with constructs of VR tourism attribute, presence during VR tourism experience, perceived value of VR tourism, satisfaction with VR tourism experience and VR tourists’ subjective well-being. Firstly, built on Philip and Hazlett’s proposed the hierarchical structure model, the PCP attributes of VR tourism was developed [20,21,22]. Our results indicate that VR attribute positively influences presence during VR tourism experience. Specifically, the pivotal attribute has the strongest impact on presence among the other two sorts of attributes, with the core attribute being the second strongest and the peripheral attribute ranking the last. This finding accords with Philip and Hazlett’s previous judgements, which considered the pivotal attribute to be the center-level, the core attribute to be the median and pivotal to be the exterior [21]. In practice, the three levels are usually reduced to two, namely the peripheral and core attribute. Similarly, our findings are also consistent with results of prior research that reported the stronger influence of the core attribute on tourist experience [24,40]. There is no consensus on the attributes of VR tourism in the related literature. Elements of VR tourism products or services determining presence during VR experiences were also revealed. To put it another way, VR tourists place more weight on core attributes (e.g., “VR product is easy to use”) and pivotal attributes (e.g., “VR product fits travel needs”) than peripheral attribute (e.g., “user interface design”), while peripheral attribute is not dispensable. Secondly, presence during VR tourism experience is demonstrated to positively affect two constructs of perceived value to almost the same degree. Presence denotes a psychological effect of “being there” in virtual space. This shows the dynamic process that VR tourists subjectively select their expected information during VR travel to receive the desired emotional and functional values which have equal importance to them. The impact of presence during VR experience on emotional response, perceived effectiveness and usability has been evaluated in the extant literature [41,42,43,44]. However, this is the first study to explore the relationship between presence and perceived value in the domain of VR tourism. Filling this gap is important not merely due to extending the VR tourism literature by covering a relationship that tends to be ignored, but also because perceived value is beneficial to predicting the effectiveness and performance of VR tourism [34,35]. Thirdly, a significant positive effect of perceived value on satisfaction confirms that the extent to which VR tourists’ emotional reactions and VR tourism products or services influence their overall evaluation of VR tourism. Consistent with clearly certified correlations in previous research, emotional value was found to be more significant than functional value in influencing VR tourists’ satisfaction [24,40]. This result suggests that focusing on the emotional design of VR tourism tends to make it easier to foster satisfaction in VR tourism and further improve residents’ well-being. Finally, satisfaction positively leads to subjective well-being, as explained by 61.3% of the sample. It is important to note here that residents’ subjective well-being will be improved if VR tourism can be served as their satisfactory means of leisure activities. Therefore, the critical role of VR tourism in improving residents’ well-being has been proved.Like all studies, the present study has some limitations that warrant consideration in future research. These investigations were only performed in the largest metropolises in China (Beijing, Shanghai, Guangzhou and Shenzhen) and the future research should consider a comparative study in other geographical areas. Further, despite the PCP attribute model provides a framework for exploring attributes of VR tourism, the importance and performance of each specific attribute are not assessed. Evaluating each specific attribute may help VR tourism producers understand their products and VR tourist feedback. Importance-performance matrix should be employed for future investigation. Besides, the survey was conducted among respondents who have used QJK during the pandemic, while users’ acceptance of VR tourism was not analyzed. Future researchers could explain VR travel intentions and subsequent behaviors with models like UTAUT model (the unified theory of acceptance and use of technology model). Future researcher should consider to perform psychological experiments onsite among those who have no experience of VR travel and thereby recommend effective marketing strategies for tourism operators who are eager to see the recovery from recession of tourism industry. Likewise, the same patterns of experiments should also be carried out among international tourists after the pandemic. VR tourism offers opportunities for marketers to communicate their tourism products to potential visitors and enhances mutual understanding between tourists from different countries as they trust better their peers than marketers [6,62].In the landscape of the COVID-19 pandemic, the tourism industry has been struggling due to the recession and the postponement of trips to tourist destinations. Residents’ well-being, to some extent, has been impacted directly or indirectly due to the spread of the coronavirus. VR tourism, as a form of leisure activity in daily life, provides an effective coping strategy to enhance residents’ levels of well-being and destination recovery. In such a context, an investigation was conducted among residents who used QJK, the largest tourism-related VR website in China, during the pandemic. The aim of this paper is to provide empirical evidence to prove the effectiveness of VR tourism in promoting residents’ subjective well-being. Hence, an integrated model with the constructs of PCP attributes, presence, perceived value, satisfaction, and subjective well-being was proposed and tested. The results indicate that the PCP attributes of VR tourism have a positive effect on presence during VR experience, which positively influences perceived value. The value of VR tourism as perceived by VR tourists results in their satisfaction. Eventually, residents’ subjective well-being is improved due to their satisfaction with VR tourism. Based on our findings, suggestions for policy makers, residents and tourism operators are offered as follows:Policy makers should make constructive use of leisure activities associated with a high level of residents’ well-being such as VR tourism while they are striving for economic development and social stability. The local government may collect applicable information and data via big data about VR tourism to make the city more livable, workable and sustainable. In particular, VR tourism can facilitate the disabled with access to destinations in VR, which will, to some degree, contribute to realizing government’s goal for social equity. Meanwhile, we recommend that local residents accept and enjoy virtual travel as VR tourism has great potential to improve their well-being, provides all sorts of travel related information to help them pursue their interests, saves time and money, and allows them to connect with friends while traveling in VR. For destination suppliers, VR tourism is capable of retaining the demands of future tourists and thereby provides practical solutions for destination recovery after the pandemic because VR tourism is associated with real visitation and intention to travel [6,12]. Therefore, it is suggested that destination suppliers cooperate with VR tourism developers. VR tourists favor the VR tourism products and services that present them with a high degree of presence. Thus, developers should highlight core and pivotal attributes when they design VR projects. Subsequently, VR tourists could achieve intensive presence and perceived positive values that will result in their satisfaction with the destination and VR tourism. All in all, VR tourism is supposed to be applied in multiple sectors for various purposes.Writing—original draft preparation, review and editing Y.L.; Methodology and conceptualization H.S.; Data collecting R.G. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Informed consent has been obtained from all subjects involved in this study to publish this paper.The dataset used in this research are available upon request from the corresponding author. The data are not publicly available due to restrictions i.e., privacy or ethical.There are no conflicts of interest to declare.Measurement Items.B1. The website link of QJK: http://www.quanjingke.com/.The QR (quick response) code of QJK’s mobile device Apps.Conceptual framework for PCP (pivotal, core, peripheral) model of VR (Virtual Reality) tourism attribute.Proposed conceptual model Notes 1: PEA = peripheral attribute; CA = core attribute; PIA = pivotal attribute; PRE = presence; FV = functional value; EV = emotional value; SAT = satisfaction; SWB = subjective well-being.Structural model results. Note 1: *** p < 0.001. Note 2: Values not in parentheses are standardized parameter estimate; values in parentheses are t values. Note 3: PEA = peripheral attribute; CA = core attribute; PIA = pivotal attribute; PRE = presence; FV = functional value; EV = emotional value; SAT = satisfaction; SWB = subjective well-being. Note 4: S-B χ2 (df) = 886.124; Normed S-B χ2 = 1.831; CFI = 0.964; NFI = 0.923; NNFI = 0.960; RMSEA = 0.041.Respondents’ demographic characteristics (n = 490).a USD 1 is equivalent to CNY 6.55.Results of measurement model.*: Highest correlation between pairs of construct; The values of AVE highlighted in shade are along the diagonal. Squared correlations among latent constructs are above the diagonal. Correlations among latent constructs are within parentheses. Standard errors among latent constructs are below the diagonal. Mardia’s normalized coefficient: 66.032. All standardized factor loadings are significant at p < 0.001.Standardized parameter estimates of structural model.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01020.txt

@@ -0,0 +1 @@
+Introduction. Digital impressions in implant dentistry rely on many variables, and their accuracy, particularly in complete edentulous patients, is not well understood. Aim. The purpose of this literature review was to determine which factors may influence the accuracy of digital impressions in implant dentistry. Emphasized attention was given to the design of the intra-oral scan body (ISB) and scanning techniques. Materials and methods. A Medline, PubMed and EBSCO Host databases search, complemented by a hand search, was performed in order to select relevant reports regarding the appliance of digital impressions in implant dentistry. The search subject included but was not limited to accuracy of digital impressions in implant dentistry, digital scanning techniques, the design and material of the ISBs, and the depth and angulation of the implant. The related titles and abstracts were screened, and the remaining articles that fulfilled the inclusion criteria were selected for full-text readings. Results. The literature search conducted for this review initially resulted in 108 articles, among which only 21 articles fulfilled the criteria for inclusion. Studies were evaluated according to five subjects: accuracy of digital impressions in implant dentistry; the design and material of the intra-oral scan bodies; scanning technique; the influence of implants depth/angulations on the digital impression and accuracy of different intra-oral scanner devices. Conclusions. The accuracy of digital impressions in implant dentistry depends on several aspects. The depth/angulation of the implant, the experience of the operator, the intra-oral scanner used, and environmental conditions may influence the accuracy of digital impressions in implant dentistry. However, it seems that ISBs’ design and material, as well as scanning technique, have a major impact on the trueness and precision of digital impressions in implant dentistry. Future research is suggested for the better understanding of this subject, focusing on the optimization of the ISB design and scanning protocols.Digital devices have had a widespread use in dental practice in the last few decades. CAD/CAM technology made it possible to fabricate implant-supported restorations through a digital workflow. Digital impressions transfer the intra-oral situation to a virtual model and represent the first step of the digital workflow. The accuracy of this procedure may determine the success of the treatment, since it is a crucial step to transfer the implant position correctly. If it is performed poorly, it can lead to a misfit of the final prosthesis, which may result at long last in mechanical and biological complications. Digital impressions can accelerate the data-capturing process and eliminate most of the drawbacks usually found with conventional impressions, thereby decreasing patient discomfort while improving the predictability of prosthesis design and manufacturing procedures [1,2].The acquisition of a digital impression is a very user-friendly procedure that subserves the daily clinical practice. However, behind the simplicity of this procedure there is a rather complex working mechanism [3,4]. The intra-oral scanner workflow starts by emitting a light beam (laser or structured light) towards the object to be digitized. When it reaches the object’s surface, the light beam suffers a deformation, and this optic effect is captured by two or more cameras on the intra-oral scanner (IOS) devices’ tip. Then, a processing software is used to calculate the 3D coordinates (x,y,z), and creates point clouds and meshes [3,4,5]. The registration and subsequent stitching of these point clouds and meshes allows the tridimensional reconstruction of the scanned object, creating a reliable model [3,4,5,6].When choosing an IOS device it is important not only to consider its operational features—such as the size of the intra-oral tip, the image acquisition’s speed or the ease of manipulation—but also its accuracy. Therefore, trueness and precision must be closely considered [3,4,5,6].Trueness consists of the ability of a measurement to coincide with the real value being evaluated [5,6]. IOS’s trueness can be evaluated by superimposing a digital impression of a scanned object with a reference model of the same object, obtained by an industrial reference scanner (such as a coordinate measuring machine or an industrial optical scanner with accuracy <5 μm). Models superimposition is evaluated using reverse-engineering software in order to determine deviations mathematically [7,8].Precision is defined as the ability of consistently taking the same measurement value. An IOS should present high trueness and precision [5,6]. IOS’s precision can be evaluated by superimposing different scans of the same object performed with the same IOS device [7,8].Multiple in vitro studies have proven that IOSs are an important and reliable tool to capture high-quality impressions, that can be used to fabricate simple (onlays, inlays or single crowns) to complex (fixed partial prosthesis) prostheses in dentate patients [9,10,11,12,13,14].However, due to the digital revolution in prosthodontics over the last few years, the speed of scientific papers’ publication struggles to keep up with the industry development [15,16,17,18,19], there being, to date, a very limited number of studies investigating the accuracy of digital impressions in implant dentistry [16,18]. Using this type of technology on edentulous patients is a complex procedure [16,18].To capture the correct implant position with a digital impression it is necessary to use a specific transfer post called an intra-oral scan body (ISB) [2]. Edentulous areas can be difficult to read and mathematically interpret for IOSs, due to the lack of distinguished anatomic references, which is why having a reliable ISB design is so important to improving the accuracy of implant digital casts [16,17,18,19].There are many factors that might compromise the performance of an IOS, when reading an implant cast, and decrease its accuracy. Regarding the equipment, aspects such as the scanning technology, the state of the device, and the temperature and illumination of the room and of the reading area may affect the accuracy of the IOS readings. The operator’s skills and experience as well as the scanning technique and sequence should also be considered as accuracy-influencing factors. In vivo, patient’s movements, limited mouth opening, and an oversized tongue may render difficult the scanning procedure [20,21,22]. In vitro, the design and material of the cast, and the design of the scan bodies as well as its light reflection properties, can affect the precision of the digital impressions. Several authors support the accuracy of this type of technology for the rehabilitation of single implants. Mangano and Veronesi conducted a randomized controlled trial comparing digital and analog workflows when restoring single implants. Both techniques showed high success rates (92%) and only 8% incidence of complications. Complications were related to biological problems (peri-implant mucositis) in patients with poor oral hygiene compliance [23]. Additionally, Ender et al. conducted a clinical trial in order to compare the precision of conventional and digital impressions in vivo, obtaining results that support the use of digital impressions in implant dentistry, comparable to those achieved for conventional techniques [21].However, the extension of the edentulous space is one of the major obstacles when using digital workflow in implant dentistry. The lack of fixed anatomical reference points, such as teeth, leads to a superimposition of images by using the first image obtained as reference and stitching the following images to the previous ones. Each individual stitch represents a possibility for incurring an error, decreasing the accuracy of the digital impressions [23]. This misalignment error has an even higher impact when more than six implants are placed in the same dental arch [4,23,24,25,26,27,28,29]. Gimenez et al. contended that scanning larger edentulous areas significantly affected both linear and angular measurements, which can be imputed to the accumulative error of the stitching process [23]. It remains unclear in the literature from which exact number of placed implants is the decrease of accuracy clinically significant. The clinically acceptable degree of inaccuracy has been diversely discussed by many authors. Klineberg and Murray considered discrepancies of up to 30 μm at the implant–abutment interface as acceptable, and Jemt proposed a limit of 150 μm to prevent long-term complications. The misalignment error increases with the distance scanned, and consequently full arches represent a bigger challenge for IOSs [30,31]. The implant depth should also be considered because it is directly related to the scan body visibility, which can influence accuracy measurements. When the scan body is fully visible, determining the implant position is less prone to errors, meaning the deeper the implant is placed, the longer the scan body should be [23,32,33]. The literature is not in agreement on the potential influence of implant angulation [23,32,33], although most recent evidence reveals that the angulated position of the implants does not decrease the accuracy of implant digital casts [23].As such, many factors may have an effect on the outcome and accuracy of digital impressions; further development of the scanning devices, scanning protocols, and imaging techniques is necessary to enhance the precision of the optical acquisition of implant scan bodies. Additionally, the scan body design requires further improvement in order to enhance the accuracy of digital impressions.As digital technology becomes empowered in implant dentistry, many commercial brands developed ISBs with different designs and geometries. Generally, ISBs are composed of three distinct areas: the scan region (corresponding to the upper portion), the body (corresponding to the middle portion) and the base (corresponding to the most apical portion that connects to the implant) (Figure 1). A deeply tapered connection or mismatch in materials between the base and the implant may influence the displacement of the ISB when tightened into place [34].The scan region contains one or more scanning areas, which may have different shapes, in order to improve the accuracy of the digital impression. By incorporating an asymmetrical shape on the scan region, the surface recognition by the CAD software becomes more simple.The majority of ISBs commercially available are made of one of two different materials, polyetheretherketone (PEEK) and titanium (Figure 1), but the body of the scan body may also contain aluminum alloy and various resins. It is important to consider the machinability of these materials and the manufacturing tolerances to improve the accuracy of ISBs. The height of commercially available ISBs ranges from 3 to 17 mm [34].Usually, dull, smooth and opaque surfaces are easier to capture in a digital intra-oral impression than shiny, rough or translucent ones. Intraorally, it becomes very challenging due to the surface’s reflection created by saliva. Recent studies have indicated that deep, undercut, steep, sharp, angled, or crowded surfaces are also more difficult to scan, leading to less accurate point clouds. Gimenez et al. concluded that gingivally placed implants presented less scan deviation than subgingivally placed implants, regardless of the angle deviation (p = 0.757) [23]. It may be necessary and advantageous to create ISBs with specific characteristics for intra-oral situations. A narrow scan body, for example, may be more effective in situations with limited interproximal space, and a shorter scan body may be easier to capture in patients with complete edentulism or limited mouth opening [34].However, in cases of edentulous jaws rehabilitation, the challenge of obtaining an accurate digital impression remains. It is necessary to create an ISB design that can be easily identified by the IOS, is accessible to manipulate by the operator, and is comfortable for the patient.Therefore, the aim of this literature review consists of understanding the state of the art of digital impressions on implant dentistry, and understating which factors may contribute to decreasing or enhancing its accuracy, in order to attempt to provide the dental clinician with evidence-based guidelines when resorting to these impression techniques.A MedLine, PubMed and EBSCO Host databases search was performed by two calibrated investigators (S.M. and M.H.-C.) in order to select relevant reports regarding the appliance of digital impressions in implant dentistry, using the (MeSH) keywords relevant for the main question. The guiding question of this review is “Which factors may influence the accuracy of digital impressions on implant dentistry?” Attending to the scope of influencing factors on the main question it was not possible to formulate a PICO strategy. Since the subject of digital impressions in implant dentistry is not a very old one in the literature, no time frame was applied, analyzing all studies published until May 2020. The literature search was limited to articles published in the English language. The analysis was performed according to the guidelines and references of an integrative review. Additionally, a hand search of four journals was conducted: Journal of Prostethic Dentistry (2014–present), Journal of Prosthodontics (2016–present), Clinical Implant Dentistry and Related Research (2016–present) and Clinical Oral Implants Research (2015–present).The database search included but was not limited to the accuracy of digital impressions in implant dentistry, digital scanning techniques, the design and material of the ISBs and the depth and angulation of the implants. The employed search terms were as follows: (implant digital impressions) AND (accuracy) AND (intra oral scan body) AND (scan body design) AND ((digital scanning technique) OR (digital scanning protocol)) AND (implant depth) AND (implant angulation) AND (intra oral scanner)). However, no studies evaluating all of these features in relation to digital impressions in implant dentistry were identified.Therefore, the database search was expanded in order to include any articles regarding digital impressions in implant dentistry, with fixed partial dentures (FPD) or full-arch prosthesis, modifying the search terms and strategy. The used search terms were then as follows: (implant digital impressions) AND (accuracy) OR (intra oral scan body) OR (scan body design) OR ((digital scanning technique) OR (digital scanning protocol)) OR (implant depth) OR (implant angulation) OR (intra oral scanner)).Duplicated results from different databases were not considered.Inclusion criteria comprised studies at all levels of evidence, excluding expert opinion, such as experimental clinical studies, in vitro and in vivo studies. All articles evaluating at least one of the following subjects were included: digital impressions in implant dentistry, digital scanning techniques, design and/or material of the ISBs, depth and/or angulation of the implants and performance of different IOS devices.Exclusion criteria comprised multiple publications based on the same population and with wrong study designs. Experimental clinical studies, and in vitro and in vivo studies, that analyzed the accuracy of digital impressions only on teeth, not considering impressions on implants, were excluded.Out of 108 results, the articles were initially analyzed considering their title and abstract, excluding 79 articles because implant impressions were not considered. A full-text analysis of the 29 remaining articles were performed, excluding 8 articles, due to a lack of information regarding the obtaining method of the reference models (n = 3), the conventional impression materials used (n = 4) and the digital files superimposition technique (n = 1). At last, 21 articles were selected for this review (Figure 2). From a journals manual search, 9 articles were selected (2 from Journal of Prosthetic Dentistry, 2 from Journal of Prosthodontics and 5 from Clinical Oral Implants Research). However, they were all duplicated from the database search and consequently not considered. All studies from the above-mentioned search scheme were analyzed by two calibrated reviewers (S.M., M.H.C.), and screened with the inclusion/exclusion criteria.It was not possible to perform the statistical analysis due to the report variability and the limited number of identified studies.From each study, the following data were extracted:Study design—randomized/nonrandomized controlled study, experimental study;Study setting—in vivo/in vitro;Type of impressions—digital/conventional;Type of arch—single-unit case, partially edentulous, completely edentulous;Type and number of implants placed;Implant depth and angulation;Type and design of the ISBs;IOS used;Scanning technique;Outcomes.A total of 21 articles were reviewed in the present study: 18 in vitro studies, 1 randomized in vitro study and 2 comparative clinical studies (Table 1 and Table 2). In total, 20 articles evaluated the accuracy of digital impressions in implant dentistry using ISBs; 5 considered the accuracy of digital impressions regarding the design and material of the ISBs; 6 focused on the accuracy of the scanning technique; 6 compared the accuracy of different IOS devices; and 8 assessed the accuracy of digital impressions concerning the depth/angulation of the implant (Table 3).In total, 13 studies refer to completely edentulous arches with two implants (1 study), three implants (1 study), four implants (3 studies), five implants (3 studies) and six implants (5 studies). Six studies examined partially edentulous arches with one implant (one study), two implants (two studies), three implants (two studies) and with two and five implants, respectively (one study). One study evaluated partially and completely edentulous arches, with three and six implants, respectively. One study examined completely edentulous arches with no implants placed, in order to determine which IOS device and scanning strategy presented higher accuracy.All studies reviewed are summarized in Table 4, Table 5 and Table 6.Twenty studies examined the accuracy of digital impressions in implant dentistry. The digital impressions accuracy outcome was evaluated by measuring linear and angular deviations or tridimensional surface deviations between reference models and test models, or by examining the fit of frameworks on test models that were fabricated on master models.To assess linear and angular distances between implants, master models and test models were measured with coordinate measuring machines (CMM) [23,35,41,43,49]. Virtual measurements of implant distances and angulations were calculated after performing optical impressions with multiple high-precision reference scanners, such as IScan D101 (Imetric 3D Gmbh, Courgenay, Switzerland) [36], IScan D103i (Imetric, Courgenay, Switzerland) [23], IScan D104i, (Imetric, Courgenay, Switzerland) [36,42], Lava Scan ST (3M ESPE, Seefeld, Germany) [8,44], D250 (3Shape, Copenhagen, Denmark) [30,38], D800 (3Shape, Copenhagen, Denmark) [40], E3 scanner (3Shape Copenhagen, Denmark) [43], Activity 880 (Smart Optics, Germany) [45], ScanRider (Italy) [24], ATOS Compact Scan 5M (GOM GmbH, Germany) [39], COMET L3D (Carl Zeiss Optotechnik GmbH) [46] and ATOS So4 II (GOM GmbH, Germany) [47]. The STL digital values were loaded into reverse-engineering software such as Rapidform (Rapidform, INUS Technology Inc, Seoul, Korea) [19,30,35,38,40], Geomagic Qualify 12.0 (Geomagic, Morrisville, NC, USA) [8,24,31,36,42,43,44,45], Mimics (Materialise, Leuven, Belgium) [23], Rhinoceros 5.0 (Robert McNeel & Associates, Seattle, WA, USA) [41], Gom Inspect Professional (GOM GmbH, Germany) [39] and ATOS Professional Software (V7.5 SR2, GOM GmbH, Braunschweig, Germany) [48], and were superimposed with their respective STL master models in order to evaluate tridimensional deviations.Attending to accuracy, it was concluded that it is viable to use a three-dimensional acquisition technology as an alternative to conventional impression procedures [23,35,42]. Giménez-González et al. registered mean linear and angular deviations for the TrueDefinition IOS from CMM measurements, from 5.38 ± 12.61 μm to –26.97 ± 50.56 μm and from 0.16° ± 0.04° to –0.43° ± 0.1°, respectively [23]. Vandweghe et al. evaluated the accuracy of four different IOS when applied for implant impressions in edentulous jaws, and concluded that the mean trueness was 0.112 ± 0,025 mm for Lava COS, 0.035 ± 0.012 mm for 3M TrueDefinition, 0.028 ± 0.007 mm for Trios and 0.061 ± 0.023 mm for Cerec Omnicam. The mean precision was 0.066 ± 0.025 mm for Lava COS, 0.030 ± 0.011 mm for 3M TrueDefinition, 0.033 ± 0.012 mm for Trios and 0.059 ± 0.024 mm for Cerec Omnicam [42]. Ciocca et al. revealed that the mean 3D position error of the digital impression was 0.041 ± 0.023 mm to 0.082 ± 0.030 mm, which is in agreement with former studies and indicates a clinically acceptable level of accuracy [43]. Menini et al. went further and compared the accuracy of conventional impression techniques with digital impressions on multiple implants, by analyzing the passive fit of a full-arch implant-supported prostheses. The Sheffield test revealed a mean gap of 0.022 ± 0.023 mm for the conventional impression and 0.015 ± 0.011 mm for the digital impression, suggesting a better accuracy of digital impressions compared to conventional ones [41]. In fact, Ribeiro et al. also concluded that for a model with four axial implants, the deviations in the digital impressions were smaller than those related to the conventional techniques [44]. However, scanning accuracy has shown to be decreased when digitizing a fully edentulous patient, compared to scanning an area of more limited extent [25]. Adriessen et al. concluded that based on the intra-oral scans obtained, the distance and angulation errors were too relevant to the manufacturers’ well-fitting frameworks for implants in edentulous mandibles. Out of 21 intra-oral scans, 5 scans presented an intra-implant distance error higher than 100 μm, 3 scans demonstrated intra-implant angulation error higher than 0.4°, with only 1 scan performing both intra-implant angulation and intra-implant distance error acceptably (lower than 0.4° and 100 μm, respectively). The lack of anatomic references for scanning on edentulous jaws appears to be the main reason for the unreliable scans [8].Out of the 21 studies reviewed, only 19 used ISBs on their protocol.Despite 16 studies mentioning the shape and dimensions of the ISBs, most studies did not examined the influence of the design of the ISB on the accuracy of digital impressions [19,23,25,31,32,36,39,40,41,42,43,45,46,47,48,49]. Only four studies focused on this subject [40,47,48,49].Three studies used prototypes [23,35,46], while 13 studies used commercially available ISBs.All authors used screw-retained ISBs that would correspond to impression copings in conventional impressions.Regarding dimensions, the ISBs examined had a height range between 8 and 15 mm and a diameter range varying from 4 to 5 mm.Among all ISBs, the most commonly used design was a cylindrical shape [19,23,25,31,36,40,41,43,47,48]. Another variation of this ISB design were examined as well, having a cylindrical design with a partially beveled upper part [32,39,40,42,45,46,47,48,49].Other designs were examined, such as flat cylinder with ball top, rectangular, cylinder with triangular region, tapered flat cylinder [47], uneven shape with bulges and indentations (cylindrical in the cervical area and slightly oval in the coronal area) and cylindrical shape with one retraction and a slightly enlarged coronal diameter [48].All four studies that directly evaluated the influence of the ISB design on the accuracy of digital impressions agreed that the precision of scanning is dependent on the ISB surface geometry and design [31,47,48,49]. Mizumoto et al. examined five different ISB designs (Table 5). Resorting to a structured blue light industrial scanner, ISBs were scanned on an edentulous maxillary model with four dental implant analogs. Five scans of the model were performed with an IOS, applying different scanning techniques. The scans were then superimposed on the master reference model. The distance deviation and angular deviation of the ISBs was calculated. Statistical analysis was performed by using a two-factor ANOVA to analyze the influence of ISB and scanning technique on the trueness and scan time, with subsequent Tukey honestly significant difference or Bonferroni-corrected Student t-tests. The ISB design had a significant effect independently (p = 0.031). A statistically significant interaction was found between the effects of the ISB design on angular deviation (p <0.001) [47].Additionally, the distance between ISBs seems to negatively impact the accuracy of digital impressions [31], and scanning time may be influenced by the ISB design [47]. Flügge et al. developed an in vitro study using two models with a different number and distribution of ISBs, produced from conventional implant impressions. These models were scanned with three different IOSs and a dental lab scanner, performing ten scans for each model and IOS. The distance and angulation between the respective ISBs were measured. The comparison of results with analysis of variance allowed us to conclude that the distance of a single tooth space and a jaw-traversing distance between ISBs revealed significantly different results for distance and angle measurements between the scanning systems (p < 0.05) [31]. The same author concluded with this study that the precision of ISB scanning was not significantly influenced by the detachment and repositioning of the ISB [31].Regardless, 14 studies mentioned the material of the ISBs, but most studies did not examine its influence on the accuracy of digital impressions [19,23,25,32,35,36,39,41,42,44,46,47,48,49]. Only one study was focused on this subject [46].The most commonly used ISB material was PEEK [19,23,25,36,38,40,46,47,48], but other materials were also examined, such as ceramics [35], metal (inox) [39], a non-specified polymer [32,39], titanium alloy [44,46], hybrids containing both titanium and PEEK [46,47,49], or PEEK and metal [47,48].When comparing which ISB material presented a better accuracy performance, a study by Arcuri et al. concluded that PEEK was the most accurate material, followed by titanium and hybrid PEEK with titanium, respectively (Table 6) [46]. To come to this conclusion, linear and angular deviations were assessed. Considering the angular deviations, the material of the ISBs significantly influenced the expected value (p = 0.0232). In multivariate analysis, when the absolute values of the linear discrepancies were summed up to obtain a global measure of the linear absolute error and were considered as the response variable, a significant impact of the material of the ISBs was identified (p < 0.0001) [46].Out of the 21 studies reviewed, only 15 studies mentioned the scanning protocol used. Four studies stated they followed the IOS manufacturer’s protocol, without specifying the procedure [31,32,42,44]. Eleven studies reported in detail the scanning technique applied [19,23,25,37,39,41,43,45,46,47,48]. However, only six studies evaluated the influence of the scanning technique on the accuracy of digital impressions. All authors concluded that the scanning protocol may influence the performance of the IOS device, and subsequently the accuracy of the digital impression [19,23,36,37,47,48].When comparing two scanning strategies, Motel et al. achieved a significantly higher accuracy overall (p = 0.031) using a one-step scanning strategy with integrated ISBs (scanning both model and ISBs placed at one single time) than with using a two-step technique, which assumed scanning the model two times: first, without ISBs, and then with the ISBs positioned on the model [48].Mizumoto et al. found a statistically significant interaction between the effects of the scan body and technique on the angular deviation (p < 0.001) [47].Studies that compared digital impressions with conventional impressions claim that the splinted impression technique is more accurate compared with the non-splinted technique in edentulous patients, when using a conventional impression protocol. Papaspyridakos et al. compared splinted and non-splinted, open-tray techniques to fabricate casts that were superimposed onto reference models by optical scanning acquisition of the xyz coordinates of the implant positions for each individual cast that was realized. The paired t-test and Wilcoxon’s signed ranks test were used to compare the 3D discrepancies within and between splinted and non-splinted techniques, respectively. Significant difference was found at the x- and y-axis, and the 3D parameters, between the splinted group and non-splinted groups (p < 0.05), but not in the vertical z-axis (p > 0.05). Within subject, global 3D discrepancies between splinted impressions and non-splinted impressions were significantly different (p < 0.05), confirmed by the clinical observation of the fitting. The splinted technique generated more accurate master casts than the non-splinted technique, but in implant dentistry, digital impressions also presented accurate results [36].When using a digital workflow and following the IOS scan protocol, a study by Giménez-González et al. revealed that the distance and angular deviations augment throughout the arch, meaning that the first scanned quadrant will always achieve better accuracy than the second quadrant. The scanning technique should take this fact into consideration [23].Mizumoto et al. revealed that scanning techniques with different surface modifications, such as placing glass fiduciary markers on the edentulous ridge, marking the ridge and palate with pressure-indicating paste or splinting the ISBs with floss, resulted in similar distance deviations as the scanning technique without any modifications (p = 0.076) [47].Out of 21 studies, 11 studies analyzed axial implants; 8 studies compared angulated implants with axial implants and 2 studies did not examine implants in their protocol.Arcuri et al. evaluated the influence of the ISB position, and consequently the implant depth/angulation, on digital impressions via scanning and edentulous maxillary models with six implants placed with different depths and angulations (Table 4). The 45 scans obtained were superimposed using a best fit algorithm to a reference model, obtained with an extraoral optical scanner. Considering the angular deviations, the position of the ISBs significantly influenced the expected value (p < 0.0001). In multivariate analysis, when absolute values of the linear discrepancies were summed up to obtain a global measure of the linear absolute error and considered as the response variable, a significant impact of the position of the ISBs was identified (p < 0.0009). Therefore, it was suggested that implants’ angulation may decrease the accuracy of digital impressions [46].Gedrimieni et al. also claimed that the angulation between implants affects the ISBs positioning and, depending on its design, may interfere with the accuracy of digital impressions [45]. However, the majority of the studies supported the theory that the accuracy of digital impressions was not influenced by different implant angulations for completely edentulous patients [23,32,39,42,44,48].Most studies did not evaluate the impact of the implants’ vertical positioning on the accuracy of the digital impressions. Arcuri et al. also examined the implants’ depth influence using digital impressions, positioning the implants equigingivally (3 and 6 mm subgingivally). It was concluded that implant depth did not affect the final accuracy of digital impressions [46].However, another study by Giménez-González et al., wherein implants were placed equigingivally (2 and 4 mm subgingivally), claims that the amount of visible ISB affects the accuracy of digital impressions, and so the depth of the implant should be taken into consideration when choosing the ISB design [23].Several devices for intra-oral optical scanning were analyzed in the reviewed studies.Nine different IOS devices were used, as follows: Comet VZ250 (Steinbichler Optotechnik GmbH, Neubeuern, Germany) [35], Cerec Bluecam (Sirona, Bensheim, Germany) [19,37], Cerec Omnicam (Sirona, Bensheim, Germany) [25,39,42], iTero (Cadent, San Jose, CA, USA) [8,19,31,37,38], LavaCOS (3M Espe, St. Paul, MN, USA) [19,37,42], ZFX Intrascan (Zimmer, Dachau, Germany) [37], Trios (3Shape, Copenhagen, Denmark) [22,31,38,42,45,46,47,48], TrueDefinition (3M Espe, St. Paul, MN, USA) [23,24,31,39,41,42,43,44] and CS 3600 (Carestream, Rochester, NY, USA) [25].Six studies compared the accuracy of some of the above-mentioned IOS devices. The results of the different studies are not in consensus.In a study by van der Meer et al., when compared with iTero and Cerec Bluecam, LavaCOS resulted in the most accurate of all three scanners tested when considering mean distance errors in completely edentulous patients [19].Contradictorily, high deviations for trueness and precision for Lava COS (p = 0.169) were demonstrated by a study of Vandweghe et al., which compared this IOS device with Trios (p < 0.001) and True Definition (p < 0.001). These two IOS demonstrated an acceptable accuracy for large-span implant-supported reconstructions [42]. Cerec Omnicam was less accurate compared to True Definition (p < 0.001) and Trios (p < 0.001), but no difference was found with Lava COS (p = 0.169). The mean trueness was 0.112 ± 0.025 mm for Lava COS, 0.035 ± 0.012 mm for True Definition, 0.028 ± 0.007 mm for Trios and 0.061 ± 0.023 mm for Cerec Omnicam. The mean precision was 0.066 ± 0.025 mm for Lava COS, 0.030 ± 0.011 mm for True Definition, 0.033 ± 0.012 mm for Trios and 0.059 ± 0.024 mm for Cerec Omnicam [42].In a study by Imburgia et al., CS3600 and Trios performed better, showing improved accuracy results in a comparative study with four different IOS devices (CS3600, Trios, Cerec Omnicam and True Definition), on partially edentulous and fully edentulous models. Considering the same type of model used in the previous study (fully edentulous model), the mean trueness was 106.4 ± 23.1 μm for True Definition, 67.2 ± 6.9 μm for Trios, and 66.4 ± 3.9 μm for Cerec Omnicam. CS 3600 had the best mean trueness (60.6 ± 11.7 μm). In the fully edentulous model, the mean precision was 75.3 ± 43.8 μm for True Definition, 31.5 ± 9.8 μm for Trios, 57.2 ± 9.1 μm for Cerec Omnicam and 65.5 ± 16.7 μm for CS 3600 [25]. Amin et al. concluded that True Definition had significantly less 3D deviations when compared with the Cerec Omnicam [39].Flügge et al. supported the higher precision of the IOS devices Trios and True Definition in comparison with iTero [30]. Amin et al. concluded that True Definition had significantly less 3D deviations when compared with the Cerec Omnicam [39].When scanning edentulous jaw models with no implants placed, Patzelt et al. did not find statistically significant differences between Cerec Bluecam, LavaCOS, iTero and ZFX Intrascan, and the use of four IOSs was feasible. Nevertheless, this study concluded that IOS devices need some improvement before the recommendation use of these scanners for the digitization of edentulous jaws without any references, such as implants, in vivo [37].This literature review advocated to analyze the state of the art of the accuracy of digital impressions in implant dentistry, attending to the multiple variables that may have an impact on it, such as the geometry and material of ISBs, the scanning protocol, the implants depth and/or angulation, and the IOS device used.It is important to highlight that limited high-quality evidence is available on this matter, and the interpretation of the results of this review should take into consideration the study settings presented. Therefore, the limitations of this review consist mostly of the inconsistency of study designs and protocols among the selected articles, not allowing a direct comparison of the obtained results. The risk of bias of the reviewed studies was also not assessed, which can be considered as another limitation of this literature review. However, within the mentioned limitations, the present review summarizes the state of the art of digital impressions in implant dentistry, providing the clinician some practical information concerning the factors that may influence the accuracy of this digital workflow.In digital dentistry, the frequency of scientific publications and evidence-based articles is significantly lower than the frequency of IOS’s hardware and software updating, which creates a temporal gap and hinders the establishment of good practice guidelines.The current evidence available on the accuracy of digital impressions in implant dentistry is mostly presented in experimental studies. Consequently, this review was mainly based on experimental studies with a low scientific evidence level. As the majority of studies on digital impressions in implant dentistry are in vitro, it is important for the clinician to carefully analyze their informative value. Only two of the reviewed studies examined the accuracy of digital impressions in vivo [8,45]. However, in vivo studies rarely present the numerical values for the accuracy of implant impressions, which excludes the possibility of a direct comparison with the outcomes of in vitro studies, since in these ones the accuracy is reported with value measurements of the deviation between reference models and test models. Further studies in vivo executed with reliable methods for outcome reporting are required.Regardless of the study setting, the comparison of outcomes resulting from conventional and digital impressions in implant dentistry suggests that digital implant impressions are as accurate as conventional implant impressions, for fixed partial prosthesis. In full-arch rehabilitation cases, the current literature does not yet provide high-quality evidence to support the selection of implant digital impression protocols over conventional techniques. However, it presents very accurate results from the in vitro studies [23,35,41,42,43,44].Several authors have agreed on the viability of using a tridimensional acquisition technology as an alternative to conventional impression procedures [23,35,42]. Menini et al. declared that digital impressions can be used for the fabrication of full-arch implant-supported prostheses, providing a satisfactory passive fit [41]. On the other hand, Andriessen et al. concluded that the distance and angulation errors were too significant to fabricate fitted frameworks on implants in edentulous mandibles, probably due to the lack of landmarks for scanning [8]. There are some factors that may influence the accuracy of digital impressions and that still require further investigation. For instance, many authors support the accuracy of digital impressions for the rehabilitation of single implants; however, the extension of the edentulous space and the increased number of implants are major obstacles to a full digital workflow in implant dentistry. Giménez-González et al. reported that the higher the edentulous span scanned, the more significantly affected both linear and angular measurements become, which can be imputed to the accumulative error of the images-stitching process [23]. However, it remains unclear in the literature with which exact number of placed implants does the decrease in accuracy become clinically significant [4,25,26,27,28,29].Multiple devices for intra-oral optical scanning were analyzed in the reviewed studies, using different types of technologies. Lava COS by 3M ESPE (Seefeld, Germany) captures data in a video sequence using the principle of active wavefront sampling with structured light projection. It requires the light powder dusting of the scanning areas, in order to minimize light reflection and locate reference points for the IOS [42]. True Definition, also by 3M ESPE (Seefeld, Germany), an upgraded version of the Lava COS, is a structured light scanner which uses a pulsating visible blue light, and also works under the principle of active wavefront sampling, generating a 3D video technology. Light powder dusting with titanium oxide powder is still required [25,42]. Cerec Omnicam by Sirona (Long Island City, NY, USA) is a structured light scanner, based on the principle of confocal microscopy and active optical triangulation. This IOS does not require the powder dusting of the scanning area and also provides color information [25,42]. The Trios 3 IOS by 3Shape (Copenhagen, Denmark) is a structured light scanner, which works under the concept of confocal microscopy and ultrafast optical scanning, capturing continuously 2D images from different positions, to create a tridimensional model. Powder dusting of the scanning area is not necessary and also provides color information [25,42]. CS 3600 by Carestream (Rochester, NY, USA) is a structured LED light scanner, that works through active speed tridimensional video. It does not require powder on the scanning surface and provides high-quality color images [25].Regarding the accuracy of different IOSs, studies by Vandweghe et al. and Imburgia et al. were consistent with the results obtained when testing the precision of True Definition (0.030 ± 0. 011 mm and 0.075 ± 0.044 mm, respectively), Trios (0.033 ± 0.012 mm and 0.032 ± 0.010 mm, respectively) and Cerec Omnicam (0.059 ± 0.024 mm and 0.057 ± 0.009 mm, respectively). When testing precision, only Cerec Omnicam obtained consistent results in both studies (0.061 ± 0.0 2 3 mm and 0.066 ± 0.004 mm, respectively). Nevertheless, the precision results obtained for True Definition (0.035 ± 0.012 mm and 0.106 ± 0.023 mm, respectively) and Trios (0.028 ± 0.007 mm and 0.067 ± 0.007 mm for Trios, respectively) did not present a significant difference, and are both clinically acceptable. Therefore, the choice of which IOS to use may not have a significant impact on the accuracy of the digital impressions, since all the latest devices present clinically acceptable results. The scanning protocol may influence the accuracy of digital impressions, and it should be taken into consideration according to the recommendations of different IOSs [25,42].Giménez-González et al. relate that the scanning protocol followed can affect the accuracy of digital impressions. It was concluded that the distance and angular deviations were significantly increased throughout the arch from the starting point, suggesting that in partial restorations the scanning protocol should start at the area of the restoration in order to achieve the most accurate result. Additionally, the design of the ISB should be attended because the amount of visible ISB can influence the accuracy of the digital impressions. According to Giménez-González et al., when the implants are deeply placed, longer ISBs designs should be preferred, which means that the depth of the implant can also affect the accuracy of the digital impressions [23]. Although some authors have analyzed the influence of ISB design in the accuracy of digital impressions, it was not conclusive which geometry is more accurate. It appears that having an asymmetrical shape may improve the accuracy results; however, additional studies are necessary to investigate this finding as it applies to ISBs. Additionally, the ideal ISB surface material requires further investigation.Regarding the implant angulation, the majority of the studies supported the theory that the accuracy of digital impressions was not influenced by different implant angulations for completely edentulous patients [23,32,39,42,44,48], but it should be taken into consideration that most of these studies were developed under laboratorial ambience, with optimal environment conditions, with stabilized models and without the patient-related features that may affect the digital impression, such as saliva and head movement. Further in vivo studies are required to assess the influence of implant angulation on the accuracy of digital impressions.Currently, there is a lack of information on these topics, and more studies are required to determine the relationship between ISBs’ features, implant depth/angulation, scanning protocol and digital impression accuracy.Based on the limited evidence available for this review, and considering the limitations mentioned, some preliminary conclusions can be drawn.Evidence suggests that digital impressions are an accurate procedure in implant dentistry.Regardless of the IOS device used, the scanning protocol can influence the accuracy of the digital impressions.Implant angulation seems to have no effect on the accuracy of the digital impressions. On the other hand, implant depth may affect the accuracy of the procedure. However, clinical guidelines cannot be drawn based on the presented data.ISBs are implant position transfer devices that are commercialized in multiple shapes, geometries and materials.The design and material of ISBs may influence the accuracy of digital impressions.Clinical guidelines cannot be drawn based on the current data. Further investigations focusing on the in vivo use of digital impressions in implant dentistry are required. Clinical studies and RCTs on this matter are suggested.Conceptualization, S.M., P.R. and M.H.-C.; methodology, S.M. and M.H.-C.; software, B.F.L.; validation, C.F. and J.V.R.-S.; formal analysis, P.R. and B.R.-C.; investigation, S.M., P.R. and M.H.-C.; resources, S.M. and M.H.-C.; data curation, S.M. and M.H.-C.; writing—original draft preparation, S.M.; writing—review and editing, S.M., J.V.R.-S. and M.H.-C.; visualization, S.M., M.H.-C., P.R., C.F., B.F.L., J.V.R.-S. and B.R.-C.; supervision, M.H.-C. and J.V.R.-S. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.Not applicable. The authors declare no conflict of interest.ISBs (intra-oral scan body ) have three regions: 1—scan region, 2—body and 3—base. (a) ISB Elos Accurate® Scan Body (Elos Medtech AB, Gothenburg, Sweden), made of PEEK. (b) ISB Klockner KL-1 (Klockner Implant System, SOADCO, Andorra), made of titanium.Flow chart presenting the screening of articles related to digital impressions in implant dentistry to be included in this review.Study design of the included articles.Study setting of the included articles.Study main subjects.Summary of the reviewed studies of digital impressions in implant dentistry.Summary of the reviewed studies of digital impressions in implant dentistry (additional columns).Summary of the reviewed studies of digital impressions in implant dentistry (additional columns).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01021.txt

@@ -0,0 +1 @@
+The Mindful Eating Questionnaire is a reliable tool for the assessment of mindful eating behavior among the general population. This study aimed to determine the reliability and validity of The Malay Mindful Eating Questionnaire (MEQ-M) in a sample of overweight and obese adults. This is a cross-sectional survey which involved 144 overweight and obese adults in a selected public university. After linguistic validation of the Malay version of the MEQ, exploratory factor analysis (EFA) with varimax rotation was performed on the scale constructs. The psychometric properties of the MEQ were assessed through Cronbach’s alpha and intraclass correlation coefficient (ICC) analysis. The EFA of the MEQ produced a seven-dimensional model (58.8% of overall variances). The concurrent validity analysis between total MEQ scores and total Mindfulness Attention Awareness Scale (MAAS) scores indicated a weak non-significant correlation (p = 0.679). The internal consistency reliability of the MEQ was reasonable (Cronbach’s α = 0.64). The agreement stability of the MEQ over eight weeks was poor (ICC = 0.10). In conclusion, the psychometric properties of the Malay-translated MEQ are acceptable through construct validity and internal consistency reliability tests. This instrument may be used for assessing mindful eating habits in the Malaysian population, especially among overweight and obese adults.Mindfulness is commonly understood as the ability of being open, accepting, and present in the moment [1]. Mindfulness trainings such as Mindfulness-Based Stress Reduction (MBSR), Mindfulness-Based Cognitive Therapy (MBCT), Acceptance and Commitment Therapy (ACT), and Dialectical Behavioral Therapy (DBT) are often described as interventions that focus to promote non-judgmental and moment-to-moment awareness of the present experience. The role of mindfulness in encouraging behaviors related to healthy body weight maintenance and reducing overweight and obesity has gained attention among scholars [2]. This practice has been beneficial to control food cravings, portion size, body mass index, and body weight [3]. Mindful eating refers to being conscious of physical sensation and emotion while eating or in a food-related environment [1]. It helps in improving one’s sensitivity to the physical cues of hunger, satiety, eating speed, and the food atmosphere [3]. These cues are crucial to self-regulate one’s desire to consume high-calorie foods. Studies have shown that mindful eating helps in reducing negative eating behaviors, sweets consumption, and serving sizes of energy-dense foods [4,5,6].There are many factors associated with eating behaviors, such as physiological (i.e., chronotype), social (i.e., coworker influence), environmental, and psychological (i.e., stress, mood) [7,8,9]. It has been hypothesized that obesity-related eating behaviors are partially associated with the inability to identify and respond to internal cues of hunger and satiety [10,11] (pp. 75–91), [12] (pp. 913–933). This lack of response to internal cues is correlated with increased overeating episodes and a greater risk for weight gain [13,14,15]. In addition, emotional dysregulation has been associated with emotional and stress eating [16,17,18,19]. Studies have shown that compulsive overeating and higher preference for high calorie, fat, sugar, and/or high sodium foods are the result of negative emotions and acute stress [20,21,22]. Emotional eating has been shown to be a strong indicator of obesity and is negatively associated with weight loss and its maintenance [23,24,25]. Furthermore, a restricted diet and increased physical activity can result in physiological discomforts that may impose an added barrier to long-term weight loss [24].There are several tools available in measuring eating behaviors such as Night Eating Syndrome Questionnaire (NEQ), Three-Factor Eating Questionnaire (TFEQ) and Binge Eating Scale (BES). The Mindful Eating Questionnaire (MEQ) is the first scale developed by Framson and colleagues that measures mindful eating [1]. It is a 28-item self-report instrument that consists of five mindful eating domains: awareness, disinhibition, distraction, emotional response, and external cues. This questionnaire has been validated previously among healthy adults aged 18 to 80 years old. The study showed good internal consistency with a reliability of 0.64 for the MEQ score [1]. In addition, each subscale had internal consistency ranging from 0.64 to 0.83. It was also reported that there were modest (0.14) to moderate (0.47) correlation among all subscales with the exception of correlation between external cues and emotional response. Another validation study among overweight and obese pregnant women yielded the same five domains of the MEQ [26]. It was found that the MEQ has poor internal consistency reliability of the summary score (0.56). As for the subscales, only the external cues subscale was not internally consistent with Cronbach’s alpha of 0.31. Its reliability was further supported by test-retest analysis, where the total and subscale scores were ranged between 0.62 to 0.85. To add, positive correlations were also observed between the MEQ subscales and the Mindful Attention Awareness Scale (MAAS) [26]. Another study of the Persian version of MEQ among women seeking weight loss reported satisfactory internal consistency for the total score and the subscales (0.73–0.81) and satisfactory test-retest reliability ranging from 0.73 to 0.91. [27]. Its construct validity analysis resulted in five domains which were similar to the original study. Contrary, the Italian version of the MEQ resulted in a 20-item pool where only two domains emerged; awareness and recognition [28]. Clementi et al. also found that both domains have satisfactory internal and test-retest reliability, and were associated with general mindfulness. Meanwhile, the MAAS is a standard tool used to assess mindfulness in everyday life among the general population [29]. It consists of a 15-item self-reported single-factor scale that focused on the mindfulness construct’s attention awareness component.The prevalence of overweight and obesity had significantly increased between 1976 and 2016 globally, in which half of them (52%) were adults over 18 years of age [30]. In Malaysia, the National Health and Morbidity Survey (NHMS) reported that the prevalence of overweight and obesity among adults increased by 5% from 2011 to 2015 [31,32]. Moreover, the prevalence by age in 2015 showed an increasing trend from 34.8% among 18–29 years to 60.2% among 50–59 years. Obesity is associated with increased risk of many non-communicable diseases (NCDs), including diabetes, cardiovascular disease, depression, some cancers, and respiratory disease [33,34,35,36,37]. Moreover, it negatively impacts bone health, quality of life, and functional capacity [38,39,40]. Consequently, obesity is also associated with expensive health care costs [41,42,43]. One study suggested that obese adults have difficulty in reflecting on the impact of obesity on their social and relational functioning despite having psychological difficulties [44]. Considering the increasing trend in obesity prevalence in Malaysia, a locally validated instrument is essential in conducting research and intervention activities. To the best of our knowledge, no measure has been carried out to assess mindful eating behavior in Malaysia’s overweight and obese adults. Moreover, there is currently no instrument measuring mindful eating (in general) in the Malaysian context using its local language. Several studies have shown that eating mindfully was associated with a lower BMI [45,46] and reduced body weight [47]. Thus, the objective of our present study was to determine the reliability and validity of the MEQ-M in a sample of overweight and obese adults. This is the first study to examine the reliability and validity of the Mindful Eating Questionnaire (MEQ) among this population. We hypothesized that the MEQ would have similar results to a previous study [40] where the questionnaire would be valid and reliable among overweight and obese Malaysian adults. As mindful eating was generally associated with general mindfulness in previous studies, we hypothesized that the MEQ-M total score would be positively correlated with the Mindful Attention Awareness Scale (MAAS).The MEQ contains five subscales: awareness (seven items), distraction (three items), disinhibition (eight items), emotional responses (four items), and external cues (six items). The eating behaviors are rated on a four-point Likert scale; 1—never/rarely, 2—sometimes, 3—often, and 4—usually/always. Reverse scoring was applied to questions 1, 2, 6, 7, 9, 11, 17, 18, 19, 27, and 28.A back-translation method was used to create the Malay-translated version of the MEQ [48]. The original version of the MEQ was first translated into the Malay language by two authors who are bilingual (English and Malay). The translated version was then piloted to 10 university staff members to test for clarity. Some unclear terms and phrases were noted. The questionnaire was then carefully checked for clarity, accuracy, the language’s suitability, and linguistic errors by two independent researchers. Once clarity and accuracy had been established, the questionnaire was then back-translated from the Malay version to English by an independent translator. The revised and modified translated items are available in the Appendix A.This study was a cross-sectional survey to assess the psychometric properties of the Malay version of MEQ conducted on 144 overweight and obese working adults conveniently recruited in a selected public university. These participants were recruited by our researcher from a health screening program held by the university among the staff. The sample size was determined based on a 5:1 ratio, where the sample size is expected to be a least five times the total number of items in the questionnaire [49] (pp. 86–99). Respondents were eligible if they met each of the following criteria: BMI ≥ 25.0 kg/m2, age 18–59 years, and no chronic diseases, such as cancer, kidney diseases, or heart diseases. Exclusion criteria were pregnant and/or breastfeeding women, having any severe mood disorder controlled by pharmaceuticals, and the use of pharmaceutical weight control. The socio-demographic information of the participants’ age, gender, educational level, monthly household income, and types of work were collected through a questionnaire. The BMI of the participants were measured using the TANITA Body Composition Analyzer (model TBF 300, Tanita Corporation, Tokyo, Japan). The Malay-translated MEQ was distributed via a Google Form. Ethical approval was obtained before data collection from the Universiti Kebangsaan Malaysia Medical Research Ethics Committee (UKM PPI.800-1/1/5/JEP-2019-391). Respondents were briefed on the purpose of the study and written consent was obtained.Factor analysis enables the determination of the underlying subdomains of a questionnaire [50]. Exploratory factor analysis (EFA) was recommended for establishing equivalence and factor structure validation of the translated and adapted questionnaires performed in different sample populations [51]. In this study, the MEQ factor structure was determined by using principal component analysis with varimax rotation [52]. This rotation produces a simpler solution and uncomplicated interpretation while maximizing the total variances of the squared loadings correlation between variables and factors. An eigenvalue of >1, a factor loading of ≥0.4 and a scree plot were applied for this study (Figure 1) [26].The Mindful Attention Awareness Scale (MAAS) is a 16-item tool that measures the frequency of mindfulness in daily life using general and situation-specific questions [53]. This instrument uses a six-point Likert scale from 1 (almost always) to 6 (almost never), in which the mean score can range from 1 to 6. Higher MAAS scores indicate greater mindfulness. This questionnaire was translated into the Malay language and validated by Zainal and colleagues [29].The MEQ’s internal consistency or homogeneity was assessed using the coefficient Cronbach’s alpha and McDonald’s omega (range 0–1). A coefficient value of ≥0.70 indicates a satisfactory internal consistency [54]. In addition, the internal consistency of the MAAS was carried out.The objective of this analysis is to measure an instrument’s or test’s stability over time. This is performed by administering the same test to the same subject at two different points. For this study, the intraclass correlation coefficient (ICC) was used to estimate the reliability of the scale. The interpretation of the agreement levels by ICC are as follows: 0.0–0.2 as small, 0.21–0.40 as fair, 0.41–0.60 as moderate, 0.61–0.80 as substantial, and 0.81–1.0 as almost perfect [55]. Statistical analysis was performed using Statistical Package for the Social Sciences 25.0 (SPSS, Inc., Chicago, IL, USA), in which significance was set at p-values < 0.05. All variables were tested for normality using the Kolmogorov–Smirnov, skewness, and kurtosis analysis. Since our data were normally distributed, Pearson’s correlation test was employed. Participants with incomplete responses of the MEQ-M and MAAS were excluded from the analysis.In total, 144 participants (41 males, 103 females) were included in this study. The mean age of the participants was 40.3 ± 6.9 years. The socio-demographic of the participants are presented in Table 1.The MEQ was analyzed by principal component factor analysis with varimax rotation. The overall Kaiser-Meyer-Olkin measure of sampling adequacy was 0.804. Bartlett’s test for sphericity produced a significant result (p < 0.001), indicating that the variables were correlated with one another. Thus, our preliminary analyses confirmed the appropriateness of principal component factor analysis for the data. The Malay-translated questionnaire had seven factors. The percentage of variances explained by rotated factor matrices ranged from 4 to 14% per factor, with seven factors explaining 58.8% of the overall variance. Percentages refer to the variance explained by each factor as follows: Factor 1, 14.5%; Factor 2, 12.0%; Factor 3, 8.5%; Factor 4, 8.1%; Factor 5, 6%; Factor 6, 5.4%; and Factor 7, 4.5%. All items loaded 0.40 or above (Table 2).The concurrent validity of the Malay-translated MEQ was calculated using Pearson’s correlations with MAAS. The correlational analysis results between total MEQ-M scores and total MAAS scores indicated a weak non-significant correlation (p = 0.679). Only factors 1, 2, and 7 were significantly correlated with MAAS (Table 3). Factor 1, 2, 4, and 5 were positively correlated with MAAS, whereas factors 3, 6, and 7 were negatively correlated with MAAS. The MEQ subscale that showed the highest correlations with MAAS measures was factor 2.Cronbach’s alpha for the MEQ was 0.64, which indicates reasonable reliability. In contrast, the omega from McDonald’s reliability test was lower. Further analysis was carried out by eliminating the External Cues subscales, which increased the Cronbach’s alpha and McDonald’s omega reliability value to 0.71 and 0.61, respectively. As for the subscales, Cronbach’s alpha values ranged from 0.27 to 0.70, whereas McDonald’s omega values ranged from 0.58 to 0.80. The reliability coefficient for the subscales were similar from both tests, except for Factor 1 and Factor 2, which showed higher reliability from McDonald’s omega test. With the elimination of item 3 from the Factor 1 subscale, the alpha value increases to 0.78 from 0.52. Other than that, the elimination of item 26 resulted in improving the Factor 2 subscale from 0.616 to 0.683. The exclusion of item 14 from the Factor 5 subscale improves the alpha value from 0.54 to 0.67. As for the test-retest reliability, the ICC for the summary score of the MEQ-M was 0.104, which means that the items have a small agreement with each other. The ICC for each subscale is presented in Table 4. As for the MAAS, the Cronbach’s alpha was 0.88, which showed satisfactory internal consistency. The test-retest reliability coefficient was 0.295, which indicates a fair agreement between the scores.Table 5 shows the correlation between sociodemographic characteristics and the total score of the MEQ-M and its subscales. Overall, there is no significant association observed with the exception to the Environmental Disinhibition subscale with BMI and gender (p = 0.013 and p = 0.044, respectively). Other than that, significant association was observed between age and two subscales (Environmental disinhibition and Emotional response) with p-values = 0.002 and 0.006, respectively. Furthermore, the questionnaire has poor correlation with all sociodemographic characteristics.The purpose of this study was to analyze the psychometric properties of the MEQ in overweight and obese Malaysian adult samples. This paper reports the translation procedure, validity, and reliability of the Malay-translated MEQ.Construct validity (EFA) results were inconsistent with findings reported in previous studies [1,27]. These studies found that the 28-item MEQ had a good fit for a five-dimensional factor structure. On the contrary, we found a seven-dimensional model from the exploration. The seven factors were labeled as Factor 1 (environment disinhibition), Factor 2 (emotional response), Factor 3 (taste awareness), Factor 4 (emotion awareness), Factor 5 (portion disinhibition), Factor 6 (external cues of food), and Factor 7 (external cues of place). Each factor was loaded strongly with factor loadings from 0.43 to 0.79. Items 7, 13, and 26 had multiple cross-loading. If a more stringent factor loading was used at the cut-off point of 0.5 (25% shared variances), items 4 and 27 of the MEQ would be dropped and, it would become a 26-item scale. This yielded a six-dimensional model for the Malay-translated MEQ. A possible reason that this study produced a seven-dimensional model, in contrast to other studies, is due to cultural differences. Our population study might have had a different understanding of mindfulness. Studies have shown that cognitive and reasoning styles are different across cultures and may affect how questions are viewed and answered [56,57]. Language and cross-cultural variations may influence the respondents’ responses and affect the analysis of the questionnaire’s psychometric properties [58].The Cronbach’s alpha and McDonald’s omega analysis showed that the Malay-translated MEQ was a reliable tool to be used among this group. The internal consistencies of the MEQ-M factors ranged from 0.54 to 0.70 for Cronbach’s alpha and 0.58 to 0.80 for McDonald’s omega, except for the External Cues subscales. This indicates that each factor’s items are moderately cohesive with each other in measuring specific mindful eating behavior. The low internal consistency of Factor 6 may be due to the small number of items. The internal consistency of Factor 7 could not be tested as there was only one item (Item 8) extracted. When item 8 was analyzed together with items 4 and 23 from Factor 6, the Cronbach’s alpha produced was 0.441. Exclusion of item 4, 8, and 23 (External Cues subscales) from the analysis has improved the reliability of the MEQ-M. Similarly, the external cues subscale was found to be invalid in a study among overweight and obese pregnant women [26]. The MEQ-M summary scale was 0.64, which is comparable to that reported by Abbaspoor (0.66) and Framson (0.64). In contrast, our findings showed better reliability compared to Apolzan and colleagues [26].The test-retest reliability of the MEQ with an eight-week interval showed only a small agreement between the items. Three subscales (environmental disinhibition, emotional response, and taste awareness) achieved a fair correlation (0.26 to 0.45). In contrast, emotional awareness, portion disinhibition, external cues of food, and external cues of place subscales had a small correlation coefficient. We also found that portion disinhibition and external cues subscales had a negative ICC value due to a negative average covariance among items. This may have affected the overall ICC of the MEQ (0.10). A small ICC means that the items’ measurement was not stable over time. The Iranian version of the MEQ test-retest reliability had a high level of correlation, where all subscales’ ICC was ≥0.7 [27]. This discrepancy may be due to the different sample populations tested only among women with normal BMI. The ICC could be improved if a sizeable heterogeneous sample in terms of BMI was included. Moreover, they retested the questionnaire in a shorter period (four weeks) compared to the present study.In correlation to MAAS, the MEQ correlated positively except for Factor 3, 6, and 7. However, the correlation was weak. This may suggest that individuals with higher mindfulness may or may not tend to eat mindfully. In contrast, MEQ was significantly correlated with MAAS among pregnant women [26]. The differences may be due to the new factors or domains produced from our study compared to the original MEQ which consists of only five factors. Other than that, the MAAS questionnaire focused exclusively on the attention awareness component of the mindfulness construct, whereas the MEQ includes constructs of eating behaviors and mindfulness. In our cultural context, this may reflect that these measures are not related to each other. Our study also found that there was no significant correlation between BMI and the MEQ-M’s subscales, except for the inverse relationship with the Environmental disinhibition subscale. This suggests that higher BMI was correlated with lower ability to stop eating even though they already feel full. Similarly, the subscales of the Iranian version of the MEQ also showed no correlation with BMI apart from the Awareness subscale [27]. In contrast, Framson et al. found that the overall scores of the MEQ was negatively associate with BMI [1].This study was the first to establish the validity and reliability of the MEQ-M among overweight and obese healthy working adults. Since this study was conducted among overweight and obese adults, this limits the generalizability of the results to the general population. Furthermore, this instrument was a self-reported questionnaire. The respondents might have had difficulties with being aware of their eating experience when that questionnaire was given. They might have provided answers describing what they should do instead. In addition, there was no concurrent validation of the MEQ-M with other eating psychopathology, such as the binge eating scale and three-factor eating questionnaire, which may provide additional support for MEQ-M’s validation.The study findings showed that the psychometric properties of the Malay-translated MEQ are acceptable through construct and internal consistency reliability. This instrument may be used for assessing mindful eating habits in the Malaysian population, especially among overweight and obese employees. However, the measurement by domains of the MEQ-M should be considered cautiously as it differs from the original domains. As the Malay-translated MEQ was only tested among overweight and obese university employees, we suggest subsequent studies among employees with normal BMI to further test its validity and reliability, especially with regards to its stability over time.S.M.A.B. contributed to the conceptualization, data curation, formal analysis, investigation, methodology, and writing—original draft; Z.A.M. was involved in the conceptualization, funding acquisition, writing—reviewing and editing, and supervision; M.A. was involved in the validation, writing—reviewing, and editing process; N.B.A.K. was involved in the reviewing and editing of the manuscript; W.N.K.I. contributed to the data curation, formal analysis, and investigation process; A.F.M.L. and S.S. were involved in the reviewing and editing of the manuscript. All authors have read and agreed to the published version of the manuscript.This study was supported by an internal research grant (DCP-2018-005/1) from the Universiti Kebangsaan Malaysia.The study was conducted according to the Declaration of Helsinki and it was approved by Universiti Kebangsaan Malaysia Human Research Ethics Committee (UKM PPI.800-1/1/5/JEP-2019-391).Signed informed consent was obtained from all participants prior to data collection.The data presented in this study is a part of an ongoing doctoral research of S.M.AB. Hence, we could not publicly release the data. However, it is available upon request from the corresponding author (Z.A.M.). The authors are grateful to all participants for their time and support. In addition, we would like to thank John W. Apolzan for his guidance on using the MEQ in this study.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.The Malay-translated Mindful Eating Questionnaire (MEQ).Scree plot of the eigenvalues for the principal component analysis of the Mindful Eating Questionnaire (MEQ).Characteristics of the study population, n = 144.Factor loadings from the Malay Mindful Eating Questionnaire (MEQ-M) principal component analysis.Correlations between MEQ-M and Mindfulness Attention Awareness Scale (MAAS).a Correlation is significant at the 0.05 level (two-tailed). b Correlation is significant at the 0.01 level (two-tailed).Descriptive statistics, Cronbach’s alpha, and intraclass correlation coefficient (ICC) of the MEQ-M subscales and the MAAS.NA: Not available.Correlation between sociodemographic characteristics and the MEQ-M (Pearson’s r).** Correlation is significant at the 0.01 level (two-tailed). * Correlation is significant at the 0.05 level (two-tailed).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01022.txt

@@ -0,0 +1,4 @@
+The aim of this narrative review is to introduce the reader to Bayesian methods that, in our opinion, appear to be the most important in the context of rare diseases. A disease is defined as rare depending on the prevalence of the affected patients in the considered population, for example, about 1 in 1500 people in U.S.; about 1 in 2500 people in Japan; and fewer than 1 in 2000 people in Europe. There are between 6000 and 8000 rare diseases and the main issue in drug development is linked to the challenge of achieving robust evidence from clinical trials in small populations. A better use of all available information can help the development process and Bayesian statistics can provide a solid framework at the design stage, during the conduct of the trial, and at the analysis stage. The focus of this manuscript is to provide a review of Bayesian methods for sample size computation or reassessment during phase II or phase III trial, for response adaptive randomization and of for meta-analysis in rare disease. Challenges regarding prior distribution choice, computational burden and dissemination are also discussed.A disease is defined as rare depending on the prevalence of the affected patients in the considered population. In the United States, a disease is rare if it affects fewer than 200,000 people in the U.S. [1] (or about 1 in 1500 people); in Japan, if it affects fewer than 50,000 patients in Japan (or about 1 in 2500 people); and in the European Union if the prevalence is no more than 5 per 10,000 (that is, fewer than 1 in 2000 people), but the definition excludes diseases that are not also life-threatening, chronically debilitating, or inadequately treated [2]. There are between 6000 and 8000 rare diseases [3], 71.9% of which are genetic and 69.9% which exclusively affect paediatric populations, and it is estimated that the global population prevalence of rare diseases is of 3.5–5.9%, which implies that 263–446 million persons are affected at any stage in their life [4]. The usual level of rigorous clinical trial evaluation of treatments is required in rare diseases just as much as in more common ones. Although in some cases, particularly in phase II trials, single-arm trials might be considered (see, for example, Grayling et al. [5]), randomized controlled trials are to be preferred when this is possible. For example, the European regulatory guidance [2] affirms that “patients with [rare] conditions deserve the same quality, safety and efficacy in medicinal products as other patients; orphan medicinal products should therefore be submitted to the normal evaluation process"; this is also in agreement with U.S. guidance [6].The main issue in drug development for rare diseases is linked to the challenge of achieving robust evidence from clinical trials in small populations when trial sample sizes are necessarily limited [7]. Even if for some rare diseases the population size is relatively large (for instance, Friedreich Ataxia in the EU) [8], the majority of rare diseases are less frequent [9]. Small population clinical trials have been the focus of much methodological research activity in the last two decades. From a regulatory perspective, the European Medicines Agency (EMA) described a methodological framework, summarizing several possible approaches, in the guidance “Guideline on Clinical Trials in Small Populations” [10] and the Food and Drug Administration (FDA) in the draft guidance on rare disease [6]. The European Union’s Seventh Framework Programme for Research, Technological development and Demonstration (EU FP7), acknowledging the need for additional methodological research work, funded three projects in 2013; the Integrated Design and Analysis of Small Populations Group Trials (IDeAl) project (www.ideal.rwth-aachen.de), the Innovative Methodology for Small Populations Research (InSPiRe) project (www.warwick.ac.uk/inspire), and the Advances in Small Trials Design for Regulatory Innovation and Excellence (Asterix) project (www.asterix-fp7.eu) [8,11,12,13].The drug development process involves on-going learning as data are observed through the series of clinical trials, and, above all in rare diseases, there is a considerable effort to optimize this learning process [14,15]. A better use of all available information can help the process and Bayesian statistics provides an opportunity to do this in a formal way (at the design stage, during the conduct of the trial, and at the analysis stage) [16,17]. Like drug development, the Bayesian approach can be seen as an on-going learning process: it starts with a prior belief (quantified as a prior distribution for the unknown model parameters), which is then updated with the new evidence (likelihood data from the new trial/experiment) to yield a posterior belief (expressed as a posterior probability distribution for the unknown model parameters). In this way, Bayesian statistics provides a mathematical method for calculating the predictive probabilities of future events, given the actual trial and the knowledge from prior trials. Moreover, a formal Bayesian analysis can incorporate different utilities or prior beliefs coming from different stakeholders and quantify how these could impact potential decision-making.Bayesian methods and designs are well established and mostly accepted, by both clinicians and regulatory agencies, in early phase clinical trials. Due to the greater flexibility, in both design and analysis, of the Bayesian paradigm with respect to the frequentist one and since type I and II errors do not have to be controlled at this stage, Bayesian adaptive designs are mostly chosen for these stages [18]. As early phase trials in all diseases use small sample sizes, designs, specifically developed for rare diseases are unnecessary. Thus, in this manuscript we will focus on novel Bayesian approaches firstly developed for confirmatory/randomized trials in the rare disease setting, where more conventional approaches may be unfeasible.The aim of this narrative review is to show to the reader Bayesian methods that, in our opinion, appear to be the most important in the context of rare diseases. Its purpose is not to present a comprehensive compendium of Bayesian statistics in rare diseases, but to give a starting point for the reader on some uses of novel Bayesian methods in this field. All methods are presented in a general way, without mathematical formalism, so that a reader who already have a basis of Bayesian statistics can understand the general idea, along with the corresponding principal(s) reference(s), in such a way the reader can find the details for methods they wish to explore in more detail. In the following sections, three specific topics dealing with the application of a Bayesian design are introduced. The first topic, in Section 2, regards methods for sample size computation or reassessment during randomized phase II or phase III trial. The second topic, a recent method proposed for response adaptive randomization, is presented in Section 3. Section 4 presents the fourth topic of Bayesian meta-analysis methods developed for evidence combination in rare diseases. Finally, we discuss frequent challenges faced when choosing a Bayesian design; in particular, the priors distributions choice, which include the issue of the quantity of information and commensurability between prior information and actual data; the computational burden required; and dissemination issues.Usual sample size determination approach focus on frequentist properties, that is, type I error and study power. However, in a rare disease setting, accruing the number of patients required to perform a fully-powered significance test after the trial may be infeasible. Increasing the allowed type I error can be a solution to reduce the required sample size. However, as recalled in the introduction, this practice is not generally supported by regulatory agencies. As an alternative, some authors have proposed the use of a Bayesian decision-theoretic framework [19]. A Bayesian decision-theoretic approach can be applied when we would like a treatment recommendation not based on type I error but on maximizing an expected gain for the total population. According this approach, we can compare the costs of clinical trial evaluation with the potential benefits to current and future patients, assessing how the cost-benefit balance differs between large and small patient populations when, in the latter, patients recruited to a clinical trial could be a substantial proportion of the population. The design of the study, including the sample size, can then be chosen on the basis of the expected gain, with the sample size that maximizes the expected gain chosen for the clinical study. Here the concept of a “gain” is interpreted very broadly and can be defined from the patient, sponsor, regulatory, public health, or society perspective, or from a combined perspective. Stallard et al. [20] have shown that for a wide range of distributions, including those for continuous, binary or count responses, and gain function forms, the optimal trial sample size is proportional to the square root of the population size, with the constant of proportionality depending on the gain function form and prior distribution of the parameters of the distribution of the data. A smaller sample size may thus be appropriate for a trial in a rare disease than in a more common one.Bayesian statistics can also be adopted to overcome some challenges in the calculation of sample sizes from the frequentist perspective. For example, for normally distributed outcomes, values for variances need to be specified, but, especially in the case of small populations, may be based on very little information, for example, that from only one very small pilot study. When using a Bayesian approach, the aggregation of prior information on the variance with newly collected data is more formalized. Brakenhoff et al. [21] proposed a framework incorporating the employment of power priors in order for operational characteristics to be controlled in case of prior-new data conflict.Bayesian group sequential designs could also be used to provide interim stopping criteria, based on efficacy and/or futility. Even if the frequentist operating characteristics of these designs are usually checked, they are not designed to optimize them. A practical guide for their implementation and reference for software can be found in Gsponer et al. [22].
+      While randomization is the established method for obtaining scientifically valid treatment comparisons in clinical trials, as the trial progresses, increasing evidence may suggest that one study group is responding or doing much better than another. As a consequence, novel randomization methods, such as response adaptive randomization [23] (RAR), have been proposed to address this ethical question continuously updating assignment probabilities based on response of the different groups to their respective treatments so as to allocate more patients to better-performing treatments. Both, frequentist and Bayesian approach can be applied, however, the latter one has gained more popularity due to its flexibility [24,25]. In the same manner as the previous decision theoretic idea, this approach could be considered in rare disease setting, where future patients in the general population is limited, to balance the benefits to current trial patients and future ones. Nonetheless, standard adaptive randomization may lead to estimation bias [26], with the potential for the trial to reach an erroneous conclusion. Therefore, novel and calibrated RAR approaches should be preferred. The small sample sizes in trials in rare diseases may also mean that it is possible to calibrate RAR methods in a way that would be infeasible in larger trials.A recent paper suggests a novel randomized response-adaptive design specifically developed for a rare disease trial [27]. It uses the framework of finite-horizon Markov decision processes and dynamic programming (DP) to recruit more patients to the more beneficial arms while guaranteeing a minimum sample size to each treatment arm. The authors show that the design has good operating characteristics, in term of (i) the percentage of patients allocated to the superior arm, which is much higher than in the traditional fixed randomized design; (ii) the power, which is higher than optimal DP; and (iii) bias and mean square error of the treatment effect estimator, which are small.Meta-analyses are used to combine evidence from multiple studies. Differences in study characteristics, such as trial design and study populations, can bring to heterogeneous treatment effects and these must be accounted for in the meta-analysis formulation. To deal with the between-trial heterogeneity, random-effects meta-analysis has become the gold standard, and the most used method is the normal–normal hierarchical model (NNHM) [28]. In a rare disease, the limited number of trials and their small sample size may impact the validity of usual frequentist meta-analysis methods. A Bayesian approach offers another way to perform random-effect meta-analyses within the NNHM framework. One of the advantages is that the solution remains coherent for small numbers of studies, although careful prior specification is required. Friede et al. [29] showed that, when doing meta-analysis with only two studies, Bayesian random-effects meta-analyses with priors covering plausible heterogeneity values offer a good compromise. They compared the Bayesian method to the NNHM, to the Hartung-Knapp-Sidik-Jonkman method (HKSJ) and to the modified Knapp-Hartung method (mKH). On one hand, the coverage of the standard method, based on normal quantiles, was unsatisfactory; on the other hand, very large (therefore uncertain) confidence intervals resulted from the HKSJ and mKH. An acceptable trade-off between these two extremes was achieved, in general, by Bayesian intervals that showed suitable characteristics. Usually, the Bayesian approach is computationally more demanding. However, optimized free software are available, such as the bayesmeta R package, which uses a general semi-analytical approach to solve the meta-analysis problem via the DIRECT approach [30] and provides an efficient and user-friendly interface to Bayesian random-effects meta-analysis [31].When dealing with binary outcomes, the binomial-normal hierarchical model is usually preferred to the NNHM, which then relies on asymptotic approximations. A challenge in this setting in rare diseases is that we could face the probability to have no events due to the small sample sizes. Frequentist approaches are known to induce bias and to result in improper interval estimation of the overall treatment effect in a meta-analysis with zero events [13]. On the other hand, Bayesian models are known for being sensitive to the choice of heterogeneity prior distributions in sparse settings, therefore, the need to identify priors with robust properties is crucial. Pateras et al. [32] proposed a general way to set prior distributions. Via simulations, they showed that a uniform heterogeneity prior, bounded between -10 and 10, on the log heterogeneity parameter scale shows appropriate 95% coverage and induces relatively acceptable under/over estimation of both the overall treatment effect and heterogeneity, across a wide range of heterogeneity levels.The Bayesian meta-analysis approach also allows implementation of a number of more advanced analysis strategies. A series of studies may be used to inform the analysis when the focus is not on an overall synthesis, but rather on a particular study that is to be viewed in the light of previously accumulated evidence. For example, Wandel et al. [33] used a Bayesian meta-analytic approach to inform a phase III study with phase II data. They investigated the use of shrinkage estimates to support data from a single trial in the light of external information. The method allows quantifying and discounting the phase II data through the predictive distribution relevant for phase III. Bayesian meta-analysis approach can also be adapted to incorporate external information from historical controls [34] or borrow information from other arms in a randomized control trial, for example, in a basket design [35,36]. Such approaches could prove very valuable in the setting of rare diseases where trials are necessarily small.As stated above, the Bayesian approach can be more flexible than the frequentist counterpart. However, the flexibility comes along with a number of possible challenges. Even if Bayesian methods can bring substantial benefits, their validity and effectiveness require expertise and care. In the following, we will describe some points that should be addressed when planning a Bayesian analysis.In Bayesian statistics, external information can be easily incorporated into the prior distributions. An informative prior distribution for the unknown parameters could be determined through elicitation of expert knowledge, from data from other trials or from a search of the literature to identify results obtained in trials of similar drugs, or the same drug in a different population, via the so called “extrapolation". Extrapolation approaches are well known in paediatrics, where the proper dosage for children is estimated starting from adults’ data, and in bridging studies, where the drug is tested in a new geographical population, for example, in Asian, given the results in a previous one, for example, Caucasian. This concept can be translated in rare disease, since rare diseases prevalence may vary by continent (i.e., IgA nephropathy is rather rare in the EU but more frequent in Asia and Africa) and we can be to adopt proofs of efficacy from the larger populations to the smaller one [8].However the prior distribution is obtained, the use of an informative prior to make inferences about medical treatments based on small sample size trials remains inherently controversial, however. Choice of a prior distribution must therefore be done carefully, since the use of informative priors may be seen as introducing bias into posterior inferences and inflating type I error rates. This is a general problem common in many different fields, and several authors have addressed the issue of eliciting experts’ opinions, building priors upon the elicited values, and performing Bayesian analyses using the resulting priors. See O’Hagan et al. [37] for a complete review. The elicitation needs to be made as meticulous and objective as possible to catch expert expertise. One way is to follow a recognized protocol that is designed to address and minimize the cognitive biases [38]. In the following, we summarize two approaches that have already been used in the context of rare disease.The first approach describes how to obtain a consensus between experts. This research was motivated by the design of the MYPAN trial, a multicentre RCT comparing mycophenolate mofetil (MMF) with cyclophosphamide (CYC) for the treatment of polyarteritis nodosa, a rare and serious inflammatory blood vessel disease in children [39]. The authors proposed to add priors on the probability of success of one arm and on the log-odds ratio of the probabilities. Then, a behavioral aggregation process, by which experts interact to reach a mutually agreeable consensus through constructive discussions, was chosen for systematic elicitation from clinicians of their beliefs concerning treatment efficacy. In particular, experts’ individual prior beliefs were obtained at the beginning of the process; then, the full group was asked to reach a consensus. The results are then used to establish Bayesian priors for unknown model parameters and the authors have also considered the possibility of considering results from related trials. A similar strategy was used in a trial of adalimumab versus pamidronate for children with CNO/CRMO [40].The second approach focuses on reflecting, when eliciting experts’ opinions, how these depend on differences in experience, training and medical practice [41]. Motivating by a 70-patient randomized trial to compare two treatments (the same described in the first approach) for idiopathic nephrotic syndrome in children (NCT 01092962), the authors proposed a Bayesian methodology for constructing a bivariate parametric prior starting from elicited graphical information. The method involves four steps: (i) each physician builds manually two histograms, one for each treatment parameter using the “bins-and-chips” graphical method of Johnson et al. [42]; (ii) then, for each physician and each treatment parameter, a marginal prior, characterized by location and precision hyperparameters, is fitted to the elicited histogram; (iii) a bivariate prior is built by averaging the marginals over a latent bivariate distribution; (iv) finally, an overall prior is obtained as a mixture of the individual physicians’ priors. The approach also suggests a framework for performing a sensitivity analysis of posterior inferences to prior location and precision.Incorporating external information, whatever the source type (other trial, experts’ elicitation, etc.), has to be done properly, as the information can be in conflict with the actual data or the amount of information can overwhelm trial data. Several methods that allow prior information to be incorporated if it is in accordance with the trial data and otherwise to be down-weighted have been proposed [34,43,44]. Moreover, the effective sample size allows to quantify the information in the prior to be specified in term of the number of hypothetical patients used to build the prior [45]. Different prior building approaches may be used for different parameters; for example, historical control can be included via a power prior approach and experts’ opinion can be used for the new treatment effect. An adaptation of the power prior approach, that is useful particularly for borrowing evidence from a single historical study, was proposed in rare disease setting [46]. Borrowing information from a historical trial is often related the type I error inflation. By determining the amount of similarity between the new and historical data, this method uses predictive probabilities and is parameterized in order to control the type I error.Estimation of posterior distributions can be challenging when prior distributions do not have simple conjugate forms. Specific Markov chain Monte Carlo algorithms, such as the Gibbs sampling or the Hamiltonian Monte Carlo, can be used to obtain an approximation of posteriors. Even if freely available software and the increasing computational power of computers may help the Bayesian implementation, writing, coding and testing the models usually requires a bigger effort than choosing a frequentist approach. In general, the more complex the model or the prior distribution, the longer the computational time to obtain the result. Validation of the method via simulation is one of the common way used in Bayesian setting. Simulating several possible scenarios can allow the user to calibrate model parameters (i.e., the quantity of information in the prior distribution) to obtain desired operational characteristics, such as the type I error control or the power. Choosing a fast and reliable approximation method is, however, crucial when simulations are required.Even if the Bayesian approach has been shown to capture the thinking behavior of clinicians [41], Bayesian methods and results are sometimes still viewed with suspicion by clinicians and traditional statisticians. The influence of the prior distribution may be considered disturbing and the lack of p-values can give the feeling that regulatory agencies will not consider the results obtained. In effect, Bayesian methodologies are usually less discussed in public regulatory guidelines than the frequentist counterpart. However, the FDA Guidance for the Use of Bayesian Statistics in Medical Device Clinical Trials [47] shows the regulatory agencies efforts in this direction. The guidance gives several Bayesian insights that could be used in general, not only in medical device field. Several sections explain how to well plan a Bayesian clinical trial, what to consider when choosing prior distributions, and how to analyze the data. Then, in the technical details sections, the FDA points out the importance of simulations to obtain operating characteristics of the planned design, to assess the type one error rate, power, etc. While also other recommendations for Bayesian analyses have been developed [17,48,49], they were primarily addressed to researchers, not to readers unfamiliar with Bayesian approaches [50]. Therefore, efforts are needed to well explain the Bayesian philosophy to non-statisticians. An example is given in Ferreira et al. [51] and Ferreira et al. [50], where the authors help clinicians interpretation of Bayesian clinical trial though a side-by-side comparison with the frequentist approach. On one hand, they teach how to transfer frequentist ideas, such as the p-values or hypotheses testing, to the Bayesian framework, such as posterior probabilities and Bayes factor, and, on the other hand, they give insights on what to check when reading a Bayesian report.The aim of this article has been to review the use of Bayesian methods in confirmatory trials in rare diseases, though many of the approaches described could also be applied in clinical trials in other more common disease areas.The formal Bayesian approach permits the incorporation of accumulating information into the analysis of the actual trial and, therefore, the updating of belief. This feature is extremely attractive in rare disease setting, where usually sample sizes and the opportunities to performs clinical trials are limited. Incorporation of previous information should be strongly considered and the Bayesian approach, with its flexibility, could be seen as a future gold standard in this field. As shown in the manuscript, the accumulated information can be used in the prior distribution settings, in sample size optimization and/or in randomization. Depending on the trial, where and when using this kind of information is used has to be carefully chosen and simulations are strongly suggested to evaluate method performances.Conceptualization, M.U.; writing—original draft preparation, M.U.; writing—review and editing, N.S. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.We would like to thank the anonymous reviewers for their constructive feedback.
+      The authors declare no conflict of interest.The following abbreviations are used in this manuscript:
+      CYCcyclophosphamideCNO/CRMOchronic recurrent multifocal osteomyelitisDPdynamic programmingEMAEuropean Medicines AgencyFDAFood and Drug AdministrationHKSJHartung-Knapp-Sidik-Jonkman methodmKHmodified Knapp-Hartung methodMMFmycophenolate MofetilNNHMnormal-normal hierarchical modelRARresponse adaptive randomizationPublisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01023.txt

@@ -0,0 +1,26 @@
+To minimize the damage from contaminant accidents in rivers, early identification of the contaminant source is crucial. Thus, in this study, a framework combining Machine Learning (ML) and the Transient Storage zone Model (TSM) was developed to predict the spill location and mass of a contaminant source. The TSM model was employed to simulate non-Fickian Breakthrough Curves (BTCs), which entails relevant information of the contaminant source. Then, the ML models were used to identify the BTC features, characterized by 21 variables, to predict the spill location and mass. The proposed framework was applied to the Gam Creek, South Korea, in which two tracer tests were conducted. In this study, six ML methods were applied for the prediction of spill location and mass, while the most relevant BTC features were selected by Recursive Feature Elimination Cross-Validation (RFECV). Model applications to field data showed that the ensemble Decision tree models, Random Forest (RF) and Xgboost (XGB), were the most efficient and feasible in predicting the contaminant source.When accidental spills of contaminant occur in natural rivers, a rapid response is necessary to minimize the damage to both aquatic life and humans who depend on the river as a water resource. Contaminant accidents in rivers are risky and urgent problems that occur frequently, mainly by transportation accidents or industrial waste [1,2,3]. In this respect, quick identification of the contaminant source plays a significant role in protecting river systems and environmental forensic by providing information of the contaminant source, such as spill location, spill mass, and release history. However, inverse tracking of the contaminant source is a problem, due to the lack of observed data and complexity of the mixing processes in a natural river. In order to overcome this limitation, a number of methods for the identification of contaminant sources have been suggested, mainly in the groundwater system; these use various techniques, such as optimization, geostatistical simulations, analytical solutions, and data-driven models [4,5,6,7,8,9,10,11,12,13,14,15]. Although contaminant source identification problems in both rivers and groundwater have a similar purpose, applying the methods developed for groundwater to rivers is challenging, due to the difference in flow and mixing characteristics between groundwater and rivers. Specifically, a quick response is more crucial in rivers than in groundwater, since the contaminants are transported more rapidly in rivers than in groundwater.Among many inverse tracking methods used in the groundwater system, the optimization method was frequently used in river systems, which iterates the calculations based on the advection–diffusion process to reach the global solution of contaminant source as an ill-posed problem. Parolin et al. [16] carried out a hybrid heuristic algorithm, which included the Luus–Jaakola method (LJ), particle collision algorithm (PCA), ant colony optimization (ACO), and golden section method (GS), to identify the spill location and intensity of contaminant source in an estuary. Zhang and Xin [17] used the basic Genetic Algorithm (GA) to identify the spill location and spill mass of contaminant sources in a small straight river. However, these optimization approaches have limitations of high uncertainties in their deterministic processes and the data used in the optimization [18]. Thus, stochastic methods based on Bayesian inference were suggested to overcome the disadvantage of deterministic optimization. Yang et al. [3] combined the Differential Evolution Algorithm (DEA) and Metropolis–Hastings–Markov Chain Monte Carlo (MH–MCMC) to enhance the optimization process with noise immunity. Nevertheless, computational loads of these methods to predict the contaminant source were too expensive to apply in real-time, and high inverse uncertainty occurred according to the objective function in the optimization procedure [18].Cheng and Jia [19] suggested a backward location probability density function (BL–PDF) to identify the spill location. They evaluated the proposed method regarding noise, and validated the model with the data from the real dye tracer test performed in the natural river, which is a significant process to test field applicability. Ghane et al. [20] also applied the backward probability method further to predict the release time, while Boano et al. [21] employed a geostatistical method to recover the release history under the assumption that the spill location was known. In order to improve the performance of the stochastic model, the Ensemble Kalman filter was coupled with backward location probability [2]. In terms of the uncertainty of the identification results, these stochastic-based methods were proven to be more applicable to the contaminant source identification problems than deterministic-based methods [2,3,20,22].Despite the valid performance of these stochastic methods, reflecting the complex mixing characteristics in inverse tracking models is very intricate, because the advection–diffusion process contains many problems of spatial and temporal scale. For this reason, data-driven approaches using contaminant spill scenarios to identify the location of the contaminant source were recently presented. The data-driven model extracts the scenario that best matches the observed data, which is obtained downstream of the spill location. This approach has the advantage that the scenarios would include the river mixing mechanisms via model parameters, and the spatial and temporal scales would be explicitly calculated. Telci and Aral [23] simulated contaminant spill scenarios in the Altamaha River, USA, and they developed a sequential feature selection algorithm using the scenarios, which sequentially eliminates potential spill locations in the scenarios. Kim et al. [24] and Lee et al. [25] employed the Random Forest (RF) method to build a spill location predictor, using the same contaminant spill scenarios used by Telci and Aral [23]. Compared to the other methods, the data-driven models require a low computational load for prediction, even though the training process requires a large dataset [26]. In this regard, the data-driven models are more feasible for the real-time prediction of a contaminant spill, facilitating a quick response to river spill accidents. As RF was used above, Machine Learning (ML) techniques have been widely utilized in data-driven models to investigate the complex functional relations in water resources [27,28,29,30,31,32,33].A significant factor in determining the performance of the contaminant identification model using a data-driven model is the reality of the contaminant scenarios. However, the previous studies [23,24,25] have a disadvantage, since the Storm Water Management Model (SWMM), which assumed the Continuous flow Stirred Tank Reactors (CSTR), was used as a contaminant transport model [23,34]. Such a simplified model would be incapable of accurate simulation of complex hydrodynamics and contaminant transport in rivers.This study presents an enhanced framework for the identification of a contaminant source in rivers. The first objective of the proposed framework was to generate realistic contaminant spill scenarios. For this objective, the Transient Storage zone Model (TSM) was used as a contaminant transport model to generate the contaminant spill scenarios. This model has been successfully used to reproduce the breakthrough curve (BTC), which is a time-concentration curve of the contaminant that represents the mixing processes with advection, shear dispersion, and storage effect in the river [35,36,37,38,39]. The second objective was to develop Machine Learning (ML) models for the identification of both spill location and mass of the contaminant source in rivers. The contaminant spill scenarios calculated by TSM were used as training and validation dataset. In this procedure, 21 features extracted from the BTCs of spill scenarios were used to predict the contaminant source by the six ML models. The optimal BTC features of both spill location and spill mass predictions were selected by Recursive Feature Elimination Cross-Validation (RFECV), which selected features recursively according to the feature importance of the ML model. Finally, the proposed models were applied to the field tracer data obtained in the river in order to ascertain the field applicability.The Figure 1a shows a flowchart of the development of the proposed framework of the Inverse Tracking Model (ITM) to identify the spill location and mass of a contaminant source. The framework consists of four steps: hydrodynamic calculation, contaminant transport simulation, BTC analysis, and ML modeling.Section 2.1 describes the first and second steps, in which the Contaminant Accident Scenarios Data-Base (CAS DB) is developed by numerical models of river hydrodynamics (HEC-RAS) and contaminant transport (TSM). Section 2.2 explains the third step, which includes the BTC analysis. This step features the BTCs of monitoring points to build the ITM. The BTC features serve as training and validation dataset of the ITM, instead of the BTC itself. Section 2.3 describes the last step of the ML process in detail. The ITM uses the classification and regression model of ML to build models that predict the spill location and mass of the contaminant source, respectively. In this process, the optimal ML model and BTC features are selected through RFECV.Figure 1b indicates the application process of the proposed ITM. When the BTC is detected from the sensor at the downstream of the spill location, the BTCs serve as the input data of the ITM. Then, this observed BTC is reduced into BTC features, which are substantial input variables of the ITM. Upon receiving the input data of the BTC features, the spill location is first predicted, and then the spill mass is predicted by adding the spilled distance to the BTC features through the predicted spill location.In most of the natural river, various types of transient storage zones, called dead zones or stagnant zones, exist along and across the stream, of which the effects cannot be modeled by the conventional one-dimensional advection–dispersion equation (1D ADE) [40]. In contrast to the main free-flowing water zone where the advection and dispersion mechanisms are dominant, the storage zone that is created by various channel irregularities, such as pools, side pockets, vegetation, hydraulic structures, and hyporheic zone, can be defined as an area where the flow is stagnant or recirculated. With respect to contaminant transport, the storage zone effect induces the shape of the BTC to present a steep slope in the rising limb, and a long tail in the falling limb. This skewness of the BTC arises due to transient trapping of contaminants in the storage zone. Since each stream has its own storage zone characteristic, the observed BTC represents the mixing properties of the stream. Therefore, the TSM generates a more realistic BTC for the non-Fickian transport processes than the 1D ADE-based model, by reflecting the storage zone effect [35,41,42].The TSM consists of two equations: the main free-flowing water zone equation, and the storage zone equation. The equations are modified version of ADE to describe the storage effect by conceptually dividing the area into the main flow zone area (AF), and the storage zone area (AS). It also exhibits a mass exchange rate (α), which is a first-order mass transfer between the main flow zone and the storage zone. Based on the assumption of steady uniform flow, conservative solute, and completely mixed storage zone, the equations are given as [35,37]:(1)∂CF∂t=−QAF∂CF∂x+1AF∂∂x(AFKF∂CF∂x)+qLAF(CL−CF)+α(CS−CF)
+          (2)dCsdt=αAFAS(CF−CS)
+          where CF, Cs, CL are the concentration of the main flow zone, storage zone, and lateral flow concentration, respectively [kg/m3]; t is time [s]; x is distance [m]; Q is the volumetric discharge of the stream [m3/s]; KF is the longitudinal dispersion coefficient of the main flow zone [m2/s]; qL is lateral inflow rate [m2/s]; AF and AS are the cross-sectional area of the main flow zone and storage zone, respectively [m2]; and α is the exchange rate of the storage zone [1/s].In real river systems, KF, AF, AS, and α in TSM equations are unmeasurable parameters, because the storage zones in each stream vary significantly. Thus, in most studies, the exact values of these four parameters were estimated using the optimization method from field tracer data [43,44,45]. With respect to spill scenarios, Rivord et al. [46] employed One-dimensional Transport with Inflow and Storage (OTIS) [37] to model contaminant spills in the Truckee River. They considered only the dispersion process using dispersion coefficients (KF) estimated by empirical equations with streamflow (Q), reach slope (S), and cross-sectional area (A). Although they estimated KF under various streamflow conditions using the empirical equation, when storage effects were not considered, their results showed relatively large errors.To overcome this limitation, empirical equations for TSM parameters have recently been derived from a meta-analysis of river mixing tracer tests [39,47]. From these equations, the TSM parameters can be estimated using easily measurable hydraulic and geometry variables. Thus, in this study, Principal Component Regression (PCR)-based empirical equations for TSM parameters were used to estimate TSM parameters. Equation (3) gives the equation, while Table 1 gives the derived power [39]:(3)(KFhU*,AFWh,ASWh,αU*/h)=exp(a)(UU*)b(Wh)c(Sn)d
+          where W is the channel width [m]; h is the mean flow depth [m]; U is the mean flow velocity [m/s]; Sn is the channel sinuosity; and U* is the shear velocity, which is estimated from the following equation: U*=ghS0, where g is the gravitational acceleration [m/s2] and S0 is the mean bottom slope.In this study, MATLAB-based TSM code was employed [48]. This model used the finite difference method and the Crank–Nicolson method, similar to the OTIS by Runkel [37].In order to generate scenarios under various streamflow and spill conditions, it is necessary to simulate a wide range of contaminant spill and flow cases in the range that may occur in the river system. Accordingly, the streamflow scenarios were generated by estimated streamflow distribution from the historical data of the study site. In this study, the 1-D Hydrologic Engineering Center-River Analysis System (HEC-RAS) (US Army Corps of Hydraulic Engineers, Washington, DC, USA), was used to calculate hydraulic and geometric information from the streamflow scenarios for the input variables of TSM empirical equations. The HEC-RAS calculates 1-D unsteady flow by solving the Saint Venant equations according to input data of initial flow rate, lateral flow, topographic data of cross-sectional shapes, and roughness coefficient [49]. In this framework, the flow regime was assumed to be steady uniform flow within sub-reach, and steady nonuniform flow considering lateral inflow from a tributary.The contaminant spill conditions were assumed to be an instantaneous injection with conservative contaminants that do not decay. The spill mass was generated randomly from a uniform distribution. In particular, the 1D ADE analytical solution of the instantaneous injection was applied to the upstream boundary concentration [40]. From this approach, spill mass can serve as an input variable of TSM simulation. Due to the initial condition given by CF(0,t)=∞, the upstream boundary condition was assumed to be that shortly after the contaminant spill, the storage zone effect does not exist. Thus, the upstream boundary concentration profile at 10 m away from the spill location as CF(10,t) was used for the initial boundary condition:(4)CF(x,t)=M4πKFtexp[−(x−UFt)24KFt]
+          where M is the spill mass [kg].In order to build the ITM, a large number of contaminant spill scenarios were required. Thus, we developed a CAS simulator using the Parallel for Loops (parfor) in MATLAB’s Parallel Computing Toolbox, which provides more efficient simulation using shared-memory parallelization of the calculations on multicore-processor CPUs. In CAS, a large amount of scenario cases were simulated, according to spill locations and streamflow scenarios.Figure 2a shows a hypothetical breakthrough curve (BTC) of the in situ river monitoring sensor from an instantaneous injection [50,51]. The BTC, which is a temporal distribution of contaminant concentration obtained from the monitoring sensor, consists of a rising limb, falling limb, and tail, as depicted in Figure 2a. In this study, the tail in the falling limb is defined as the portion of which the concentration is below the value of 0.1 maximum concentration of BTC as shown in Figure 2a. Although the ideal shape of BTC based on ADE is a bell shape, the actual shape of BTC in rivers is asymmetrical due to the complexity of the flow mechanism and the river morphology, including the storage effect in natural rivers. Furthermore, the BTC implies hydraulic and geometry characteristics due to the passive behavior of contaminants in the stream when the contaminants reach the in situ sensor. For this reason, the BTC can be used as relevant information to track contaminant source inversely. Therefore, in this study, the various features were extracted from the BTC, and those features served as input variables of the ML modeling for the development of the ITM, as shown in Figure 1. This approach enhances the accuracy of the ML model by removing the irrelevant information of the BTC, which also makes the models more efficient by reducing the dimension of input variables. Consequently, in this study, the BTC was characterized by 21 features, as shown in Figure 2 and Table 2.The features are categorized into shape, concentration, slope, time, integral, derivative, and phase features, as shown in Figure 2b–d. First, the shape features, which are widely used for analyzing the BTC [23,24,52,53], were calculated from the third and fourth temporal moments that indicate the asymmetry and peak of the BTC. The equations to calculate the features and the temporal moment are as follows:(5)mk=∫0∞tkC(x,t)dt
+        (6)Sk=m3m23/2
+        (7)K=m4m22−3
+        (8)σ=m2
+        where m is the temporal moment and k is the degree of the moment; the other notations are given in Table 2. The temporal moments were estimated using the trapezoid rule [23].Secondly, the slope features were applied to the segments of BTC, of rising limb, falling limb, and tail. The slope of the rising and falling limb was calculated by dividing the maximum concentration by the time variation of each part. These features indicate how quickly the contaminant increases and decreases. Thus, if advection is more dominant than dispersion, the peak concentration is increased, and the retention time is decreased, which is equivalent to the slope being increased. In particular, the magnitude of the storage zone effect from the contaminant retention is featured as the power-law shape, described in previous studies [54,55]. For this reason, the tail slope was calculated by the power of the equation from the power-law regression.Next, the time features include the standard deviation and duration of concentration. The Standard deviation quantifies the variance of BTC, which is calculated from the second moment, as shown in Equation (8). Moreover, durations refer to the time needed for the concentration to reach a specific percentage of maximum concentration, and the width of the rising and falling limb. The duration of a specific concentration indicates for how long the concentration stays above the reference concentration. The integral features are the area of each part of the BTC. In addition, we suggest a critical area where the maximum concentration passes from half the maximum concentration in the rising limb, which is defined as the most damaging area.On the other hand, the derivative and phase features were estimated in phase space, which generates the novel features from the time dependence of the scalar quantity [56]. The phase space was defined so that the concentration and the first derivative are coordinated, as shown in Figure 2c [57]. In this space, the absorption and desorption processes in chemical sensors were characterized. Therefore, we employed the maximum derivative value and the area of positive value in the phase space as features of the rising limb; moreover, we selected the minimum derivative value and the area of negative value in the phase space as features of the falling limb. The phase features can be defined as Equation (9):(9)P=∫C(ti)C(ti+1)DdC
+        where C(ti+1) and C(ti) are the concentration at time ti+1 and ti.In this framework, we focused on the optimal BTC features and ML models to predict the spill location and spill mass. We conducted six ML models, which consisted of three decision tree-based models: Decision tree (DT), Random Forest (RF), and XGBoost (XGB); two linear models: Ridge and linear Support Vector Machine; and a nonlinear SVM using the Radial Base Function (RBF) kernel. For the prediction of both spill location and mass, predictors were separately developed by classifiers and regressors of the ML models. First, the spill location predictor was developed by using a classification model, because the spill location is labeled as discrete integers, as shown in Figure 3c, which present the potential spill location. In contrast, the spill mass is represented by continuous values as quantities. Thus, the spill mass predictor was developed by using the regression algorithms. Additionally, although both predictors were trained by using the same BTCs at the monitoring point, the optimal BTC features to predict two target variables were investigated separately. All of these models were implemented as both regressors and classifiers using the Scikit-learn and XGBoost libraries in Python 3.7 (Python Software Foundation, Beaverton, OR, USA).The DT is a non-parametric model, and is used as both a classification model and a regression model [58]. This model divides the space of the input variable into multiple hierarchies according to the value of the output variable based on the tree structure. The prediction is performed by taking the mode or average of the output variables through the hierarchy. In the training process of this model, the splitting variables and the split nodes are determined by the Gini index, as given in Equation (10):(10)Gini=∑k=1Kp^mk(1−p^mk)
+          where K represents the number of classes in the label and p^mk is the proportion of the kth class in the node m.This model has the following advantages: (a) Ease of investigating the process of prediction; (b) Insensitivity to noise and truncated data; and (c) High efficiency—it takes a short time to build the model and gives a short-term prediction. Due to these reasons, DT is suitable to be applied to a chemical accident response system that requires rapid forecasting.In this study, the advanced DT-based algorithms, such as RF and XGB, were also developed to overcome the disadvantage of DT having a high variance of prediction. RF consists of ensemble learning by combining a large collection of DTs, and obtains the results by averaging or voting [59]. Specifically, each DT predictor of RF is developed from a random selection of samples and variables. This process is based on the Bagging (abbreviation for bootstrap aggregation) method proposed by Breiman [60]. It generates the sample by a bootstrap sampling, which samples randomly with replacement. Thus, the Bagging method with randomization reduces the variance of RF by reducing the correlation between the trees. With respect to regression, this model is performed by averaging the predictions of each DT. Otherwise, the classification model is performed by obtaining the majority class vote from the results of each DT.On the other hand, Chen and Guestrin [61] recently suggested XGB to improve the performance of DT. This model is also an ensemble learning method of DT, and appeared as the top model in various machine learning comparison studies [62,63]. The difference from RF is that XGB is based on the gradient boosting method. In the gradient boosting method, each DT of XGB is developed at an iteration to reduce the error. Thus, XGB integrated multiple DTs into one strong predictor having sequential structure with randomization. In comparison with the conventional gradient boosting method, XGB is the stepwise forward additive model by including a regularization term in the objective function. In addition, it automatically utilizes the multicore and distributed settings for an efficient training process [64,65].In the XGB, additive functions to predict the output voted or averaged by a collection F of k trees can be written as:(11)y^=∑kKfk(xi),fk∈FThe objective function with loss function and regularization term is used to correct the previous DT through the iteration for optimization, which is given by:(12)L(φ)=∑il(y^,yi)+∑kΩ(fk)
+          where l is a loss function that measures the error between the prediction value (y^i) and the target value (yi), and Ω(f) is a regularization term that describes the complexity of DT fk, which is defined as:(13)Ω(f)=γT+12λ‖w‖2Due to the complexity of learning all DT parameters at once, the prediction value (y^i) is given from additive training, which adjusts the current state for the iteration t from the previous iteration t − 1:(14)y^i(t)=∑k=1fk(xi)=y^i(t−1)+ft(xi)
+          where γ is the complexity of tree leaf in the DT, T is the number of leaves in the DT, λ is the scale parameter, and w is the scores vector of leaves in the DT. By substituting Equation (14) into Equation (12), the objective function is described as Equation (15). Then, the objective function can be simplified to Equation (16), by taking the second-order Taylor expansion:(15)L(t)=∑i=1nl(yi,y^i(t−1)+ft(xi))+Ω(ft)
+          (16)L(t)≅∑i=1n[gift(xi)+12hift2(xi)]+Ω(ft)
+          where gi=∂y^(t−1)l(yi,y^(t−1)) and hi=∂y^(t−1)l(yi,y^(t−1)).The SVM is a widely used algorithm for both classification and regression. The SVM uses a hyperplane determined by support vectors to classify labeled datasets, which determines the decision boundary of all classes [66]. An optimal hyperplane is a classification plane obtained from the maximum classification margin. It can be obtained from the decision function of SVM in Equation (17). The margin is 2‖w‖, which can be maximized by minimizing the ‖w‖2. Thus, the optimization problem can be transformed into a dual problem through the Lagrange optimization method (Equation (18)):(17)f(x)=w⋅x+b
+          (18)L=argLmax(∑i=1nαi−12∑i,j=1nαiαjyiyjxixj)
+          where αi refers to the Lagrange multipliers, and the constraints are αi≥0 and ∑i=1nαiyi=0.On the other hand, SVM can be transformed into a nonlinear predictor by mapping the features into a higher dimension space. This new space can be approximated by replacing the x in Equation (18) by the kernel function K(xi,xj):(19)L=argLmax(∑i=1nαi−12∑i,j=1nαiαjyiyjK(xi,xj))In this study, the Radial Basis Function (RBF) was used as the kernel function. The RBF can be written as follows:(20)K(xi,xj)=exp(−γ‖xi-xj‖2)Support Vector Regression (SVR), which was developed by Vapnik et al. [66], is a revised version of SVM to apply for the regression problem. The difference from SVM is that SVR solves Equation (17) to find an f(x) having at the most ε deviation from the target value yi. More detail of this regularization problem can be found in Awad and Khanna [67].The Ridge regression model is a regularized linear regression. This model reduces the overfitting results by adding the regularization term into the weight coefficient. Since the overfitting increases the weight coefficient, Ridge regression can obtain a more accurate weight coefficient that indicates feature importance. In Ridge regression, the regularization is performed by minimizing the squared sum of weights with the squared sum of errors:(21)w=argwmin(∑i=1nei2−λ∑j=1mwj2)
+          where w is the weight coefficient, e is an error, and λ is the scale parameter of regularization.In this study, the six ML models mentioned above were divided into two groups by the feature importance metrics: mean decrease impurity and weight coefficient. First, the mean decrease impurity was used in the DT-based models (DT, RF, XGB). In the single DT model, the amount of performance improvement in each split node was calculated by the mean decrease of the node Gini index (Equation (5)) classification. The regression performance was obtained from the mean residual sum of squares. In ensemble DT models, the feature importance of all DTs within the model were averaged. A detailed theoretical background can be found in [68]. Second, indicating the feature importance of SVM and Ridge, the square of the weight coefficient in Equation (12) is the distance of each variable margin in the classification model. In terms of the classification, this means that the bigger the margin, the more precisely the significant variable is classified. From the aspect of regression, a weight coefficient wi of variable i quantifies the effect on the prediction y^i, which indicates the feature importance of the regression predictor.In the suggested BTC features, not all features are relevant to predict the spill location and spill mass. The redundant features may increase the modeling complexity, as well as leading to a decrease in the accuracy of ML models [69]. Moreover, excluding the redundant features is necessary to clarify the relationship between the BTC features and the contaminant source. Note that the information of the BTC implies the hydraulic and geometry characteristic of the transported reach in the river, which dominates the mixing characteristics of contaminants. Therefore, to predict the contaminant source, we can expand the significant BTC features to the dominant hydraulic and geometry factor.In this study, recursive feature elimination cross-validation (RFECV) was employed to select the optimal feature sets of each model. RFE is a greedy algorithm to rank the features using the particular feature importance criteria of each model. This algorithm starts with a full set of features; it then removes the redundant feature repeatedly, until the model performance becomes poor. Then, the remaining features are selected as an optimal feature set. In addition, RFECV improves RFE with N-fold cross-validation, which can reduce the bias of the selected optimal feature set. As feature importance in RFECV, we utilized the feature importance criteria of each model for training each model by each selected feature set. RFECV was implemented using Scikit learn library in Python 3.7.Due to the different tasks of approximating a mapping function, classifier and regressor were judged by different types of criteria. With respect to classifiers, accuracy, specificity, and sensitivity were used to measure the modeling performance, as shown in Equations (22)–(24):(22)Accuracy=TP+TNTP+TN+FP+TN
+        (23)Specificity=TPTP+FP
+        (24)Sensitivity=TPTP+TNThe number of true negatives (TN), false negatives (FN), true positives (TP), and false positives (FP) were used as the main components of the suggested criteria. The accuracy, specificity, and sensitivity show the overall ratio of accurate, negative, and positive prediction, respectively [27,70].In the case of the regressor, R2 (coefficient of determination), Root Mean Square Error (MSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) were utilized to measure the quantitative error. The formulae are listed in Equations (25)–(28):(25)R2=1−∑i(yi−yi^)2∑i(yi−y¯)2
+        (26)RMSE=1n∑i(yi−yi^)2
+        (27)MAE=1n∑i|yi−yi^|
+        (28)MAPE=1n∑i|yi−yi^yi^|×100%
+        where yi is the actual value and yi^ is the prediction value. The RMSE is the square root of MAE, which has consistent units of target variables. The MAE is similar, which is calculated by the sum of the absolute error. The MAPE indicates a relative error, which is usually reported as a percentage. Regressors are ensured as better models when these criteria represent smaller values.The study site to apply the ITM framework in this study is the Gam Creek in Gimcheon City, South Korea. This river is located in the vicinity of an industrial complex, which poses a high risk of pollutant spill accidents. In addition, Figure 3 shows that it joins with the Nakdong River, where a large number of people and agriculture depend on the river as a water source. In terms of morphology, the Gam Creek is a typical braided river, of which the bed material is composed of sand substrate, and Figure 4 shows that the river contains plenty of storage zones, such as sand bars, vegetation, and side pockets.The tracer tests used for field validation of the ITM framework were conducted under different streamflow conditions in October 2019 and June 2020. Figure 3 shows that the tests were conducted in the reach of Point 16 (injection point) to Point 20 (monitoring point). A fluorescent dye, Rhodamine WT, was used as a tracer material, which is a widely-used conservative tracer [43,71,72,73]. In Test 1 and Test 2, 15 and 7.5 L, respectively, of 20% Rhodamine WT solution were injected. Multiple point injection, according to the lateral direction of the channel, was conducted to achieve full mixing conditions in the horizontal and vertical direction for one-dimensional mixing conditions in the real stream. In addition, the distance between the injection point (IP) and Section 1 (S1) was estimated using Equation (29) [74]:(29)L0=0.1(1n)2UW2Ez
+      where L0 is the distance from the injection point for complete mixing on cross-section, n is the number of injection points in the lateral direction, and Ez is the lateral mixing coefficient, which is estimated from Ez=0.15hU* [40].The Rhodamine WT was measured using YSI-600OMS fluorometry, and the concentration was calibrated using known concentration solutions in the range of 0 to 200 ppb. In order to obtain cross-sectional average concentrations, three or four sensors were installed laterally at uniform distance at all sites. Then, cross-sectional average concentrations were obtained by averaging the concentration data from all sensors in each section. Figure 4 is a photograph of Test 1 taken from a UAV, which was taken immediately after Rhodamine WT injection. In this figure, the anomalous spatial distribution was visualized with the storage zone effect from the sand bar, side pockets, and bridge piers. Due to these storage zone effects, the cross-sectionally averaged BTCs of Rhodamine WT showed a highly skewed and long-tailed shape. The discharge, velocity profiles, and water depth were measured using a Sontek Flowtracker acoustic Doppler velocimeter. The bottom slope was measured using a Sokkia GRX1 as Real-Time Kinematic-Global Positioning System (RTK-GPS). Table 3 shows the summarized hydraulic and geometry conditions of the field tracer tests. In Test 1, the discharge (Q) was six times larger than Test 2, so the mean width (W) and mean velocity (U) in Test 1 were greater than in Test 2. The tracer mass (M) was injected at twice the amount of Test 2 in Test 1. Figure 5 shows the BTCs of Test 1 and Test 2 at different distances downstream of the injection point. Although more tracers were injected in Test 1, the peak concentration was higher in Test 2 since the mass was diluted a lot due to the high discharge. Furthermore, the advection was more dominant, and dispersion was less than in Test 2, due to the high mean velocity.In order to generate a training dataset for the proposed framework, breakthrough curves for the chemical accident scenarios in Gam Creek were created using CAS with TSM. Figure 3c shows that the spill scenarios were developed at 30 potential spill locations along the Gam Creek. The Hwangsan Bridge and Gampo Bridge were used as monitoring points in the Gam Creek to build the two inverse-tracking models represented as Model 1 and Model 2, respectively. For various flow conditions, 450 streamflow scenarios were generated from the log-normally fitted distribution using 10 years of historical streamflow data from an observation station located at Daedong Bridge. The streamflow data from 1 January 2010 through 31 December 2019 was obtained from the GIS-based Water Resources Management Information System (WAMIS) in South Korea. Using these sampled streamflow scenarios as input variables, the HEC-RAS was simulated to calculate hydraulic and geometry variables (U, U*, A, h) for estimation of the TSM parameters by Equation (3). The river geometry data and the Manning’s n coefficient of each cross-section were collected from the Master plan reports of Gam Creek [75]. The constructed HEC-RAS geometry consisted of 180 cross-sections within 39 km reach length. Manning’s n coefficient ranged (0.024–0.033). The sinuosity (Sn), which is a constant value, regardless of flow condition, was also estimated by the HEC-RAS geometry.With respect to contaminant transport simulation, the total number of chemical accident scenarios was 13,500, which represented 30 potential spill locations for 450 streamflow scenarios. The spill mass was given to each scenario simulation from a randomly sampled value in the range of 0 to 10 ton. The spilled contaminants were assumed to be a conservative constituent that did not decay. In order to prepare simulation of chemical accident scenarios with the TSM model, the total model domain needs to be divided into sub-reaches having the same TSM parameter set. In this study, 48 sub-reaches were constructed by dividing the reaches into sections considering the river flow and geometric conditions, such as velocity, water depth, width, sinuosity, bridge, and tributaries. To achieve this, the averaged hydraulic and geometry variables were calculated to estimate the TSM parameters of the sub-reaches using empirical equations for TSM parameters (Equation (3)). Table 4 gives the statistics of the estimated reach averaged hydraulic, and geometry variables that served as input variables of Equation (3). In addition, Table 5 gives the estimated TSM parameters of each sub reach according to the streamflow scenarios. Notably, reasonable range values were calculated when compared with TSM parameters reported in previous studies [38,39,46].Moreover, the Froude number (Equation (30)) of all streamflow scenarios represents that only subcritical flows were generated. Among the sampled streamflow values, the flow condition was only close to the supercritical flow with a Froud number of 0.94 at the maximum value of 129.51 m3/s. With the recognition that supercritical flow occurs at flood season, future studies should consider the hydrodynamic simulation with unsteady flow with the precipitation. In terms of numerical stability, it is necessary for reliable results to estimate the numerical error of simulated chemical accident scenarios. Silavwe et al. [76] suggested that the Peclet number (Equation (31)) of the Crank-Nicolson method-based 1D ADE should be less than 2 to avoid numerical error. Additionally, Choi [77] performed a numerical error test with the same TSM model as this study. The numerical error test showed that when the Peclet number did not exceed 5, oscillation-free solutions were obtained. Based on these results, the generated chemical accident scenarios were numerically stable due to the Peclet number of simulated chemical accident scenarios being in the range (0.20 to 2.23), as shown in Table 5. In Table 5, Froude number and Peclet number are defined as:(30)Fr=UFgh
+        (31)Pe=UF⋅ΔxKFFor training datasets to build predictors for both spill location and spill mass, the BTCs at two monitoring points were extracted from chemical accident scenarios of Gam Creek. Then, from these BTCs, BTC features were extracted and labeled with their spill location and spill mass for supervised learning. The development of the suggested framework consists of two steps. First, RFECV was used to identify the optimal feature subset of the ML algorithms and develop predictors for spill location and spill mass. In this step, 80% of the BTC features dataset was used as a training dataset and 20% was used as a test dataset. Second, five-fold cross-validation was conducted on the dataset to compare the performance of each ML model by optimal feature subset selected by RFECV. In this study, using the Ridge, DT, RF, XGB, and SVM classifier, two inverse tracking models were built, depending upon the monitoring points: Model 1 (Gampo Bridge) and Model 2 (Hwangsan Bridge), as shown in Figure 3c. Field application of the trained ML models for spill location and spill mass using field tracer test data is described in Section 5.In order to investigate the relevant BTC features for inverse tracking the spill location and spill mass of contaminant source, the importance of BTC features was estimated using the suggested ML models. All the feature importance was calculated to relative importance in the range 0 to 1. The feature importance of Model 1, which covers a more extended domain than Model 2, was plotted in Figure 6. In this figure, the first three bars are DT-based models using reduction of the Gini index as feature importance criteria, while the next two bars are Ridge and SVM using the weight coefficients as feature importance criteria. The feature importance values obtained by the reduction of the Gini index and the weight coefficients tended to be inversely proportional.Figure 6a shows that for spill location prediction, the slope of the tail (St) proved to be the most crucial factor for the DT-based model. This feature represents the magnitude of the storage zone effect. The increase in the storage zone effect induces the long-tailed BTC, due to the trapping effect [78,79]. The duration above 50% and 10% of Cmax (T50, T10) were relatively important for SVM. The T50 depends on dispersion, while the T10, which is the time length of the tail, is largely affected by the storage effect. The maximum derivative (Dmax) was relatively important for Ridge and XGB. This feature represents the derivative value when the concentration increases most rapidly in the rising limb of BTC. This feature is dominantly affected by the advection. However, compared to the DT-based model, the importance of the features was generally low in general in the SVM and Ridge. All of the feature importance of SVM and Ridge was under 0.2.Figure 6b demonstrates that in spill mass prediction, the maximum concentration (Cmax) was the most important factor for the DT-based models. The distance and the falling limb area of the phase space (Pf) were also important features for the DT-based models. When the contaminant is spilled into the river, the Cmax of the contaminant cloud decreases as it is transported downstream from the spill point. Thus, the distance and Cmax can be judged as complementary factors to predict the spill mass. Additionally, the falling limb area of the phase space (Pf) represents the concentration reduction rate, which can be affected by the velocity and the storage zone. However, the slope of rising limb (Sr) and area of falling limb (Af) were most important for SVM and Ridge regression models. Furthermore, SVR has more highly important features than Ridge regression, such as maximum derivative (Dmax), and total area (A). This can be explained by the different method of regularization of both models, as described previously. Since the Ridge regression regularizes the weight coefficient (Equation (16)), the feature importance can be underestimated. Consequently, the Sr was the most important feature for spill mass prediction for SVM and Ridge regression models, and the Cmax is the most important feature for spill mass prediction for tree-based models.In Figure 7a,b, RFECV with five-fold cross-validation was conducted based on accuracy as a score to identify the optimal feature subset. Table 6 represents the optimal hyperparameter set and selected optimal features. The best hyperparameter was investigated by grid search in the range based on previous study [80,81]. Parameters not listed followed the default settings of the Scikit-learn and Xgboost libraries [61,82].The model performances were investigated through the three performance criteria described in the previous section. Table 7 represents the five-fold cross-validation results with all performance criteria as averaged values. From these results, DT ensemble models, RF and XGB, outperformed in all performance criteria: accuracy, sensitivity, and specificity all scored around 0.97, respectively. Meanwhile, Ridge and SVM-linear showed weak performance and produced a low-performance score. Moreover, the RF model not only showed the best performance, with an accuracy of 0.97, but also used only three and four variables as optimal features for Models 1 and 2, respectively. However, as the number of selected features grew, it showed overfitting. The results of SVM-RBF with a feature subset selected from SVM-linear showed that its performance was almost the same as the DT-based model, which is a significant improvement over the SVM-linear model. For most ML models, Model 2 showed better performance than Model 2, which means that the shorter the length of the model domain, the better the model performance.Spill mass models for Model 1 and Model 2 were also built by the Ridge, DT, RF, XGB, and SVR regression models, according to the monitoring points shown in Figure 2. Similar to the evaluation processes in the spill location models, RFECV and five-fold cross-validation were applied to find the optimal feature subset and, thus, optimal ML models. The results showed that among the ML models, RF shows the best accuracy of 0.97 (R2). RF also selected the smallest number of features: seven and six features in Models 1 and 2, respectively, as shown in Table 8. Unlike the spill location predictors, RF and XGB showed similar performance without overfitting, according to the number of selected features. In addition, the DT-based models also outperformed linear models, as shown in Figure 8. Table 9 summarizes the results of regression performance from the averaged five-fold cross-validation results with the four performance criteria. This table shows that RF and XGB outperformed the other ML model performances.The developed ML models were validated using the field tracer data obtained at Gam Creek. Among the measured Rhodamine WT concentration curves shown in Figure 5, the curves measured at Gampo Bridge were used as BTCs of the monitoring point of Model 2. Since the two tracer tests performed with different spill mass condition, the arrival time of Test 1 is earlier than Test 2 due to the faster flow condition, and the maximum concentration of Test 1 is lower than Test 2, because the flow rate of Test 1 was approximately five times that of Test 2. Compared with the synthetic BTC, the real BTC contained fluctuations due to channel irregularities and measurement error, as shown in Figure 5. Thus, this can cause a discrepancy with the BTC features of synthetic BTC, the validation of ML models with a field test is necessary.Figure 9 presents the prediction probability of ML models according to the potential spill locations. In this figure, we compared the ensemble DT-based models, RF and XGB, and SVM. In order to estimate the prediction probability, ensemble DT-based models estimate the mean predicted probabilities of the trees. The location probability of a single DT is the fraction of samples of the same location in a leaf. In SVM, the prediction probability was estimated by using Platt scaling, which fits the SVM output into probabilities by using an additional sigmoid function [83]. Both processes were achieved using the predict_proba (X) function in the Scikit-learn, which is a Python-based machine learning library.The results show that only the SVM-RBF and RF predicted the correct spill location, showing Point 15 with the highest probability. In the case of Test 1, RF predicted the true spill location with 61% of probability, indicating a higher probability than SVM-RBF of 50%. In the case of Test 2, the SVM-RBF predicted the true spill location as a probability of 55%. This is higher than RF, which had a probability of 34%. This result was obtained because the slope of the tail and the time features, which are a value for time without a concentration value, was important for the prediction of the spill location.On the other hand, the SVM-linear predicted the wrong location, and showed low probabilities for all locations. It can be seen that the linear model yields underfitting results, because the spill location and BTC features have a non-linear relation. However, the XGB, which showed similar accuracy to RF when validated with synthetic BTC, was rather poor in predicting the spill location. The results of XGB showed that in both cases, point 20, the closest location to the monitoring point (Gampo Bridge), was predicted as the spill location with 94 and 62% probabilities, respectively. This result implied that the trained model was overfitting. Additionally, it can be seen that RF is less sensitive to data noise than XGB because RF largely depends on time and slope features (St, Tr) that are less affected by noise. In conclusion, the XGB built a model that was too fit for the scenario-based training data set, and was not suitable in handling the field data. Hence, the parallel bagging method is more suitable to the application with field data, including more noise than the sequential boosting method in DT-based models. The noise is decreased in the bagging method by aggregating the single DT predictors in parallel.In summary, both SVM-RBF and RF possess stable predictions, even with real concentration curves from field tests. However, it can be concluded that RF is not only the most accurate, but also the most efficient, with the smallest number of BTC features, namely, 3–4, as compared to SVM by utilizing all BTC features.The spill mass predictors were also validated with the BTCs of the field tracer tests. The true spill mass values of Tests 1 and 2 were 3.48 and 1.74 kg, respectively. Table 10 demonstrates the true spill mass and estimated mass from RF, XGB, SVR-linear, and SVR-RBF. Additionally, the percent errors were used for comparison between ML models, as listed in Table 10. The prediction results show that for both tests, the XGB produced the smallest errors, while the estimations of the other models were found to involve high errors. Specifically, the SVR-linear diverged during the prediction. SVR-RBF showed better prediction results than the linear model, but both tests showed high errors. This means that the linear model is incapable of prediction with noisy data. Additionally, the SVR-RBF, which is well fitted with the BTCs of the scenarios, has no margin to be applied with the noisy data. In the case of RF, this model highly underestimated the spill mass close to 0 kg. From this result, it is evident that RF has low noise immunity, since the number of optimal features is small. In other words, RF is the same advanced DT model as XGB, but this model depends on only eight features, as described in Table 8. Thus, the high dependency on small features causes low noise immunity.Consequently, XGB is the most feasible ML model for the prediction of spill mass in the field. In contrast to the spill location prediction, the boosting method of XGB showed a better result than the bagging method of RF. Additionally, the results show that the larger the number of optimal features, the better for spill mass prediction to apply in the field.In this study, a practical framework of the Inverse Tracking Model (ITM) was developed to predict the spill location and mass of contaminants accidentally released into the river. In this framework, the numerical model of TSM was used to simulate the realistic BTCs of contaminant spill scenarios via reflecting a wide spectrum of river flow and mixing processes. From the contaminant spill scenarios, 21 features were extracted from the BTCs of a monitoring point, which indicate various characteristics of BTC. To build the optimal ML models for spill location and mass, we applied six ML models, and selected optimal BTC features using RFECV. The application and validation of the proposed framework were performed in Gam Creek, South Korea. From the results, the key conclusions and suggestions are as described below.In the development of spill location predictors, the ensemble DT-based model, RF and XGB, outperformed other ML models. Furthermore, RF was the most efficient model, with a minimum number of optimum features. Among features of BTC, the slope of the tail (St), which characterizes the storage zone effect, played a significant role in predicting the spill location. From this result, it is evident that the tail of BTC implies the characteristic of the reach where contaminant transported due to the storage zone distributed in the reach. The SVM-RBF showed less accurate results than DT-based models in scenario-based validation results. In the development of spill mass predictors, RF and XGB showed better performance than the other ML models.In the field application, for the prediction of spill location, the SVM-RBF was less affected by data noise of measured BTC from tracer tests than DT-based models due to the uniformly distributed BTC importance in field application. Nevertheless, from the aspect of the number of optimal features, RF was considered to be the most accurate and economical for the spill location prediction. For the prediction of the spill location, the XGB showed better field applicability than RF. In other words, the boosting method was more appropriate than the bagging method in the prediction of spill mass. Moreover, it could achieve more noise-immune models when using all BTC features.The proposed framework has an advantage in that only the observed BTC is needed to predict the contaminant source characteristics, with no requirements of hydraulic or geometry information. However, it also has the limitation that the range of potential spill mass values to build the model is uncertain.For future studies, some potential to improve the framework exists. First, the pulse injection should be taken into account for more various contaminant spill cases. Second, the unsteady flow with rainfall–runoff needs to be added into the contaminant spill scenarios. These improvements can be accomplished by minor modifications. Despite some remaining work for future study, the proposed framework will provide a practical and rigorous model for real-time application as a river accident response system.Conceptualization, S.K. and I.W.S.; methodology, S.K.; software, S.K.; validation, S.K., I.W.S. and H.N.; formal analysis, S.K.; investigation, S.K.; resources, S.K., S.H.J., D.B. and H.N.; data curation, H.N., S.H.J., D.B. and S.K.; writing—original draft preparation, S.K. and H.N.; writing—review and editing, S.K., I.W.S. and H.N.; visualization, S.K.; supervision, I.W.S.; project administration, I.W.S.; funding acquisition, I.W.S. All authors have read and agreed to the published version of the manuscript.This research was supported by a grant from the Korea Agency for Infrastructure Technology Advancement (KAIA), funded by the Ministry of Land, Infrastructure and Transport (Grant 20DPIW-C153746-02), and the BK21 PLUS research program of the National Research Foundation of Korea.Not applicable.Not applicable.This research work was conducted at the Institute of Engineering Research and Institute of Construction and Environmental Engineering in Seoul National University, Seoul, Korea.The authors declare no conflict of interest.Schematic of the Inverse Tracking Model (ITM) framework: (a) development, and (b) application; HEC-RAS (Hydrologic Engineering Center’s River Analysis System) and TSM (Transient Storage Zone model) are hydrodynamic and contaminant transport model; CAS is Chemical Accident scenario Simulator; DB is Data Base; BTC is BreakThough Curve; DT is Decision Tree; RF is Random Forest; XGB is Xgboost; Ridge is Ridge regression model; SVM is Support Vector Machine.(a) Typical BTC and BTC features for (b) shape, time, and slope, (c) derivative and phase, and (d) concentration and area.Site map showing (a) the location of the Gam creek region in South Korea; (b) the model domain and the location of urban areas that would potentially be subject to significant damage from the spilled contaminants near the region; (c) the potential spill locations and monitoring points developing ITM in the Gam Creek region; (d) the tracer test reach.Photograph of the behavior of a tracer cloud that depicts the storage zone effect in the Gam Creek test reach.BTCs of (a) Test 1 and (b) Test 2 at different distances downstream of the injection point.BTC feature importance by each ML model from Model 1 for (a) the prediction of the spill location and (b) the prediction of the spill mass.RFECV results of (a) Model 1 and (b) Model 2 for the spill location for finding the optimum number of features for each ML model.RFECV results of (a) Model 1 and (b) Model 2 for the spill mass for finding the optimum number of features for each ML model.Predicted probability of spill location by using ML models using the measured BTC at the monitoring point (Gampo Bridge); The red boxes show the two highest values.Derived power of TSM empirical equations by PCR.Symbols and descriptions of BTC features.Experimental condition of the field tracer tests in Gam Creek, South Korea.Statistics of the estimated hydraulic and geometry variables of 48 sub reaches from HEC-RAS model simulations according to the streamflow scenarios in Gam Creek.Statistics of the estimated TSM parameters of 48 sub reaches and the estimated non-dimensional parameters of the CAS scenarios in Gam Creek.Hyperparameter and optimal feature subset from RFECV of the spill location predictor.Abbreviations: DT is Decision Tree; RF is Random Forest; XGB is Xgboost; SVM is Support Vector Machine; RBF is Radial Basis Function.Validation results of the spill location prediction models.Hyperparameter and optimal feature subset from RFECV of the spill mass predictor.Validation results of the spill mass prediction modelsAbbreviations: RMSE is Root Mean Square Error MSE is Mean Square Error; MAE is Mean Absolute Error; MAPE is Mean Absolute Percentage Error.Predicted spill mass of ML models using the measured BTCs of Tests 1 and 2.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01024.txt

@@ -0,0 +1,17 @@
+The non-stationarity, nonlinearity and complexity of the PM2.5 series have caused difficulties in PM2.5 prediction. To improve prediction accuracy, many forecasting methods have been developed. However, these methods usually do not consider the importance of data preprocessing and have limitations only using a single forecasting model. Therefore, this paper proposed a new hybrid decomposition–ensemble learning paradigm based on variation mode decomposition (VMD) and improved whale-optimization algorithm (IWOA) to address complex nonlinear environmental data. First, the VMD is employed to decompose the PM2.5 sequences into a set of variational modes (VMs) with different frequencies. Then, an ensemble method based on four individual forecasting approaches is applied to forecast all the VMs. With regard to ensemble weight coefficients, the IWOA is applied to optimize the weight coefficients, and the final forecasting results were obtained by reconstructing the refined sequences. To verify and validate the proposed learning paradigm, four daily PM2.5 datasets collected from the Jing-Jin-Ji area of China are chosen as the test cases to conduct the empirical research. The experimental results indicated that the proposed learning paradigm has the best results in all cases and metrics.Pollution of the environment is one of the most serious issues facing humankind today, and badly polluted air can cause great damage in economics and people’s lives. According to the World Health Organization (WHO), it is known that almost 3 million children die every year from a range of problems caused by air pollution [1]. With the process of industrialization and urbanization, the air pollution is becoming increasingly serious and the hazy weather has grown rapidly, especially in developing countries. In recent years, the foggy weather in many areas of China have become increasingly serious. Since the beginning of 2013, sustained haze weather has turned Beijing-Tianjin-Hebei (Jing-Jin-Ji region) into heavy pollution region. Fine particulate matter is one of the key contributors that leading to air pollution and hazy weather. It carries many adverse health effects, such as respiratory diseases and premature death [2].Recently, increasingly countries have set up environmental monitoring systems, which can provide a large amount of PM monitoring data. However, PM data are affected by many factors and fluctuates greatly over time, making it very challenging to predict. Therefore, many models and tools have been developed to predict PM2.5 and other air pollutant concentrations to improve the accuracy of the predictions. These models can be generally categorized into physical, statistical and hybrid models. For example, physical methods can be used to simulate the processes of emissions, diffusion and transfer of pollutants through meteorological, emission, and chemical models [3,4,5]. Statistical methods which mainly include autoregressive integrated moving average model (ARIMA), artificial neural networks (ANN) and multiple linear regression (MLR) [2,6,7,8,9], have been broadly applied to the pollutant concentration prediction. For instance, Ref. [10] proposed a forecasting model based on MLR and bivariate correlation analysis to predict the annual and seasonal concentrations of PM10 and PM2.5. Ref. [11] studied the effects of meteorological factors on ultrafine particulate matter (UFP) and PM10 concentrations under traffic congestion conditions using the ARIMA model. However, in practice, most pollutant sequences are non-linear and irregular, which may involve the problem of non-linear dynamical systems, so these linear algorithms are still problematic in predicting PM concentration. On the contrary, using artificial neural network models to predict pollutant concentration can overcome the limitations of traditional linear models and handle nonlinear problems well. [12] developed extended model based on long-term and short-term memory neural network. The model takes into account the spatiotemporal correlation to predict the pollutant concentration and shows excellent performance. [13] applied cuckoo search (CS) to optimize BPNN to predict PM concentrations in four major cities in China.Recently, to predict air quality more accurately, many hybrid models have been proposed based on ensemble learning paradigms, data preprocessing techniques and heuristic algorithms. For example, Ref. [14] developed a new prediction model based on the multidimensional k-nearest neighbor model and the ensemble empirical mode decomposition (EEMD) method. Ref. [15] developed a novel hybrid model based on wavelet transform (WT) and stacked autoencoder (SAE) and long short-term memory (LSTM) to simulate PM2.5 at six sites in China. Ref. [16] developed a model based on a combination of WT and neural network algorithm to decompose the PM2.5 data and then perform sub-series prediction analysis and finally data reconstruction. Ref. [17] proposed a novel PM2.5 hybrid prediction model, which includes a new pre-processing method (wavelet transform and variational mode decomposition), using differential evolution (DE) algorithm optimized BPNN to predict each decomposition sequence. The drawback of the decomposition-based prediction model is that using a single method to predict all signal sequences. Since different decomposition sequences have different characteristics, a single model does not fit all the characteristics of the decomposition sequences [18]. Thus, ensemble prediction model integrated multiple single models will help avoid the shortcomings of a single model and further improve the prediction accuracy. Furthermore, many heuristic algorithms are used to help optimize the weight coefficients of the ensemble model. [19] developed an ensemble model based on differential evolution (DE) to determine the optimum weights for electricity demand forecasting. Ref. [20] employed the cuckoo search algorithm (CSO) to optimize the weight coefficients of ensemble model. Whale optimization algorithm (WOA), proposed by Ref. [21], is a novel heuristic algorithm by imitating whale behavior in nature. However, the WOA will encounter problems such as being stuck in a local optimal solution and slow in convergence, when solving more complex problems. Thus, a new improved whale optimization algorithm (IWOA) is proposed in this study to strengthen the local seeking capability of the WOA.Through the above analysis, considering the criticality of data pre-processing and the limitations of one single prediction model, a new hybrid decomposition–ensemble learning paradigm based on variation mode decomposition (VMD) and modified whale-optimization algorithm (IWOA) is introduced. First, the original PM sequence is decomposed into different VM sequences using VMD. Then, the weight-determined ensemble model, which optimized by IWOA, is employed to forecast each decomposition component. Finally, several prediction subsets are assembled into the final prediction result.The paper is structured as follows: in Section 2, several single forecasting models, the ensemble prediction theory and VMD, are introduced. In Section 3, the proposed decomposition–ensemble model is presented. In Section 4, the study areas and the evaluation criteria are described. In Section 5, the comparative results of the proposed model and other models is in conducted. Finally, in Section 6, the conclusions the important results of this paper are explicitly introduced.Four individual forecasting models, VMD, IWOA, which employed in the suggested ensemble model, are described as follows.In latest years, many prediction models have been developed and applied to PM2.5 concentration prediction. This paper uses four popular methods, BPNN, ANFIS, ANFIS-FCM and GMDH, which show good performance in PM2.5 prediction, to construct the ensemble models.The BPNN is a multi-layer feed-forward neural network, which is widely used in many fields. The BPNN algorithm needs to find the parameter with the minimum error, i.e., the minimum value of the error between the output value and the actual value according to the negative gradient direction. The process of the BPNN is mainly divided into update and learning stages:(1)kij(t)=wij(t−1)−Δwij(t)
+          (2)Δwij(t)=η∂E/∂wij(t−1)+α⋅Δwij(t−1)
+          where η denotes the learning speed, wij represents the weights between nodes i and j, α denotes the impulse parameter, E denotes the error super curve face and t denotes the current iterative steps.Ref. [22] proposed the ANFIS that combines the blur systems and neural networks. It plays the advantages of both and makes up for the shortcomings of each. ANFIS can form an adaptive neuro-fuzzy controller by using a neural network learning mechanism to automatically retrieve rules from the input and output sample data. Through the offline training and the online learning algorithms, it can create fuzzy inferences and control the self-adjustment of rules, thereby making the system itself develop towards adaptive, self-organizing, and self-learning. ANFIS includes five-layer network, and each layer contains several node functions.Layer 1: This process is the fuzzy layer. Each node in layer 1 is adaptive, all have node function, and will generate membership of a fuzzy set.
+          (3)Oi1=μAi(x)
+          where x denotes the input to node I, and Ai denotes the language label associated with the function of this node. The “μ” denotes the membership functions for Ai, which described by generalized Gaussian functions.Layer 2: In this process every node is a circular node labeled ∏ out, i.e., ∏-norm operation:(4)Oi2=μAi(x)+μBi(y),i=1,2Layer 3: At each node in this layer, the ratio of the firing weights under the ith rule to the sum of the firing weights under all rules is calculated:(5)Oi3=w¯i=wiw1−w2,i=1,2At this level, all outputs are collectively referred to as normalized emission intensity.Layer 4: In this process, the contribution of the ith rule to the overall output is calculated:(6)Oi4=w¯ifi=w¯i(aix+biy+ci),i=1,2
+          where w¯i represents the out of layer 3, and (aix+biy+ci) is the parameter set.Layer 5: In this layer, the signal node ∑ calculate the final output as the sum of all incoming signals:(7)Oi5=∑iw¯ifi=∑iwifi∑iwiThe final output of the adaptive neural fuzzy inference system is:(8)fout=w¯1f1+w¯2f2=w1w1+w2f1+w2w1+w2f2=(w¯1x)p1+(w¯1x)q1+(w¯1x)r1+(w¯2x)p2+(w¯2x)q2+(w¯2x)r2The FCM is a type of data aggregation method. In the FCM method, each data point needs to be classified as a level assigned at the member level in the cluster. FCM divides a selection of n vector xi, (i = 1, 2, …, n) into fuzzy groups, and then finds a clustering center in each fuzzy group in a way that minimizes the cost function of the similarity measure. The above i = 1, 2, …, c represents random sampling from n points. Here is a brief introduction to the stage of the FCM algorithm. First, choosing the centers of cluster ci, (i = 1, 2, …, c) from the n points (x1, x2, x3, …, xn) randomly. Second, the membership matrix U, is calculated by using the subsequent equation as follows:(9)μij=1∑k=1c(dijdkj)2m−1
+          where dij=‖ci−xj‖ is the Euclidean distance which involves the i-th cluster center and the j-th data point, and m is the fuzziness index. Third, compute the cost function using the following formula.
+          (10)J(U,c1,…,c2)=∑i=1cji=∑i=1c∑j=1nμijmdij2Stop the process when it falls below a certain threshold. Additionally, finally, the new c fuzzy clustering center ci, i = 1, 2, …, c is calculated using the following equation:(11)Ci=∑j=1nμijmxj∑j=1nμijmThe GMDH is a series of computer-based inductive algorithms for the mathematical modeling of multi-parameter data sets. It is characterized by a fully automatic structure and parameter optimization of the model. The GMDH can be used for data extraction, knowledge detection, prediction, modeling of complex systems, optimization, and pattern recognition [23]. The GMDH algorithm features an induction procedure to classify increasingly complex multinomial models and select the best solution by the externality criterion.The GMDH pattern generally have multiple sets of inputs and one set of outputs, and is a subset of the components of the basic function:(12)Y(x1,…,xn)=a0+∑i=1maifi
+          where a denotes coefficients and f denotes the fundamental function that depends on different inputs, m represents the number of fundamental function components.The basic function (12) is called the partial model and the GMDH considers various subsets of this function and thus finds the optimal solution. The coefficients of the model are first estimated using the least squares method. Then, the number of local components of the model is gradually increased. Finally, the GMDH algorithm finds the best complexity model structure by minimizing the external criterion. This process is called self-organizing of the model.The main idea of the GMDH neural network learning algorithm is as follows: a series of source neurons are generated by performing cross-combining on each entry unit of the system, and the mean square error of the output error corresponding to each neuron is calculated; then, several outputs are chosen from the created neurons with smaller mean square error than a predetermined threshold, the selected neurons are used as the input unit of the new generation; the process of survival of the fittest and gradual evolution is repeated until the new generation of neurons is no better than the previous generation.The VMD is a non-recursive signal treatment algorithm that decomposes the original signal into a family of patterns with a specific frequency spectrum domain bandwidth [24]. During the decomposition process, each pattern can be compressively pulsed around a certain center. If the bandwidth of each pattern is required, three steps should be completed. After that, a constraint variational problem can be given. The details of VMD are described in [25].An improved heuristic algorithm IWOA is developed to enhance the performance of the ensemble model. The IWOA determines the optimal weight coefficient of the ensemble model. The basic whale optimization algorithm (WOA), chaotic local search (CLS), and the WOA modified by CLS will be described below.The WOA, put forward by S. Mirjalili in 2016, is a simulation of the hunting mechanism of humpback whales, called bubble-net feeding method. Humpback whales form distinctive bubbles by circling around their prey in a circular or “9 shaped” path during foraging. With special exercises, humpback whales first form a spiral bubble 10-15 m below their prey while swimming upstream to the surface. Then, it surrounds the prey with its flashing fins to prevent it from escaping and catch it [21]. The mathematical principles of the above humpback whale behavior are described as follows:(a) Encircling prey:The humpback whale orbits its prey, and updates its position to the best search agent as the number of iterations increases. It can be depicted mathematically as:(13)D→=|C→X*→(t)−X→(t)|
+          (14)X→(t+1)=X*→(t)−A→⋅D→
+          where X* denotes the position vector of the best solution obtained thus far, X→ is the position vector, A→ and C→ denote the coefficient vectors, and t denotes the current iteration.(b) Bubble-net attacking method:The mathematical modeling of humpback whale’s vesicular behavior is designed for the following two methods. 1. Shrinking encircling mechanism: This is a bracket predation mechanism that requires finding a new agent location, which can be anywhere between the agent’s original location and the current optimal agent location. The values of A→ in this process is in the interval [−1, 1].2. Spiral update position: A spiral equation needs to be created between the positions of both the prey and the whale to mimic the spiral motion of the humpback whale, as shown below.
+          (15)X→(t+1)=D′→⋅ebl⋅cos(2πl)+X*→(t)The probability p, a random number in [0,1], is assumed to select between the shrinking encircling and the spiral-shaped path during the optimization process.(c) Search for prey:During the search phase, the variation of vector A→ can also be used to find prey at random. Therefore, to move away from a reference whale, A→ can be utilized with the random values greater than 1 or less than −1. The mathematical model at this stage is as follows:(16)X→(t+1)=Xrand→−A→⋅D→D→=|C→⋅Xrand→−X→|
+          where Xrand→ is a random position vector (random whale) selected from the current population.As mentioned earlier, WOA was recently proposed and widely used in many fields. However, it also has shortcomings, like slow conversion in the late stage and easy to fall into a local optimum. Additionally, the chaotic local search (CLS), based on chaotic search, can effectively avoid the local optimization and converge to the global optimization. The blending of WOA and CLS can help improve global conversion and prevent falling into local solutions. To accelerate the local convergence of WOA, the chaotic local search algorithm is also applied. When WOA finishes iterating to find the best solution, the acceptance of these new solutions as determined by CLS will perform to local search a better solution close to the best solution. A logistic equation applied in CLS is defined as follows:(17)cxiiter+1=μcxiiter(1−cxiiter)
+          where cxi denotes the ith chaotic variable, iter is the iteration number. When μ=4, the above equation exhibits chaotic dynamics, cxi denotes range in (0,1) and cx0∉{0.25, 0.5, 0.75}. For more details about CLS, please refer to [26].The pseudo-code of the IWOA algorithm is outlined as follows:
+		  Algorithm: Improved whale-optimization algorithm (IWOA)Objective:Minimize and maximize the objective function f(x), xi=(xi1, xi2, …, xid)Parameters:iter-iteration number.Maxiter-the maximum number of iteration.I-a population pop.p-the switch probability1. /*Initialize a population xi=(xi1, xi2, …, xid)2. WHILEiter < Maxiter3. FORi = 1 to I Update A→, C→, l and p4. IFp > 0.55. IF(|A→|<+1)6. Update the position of the current solution by Equation (14)7. ELSE IF(|A→|>+1)8. Randomly choose a search agent9. Update the position of the current search agent by Equation (16)10. END IF11. ELSE IFp > 0.512. Update the position of the current search by Equation (15)13. END IF14. END FOR15. /*Jump out of local optimum by using chaotic local search. */16. Calculate cxiiter=xiiter−ximinximax−ximin17. Calculate the next iteration chaotic variable by Equation (16)18. Transform cxiiter+1 for the next iteration xiiter+1=ximin+cxiiter+1(ximax−ximin)19. /*Evaluate xiiter replace xiiter by xiiter+1 if the newly generation is better. */20. /*Find the current best solution gbest*/21. iter = iter + 122. END WHILEIn this part, we suggest a new hybrid decomposition–ensemble learning paradigm that integrates VMD method, several prediction models and IWOA optimization. The main process of the developed decomposition–ensemble paradigm is shown in Figure 1. The three main steps of the ensemble model are as follows:Step 1: Decomposition process:First, the features and noise of the original pollution data needed to be cleaned and processed so that an effective prediction model could be built. In this study, VMD technology was used to disaggregate the original pollution datasets into a set of VMs and the residue component with corresponding frequencies.Step 2: Ensemble forecasting and IWOA optimization:The decomposition sequences with different characteristics were obtained via the VMD process. However, different sequences had different properties, which meant that a single prediction method could no longer effectively adapt to all the characteristics of the VMs. Thus, the ensemble strategy is adopted to solve this problem, and can be described as that if there are M types of prediction methods with the correct selection of weight coefficients to solve a problem. The results of multiple models were added together. Assume that Emodel (Model = “BPNN”, “ANFIS”, “ANFIS-FCM”, “GMDH”) is the ensemble prediction result of each VM by using the above methods. Then, using IWOA to optimize the output of the Emodel, it can be expressed as
+      (18)OutputSCWOA−NNCTVMs=w1×Pmodel1+w2×Pmodel2+w3×Pmodel3…
+      where wi (i = 1, 2, …, N) is the weight coefficient of the model N. wi∈[−2, 2] is the range of weight coefficients by NNCT [27].To improve the optimal weight coefficients wi(i = 1, 2, …, N), IWOA was employed to find the optimal solution for the ensemble weight coefficients. Before optimization, the objective equation needed to be confirmed first. The objective function of this paper is set by Equation (18). When the predefined minimum value of the objective function or the maximum iterations was reached, the optimization process was terminated. Nevertheless, the search boundary of the WOA is set to [−2, 2], the nesting dimension is 5 and the maximum number of iterations is 500.Step 3: Assemble forecasting results:Through the above steps, the overall prediction results of the VMs were obtained. Then, the prediction results were combined to obtain the final result.In this paper, the PM2.5 concentration data from the Environmental Protection of the People’s Republic of China (http://www.mep.gov.cn/) were collected to verify the performance of the proposed model. The selected daily PM2.5 concentration sample data are for Beijing, Tianjin, Baoding and Shijiazhuang from 1 August 2015 to 31 August 2017. The total data number of daily PM2.5 concentration for each city were 763. In each experiment, the first 572 data (approximately 75% of the total data) of each VM were used for training subsection, and the rest were the test subsection. When all predicted VMs were integrated into the overall result, the 189 pieces of data (about 25%) in the test results were used for optimizing the weights of the ensemble model and the rest were used for model testing.To effectively assess the prediction performance of the developed model, four popular error criteria, shown in Table 1, were employed to assessment the prediction capacity of the developed model. Smaller values denote better prediction performance.Here, yn and y^n present the actual and predicted values at time n, respectively. N denotes the sample size.In the proposed VMD-IWOA ensemble model, the original PM2.5 concentration sequence is first decomposed into several independent VMs by using VMD. However, too many VMs introduce new problems. During the integrated prediction process, each VM generates estimation errors, and too many VMs cause an accumulation of errors. It also increases the time consumed in a single prediction step. To prevent the above problems, the entire VMs were restructured into three VMs and a residual.The BPNN, ANFIS, ANFIS-FCM and GMDH prediction models were applied to forecast each VM, which reconstructed in Section 5.1. Additionally, then, the ensemble model integrates the results of the four prediction models on each VM, and optimizes the weights of the four prediction results based on IWOA. Before the simulation, the parameters of the four neural network model need to be initialized. The input nodes of the neural network are set to four, the hidden nodes to nine and the output nodes to 1. Besides, the rolling single-step forecasting operation method based on PM2.5 concentration data of four cities is used to test the predictive performance. The detailed experimental parameters of the four neural networks are shown in Table 2.Table 3 shows the prediction results of the single models and the proposed ensemble model for each VM. To evaluate model performance, the RMSE was utilized as a model evaluation index. As can be seen from Table 3, each model performed optimally predictive behavior at a particular VM. For instance, the experimental results in Beijing were shown as follows: the BPNN provides the lowest RMSE values among all single models at VM2 and VM3, while at VM1 and residual, GMDH has the lowest RMSE values. The prediction results in Tianjin show that the ANFIS presents the best results at VM1. The FCM performs best at VM3. At VM2 and Residual, the GMDH provides the best results. The experimental results in Baoding show that among all of the single models, the RMSE value was lower than those of the other methods at VM1 and Residual, when the ANFIS was applied. At VM2 and VM3, the GMDH presents the optimal results. The forecasting results in Shijiazhuang reveal that the GMDH performs better than the others at VM2, VM3 and residual while ANFIS performs the best at VM1.Based on the above analysis, it can be revealed that each model has its advantages on the particular VMs. A single prediction model cannot be used to predict all decomposition signals uniformly. Thus, the most suitable model is selected according to the different conditions, which reveals that an ensemble model can incorporate the virtues of multiple individual models to overcome the limitations of individual models. Therefore, this study proposed an ensemble model based on the IWOA to seek the best weight coefficients of the ensemble model. The searching boundary is set in [−2, 2] based on the NNCT, and the RMSE criteria is used as fitness function of IWOA. Table 3 presents the best weights and final results of the ensemble model. By comparing with each single model, it indicates that the developed ensemble model can give the desired prediction results.Comparing the ensemble model with BPNN, ANFIS, FCM and GMDH, the average RMSE of four cities at VM1 was reduced by 26.10%, 2.62%, 7.80% and 3.97%, respectively; At VM2, the average RMSE of four cities was reduced by 5.81%, 11.15%, 11.51% and 3.59%; At VM3, the average RMSE of four sites was reduced by 7.19%, 58.21%, 13.92% and 6.22%, respectively. For Residual, the average RMSE of four sites was reduced by 17.79%, 33.09%, 22.27% and 7.51%, respectively. Consequently, it can be seen that compared with the single models BPNN, ANFIS, FCM and GMDH, the forecasting result of the ensemble model is significantly improved on each VM component.To evaluate the proposed ensemble model, three types of model comparison experiments were designed to compare the proposed ensemble model with other individual models, VMD-based models, and existing benchmark models.The experiment compares four VMD-based prediction models with the developed ensemble model. The four VMD-based models are VMD-BPNN, VMD-ANFIS, VMD-FCM and VMD-GMDH, which were constructed to emphasize important usages of the data decomposition technology. The corresponding improvement of the developed ensemble model and the VMD-based models are shown in Table 4 and Figure 2. By comparing the ensemble model with the VMD-BPNN, VMD-ANFIS, VMD-FCM and VMD-GMDH, we can conclude that the ensemble model significantly outperforms the other VMD-based models according to four evaluation criteria. For example, in Beijing, the ensemble model leads to 2.3843, 10.6660, 3.6867 and 2.1953 reductions in MAE, 5.4454, 21.3895, 11.6926 and 9.7510 reductions in RMSE, 0.3159, 11.9748, 12.3061 and 5.3553 reductions in MAPE, 5.2795, 21.3508, 11.6465 and 9.6318 reductions in TIC to compare with VMD-BPNN, VMD-ANFIS, VMD-FCM and VMD-GMDH, respectively. In addition, Figure 2 illustrates the comparison of actual values and the forecast values. The predicted results from the developed ensemble model are better than other VMD-based models.This experiment used four individual models to make comparison with the developed ensemble model. The four individual models are BPNN, ANFIS, FCM and GMDH. Table 5 indicates the comparison forecasting results between ensemble model and other single models. From Table 5, by comparing the ensemble model with the BPNN, ANFIS, FCM and GMDH, there are significant improvements in the predictions of the proposed model. For example, in Beijing, the ensemble model leads to 66.4829, 71.3965, 67.8848 and 82.5946 reductions in MAE, 65.7865, 73.3943, 7.7401and 81.1458 reductions in RMSE, 67.7547, 71.7270, 67.5358 and 83.5715 reductions in MAPE, 65.6355, 73.1804, 67.5553 and 80.7598 reductions in TIC to compare with BPNN, ANFIS, FCM and GMDH, respectively. Besides, Figure 3 presents the comparison between the actual values and the forecast values. The forecast results from the developed ensemble model are better than other single models.This part was conducted to further verify that the suggested hybrid decomposition–ensemble method can effectively improve performance prediction. Several existing models widely used in environmental prediction were applied to conduct comparative studies to access the suggested models. The existing models include two simple algorithms (i.e., ARIMA and RBFNN) and three hybrid algorithms (i.e., SSA-ENN, EEMD-GRNN and EEND-WOA-BPNN). The results of the comparative study are given in Table 6 and Figure 4. It can be seen from Table 6 and Figure 4 that the values of MAE, RMSE, MAPE and TIC of the developed model are all lower than the other existing models, which further shows the prediction performance of the developed ensemble model has obvious advantages. For example, comparing the proposed model with ARIMA, RBFNN, SSA-ENN, EEMD-GRNN and EEND-WOA-BPNN, the MAPE of Beijing was reduced by 94.01%, 91.14%, 90.22%, 86.73% and 69.20%, respectively. For Tianjing, the average MAPE of was reduced by 92.13%, 89.47%, 89.39%, 82.92% and 57.96%, respectively. For Baoding, the average MAPE was reduced by 94.01%, 91.14%, 90.22%, 86.73% and 69.20%, respectively. For Shijiazhuang, the average MAPE was reduced by 94.01%, 91.14%,90.22%, 86.73% and 69.20%, respectively.In addition, the error mean and error STD are also used to evaluate the models’ accuracy and stability, and the results shows that the developed model has higher accuracy and stability than other existing models. Therefore, it can be concluded that the proposed ensemble model can be successfully and effectively employed for PM2.5 concentration prediction compared with existing models. Furthermore, the proposed ensemble model has the following highlights compared to previous works [15,16]: 1. the data decomposition; 2. multi-model integration prediction; and 3. the optimized ensemble pattern weighting coefficients.Reliable and precise PM2.5 concentration forecasting is important for air quality early warning and pollution control. Owing to uncertainties and unstable of the PM2.5 datasets, the original PM2.5 series are very difficult to forecast accurately. Thus, it is still a challenging task to predict and simulate the PM2.5 reasonably. In this study, a new hybrid decomposition–ensemble learning paradigm, which based on variation mode decomposition (VMD) and modified whale-optimization algorithm (IWOA), is proposed to predict the PM2.5 concentration. In this developed paradigm, the VMD method was employed to decompose the original PM2.5 sequence into several VM series for forecasting. The prediction results show that the single prediction model used for pollution concentration prediction has limited capability and is not appropriate for all VMs. To this end, an ensemble model, based on four individual forecasting approaches, BPNN, ANFIS, FCM and GMDH, is proposed for predict all the VM components. Furthermore, in order to ascertain the best ensemble weight coefficients, an improved Whale Optimization Algorithm, named IWOA, is proposed and the final forecasting results were achieved by reconstructing the precise sequence. The main contributions of this paper are summarized as follows: (1) A new decomposition–ensemble learning paradigm is developed for PM2.5 concentration forecasting. (2) The VMD technique is adopted to decompose the primary PM2.5 series. (3) ANFIS, ANFIS-FCM and GMDH are utilized for PM2.5 forecasting. (4) An improved heuristic algorithm, IWOA, is developed to improve the weight coefficients of the ensemble model.To evaluate the developed model, daily PM2.5 sequence from four cities located in Jing-Jin-Ji area of China were collected as the test cases for the comparison study. The comparison results indicated that the developed ensemble model is superior to comparison models, include four VMD-based models, four individual models, two benchmark models and three existing models. Thus, the developed ensemble model provides an effective forecasting ability, especially for the highly volatile and irregular data (e.g., PM2.5 concentration) and can be a powerful tool for decision makers in air quality monitoring and early warning system.Conceptualization, H.G. and X.H.; Methodology, H.G. and X.H.; Investigation, Y.G., W.Z. and Z.Q.; Data curation, Y.G., W.Z. and Z.Q.; Writing-original draft preparation, H.G. and X.H.; Writing-review and editing, Y.G., W.Z. and Z.Q. All authors have read and agreed to the published version of the manuscript.This research was funded by the Key Project of China Ministry of Education for Philosophy and Social Science: Big Data Driven Risk Research on City’s Public Safety (Grant No. 16JZD023) and the National Natural Science Foundation project (41225018).In this paper, the PM2.5 concentration data from the Environmental Protection of the People’s Republic of China (http://www.mep.gov.cn/) were collected to verify the performance of the proposed model.This work was supported by the Key Project of China Ministry of Education for Philosophy and Social Science: Big Data Driven Risk Research on City’s Public Safety (Grant No. 16JZD023) and the National Natural Science Foundation project (41225018).The authors declare no conflict of interest.The main procedures of the combinatorial model proposed in the paper. Step 1: Variation mode decomposition (VMD) process; Step 2: Ensemble forecasting on variational modes (VMs); Step 3: Assemble forecasting results.Forecasting results of the ensemble model and VMD-based models.Forecasting results of the ensemble model and individual models.Forecasting results of the ensemble model and existing models.Four evaluation rules.Experiment parameters of artificial neural networks (ANNs).VMs forecasting results of the individual model and ensemble models in four cities.The results of the ensemble model and other VMD-based models at four cities.The results of the ensemble model and other single models at four cities.Comparison of prediction performances with existing models.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01025.txt

@@ -0,0 +1 @@
+The super-aged society of Japan is experiencing an increased demand for dental hygienists, of which there is currently a shortage. This study aimed to investigate the factors related to the working status of dental hygienists in Japan. We conducted a survey by mailing a questionnaire on employment to 1444 dental hygienists and obtained 537 valid responses. We conducted a bivariate analysis using either a chi-square test or t-test as well as a multiple logistic regression analysis to determine the factors related to working status. The overall employment rate was 68.2%, with a significant difference between age groups (p < 0.01). Approximately 80% of respondents considered working hours and human relations at the workplace to be important, and more than 70% of respondents considered wages as important. Finally, the following four variables were found to be significantly associated with employment status: training course attendance, a desire to work full-time, consideration of wages as important, and consideration of working hours as important. These findings suggest that it is necessary to improve working conditions and environments, including wages and working hours, as well as provide a more robust system of continued professional development for dental hygienists to increase the workforce.The number of teeth in the oral cavity decreases with age, and a decline in masticatory and swallowing functions, which is not seen in younger generations, is frequently observed in the elderly. In a society with a markedly aged population, it is necessary to provide new oral health services that are tailored to the oral conditions of the aged. Many studies have found that good oral health in the aged contributes to improved general health and quality of life (QOL) [1,2]. The World Health Organization (WHO) has indicated that oral health is one of the behavioral determinants of active aging [3]. Further, some epidemiological studies have reported that declining oral function may be associated with malnutrition in aged individuals [4,5]. In Japan, oral health measures for the aged play a major role in national health promotion policies, such as the establishment of goals related to the oral function of the aged in the national health plan [6]. These changes in oral health due to aging of the population have a significant impact on the work of dental professionals.Due to the aging of the global population, Federation Dentaire Internationale (FDI) launched The Oral Health for an Ageing Population (OHAP) project in 2015 to strengthen the role of the oral health community in achieving health longevity [7]. Usui et al. conducted a survey of dental experts on areas of dental need that were expected to increase in the future in Japan. The study reported that a high percentage of respondents cited “home dentistry”, “dentistry for the aged”, “feeding and swallowing”, “regenerative dentistry”, and “preventive dentistry” as the key areas [8]. Additionally, oral care services provided regularly by dental professionals have a significant effect on reducing aspiration pneumonia [9]. The provision of oral health services is fundamental to maintaining the health of the elderly.It has been reported that oral health care provided by dental hygienists is also effective in enhancing swallowing functions [10]. In a recent epidemiological study, the introduction of an oral function improvement program for the aged reduced the risk of frailty [11]. Thus, dental hygienists play an essential role in the implementation of oral health care services for the aged. As the number of elderly people increases, dental hygienists need to be adequately deployed in the community to ensure continued oral health care. In addition, it is necessary to supply sufficient professional education to dental hygienists concerning elderly care to provide safe oral health services to the aged with underlying diseases. In Japan, the percentage of elderly aged 65 years and above had already reached 28.4% in 2019. This is an extremely high percentage, even among developed countries [12] and, consequently, there is a great need to increase the workforce of dental hygienists.The prevalence of dental caries in children has decreased significantly in Japan. In addition, the proportion of elderly individuals who still have their original teeth is increasing each year. According to the 2016 Survey of Dental Diseases in Japan, 51.2% of people at 80 years of age have 20 or more of their original teeth remaining. However, the prevalence of dental caries and periodontal disease among the aged remains high. Thus, there is a strong need to expand the system for providing lifelong oral health services. Furthermore, oral health services for the aged requiring long-term care are provided by the long-term care insurance system that started in 2000. When a dental hygienist under the direction of a dentist provides oral care to aged people who require nursing care, the nursing care insurance provides additional benefits. If a dentist or dental hygienist provides technical advice and guidance on oral care to nursing staff working in a residential facility, they can also receive additional payment under the nursing care insurance system.As of 2018, the total number of employed dental hygienists in Japan was 132,635, of which only 73 were men. In Japan, the Dental Hygienist Act granted dental hygienist licenses to mainly women until 2013. As a result, 99.95% of employed dental hygienists were women. There is a significant shortage of dental hygienists who can provide adequate dental health services to all generations [13]. Thus, there is an urgent need to expand the workforce of dental health hygienists. However, the ratio of employed dental hygienists to the total number of registered dental hygienists was merely 46.0% in 2018 [14]. To address the shortage of dental hygienists in Japan, it is important to establish work policies that allow registered dental hygienists to continue working. However, few studies have explored the employment status of dental hygienists in Japan.In addition to basic conditions of employment, such as wages and working hours, related factors such as marriage and childbirth may affect employment since most dental hygienists in Japan are women. Furthermore, acquiring adequate professional education for rendering dental hygiene services is another factor affecting employment status. Therefore, the purpose of this study was to investigate factors related to the working status of dental hygienists in Japan by comparing the multiple factors between employed and unemployed dental hygienists.This study was a self-administered questionnaire survey for members of three alumni associations of dental hygiene training schools; it was not a national survey. Since all members of the alumni association were women, no gender analysis was conducted in this study. The distinction between unemployed and employed was based on whether the respondents were engaged in dental hygiene work at the time they answered the questionnaire. Those who were engaged in some occupation but not engaged in dental hygiene work were classified as not employed.The study included three alumni associations representing dental hygienist training schools in three different regions (Kanto, Kyushu, and Hokkaido) of Japan. All three alumni associations have been in existence for more than 40 years. We conducted a survey using a self-administered questionnaire to individuals aged 20–59 between August 2018 and December 2019. Since only 4.5% of employed dental hygienists in Japan are over 60 years old, the target age range for analysis was defined. The questionnaires were sent to 1444 alumni members by mail, and the secretariat of each alumni association executed the stratified random sampling method. The survey organizer received stickers printed with the sampling results, which were used to send the questionnaire to the alumni members. The respondents completed the questionnaires anonymously.Based on the national survey items concerning employment [15], we examined the following: the subjects’ basic characteristics, employment status, working condition preferences, and participation in professional skill development training programs within the past year. We obtained basic demographic information such as the respondents’ age, marital status, and number of children. The respondents were asked about their current and desired employment status (full-time or part-time). The respondents were also asked to indicate whether they assign priority to each of the following five items using a binominal scale: wages, working hours, working location, job description, and human relations at the workplace.After examining the descriptive statistics of each variable, we explored the associations between employment status and each variable by conducting a bivariate analysis using either a chi-square test or unpaired t-test. Next, we conducted a multiple logistic regression analysis (likelihood ratio, forward–backward stepwise selection method) to include the independent variables that showed a significant association to employment status in the bivariate analysis. We obtained the odds ratios, which were adjusted for the inter-relationship between the independent variables. We also included the alumni association category as an independent variable in the analysis since we incorporated responses from three different alumni associations in this study. Finally, we identified the factors that were closely related to the employment of dental hygienists in Japan after a multiple logistic regression analysis. The significance level was set at p = 0.05. IBM SPSS Statistics for Windows Version 26.0 (IBM Corp., Armonk, NY, USA) was used for statistical processing.This study was conducted as per the guidelines of the Declaration of Helsinki. The analytical data did not contain any information by which the participants could be identified. The ethics review committee at the Nippon Dental University College at Tokyo approved the study (Approval No.: Tokyo Tanrin-218).Of the 549 questionnaires collected, 537 valid responses were received (response rate: 37.2%). The age distribution of the 537 dental hygienists is shown in Figure 1. Approximately 66% of the respondents were in the 30- to 40-year-old age group. The mean age of the sample was 39.1 years (standard deviation: 8.9).Table 1 provides the demographic information of the respondents related to age distribution. The employment rate was 85.6% for those in their 20s but decreased significantly to 67.6% for those in their 30s. The employment rate remained almost unchanged for those in their 40s but decreased over time to 50.6% for those in their 50s. The rate of subjects seeking full-time work followed a similar trend to the employment rate. Moreover, nearly 70% of respondents in their 30s were married.Similarly, 65.8% of respondents in their 30s reported having children. Of the respondents in their 20s, 60% had participated in a training program, however, this rate decreased as the respondent age increased. Overall, 66.3% of respondents were married, 60.5% had children, 49.5% wanted to work full-time, and 41.5% had participated in a training program within the past year.Table 2 shows results related to variables that were prioritized by the subjects. Specifically, it provides information on the factors that the respondents selected as being important for their continued employment. More than 80% of the respondents listed “working hours” and “human relations at the workplace” to be crucial, and approximately 73% prioritized “wages”. However, 60.89% and 52.33% of the respondents prioritized working location and job description, respectively. These rates were relatively low compared to the results for other variables.We conducted a chi-square test to analyze the current employment status of dental hygienists (Table 3). The results indicated that there were significant differences between the employed and unemployed groups for the following variables: marital status, having/not having children, desire to be employed, participation in skill development training programs, and age (p < 0.01). The results also indicated significantly higher values in the employed group concerning the rate of seeking full-time work and participation in training programs. However, the unemployed group showed significantly higher values for marital status, having children, and age.We also used a chi-square test to analyze the relationship between employment status and employment priorities (Table 4). The results indicate that there were significant differences between the employed and unemployed groups with regard to wages (p < 0.01) and working hours (p < 0.05). The employed group had significantly higher values for the variables concerning essential labor elements such as working hours and wages. However, there were no significant differences between the two groups in terms of working location, job description, or human relations at the workplace.Given that the bivariate analysis indicated that age might be a potential confounding factor, we conducted multiple logistic regression analysis to determine which factors affect employment status after adjusting for confounding factors (Table 5). The dependent variable was “employment status as a dental hygienist (employed/unemployed)”. Eight independent variables were included in the analysis: age, marital status, having children, participation in skill development training, the priority assigned to wages, and priority assigned to working hours. The bivariate analysis indicated that these eight variables were significantly related to employment status. We determined that the following four items were significantly associated with employment status using the multiple logistic regression analysis: “participation in skill development training programs”, “desired employment status”, “priority assigned to wages”, and “priority assigned to working hours”. However, “age”, “marital status”, and “the presence of children” were not extracted from the multiple logistic models.The results indicated that the employment rate sharply declined from the 20s age group to the 30s age group. Subsequently, the employment rate gradually decreased with age. The Japan Dental Hygienists Association reported a similar trend in 2017 [14]. According to a nationwide survey on overall women’s employment status in Japan [16], the employment rate declined when women were in their 30s and then increased after they reach their 40s when they no longer required childcare. However, the results of the present study indicated that the employment status of dental hygienists differs significantly from this trend. Our findings suggest that the re-employment of dental hygienists does not progress as rapidly as it does in the overall population. One explanation for this may be that dental hygienists find it challenging to maintain their dental professional skills after leaving their job. While there have been three training centers established in Japan as part of a project funded by the Ministry of Health, Labour, and Welfare to support the re-employment of dental hygienists [17], the number of training centers falls significantly short of the demand for re-employment.The bivariate analysis revealed significant differences in demographics, such as age, marital status, having children, and desire to work full-time between employed and unemployed dental hygienists. Several previous studies have noted a similar trend [13,18]. Approximately 90% of dental hygienists in Japan worked in private dental clinics in 2018; however, most of these clinics do not have on-site childcare facilities. Thus, the social infrastructure needed to support the work–life balance of dental hygienists with children is still lacking.In this study, the top three most prioritized factors concerning employment were the following: “working hours”, “human relations at the workplace”, and “wages”. However, there was no significant difference in the “human relations at the workplace” variable between the employed and unemployed dental hygienists. This result suggests that “human relations at the workplace” does not determine employment. In general, job description is a significant factor in choosing an employer; however, there was no significant difference between the two groups in terms of working location or job description in this study. Although some new tasks, such as oral health care for older adults, have been developed, there are few differences in the tasks that are regularly performed at common dental clinics. This is a possible explanation as to why the number of dental hygienists who prioritized the importance of job description was relatively small.Participation in skills training programs was also related to employment status. Training institutes in Japan and Korea previously tended to emphasize dental assistance rather than dental health guidance [19]. Japan’s educational system for dental hygienists has increased the length of the required educational period since 2010 [20]. However, dental hygienists in their 30s and older have received prior education that emphasized dental assistance. In the future, job description would be of increased importance for continuous employment, depending on the increase in dental hygienists trained in the new curriculum. It is necessary to develop realistic educational programs based on dental hygienists’ actual work situations following graduation [21].The multiple logistic regression analysis set “employment status” as the dependent variable, adjusted for confounding factors such as age, and identified the following four variables as relevant factors: “participation in skill development training”, “desire for full-time work”, “priority assigned to working location”, and “priority assigned to working hours”. In a previous study [12], life events such as marriage and the delivery of a child, which have been frequently cited as factors related to employment among dental hygienists, were not extracted after adjusting for confounding factors using multiple logistic regression analysis. Our findings will be useful for increasing the number of working dental hygienists. It is also noteworthy that “participation in skill development training” and “priority assigned to wages” were extracted as variables related to employment status. It has been important for dental hygienists to have their personal, family, and public needs simultaneously met to maintain continuous employment as a dental hygienist. Participation in training programs that aim to improve professional skills should be a fundamental condition for maintaining their professional careers.Furthermore, wages, desired employment status, and working hours are common motivational factors for all workers. Dental hygienists’ professionalism is based on intrapersonal, interpersonal, and public professionalism [22]. The variables extracted in this study consisted of three domains of professionalism. Furthermore, a path analysis of work ethic factors reported that the intrapersonal factor was closely related to work ethic among employed dental hygienists [23]. There is an urgent need to foster these three professionalism factors in dental hygiene education.The findings obtained in the present study suggest that it is necessary to provide opportunities for self-improvement besides wages, working hours, and employment status. It is also necessary to frequently update knowledge and specific skills that were not included in traditional education curricula to provide adequate oral health care for older adults.It is also highly likely that the working environment surrounding female dentists is similar to that of dental hygienists. A recent study reported that the burden of childcare for female dentists in Japan is so heavy that 57.1% of female dentists in Japan would like to change their working style [24]. A similar analysis of employment-related factors among female dentists in Japan should be conducted in the future.A noteworthy aspect of this study is that professional education played a significant role in the employment of dental hygienists. Professional education has had a great impact on enhancing the identity of dental hygienists. A study on the professional role of dental hygienists in Finland reported that there was a lack of alignment between the traditional role of the dental hygienist and the evolving scope of dental practice [25]. Even though the systems for providing oral health and dental health care services are vastly different in Japan and Finland, both countries have common educational challenges. Reassessment involving the stakeholders and improvement of current training programs would be necessary in Japan as well. In addition, a competency analysis of dental hygiene work would greatly enhance the work capacity of dental hygienists and thus ensure a larger workforce [26]. Unfortunately, there is not enough competency analysis on dental hygiene work in Japan; there is an immediate need to promote such research.In Japan, the disease structure of dental diseases is changing drastically due to the rapid decline in the prevalence of childhood caries as well as the increase in the number of aged people who still have most of their original teeth. However, sufficient oral health services are not yet available for all aged people, owing to the shortage of dental hygienists. The number of dental hygienists employed in long-term care facilities accounted for only 0.3% of the total number of dental hygienists in 2018. The number of employed dental hygienists is vastly insufficient for the increasing rate of elderly people requiring long-term care.In providing oral care services to elderly people who require nursing care, medical knowledge of the physical functions of the elderly and a welfare mindset are also necessary. For dental hygienists to continue to work, there is a strong need to expand professional education programs that foster appropriate oral care skills.As this study focused on a local population, it is difficult to analyze the national situation based on the findings from this study alone. Additionally, another limitation of this study is the relatively low response rate. Despite the limitations, this study’s multiple logistic regression analysis clarified the factors related to the employment status of dental hygienists in a wide range of age groups. We elucidated the differences in each independent variable between employed and unemployed dental hygienists. In addition, many of the items used in the survey were taken from the Labor Force Survey conducted by the Japanese government and, therefore, might have been conceptualized insufficiently.The findings obtained in this study suggest that to increase the dental workforce in Japan, it is necessary to enhance working conditions and environments such as wages and working hours, as well as to provide a more robust educational system to strengthen the professional development of dental hygienists.Conceptualization and analysis of this study were done by H.M., R.T., K.O., and Y.U. The original draft was prepared by H.M. and reviewed by R.T., K.O., and Y.U. All authors have read and agreed to the published version of the manuscript.This research was funded by a Health Labour Sciences Research grant (19IA1010).The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the Nippon Dental University College at Tokyo (protocol code Tokyo Tanrin-218).Informed consent was obtained from all subjects involved in the study.The authors declare no conflict of interest.Age distribution of subjects (N = 537).Demographic results by age.Prioritized variables of the subjects according to the survey results.Difference between unemployed and employed respondents along with the five indicators.Relationship between employment status and employment priorities.Relevant factors concerning employment status using multiple regression analysis.β = beta coefficient, SE = standard error, Wald = Wald χ2 statistics, OR = odds ratio, CI = confidence interval.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01026.txt

@@ -0,0 +1 @@
+Research in instability has focused on the analysis of muscle activation. The aim of this systematic review was to analyse the effects of unstable devices on speed, strength and muscle power measurements administered in the form of controlled trials to healthy individuals in adulthood. A computerized systematic literature search was performed through electronic databases. According to the criteria for preparing systematic reviews PRISMA, nine studies met the inclusion criteria. The quality of the selected studies was evaluated using STROBE. The average score was 14.3 points, and the highest scores were located in ‘Introduction’ (100%) and ‘Discussion’ (80%). There is great heterogeneity in terms of performance variables. However, instability seems to affect these variables negatively. The strength variable was affected to a greater degree, but with intensities near to the 1RM, no differences are observed. As for power, a greater number of repetitions seems to benefit the production of this variable in instability in the upper limb. Instability, in comparison to a stable condition, decreases the parameters of strength, power, and muscular speed in adults. The differences shown are quite significant in most situations although slight decreases can be seen in certain situations.Strength training under unstable conditions, as well as destabilizing devices, have gained popularity in the last decade for athletes in order to strengthen core muscles, improve balance, proprioception and increase performance [1,2,3,4]. Muscle power is considered to be one of the main determinants of many short-term explosive sporting events [5]. Power has been defined by different authors throughout history. Bompa [6], defined muscle power as the ability to perform different actions, developing maximum strength in a short period. For this reason, power is a manifestation of the strength that most athletes in different disciplines consider to be of greater importance for the development of certain movements [7,8,9,10,11,12].A potential reason for similar training-induced adaptations observed after unstable situations compared to stable ones could be related to similar or even higher levels of muscle activation in favour of unstable conditions [13,14]. In the case of the evidence shown concerning the use of unstable situations concerning strength, power and speed, maximum tests have been used to evaluate muscle strength; and tests to determine muscle power (i.e., abdominal power test, medicine ball throwing and different types of jumps) [15,16,17,18,19,20]. However, current results indicate that at least for non-elite athletes, there is a stress/strength and power training intensity pathway that is sufficient to induce positive training adaptations. In their review, Behm and Colado [21] reported that the average strength deficit in unstable situations compared to similar stable exercises was 29%. Furthermore, in healthy young adults, strength training with low loads compared to high loads is equally effective when improving muscle strength [22,23].Because sport is not usually practised under stable conditions, such as throws, jumps, changes of direction, where the body must be stabilized while a specific action is being performed, training should try to represent the requirements of the specific sport [24,25,26,27]. Also, training under unstable conditions or unbalanced conditions can resemble training and daily activities, providing effective transference [1]. The use of unstable training has been proposed to improve movement specific effect through increased activation of stabilizers and core muscles [14,25,27].Although initially unstable surface training was reserved for rehabilitation and prevention programs to reduce the rate of injury, due to the proprioceptive overload they provide [28,29,30,31] this type of training is now included in strength and conditioning programs [10,32,33,34]. Currently, the use of these devices has been incorporated into traditional exercises to promote neuromuscular coordination and recruitment but there is controversy regarding the effects of this combination on sports performance and core stability activation [27,35].There are various ways of generating instability in the performance of exercises but the most common has been through the use of different devices [36,37,38,39]. The use of unstable devices in strength training has led to the development of numerous investigations focused on the analysis of muscle activation [40,41,42,43,44,45,46]. Nowadays, the use of different training methods that retain the stabilising capacity of athletes has become a common and frequent practice. The use of specific devices to create unstable environments, such as exercise or Swiss ball [47], the semi-sphere balance balls, like BOSU [48] and the suspension devices, like TRX [49] have been widely used in sports centres and are widely used throughout the population. Therefore, this review has focused on testing the use of these types of devices and not others, to clarify their influence on sports performance.Swiss ball are air-filled balls covered with soft elastic material with a diameter of approximately 35 to 85 cm [47]. The use of unstable surfaces, such as the exercise ball, began to be used in strength and muscle conditioning training as a method of strengthening core, stabilizing muscles, that is the musculature with a deep location which is responsible for a good body posture both in our daily life and in the practice of sports [36,50,51,52]. The response of muscle activity to this unstable surface can be variable and depends on the type of exercise or muscles being tested. Patterns of muscle activation during bench presses have reported variable results based on muscle function. In the stabilising or core muscles, it has been shown that there is an increase in the activation of the internal obliques, the external obliques and the rectus abdominis [25,53,54,55,56] and spinal erectors [56] while having minimal effect on the rectus abdominis [13,56]. In the upper extremities, compared with a stable bench press, a greater increase in anterior deltoid, pectoralis major, triceps and serratus anterior activity has been demonstrated during execution on a stability ball [30,51,55,57,58,59]. However, improvement in muscle activation has not always been evident. In the case of the central or stabilising muscles, other studies have not shown any change in the oblique and internal spinal erectors [60]. In addition, other studies have shown higher data for the stable condition as opposed to the unstable condition, in the main motor musculature responsible for movement, as in the case of the pectorals and triceps in the bench press [43], or no main motor muscle in the shoulder press [59].The BOSU Balance Trainer ® (DW Fitness, LLC, Clifton, NJ, USA), or “both sides up” is an exercise device used to improve balance, core muscle or torso strength, and proprioception created for military service veterans [61]. The flat part of the device is a 25-inch platform with two built-in handles, and the other part is an inflatable rubber dome that rises about one foot above the ground. Each side can be used in different ways to create different situations depending on the exercise.Different studies have analyzed the influence of the semi-sphere ball for training, providing a great variety of results. Authors found increased muscle activity in the rectus abdominis and external oblique in the performance of abdominal plates and gluteal bridge [53]. Anderson and Behm [14] reported increased EMG activity in the vast lateral, soleus and superficial trunk muscles, but not in the femoral biceps when comparing squatting with free weight on a stable versus an unstable surface. In bench press, greater activation of the internal oblique, spinal erector, soleus and biceps femoris was also evident [56]. However, Willardson et al. [35] compared EMG activity in the core performing 50% of 1RM in squats on a stable surface and in a semi-sphere ball and observed no differences between conditions. Authors examined the activity of the brachial triceps, spinal erector, rectus abdominis, internal oblique, and soleus while performing traditional and unstable bending in a single (hands or feet on the unstable surface) or dual (both hands and feet on the unstable surface) instability and found that the dual condition caused the highest percentage of change (>150%) for all muscles analyzed; compared with the other conditions [62].A new method available to increase muscle activation is suspension training. This type of training uses the principles of body weight and strength boosting to improve motor unit recruitment [63]. The most applied suspension device is the TRX Home Suspension Training Kit (Fitness Anywhere LLC, San Francisco, CA, USA). In suspension training, a specific device is required to create an unstable condition. This method uses a system of straps with handles on the bottom and which are attached to a single anchor point [64]. Among the different strength training possibilities, suspension training is widely applied in various contexts. It is considered an effective technique for improving neuromuscular activation that precedes the use of heavy loads in traditional exercises [65]. Besides, improvements in speed and strength indicators have been found by the use of suspension training, suggesting increased recruitment of core/stabilizing muscles [66].Regarding evidence from suspension devices, the effects of usage on both lower body muscle activity [41,67,68] and trunk stabilizing muscle [65,69,70,71] have been investigated. Clear evidence has been established regarding these devices that witnessed increased muscle activation in the stabilizing and synergistic muscles when performing exercises under these conditions [65,69,71,72]. Concerning lower body exercises, very high activation has been shown for the femoral and semitendinous biceps (>90% MVIC) [68], the hamstring, the gluteus maximus, the gluteus medius and the long adductor. Although no significant differences were found in the rectus femoris [41,67]. Also for the Bulgarian squat exercise, no difference in muscle activation was found between the stable and the suspended condition [64]. For the exercises of the upper part of the trunk and the stabilising muscles, they have been studied with the performance of the Push-up exercise. It has been shown that greater muscular activation in the core, rectus abdominis and external oblique muscles [69,73], however, the stable situation reported greater activation for the pectoral and deltoid muscles [65,69,71] the brachial triceps [65] and the clavicular portion of the pectoralis [71]. However, for the frontal plate exercise, Byrne et al. [70] reported no significant differences when studying exercise with suspension devices.Research has focused on the analysis of muscle activation, determining different considerations, claiming their strengths and weaknesses. However, when the training objective is hypertrophy, or gain in muscle mass, strength, or power, it has not been recommended that the exercises be carried out using unstable situations [5].As mentioned earlier, muscle activation has been shown to improve the stabilising muscles and reduce the main motor muscles involved in performing the task, but it is not known whether these activations can lead to improved performance.In terms of the choice of devices, the 3 devices have been widely used for both rehabilitation, proprioception, and development of muscular capacity. With any one of them, there are numerous proposals for working on muscular strength, and all of them focus on the argument of greater activation of the central or core muscles. In fact, the use of suspension devices alone cannot bring about an improvement in strength or power by itself, since one always works with one’s own body weight, without external loads that increase the intensity of the tasks to be carried out. Semi-spherical and exercise ball devices are not only an implement that increases instability, but there are concrete references that indicate that it could be a way to improve strength and power [30,43,59,74]. Although it could have certain limitations, therefore it has been decided to include it, to be able to evaluate all devices and tools that generate instability to a greater or lesser degree.Therefore, a synthesis of the literature seems necessary to determine whether performing exercises with unstable material provides additional effects on measures of speed, strength and muscle power compared to stable execution. The existing controversy regarding the use and results provided by instability training, and the variability of surfaces, devices, instability positions, samples and exercises generated a heterogeneity of results that makes this review necessary. The aim of this systematic review was to provide a scientifically based study regarding the effects of unstable devices on speed, strength and muscle power measurements administered in the form of controlled trials to healthy individuals in adulthood, apart from muscular activation. It is hypothesized that unstable devices produce similar or not excessively inferior performance improvements to stable conditions because performance with instability is very demanding on the neuromuscular system (i.e., additional stability of joints and posture during exercise is required).The present research was designed to qualitatively synthesize the available scientific evidence concerning the effect of instability in strength, power, and speed training. The stages of the review procedure and subsequent analysis of the original articles stayed within the guidelines set out in the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) [75] checklist and the Population, Interventions, Comparisons, Outcomes and Study Design (PICOS) question model for the definition of inclusion criteria.Primary and original studies to evaluate the strength, power or execution speed in instability were included. Furthermore, studies had to have been published in any language, in peer-reviewed journals with an impact factor included in the Journal Citation Reports of the Web of Science (JCR of WoS) or Scimago Journal & Country Rank (SJR of Scopus) until November 2020.According to the ‘PICOS’ question model, the inclusion criteria were: (1) ‘Population’: physically active and healthy participants (both men and women) between 18 and 65 years. This age range includes all participants considered to be of adult age; (2) ‘Intervention’: acute training effects on strength, power and/or speed of execution using a Swiss ball, semi-sphere ball or suspension devices; (3) ‘Comparison’: differences in tasks multi-articular upper or lower limb between the execution of exercises in stable conditions and execution in unstable situations; (4) ‘Outcomes’: at least one strength, power and/or speed result had to be reported in the study; (5) ‘Study Design’: descriptive and quasi-experimental research based on a comparison between stable and unstable situations.The exclusion criteria were: (1) the studies were for intervention periods, randomized control trials, and clinical trials; (2) they included patients or persons with disease or injury; (3) any data about muscle activation, because it is not a performance variable; (4) the subjects were not of adult age (under 18; e.g., children and adolescents and over 65’s as the elderly) (5) the chronic effects of the situations under investigation were assessed; (6) Any other type of unstable device other than Swiss ball, semi-sphere ball or suspension devices, because these are the most frequently used devices; (7) any measurement that includes unilateral exercises or with different types of support such as exercises executed in a monopodal position.A systematic computerized literature search of the Web of Science, PubMed and EBSCOhost with full text was conducted until November 2020 to capture all relevant articles investigating the effectiveness of instability versus stability. The following Boolean search strategy was applied using the operators ‘AND’, ‘OR’ and ‘NOT’: (‘instability resistance training’ OR ‘instability strength training’ OR ‘free-weight training’ OR ‘suspension training’ OR ‘unstable devices’) AND (‘power’ OR ‘power performance’ OR ‘speed’) AND (‘strength’ OR ‘muscle strength’ OR ‘muscle power’ OR ‘muscular power’) NOT (‘natural surfaces’) AND (‘stability balls’ OR ‘bosu’ OR ‘suspension devices’ OR ‘unstable devices’). The unrestricted language search was limited to the human species and the availability of the full text of original articles reporting on a quasi-experimental trial in an academic journal. Also, we checked the reference lists of each included article and reviewed relevant review articles to identify additional studies suitable for inclusion in the database.The authors worked separately and independently to ensure the reliability of the process and the suitable eligibility of the studies. According to the criteria for preparing systematic reviews “Preferred Reporting Items for Systematic Reviews and Meta-Analysis”—PRISMA [75], the protocol carried out in the months of July, August and September 2020 and it was made up of four stages (Figure 1): (1) Identification: the first author (M.M.) found 167 studies in the four digital databases; (2) Screening: the first author (M.M.) eliminated the duplicate files (n = 8) and excluded those considered not relevant through a previous reading of the title, abstract and keywords (n = 90). Furthermore, the first author (M.M.), jointly with the second (J.L.C.) and third (J.R.G.), rejected the studies linked to the instability according to the exclusion criteria through a full-text reading (n = 55); (3) Eligibility: the first (M.M.), second (J.L.C.) and third author (J.R.G.) eliminated full-text studies from the selection process by the eligibility criteria (n = 45); (4) Inclusion: the remaining studies (n = 8) based on the relationship between the execution of the exercises in a stable condition and their execution in an unstable condition were finally considered. An additional article was identified from the reference lists of included papers and review articles already published [24,63,76,77].A standardized form was used to extract data from the studies included in the review for assessment study quality and scientific evidence. Thus, information about (A) ‘authors and year of publication’, (B) ‘sample experience’ (C) ‘sample size and sex’ (number of players, sex), (D) ‘sample characteristics’ (age, height and weight), (E) ‘variable measured’ (strength, power and/or speed), (F) ‘type of exercise’ (G) ‘number of situations’ whether the tasks were executed by comparing only the stable condition vs. the unstable condition or whether more variations were included), (H) ‘device’ (unstable device implemented), (I) ‘training volume’ (number of sets/repeats/rest per exercise), (J) ‘intensity’ (percentage of one maximum repetition (1RM)), (K) ‘strength results’ (maximum strength (e.g., 1RM), mean strength), (L) ‘power results’ (maximum power, mean power and concentric phase power) and (M) ‘speed results’ (maximum and mean speed) were collected. The results data reflect the percentage of decrease or increase in instability concerning stability. If the included studies did not report the results (i.e., means and standard deviations) of the pre-and post-tests, the authors of those studies were contacted.The quality of all eligible cross-sectional studies was evaluated using the criteria for strengthening the reporting of observational studies in epidemiology “STROBE” [78]. The following scale was used to rate the quality of studies: (a) good quality (>14 points, low risk of major or minor bias), (b) acceptable quality (7–4 points, moderate risk of major bias), and (c) poor quality (<7 points, high risk of major bias). The score was obtained through the 22 points on the STROBE checklist. Two independent reviewers (M.M. and J.L.) conducted study quality assessment. Rating disagreements were resolved by J.R. and inter-rater reliability calculated.Scientific evidence on the sample characteristics (B, C, D), variables (E), exercise and variation (F, G) device used (H) volume and intensity training (I, J) and results for strength, power and speed is shown in Table 1, Table 2 and Table 3. Format and design, including the author and the year of publication, the sample characteristics (overall number, gender, age, height and weight), the variable measured (strength, power and/or speed), type and number of variations, the device used (exercise ball, semi-sphere ball or suspension device), training volume (number of sets/repeats/rest per exercise), intensity training (percentage of one maximum repetition (1RM)), strength results (maximum strength, mean strength), power results (maximum power, mean power and concentric phase power) and speed results (maximum and mean speed) are included.Table 1 shows scientific evidence on the sample characteristics (B, C, D) and variables (E). Format and design, including the author and the year of publication, the sample characteristics (overall number, gender, age, height, and weight), the variable measured (strength, power and/or speed).Evaluation of the characteristics of the sample: (B) Experience. The experience of the sample was quite heterogeneous, with the participants standing out trained (n = 3; 33.34%); trained in strength without experience in instability (n = 2; 22.22%), trained in strength and instability 1 year earlier (n = 2; 22.22%); recreational (n = 1; 11.11%); no previous experience in strength or instability indicated (n = 1; 11.11%); (C) Sex. The distribution of the sample was very unbalanced with more male participants (n = 158; 98.14%) than female participants (n = 3; 1.86%); (D) Characteristics of the sample. The whole sample was identified as being between 18 (lower limit) and 25 (upper limit) years of age. The height range was identified as 167 cm to 185 cm. The weight range was identified as 79 kg to 88 kg.Table 2 Shows scientific evidence on exercise and variation (F, G) device used (H) volume and intensity training (I, J). Format and design, including type and number of variations, the device used (exercise ball, semi-sphere ball or suspension device), training volume (number of sets/repeats/rest per exercise), intensity training (percentage of one maximum repetition (1RM)).According to exercise (F): The most evaluated sports task was the “bench press” in five studies (41.67%) and “squat” in four (33.33%). “Deadweight”, “plantar flexions” and “leg extension” were also evaluated (8.33% each of the exercises). (G) The number of situations. The number of comparisons between stable and unstable exercises was 100% of the situations that only compared the stable situation with an unstable one. (H) Type of device. The use of Swiss ball material was 54.55% (n = 6); the use of the semi-sphere ball was 36.36% of the studies analysed (n = 4). In only one study was a suspension device (TRX) used (n = 1; 9.09%). (I) Training volume. The number of series, repetitions and rest was quite heterogeneous. In the case of the series, only two studies are evident, comprising between three and six series. In the case of the repetitions, they varied from isometric execution to 25 repetitions, with the execution of 3–6 repetitions being the most used (60%). In terms of rest, they vary between 3 and 5 min. (J) Training intensity. In the case of the intensity of training, the percentage of load most used was 75% of 1RM (n = 3; 30%). 20% of the investigations did not use external load (n = 2). The rest of the investigations ranged from maximum repetition to 40% of 1RM.Table 3 shows scientific evidence on strength results (K), power results (L) and speed results (M). Format and design, including strength results (maximum strength, mean strength), power results (maximum power, mean power and concentric phase power) and speed results (maximum and mean speed).In performance measures, it can be seen how the use of instability decreases in some cases substantially in relation to the stable condition. Although with loads close to the RM no differences are appreciated. In terms of power, the difference seems to be slighter in stable and unstable condition and even at a higher number of repetitions the instability seems to improve power production. The execution speed also shows a lower production when instability is added.Concerning the strength parameter with a Swiss ball, the bench press exercise showed 59.4% less isometric strength in instability [13], 5.9% [82], but no differences were found in 1RM [81]. For the lower limb exercises, 70.5% less was evidenced in the unstable condition in leg extension exercises than in the stable condition while the unstable force in plantar flexors was 20.2% less than the stable condition [79]. Also with the same exercise, a decrease with the execution with the semi-sphere ball concerning the stable condition of 19% was evidenced [83]. In the case of the deadweight exercise, the decrease in the maximum isometric contraction between the stable condition and the execution with semi-sphere ball was 10.2% [80]. Regarding suspension training squat exercise with bipodal execution, in the eccentric phase, peak and average force showed a decrease of 46.8% and 13.8% respectively for the lower left limb. In the concentric phase, the use of the suspension training tool caused a decrease of 12.6% in peak force and 12.8% in mean force. For the right lower limb, in the eccentric phase, during execution with the suspension training tool, the force decreased by 42.9% and the mean force by 11.7%. In the concentric phase, during execution with the suspension training tool, the peak and average force decreased respectively by 11.9% and 13.2%. During monopodal execution, the eccentric phase in the left limb, the peak force suffered a decrease of 41.8% and the average force a decrease of 18.1%. In the concentric phase, on the other hand, the use of the suspension training tool caused a decrease of 13.5% in peak force and 15.8% in average force. For the right limb during monopodal execution, in the eccentric phase, the force has decreased by 45.1% and the average force by 17.4%. In the concentric phase, the use of the suspension training tool caused a decrease of 12.4% in the force and 14.3% in the mean force [84].For the variable of power with a Swiss ball, a decrease in the unstable situation of 9.9% concerning the stable situation has been evidenced with the chest press exercise [82], and of 10.3% in the average power, 7.3% in the maximum power and 11.5% in the power exercised in the concentric phase [85]. For the average power, with the execution of 25 repetitions, a decrease of 6.9% was found in the unstable bench press, although in the last three repetitions the average power exercised in the unstable condition was 5.6% higher than the stable condition. On the contrary, among the first three repetitions, the unstable data was 12.9% lower than the stable condition. The power exercised in the concentric phase of the bench press was reported to be 4.6% lower in the unstable bench press, although in the last three repetitions the power exercised in the concentric phase of the unstable condition was 13.2% higher than the stable condition. On the contrary, among the first three repetitions, the unstable data was 13.8% lower than the stable condition [86]. To average power, 25 repetitions showed a decrease of 19.3% in the unstable squat, although in the last three repetitions the average power exercised in the unstable condition was 21.4% higher than the stable condition. On the contrary, among the first three repetitions, the unstable condition was 17.1% lower than the stable condition. The power exercised in the concentric phase of the squat was reported to be 18% lower in the unstable squat, although in the last three repetitions the power exercised in the concentric phase of the unstable condition was 20.6% lower than the stable condition. On the contrary, among the first 3 repetitions, the unstable condition was 16.2% lower than the stable condition [86].Only one research study has been shown to consider the execution speed of strength exercises measured in instability. In the case of the speed variable with a Swiss ball, there has been a 9.1% decrease in the unstable condition concerning the stable banking press [82].The quality analysis (STROBE’ checklist) yielded the following results (Table 4): (a) The quality scores ranged from 13–16; (b) The average score was 14.3 points; (c) Of the 9 included studies, 5 (55.55%) were considered to ‘fair quality’ (13–14 points); and 4 (44.44%) were categorized as ‘good quality’ (15–16 points).By sections, the highest scores were located in ‘introduction’ (100%) and ‘discussion’ (80%) and among the highest quality studies, items no. 2 (background/rationale); no. 3 (‘Objectives—State specific objectives and/or any pre-specified hypothesis’); no. 6 (participants); no. 7 (variables); no. 8 (‘data source—procedure for determining performance measurement’), no. 11 (‘descriptive results—the number (absolute frequency) or percentage (relative frequency) of participants found in each grouping category and subcategory’); no. 12 (statistical methods); no. 15 (outcome data); no. 16 (main results); no. 18 (‘key results—a summary of key results concerning study objectives’); no. 20 (interpretation) and no. 21 (generalisability) were considered complete (100%), while the most commonly absent or incomplete item (0 points) was found in items no. 9, 10, 13 and 14 (‘main results—a measure of effect size’). The lowest scores were found in the ‘funding’ section (20%).This is the first systematic review of the literature to examine the effects of instability on measures of muscle strength, power, and speed, administered in the form of quasi-experimental studies in healthy individuals during adulthood.About the production of force, for the exercises of the upper limb, high decreases in values have been observed for the unstable condition. These differences have ranged from 20 to 75% loss in force development in unstable conditions. According to Kornecki, Kebel, and Siemieński [87], the stabilising function of the skeletal muscles is necessary for the coordinated performance of any voluntary movement, and it significantly influences muscle coordination patterns. Therefore, significant reductions in muscle production probably occurred because the muscles around the shoulder complex needed to give priority to stability over force production. Furthermore, under conditions of instability, the stiffness of the joints that act can limit gains in strength, power and speed of movement [88].However, in the data evidenced in the study by Koshida et al. [82], the losses in force values are much lower than in the rest of the studies (5.9% loss in instability) compared with 59.4% loss in Anderson and Behm [13] and in the case of Goodman et al. [81], no significant differences are observed. This inconsistency between the previous research can be attributed to the type of muscle contraction, the degree of instability during the recorded task and the equipment. In Koshida et al. [82] the bench press movement was performed in a supine position with the Swiss ball placed in the thoracic area, which provided a broader support base than for other activities performed in a sitting or standing position. Therefore, the instability imposed on the trunk stabilising muscles would probably be less significant than in previous research. Besides, both studies used dynamic contractions with an Olympic bar with weight plates, while Anderson and Behm [13] used isometric contractions with two independent handles held by straps to force the transducers into the ground. The difference in equipment could impose different levels of instability on the shoulder joint and trunk muscles. Although bilateral contractions were performed in both studies, the independence of each hand in the study by Anderson and Behm [13] may have increased the effort required to maintain balance and the need for the muscles to stabilize during maximum isometric contractions, therefore reducing the net force output. In the case of Goodman et al. [81], where no differences were found, it could be due to the use of different loads, since 1RM was used while in Anderson and Behm [13] 75% of 1RM loads were used and in Goodman et al. [81] were used 50% of 1RM. These data could indicate that the percentage of external load can influence the effect of the instability in the training.In the case of force production in the lower limbs, there have been notable decreases when comparing tasks performed in instability concerning stable conditions. These decreases ranged from 10% to 19%, so it seems that instability affects the upper body more than the lower. With the use of the semi-sphere ball, analysed in terms of strength, with the performance of a dominant hip exercise such as deadweight the decrease in strength was 10.2% [80] while with a dominant knee exercise such as squat it was 19% [83]. This could indicate that certain movements could be affected to a lesser extent depending on the instability. However, as detailed above the methods and loads used were very different. It is noteworthy that many of the studies to check force production in the lower limb using isometric contractions. However, isometric contractions are not usually used in strength training. Despite this, results obtained under isometric contractions have reported that conditions are strongly correlated with dynamic mobility performance [89]. However, due to the isometric test mode, subjects could gradually build up strength while stabilizing and maintaining balance on different surfaces. During the 3 s of maximum effort, the subjects may have been able to stabilize the limbs and trunk and therefore be able to exert a considerable amount of force in unstable conditions. We only know of one study that investigates the production of maximum force in squats on a stable and unstable surface [3]. These researchers used an inflatable balance disk and reported a decrease of approximately 46% in peak force. Although there was a greater decrease in force in that study, it could be attributed to the lack of a familiarization session, which the rest of the studies did consider appropriate.In terms of strength, there seems to be a differentiation in the data concerning the devices used. When the main movement involves the muscles of the upper body the Swiss ball has been used and in the case of the main movement being performed on the lower body, the semi-sphere ball has been used. The exceptional case was the execution of an exercise such as squatting where the instability with the device in suspension was placed in the upper body. In the case of the Swiss ball, it seems to have a greater influence on the decrease of the force values (differences of 20–75%) compared to the semi-sphere ball (differences between the and 10% and 19%) suspended device (detriments between 12–47%). However, in some cases, the Swiss ball did not produce any differences between the conditions. So, it does not seem to be a determining factor in the case of muscle strength.Concerning the production of muscular power in the upper limb, decreases in the unstable condition have been observed. Decreases with the unstable condition ranged from 7% to 17%. The data found in the studies that analyzed the bench press with Swiss ball reported a very similar percentage decrease in terms of average power (10.3% [85]; 9.9% [82]; and 12.9% [86]). These small deviations found may again be due to the different percentage of load used (50% vs. 75%) and the different volume of training applied. However, in some situations, instability has produced better power data than a stable condition. These better data have been produced in the last repetitions of the executions with high numbers of repetitions in the exercise (22–25 repetitions). The improvements observed in average power were 5.6% and in the concentric phase 13.2%.The increase in power observed could be due to previous evidence that has shown that producing a high power output with a light to moderate load would be more effective in developing maximum power than using a heavy load [90,91]. Thus, it appears that such a low rate of reduction may still allow muscle power to be gained from strength exercise in the unstable condition. The mechanism of energy production using the stretch-shortening cycle employs the energy storage capacity of several elastic components and the stimulation of the stretching reflex to facilitate muscle contraction for a minimum period. The concentric muscle action does not occur immediately after the eccentric, the stored energy is dissipated and lost in the form of heat and also the strengthening stretching reflex is not activated. Resistance to instability exercise can compromise the three phases of the stretching-shortening cycle, including the amortization phase. Around this turning point, where the eccentric phase becomes concentric, the maximum force is produced. At the same time, the subjects must stabilize the torso on an unstable surface to provide firm support for the contracted muscles. This additional task can compromise the contraction of the muscles acting on the bar. Their less intense contraction not only prolongs the change of direction of movement but, due to the lower maximum force, impairs the accumulation of elastic energy. The consequence is less speed and power in the subsequent concentric phase [92,93]. However, the subjects of the study by Zemkova et al. [86] were able to produce greater power during the executions on an unstable surface than on a stable one. This higher energy production can be attributed to the so-called ball bounce effect.In terms of power, there seems to be a differentiation in the data about the devices used. The use of semi-sphere ball seems to have a greater influence on the decrease in power since the detriments with this device varied between 15% and 22%. In the case of the Swiss ball, the decrease in power oscillated between 7% and 14%, with better power data being found in unstable conditions with this device (5–13%). However, when the instability was placed where there was no movement, as in the case of the device in suspension and the squat, improvements of between 5% and 10% in power production were shown. Therefore, placing the instability where the main movement does not occur seems to be a good option for power improvement.Finally, about the speed of execution, a decrease in the values in unstable conditions in comparison to stable conditions is observed, but the analysis of this variable has hardly been studied. According to Adkin et al. [94], a postural threat in a subject (fear of falling) will lead to a reduction in the magnitude and speed of voluntary movements. Thus, muscle stabilization seems to compromise gains in strength, power and speed of movement [95]. It should also be noted that new patterns of movement are generally learned at low speed, while sport-specific motor actions are executed at high speed [26].The great heterogeneity in terms of volume and intensity of the load is remarkable. Also, instability seems to affect the force variable to a higher degree, but with intensities close to 1RM no differences are observed. As for power, a greater number of repetitions seems to benefit the production of this variable in instability in the upper limb. Finally, speed has barely been analysed and seems to show losses of speed in instability but not excessively so.The great heterogeneity found is a limitation of the study, however, the results of this study can be applied in various ways. It would be interesting to include training in instability in athletes trained in this type of situation. All the information about this is with beginner athletes, and what is interesting about the application of instability is the individualization of the subjects. Unstable surfaces can be very interesting tools for optimising training, because although decreases in performance variables have been shown, this may not be the case for experienced athletes. As for integrated work, where besides strength, power and speed, other qualities such as balance can be analysed. Also, the angles with which the work is done in instability are different concerning stable conditions, so that other complementary muscles are worked. Finally, variety in environments, methods and exercises is one of the principles of training and these unstable situations provide it.The complexity of execution in this type of unstable situation, where the technique can be affected, is noteworthy. For this reason, the level of experience of the athletes is important to be able to apply this type of training. Besides, the population requires the help of a qualified professional who can help and direct the sessions or tasks with this type of device, with an appropriate and individualised programme according to the different users.The main findings of this review were that there is great heterogeneity in analysing the acute effects of instability on performance variables. Instability compared to a stable condition decreases the parameters of strength, power, and muscular speed in adults. The differences shown are quite significant in most situations although slight decreases can be seen in certain situations. However, for the upper limb, a greater number of repetitions seems to increase the power values in instability compared to the stable situation. The variables of force, power and speed seem to be affected when instability is implemented. However, it seems necessary to extend the investigation of instability with the performance variables because the results are very heterogeneous and there are no unified criteria to evaluate the different conditions, subjects, tasks and devices.Conceptualization, M.M. and J.L.-C.; methodology, M.M., J.L.-C. and J.R.-G.; software, M.M.; validation, J.L.-C. and I.R.R.; formal analysis, M.M.; investigation, M.M.; resources, M.M. and. J.L.-C.; data curation, M.M. and A.G.-A.; writing—original draft preparation, M.M.; writing—review and editing, J.L.-C., J.R.-G. and I.R.R.; visualization, M.M., J.L.-C., J.R.-G., A.G.-A. and I.R.R.; supervision, J.L.-C. and J.R.-G.; project administration, J.L.-C., J.R.-G. and I.R.R.; funding acquisition, A.G.-A. and I.R.R. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.We wish to thank all athletes who completed the experimental protocol. We would also like to thank the Faculty of Physical Activity and Sport Sciences (INEF) and the Universidad Politécnica de Madrid (UPM) for their support.The authors declare no conflict of interest.Flow chart illustrating the different phases of search and survey selection.Scientific evidence on the sample characteristics (B, C, D) and variables (E).M = men; W = women; a = age; h = height; w = weight; cm = centimetres; kg = kilograms.Scientific evidence on exercise and variation (F, G) device used (H) volume and intensity training (I, J).rps = repetitions; min = minutes; RM = repetition maximum.Scientific evidence on strength results (K), power results (L) and speed results (M).INS = instability; S = Swiss ball; B = Semi-sphere ball; T = Suspension device; LE = leg extension; PF = plantar flexors; EF = eccentric phase; CF = concentric phase; MVIC = maximum voluntary isometric contractions; MVC = maximum voluntary contractions; MS = mean strength; MáxS = maximum strength; LF = left foot; RF = right foot; BP = bench press; SQ = squat; MP = mean power; PMáx = maximum power.The study quality analysis (STROBE’ checklist).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01027.txt

@@ -0,0 +1 @@
+World Health Organization data show that approximately 800,000 persons die by suicide each year [1]. Moreover, suicide trends remain stable or are increasing across many nations. In light of the global concern posed by suicide, decades of research have been devoted to identifying factors that place people at risk of suicidal thoughts and behaviors (STBs). Despite these efforts, a recent meta-analysis demonstrated that most risk factors do little better than chance in predicting future STBs [2]. Suicide resilience factors were not better predictors of suicide either; however, they have been dramatically understudied according to this synthesis of the literature. Thus, this Special Issue includes articles that attempt to increase the field’s understanding of suicide resilience. Clement and colleagues [3] investigate the interrelationships between hope, optimism, hopelessness, and grit, as well as their collective relationship to suicidal ideation. In a cross-sectional sample of undergraduate college students, the authors used exploratory factor analytic techniques to determine how these risk and resilience factors relate at both the subscale and item level. The authors demonstrated a single-factor solution for these measures when subscale scores were entered; however, when all items across the scales were factor analyzed together, five separate subscales emerged that did not represent the original five scales. These newly defined subscales had differential relationships with suicidal ideation, demonstrating that a potential reconceptualization of positive future thinking styles may be needed to better understand suicide resilience.The presence of positive future thinking such as hope is a clear correlate of indicators of mental well-being in a variety of samples [4]. Russel, Rasmussen, and Hunter [5] extend research on mental well-being and risk of STBs in a longitudinal sample of adolescents assessed at baseline and 6-month follow-up. Their investigation demonstrates that mental well-being appears protective against eventual thoughts of self-harm and self-harm behaviors through a reduction in feelings of defeat and entrapment. This investigation through the lens of the Integrated Motivational Volitional model of suicide (IMV) [6] provides a useful framework of understanding how upstream suicide prevention efforts that foster broader mental well-being may impact risk of self-harm and suicide.Although a broad framework such as IMV may inform suicide prevention efforts, unique resilience factors exist in underrepresented populations [7]. The integration of these group-specific factors with information derived from majority populations may increase the relevance of suicide resilience frameworks in underrepresented groups [8]. Cramer and colleagues integrate identity characteristics into the central tenets of the preferences in information processing (PIP) model of suicide risk in a sample of adults who self-identify as members of the alternative sexuality community [9]. Higher education and monogamous relationship status were suicide protective factors. More importantly, a high need for affect approach, or a willingness to engage emotions, buffered the negative association between depression and suicide.The three investigations noted above attempt to inform the suicide resilience research literature. Despite a disparity in work that investigates suicide resilience compared to suicide risk [10], research demonstrates some progress in the creation of several evidence-based, suicide-specific interventions such as Dialectical Behaviour Therapy (DBT) [11], Brief Cognitive Behavioral Therapy (BCBT) [12], and the Collaborative Assessment and Management of Suicide (CAMS) [13]. The promising evidence of these interventions in reducing STBs is met with a paucity of research regarding the actual use of suicide-specific care in the community. Moscardini and colleagues [14] investigate the extent to which behavioral health providers routinely use suicide safety planning [15] in their everyday practice. The investigation reveals a high level of comfort and the regular use of the intervention, but an overarching desire for continued training in suicide safety planning. Similarly, the investigation shows that the frequency of utilization of the intervention varies across providers and is related to providers’ personal history of experiencing STBs but not professional exposure to patient suicide. This research indicates that, although suicide safety planning may be a vitally important skill in any behavioral health provider’s preverbal toolbox, the successful implementation of this intervention is not a given and continued education in the intervention may be needed for successful regular utilization.Finally, suicide-specific care such as suicide safety planning is not always needed in treating patients presenting for mental health concerns. Thus, although suicide prediction appears to be a difficult pursuit [16], suicide risk identification is a necessary step to determine the level of suicide-specific care that may be warranted. Cohen and colleagues [17] conducted a feasibility investigation of whether machine learning techniques with natural language processing of adolescent discussions with their therapists may improve suicide risk assessment practices. The investigation demonstrated that this practice focused on augmenting standardized clinical risk assessment efforts may provide important information for assessing risk of STBs and thus the cueing of suicide-specific care practices.In summary, this Special Issue of the International Journal of Environmental Research and Public Health details new investigations into factors that relate to suicide resilience in samples of adolescents and adults. This research indicates a need for the further refinement of our knowledge of how positive thinking styles relate to each other and thus suicide resilience, the need for upstream suicide prevention efforts that target not only risk of STBs but broader mental well-being, and the utility of incorporating identity considerations into theoretical frameworks of suicide risk and resilience. This research could inform the refinement and creation of suicide-specific interventions. This work, however, should be aware of important implementation considerations identified by providers using suicide safety planning. It is possible for behavioral healthcare providers to feel confident in their utilization of these interventions but demonstrate varying indicators of using these practices with fidelity. Finally, suicide risk assessment is needed to identify the level of suicide-specific care a patient may need. The use of natural language processing and machine learning techniques may enhance the identification of the level of care needed.The authors declare no conflict of interest.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01028.txt

@@ -0,0 +1 @@
+Background: A do-not-resuscitate (DNR) order is associated with an increased risk of death among emergency department (ED) patients. Little is known about patient characteristics, hospital care, and outcomes associated with the timing of the DNR order. Aim: Determine patient characteristics, hospital care, survival, and resource utilization between patients with early DNR (EDNR: signed within 24 h of ED presentation) and late DNR orders. Design: Retrospective observational study. Setting/Participants: We enrolled consecutive, acute, critically ill patients admitted to the emergency intensive care unit (EICU) at Taipei Veterans General Hospital from 1 February 2018, to 31 January 2020. Results: Of the 1064 patients admitted to the EICU, 619 (58.2%) had EDNR and 445 (41.8%) LDNR. EDNR predictors were age >85 years (adjusted odd ratios (AOR) 1.700, 1.027–2.814), living in long-term care facilities (AOR 1.880, 1.066–3.319), having advanced cardiovascular diseases (AOR 2.128, 1.039–4.358), “medical staff would not be surprised if the patient died within 12 months” (AOR 1.725, 1.193–2.496), and patients’ family requesting palliative care (AOR 2.420, 1.187–4.935). EDNR patients underwent lesser endotracheal tube (ET) intubation (15.6% vs. 39.9%, p < 0.001) and had reduced epinephrine injection (19.9% vs. 30.3%, p = 0.009), ventilator support (16.7% vs. 37.9%, p < 0.001), and narcotic use (51.1% vs. 62.6%, p = 0.012). EDNR patients had significantly lower 7-day (p < 0.001), 30-day (p < 0.001), and 90-day (p = 0.023) survival. Conclusions: EDNR patients underwent decreased ET intubation and had reduced epinephrine injection, ventilator support, and narcotic use during EOL as well as decreased length of hospital stay, hospital expenditure, and survival compared to LDNR patients.Although the main responsibilities of emergency physicians (EPs) when treating acute critically ill emergency department (ED) patients include initial resuscitation, stabilization, rapid diagnosis, and curative treatment, aggressive resuscitation may not be appropriate or desired when managing seriously ill patients with advanced chronic illness with trajectories of dying. Early conversations with patients and surrogates regarding aggressive resuscitative measures are critical with respect to patient autonomy and appropriately tailored care. Do-not-resuscitate (DNR) orders are an alternative for patients at the end of life [1], to prevent nonbeneficial resuscitative measures and unnecessary suffering when patients are imminently dying [2,3]. Placement of DNR orders is variable between different types of hospitals and based on differing patient demographic factors [4,5].In theory, the DNR order itself should not directly impact care until the moment of cardiac arrest. However, early DNR (EDNR: order placement within 24 h of ED presentation) was found to directly influence both resuscitative and ancillary care [6] and resulted in a decrease in potentially critical hospital interventions, with wide variability in practice patterns between hospitals [3]. EDNR often is a proxy of patient’s underlying disease, prehospital frailty, and burden of comorbidities [4], an independent predictor of 28-day mortality [7]; and a strong predictor of short-term mortality risk [8]. The aim of the study was to determine differences in patient characteristics, demographics, hospital care, survival, and resource utilization between EDNR and late DNR (LDNR: after 24 h of ED admission) among ED patients admitted to the ICU.This was a retrospective cohort study of adult ED patients (≥18 years) who presented to the ED of Taipei Veterans General Hospital (TVGH) from 1 February 2018, to 31 January 2020. This project was approved by the TVGH Institutional Research Board, which waived the need for patient consent (IRB No: 2020-11-010BC).TVGH, a 3000-bed university-affiliated medical center, has an annual ED census of 85,200 ± 1812 over the past five years. The emergency intensive care unit (EICU) is a 13-bed ICU within the ED [9] where acute critically ill patients who are not admitted to the specialized ICU immediately after initial ED resuscitation and stabilization receive intensive care. The primary goal of the EICU setting was to implement continuous emergency and critical quality of care within ED prior to available specialty ICU transfer. The operative system in our EICU is semi-open model, that both the EPs and physicians in other subspecialties cooperatively take care of all admission patients. This system was supervised by Emergency Quality Control Committee.Patients with a diagnosis meeting the criteria for acute severe critical illness (item A) and who also fulfilled two of the criteria for initiating PCC (item B) were categorized as the PC-eligible group (Table 1). The others were categorized as the PC-ineligible group. Palliative care consultation (PCC) screening was initiated for acute critically ill patients aged ≥18 years who were admitted to the EICU from 1 February 2018, to 31 January 2020. Exclusion criteria were age <18 years and medical records with incomplete or missing data. DNR orders are orders to withhold resuscitative measures including CPR, intubation, defibrillation, cardioactive drugs, or assisted ventilation. Patients who did not sign the DNR form on admission (n = 1565) and patients who signed the DNR form before admission (n = 185) were excluded (Appendix A). Patients were considered to have a preexisting DNR order if a DNR order was found in the patient’s chart dated before the day of admission. The primary exposure variable was whether an order to limit resuscitation efforts was written within the first 24 h of admission (EDNR).Utilization criteria were formulated by palliative care (PC) and hospice specialists and adopted to identify patients at high risk of poor clinical outcomes as their care commonly involves prolonged use of advanced medical resources or technologies [10]. Two trained authors entered the abstracted data for study analyses. The information, time and date of each DNR orders were collected via inpatient electronic medical record systems.Data collected were patient characteristics, hospital care, medical resource utilization, hospital length of stay (LOS), and total expenditures and in-hospital mortality. Hospital care included endotracheal (ET) intubation and ventilator support, cardiopulmonary resuscitation (CPR), cardioversion/defibrillation, epinephrine injection, vasopressor therapy, cardiac pacemaker insertion, extracorporeal membrane oxygenation (ECMO), endotracheal removal, and narcotic use.Data are expressed as mean ± SD for continuous variables and number (%) for categorical variables. Data distribution was assessed using the Kolmogorov-Smirnov test. Comparisons of numerical variables were performed using an unpaired t-test (parametric data) or Mann-Whitney U test (nonparametric data). Categorical variables were compared using the two-sided chi-square or Fisher’s exact test. Factors showing statistical significance (p < 0.05) in the univariate analysis were included in the multiple regression analysis. Survival time was calculated from the date of admission to the date of death using the Kaplan-Meier method, and the difference in survival time between the eligible and ineligible groups was compared using the log-rank test. p < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS software version 22.0 (SPSS Inc., Chicago, IL, USA).A total of 1064 patients were recruited for the study; 619 (58.2%) had EDNR and 445 (41.8%) LDNR. The screening items for PC consultation at the time of EICU admission are shown in Table 1. Patients with EDNR had more advanced cardiovascular diseases, advanced central neurological diseases, septic shock, adult respiratory distress syndrome (ARDS), multiple organ failure or impending death, and were very severely frail (all p < 0.001). The clinical characteristics of EDNR and LDNR patients are compared in Table 2. The mean age of EDNR patients were older than LDNR (80.8 vs. 77.3 years, p < 0.001). While more LDNR patients lived with family (84.9% vs. 78.2%), more EDNR patients lived in veterans’ homes (4.4% vs. 2.3%) and long-term care facilities (8.6% vs. 4.7%, all p = 0.035). Patients with EDNR had reduced length of hospital stay (17.8 ± 18.4 days vs. 30.3 ± 31.7 days, p < 0.001), and lower total hospital expenses (246,684 ± 266,447 new Taiwan dollar (NTD) vs. 468,532 ± 476,382, p < 0.001).Table 3 shows the univariate and multiple logistic regression analyses of clinical characteristics between EDNR and LDNR patients. The risk factors associated with EDNR were age >85 years (AOR 1.700, p = 0.039), living in long-term care facilities (AOR 1.880, p = 0.029), presence of advanced cardiovascular diseases (A5) (AOR 2.128, p = 0.039), patients whom medical staff would not be surprised if they died within 12 months (B1) (AOR 1.725, p = 0.004), and patients whose family requested PC (B8) (AOR 2.420, p = 0.015). The screening items (item A and item B) for assessment of palliative care consultation at the time of admission were listed in full in Table 1.Differences in hospital care between EDNR and LDNR patients are compared in Table 4. Patients with EDNR received less endotracheal tube (ET) intubation procedures (15.6% vs. 39.9%, p < 0.001), less epinephrine injection (19.9% vs. 30.3%, p = 0.009), less ventilator support (16.7% vs. 37.9%, p < 0.001), and less narcotic use (51.1% vs. 62.6%, p = 0.012).Table 5 shows multiple logistic regression analyses of hospital care between patients with mortality with EDNR and LDNR. Patients with EDNR underwent lesser ET intubation procedures (AOR 0.198, p = 0.007) and had reduced narcotic use (AOR 0.518, p = 0.001).Figure 1 shows the Kaplan-Meier curve for survival between patients with EDNR and LDNR. Patients with EDNR had a significantly lower 7-day, 30-day and 90-day survival.The study found several differences in patient characteristics, hospital care, survival, and resource utilization between EDNR and LDNR patients.Patient characteristics that predict EDNR were age >85 years, living in long-term care facilities, presence of advanced cardiovascular diseases (A5), “medical staff would not be surprised if the patient died within 12 months of this episode” (B1), and patients’ family requesting PC (B8).Compatible with our finding that age >85 years was an independent risk factor for EDNR, other studies also found that older patients were more likely to have an EDNR order [11,12,13]. Age was a powerful predictor of an explicit DNR directive in all categories of patients older than 50 years of age [13]. Age ≥ 80 years was an independent risk factor for DNR orders after controlling for comorbid conditions [14]. Other than being associated with more comorbidity, functional impairment, and higher mortality [14], older patients may have an opportunity to discuss with their physicians and families about advance directives and may be more likely to have accepted and expected their own mortality [15]. However, if decisions on EDNR are based purely on the patient’s chronological age without factoring in survival, quality of life, or patients’ wishes, it may be constituted as ageism. Our study confirmed that age is an important factor for EDNR in critically ill patients, but whether ageism, withholding treatment solely on the basis of age, plays a part in the decision-making process remains unclear.Our study found that patients living in long-term care facilities were more likely to have EDNR orders (AOR 1.880). Of the 74 patients from nursing homes in our study who signed a DNR order at the ED, 53 patients (71.62%) had EDNR. Similarly, a Danish cohort of patients with community-acquired pneumonia (CAP) found patients with EDNR were older and more frequently nursing home residents (41% vs. 6%, p < 0.001). [16] Marrie et al., found that coming from a chronic care facility or a nursing home was a major demographic associated with DNR upon admission, and more than half (53.8%) from institutions had a DNR order in place on admission [17]. This may reflect nursing home policies or a greater awareness among this group to have advanced directives. However, the prevalence of DNR directives among Taiwanese nursing home residents was lower than that in other countries [18]. The EDNR status associated with long-term care facilities may be due to physicians’ awareness of the poor outcomes of resuscitation for nursing home patients and lower odds of achieving return of spontaneous circulation (ROSC) [19] and more likely to approach family and surrogates early with DNR discussion.Our study found that presence of advanced cardiovascular diseases (chronic heart failure (CHF, New York Heart Association III or IV), chest pain, or dyspnea while performing minimal exercise or on minimal exertion, or devastating inoperable peripheral vascular diseases) were a strong predictor of EDNR (AOR 2.128). EOL discussion in patients with heart failure (HF) is of particular importance because patients often experience repeated hospitalizations and a progressive decline in quality of life as they approach death [20]. However, the waxing and waning pattern typical in HF makes it difficult to accurately prognosticate expected survival, rendering it difficult for physicians to approach patients and surrogates with DNR discussions. Moreover, patients with HF were found to have frequent changes in code status, underscoring the importance of periodically reviewing resuscitation preferences as advocated by the American Heart Association [21]. A study found that three-quarters of community patients with HF elect DNR before death; however, changes in resuscitation preference are often made in the hospital in the final days to weeks of life [22]. This discordance with our finding that patients with HF are at higher risk for EDNR may be because our study included patients who were older (mean 80.8 ± 14.2 years) and had more advanced disease (CCI ≥ 7 47.8%). This may explain the early DNR order in CHF patients with advanced age that tend to underestimate their life expectancy [23].The decision to forgo resuscitative measures should reflect patient values and preferences. However, physicians’ judgments on patient condition and survival may have a direct impact on patients’ preference for DNR decisions. Many patients and surrogates require a discussion on prognosis with their physicians prior to making a DNR decision [24]. Our study found that the factor “medical staff would not be surprised if the patient died within 12 months of this episode” was a risk factor for EDNR. In line with our study results, a multicenter study found that the one of the strongest predictors of DNR directives were physician prediction of low probability of survival. It is not only physician predictions of high likelihood of death that were associated with DNR order but also moderate likelihood of death [12]. The question “should physician’s judgments on patient survival influence DNR decision?”, is an ethical dilemma.Our study found that “patients’ family requesting palliative care” is a predictor of EDNR. Inability to participate in decision-making was a strong predictor of a DNR directive during the first 24 h of ICU admission [13]. Patients who were unable to participate in decision-making were significantly more likely to have a DNR directive than a resuscitate directive [13]. A Taiwanese study revealed that the prevalence of DNR directives among Taiwanese nursing home residents was lower than in other countries, with 91% of the directives being put in place by family surrogates [18]. Other studies have shown that as many as 40% of hospitalized adults are unable to make their own medical decisions [25], with DNR decisions being made by family one-third of the time [26]. This is consistent with our finding that 59.4% of patients with DNR and 68.5% with EDNR were categorized as requiring assistance in terms of medical decisions, who were unable to participate in DNR decision making.Our study found that EDNR is associated with decreased ET intubation, epinephrine injection, ventilator support, and narcotic use, but no difference in CPR, cardioversion, vasopressor use, cardiac pacemaker insertion, ECMO, intraaortic balloon pumping (IABP), or withdrawal of ET tube was found compared to that in LDNR. This is similar to a sepsis study where the DNR group did not receive less ancillary care of the central line, vasopressors, blood transfusion, emergent hemodialysis, or surgery [11]. Our findings are also comparable with a study where chronic obstruction pulmonary disease (COPD)decedents with EDNR were less likely to undergo invasive mechanical ventilator support during their terminal hospitalization [27]. In theory, the DNR order itself should not directly impact care until the moment of cardiac arrest. However, one study found that EDNR directly influenced both resuscitative and ancillary care, with fewer invasive interventions being performed in the last week of life, including dialysis, mechanical ventilation, feeding tubes, and CPR, compared to those with LDNR and no DNR [6]. Another study on out-of-hospital cardiac arrest (OHCA) patients found that EDNR is associated with a significant decrease in potentially critical therapeutic options, including cardiac catheterization, bypass surgery, and blood transfusion after resuscitation, and is associated with less aggressive hospital care, fewer potentially beneficial procedures, and worse survival [3]. The impact of EDNR on both resuscitative measures and ancillary care may be that patients with DNR, are also less likely to receive nonbeneficial aggressive care at the end of life [28] and are more likely to receive care consistent with their preferences [29]. This wide variability in practice patterns between hospitals and physicians suggests a lack of standardized approach to the EDNR order and subsequent resuscitative measures and ancillary care. This further emphasizes the importance of communication between physicians and patients to align care with treatment goals. Physicians should be careful not to interpret DNR, which is “do not perform CPR in the event of cardiac arrest” as “do not actively treat this patient.”In contrast to a study in which nurses were more comfortable giving opioids for pain management at the EOL [6], our study found that EDNR is associated with decreased narcotic use. We hypothesize that the shorter duration from DNR placement to death and shorter hospital LOS associated with EDNR allowed less time for physicians to address the family members on patients’ comfort during the care. This certainly leaves room for improvement in patient comfort during hospital care, especially in patients with an EDNR order.Patients with EDNR had lower 7-, 30-, and 90-day survival; our finding is compatible with another study where EDNR was found to be an independent predictor for 28-day mortality [7]. In a Danish cohort of patients with community acquired pneumonia (CAP), EDNR was associated with higher mortality after adjustment for clinical risk factors [16]. Among intracerebral hemorrhage (ICH) patients, EDNR is an independent predictor of poor outcome [30,31]; 2.6 times more likely to die than those without DNR order [30]. OHCA patients with EDNR usually die in the hospital without discharge to home within one day of admission [3]. Moreover, EDNR order is often a proxy of the patient’s underlying disease, prehospital frailty, and burden of comorbidities [4] that is not captured by the established risk factors for mortality. Hence, EDNR may function as a composite of prognostic variables, resulting in a strong prediction of short-term mortality risk [8]. However, other studies have argued otherwise. In one study, patients with LDNR (written on day 6 or later) were twice as likely to die in the hospital than patients with EDNR. [32] Marrie et al., hypothesized that EDNR reflects comorbidities and the general health status of a patient, while LDNR represents a lack of response to treatment and comorbidity [17]. The LDNR discussion that occurred later may be because patients are not responding to treatment and at imminent risk of death. A study on sepsis, in fact, found better outcomes in the EDNR group than that in the LDNR group [11]. It remains unclear if the EDNR in our study directly reflects general health status or LDNR directly reflects a lack of response to treatment. The nature and breadth of the discussion leading to EDNR, LDNR, and subsequent withdrawal of care are beyond the scope of the study, leaving room for further research.Patients with EDNR had decreased hospital length of stay (LOS) (17.8 ± 18.4 days vs. 30.3 ± 31.7 days) and decreased total medical expenditure (246,684 ± 266,447 NTD vs. 468,532 ± 476,382) compared to LDNR patients, which is compatible with other studies. One study on COPD patients found that the average total medical cost during the last hospitalization was nearly twofold greater for LDNR than for EDNR decedents [27]. We further stratified patients into mortality and survival groups. Among patients with mortality, the hospital LOS in EDNR patients was shorter than LDNR. (11.7 ± 16.4 days vs. 25.2 ± 24.5 days, p < 0.001). The lower probability of survival in EDNR compared to LDNR (Figure 1) may explain the shorter hospital LOS. Surprisingly however, among patients who survived, EDNR had a shorter hospital LOS (faster recovery) than LDNR (22.9 ± 18.5 vs. 3.4 ± 36, p < 0.001). The finding agrees with Marrie et al.’s hypothesis that LDNR may be a result of poor response to treatment and comorbidity [17] hence patients with LDNR who survived may have a prolonged hospital course.Our study has several limitations. First, as a retrospective study, it is liable to underreport due to missing or incomplete medical records. Second, although the inclusion criteria were strictly followed, there still may exist confounding discrepancies between the criteria and clinical conditions of patients recruited. Third, although the study determined the differences in hospital care and outcome between EDNR and LDNR, it did not stratify patients into subgroups based on diagnoses such as ICH, OHCA, CAP, etc. Fourth, the study determined the patient characteristics associated with EDNR, but the rationale for selecting EDNR and LDNR remains unclear in these patients. Fifth, the study did not assess psychosocial aspects such as patient and surrogate viewpoints, their satisfaction or dissatisfaction, and patients’ quality of death associated with EDNR.Physicians should understand the potential impact on hospital care and survival associated with EDNR to tailor care with treatment goals. Patients who have not had this discussion should be made aware that EDNR is associated with decreased ET intubation, reduced epinephrine injection, ventilator support, and narcotic use during hospital care, decreased length of hospital stay, hospital expenditure, and decreased survival.While EDNR orders may be appropriate for guiding subsequent treatment, EDNR may not be entirely appropriate in certain groups of patients in whom long-term prognosis is difficult to ascertain and 24 h may be premature to make this decision.J.C.-Y.C. interpretation of data, drafted the article, revise it critically for important intellectual content; C.Y.: concept and design of the study, acquisition of data; L.-L.L.: concept and design of the study, acuqisiton of data; Y.-J.C.: concept and design of the study, analysis and interpretation of data; H.-H.H.: concept and design of the study, analysis and interpretation of data; J.-S.F.: concept and design of the study, interpretation of data; T.-F.H.: concept and design of the work, interpretation of data; D.H.-T.Y.: concept and design of the work, interpretation of data, revise it, critically for important intellectual content, approved the published version. All authors have read and agreed to the published version of the manuscript.This study was supported partly by research grants MOST107-2314-B-075-053 and MOST 108-2314-B-075-034 (D.H.-T.Y.) from the Ministry of Science and Technology, Taiwan, and 107VACS-002, V108C-123, and V109C-046 (D.H.-T.Y.) from Taipei Veterans General Hospital, Taiwan.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Taipei Veterans General Hospital, IRB No: 2020-11-010BC.The study was approved by the Institutional Review Board of Taipei Veterans General Hospital, which waived the need for patient consent.All data submitted comply with Institutional or Ethical Review Board requirements and applicable government regulations. For further information, please contact David Hung-Tsang Yen (hjyen@vghtpe.gov.tw).The authors declare that there is no conflict of interest.Flow shart on patient selection.The cumulative survival curve of early DNR patients and late DNR patients.The screening items for needs assessment of palliative care consultation among 1064 patients at the time of admission.Results expressed as number (%) for categorical variables; Of the 1064 screened patients, 36.7% (390/1064) has one item, 27.1% (288/1064) has 2 items, and 1.7% (18/1064) has 3 items of acute critical and life-limiting illnesses. Of the 1064 patients having one or more items of acute critical and life-limiting illnesses, 1.2% (13/1064) has one item, 9.4% (100/1064) has 2 items, 28% (298/1064) has 3 items, 22.9% (244/1064) has 4 items, and; 3.3% (35/1064) has 5 or more items of the unmet palliative care needs; COPD = chronic obstructive pulmonary disease; NYHA = New York Heart Association; ARDS = adult respiratory distress syndrome; CSHA-CFS = Chinese-Canadian study of health and aging clinical frailty scale ADL = activities of daily living; BMI = body mass index; * p < 0.05 is considered statistically significant using chi-squared analysis or Fisher’s exact test.Comparison of clinical characteristics between early DNR and late DNR patients.Results expressed as number (%) for categorical variables and mean ± standard deviation for numerical variables; TTAS = Taiwan Triage and Acuity Scale; ED = emergency department; APACHE = Acute Physiology and Chronic Health Evaluation; ICU = intensive care unit; * p < 0.05 is considered statistically significant using Mann-Whitney U test or chi-squared analysis.Univariate and multiple logistic regression analyses of clinical characteristics between early DNR and late DNR patients.TTAS = Taiwan Triage and Acuity Scale; APACHE = Acute Physiology and Chronic Health Evaluation; OR = odds ratio; 95% CI = 95% confidence interval; AOR = adjusted odds ratio; * p < 0.05 is considered statistical significance in regression model.Comparison of hospital care in hospitalization between 282 early DNR patients with mortality and 198 late DNR patients with mortality.Results expressed as number (%) for categorical variables; ET = endotracheal; CPR = cardiopulmonary resuscitation; ECMO = extracorporeal membrane oxygenation; IABP = intra-aortic balloon pump; * p < 0.05 is considered statistically significant using chi-squared analysis or Fisher’s exact test.Multiple logistic regression analyses of hospital care between early DNR and late DNR patients with mortalityET = endotracheal; OR = odds ratio; 95% CI = 95% confidence interval; AOR = adjusted odds ratio; * p < 0.05 is considered statistical significance in regression model.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01029.txt

@@ -0,0 +1 @@
+While COVID-19 infection and mortality rates are soaring in Western countries, Southeast Asian countries have successfully avoided the second wave of the SARS-CoV-2 pandemic despite high population density. We provide a biochemical hypothesis for the connection between low COVID-19 incidence, mortality rates, and high visceral adiposity in Southeast Asian populations. The SARS-CoV-2 virus uses angiotensin-converting enzyme 2 (ACE2) as a gateway into the human body. Although the highest expression levels of ACE2 are found in people’s visceral adipose tissue in Southeast Asia, this does not necessarily make them vulnerable to COVID-19. Hypothetically, high levels of visceral adiposity cause systemic inflammation, thus decreasing the ACE2 amount on the surface of both visceral adipocytes and alveolar epithelial type 2 cells in the lungs. Extra weight gained during the pandemic is expected to increase visceral adipose tissue in Southeast Asians, further decreasing the ACE2 pool. In contrast, weight gain can increase local inflammation in fat depots in Western people, leading to worse COVID-related outcomes. Because of the biological mechanisms associated with fat accumulation, inflammation, and their differential expression in Southeast Asian and Western populations, the second wave of the pandemic may be more severe in Western countries, while Southeast Asians may benefit from their higher visceral fat depots.A second wave of the COVID-19 disease has rapidly been gaining momentum since early fall 2020. Although governments and health systems seem to be more prepared for preventing and handling the pandemic, outbreaks of COVID-19 intensify, and many parts of the Western world have experienced a full-blown second surge of the pandemic. Once again, Europe and the United States are the epicenters of the disease, with lockdown restrictions being reimplemented as the effects of the second wave become more severe. In contrast, China has avoided its second wave so far, and looks stable with its infection and mortality rates [1]. Moreover, the whole region of Southeast Asia and Oceania have been hit by the coronavirus pandemics much less than other parts of the world. Countries such as Taiwan and Vietnam have successfully avoided the second wave of the SARS-CoV-2 infection in spite of their high population density.Overall, population density is not related to COVID-19 infection rates (τ = 0.02, p = 0.72, n = 99 nations, Figure 1A) nor to mortality rates cross-nationally in North America, East Eurasia, and West Eurasia (τ = −0.08, p = 0.24, n = 99 nations, Figure 2A), suggesting that factors other than population density play a more prominent role in the spread and severity of the COVID-19 disease [2]. Within world regions, population density was negatively associated with COVID-19 cases (per 1 million inhabitants) in North America (τ = −0.48, p = 0.009, n = 16), but positively in East Eurasia (τ = 0.13, p = 0.40, n = 21) and West Eurasia (τ = 0.29, p < 0.001, n = 62) (Figure 1B). Only in North America and West Eurasia were the associations statistically significant, though the effects were in opposing directions. We found that population density is not related to COVID-19 mortality (Figure 2A). Within world regions, population density was negatively associated with COVID-19 deaths (per 1 million inhabitants) in North America (τ = −0.47, p = 0.01, n = 16) and East Eurasia (τ = −0.03, p = 0.83, n = 21), but the association with COVID-19 deaths was positive in West Eurasia (τ = 0.16, p = 0.07, n = 62) (Figure 2B). Only in North America was the association statistically significant. These analyses were done in a sample of 99 countries from East Eurasia, North America, and West Eurasia, as those countries are the focus of the rest of this article. We used Kendall’s Tau to analyze these bivariate correlations because the data were non-normally distributed and had relatively small sample sizes. A larger sample of nations from all world regions yielded smaller, statistically non-significant correlations between population density (log) and cases per 1 million population (rs = 0.12, p = 0.11, n = 176), and between population density (log) and deaths per 1 million population (rs = 0.05, p = 0.53, n = 176). Data on total COVID-19 deaths and COVID-19 tests per 1 million inhabitants up to 10 January 2021 were collected from the Worldometer site (https://www.worldometers.info/coronavirus/#countries). Data on population density (2018) were collected from World Bank statistics. In this Perspective article, we argue why human populations differ in SARS-CoV-2 infection prevalence and severity based on their differential build-up of energy reserves into subcutaneous and visceral fat depots, and why the second wave of COVID-19 may be more severe in Western countries. While obesity is traditionally considered as a disorder of the energy homeostasis system, we suggest that certain types of fat may have a positive effect on COVID-19 infection rates and outcomes. This suggestion is based on the idea that organisms store body fat as a hedge against diseases to allow them to survive periods of pathogen-induced anorexia [3] and that visceral fat depots can help detect and eliminate pathogens and maintain immune homeostasis of the gut microbiome [4].Stringent quarantines, city lockdowns, local public health measures, and mandatory quarantines to ban or restrict international traffic and/or domestic traffic are considered the main reasons that stopped the SARS-CoV-2 virus’s spread in China, South Korea, and Singapore [5]. While Southeast Asian countries and New Zealand adopted the more stringent suppression strategy from the beginning of the pandemics [6], most Western countries (with the exception of the countries that were hit hardest during the first wave, such as Spain and Italy) initially focused on measures that only mitigated the spread of the virus and decreased transmission rate [5]. Although many European countries and the USA have started to follow more stringent suppression strategies since spring 2020, this change did not allow them to avoid the second wave of the pandemic, minimize mortalities, nor improve their economies. While Southeast Asian countries such as Vietnam, Laos, and Cambodia took serious measures to prevent the SARS-CoV-2 virus from spreading, they have never implemented the extreme quarantines characteristic of China and Italy. However, these countries have had consistently low incidence and mortality rates during the second wave (Figure 3), which is puzzling and requires further explanation. Understanding the success of these Southeast Asian countries is important, because the testing and especially production and implementation of effective vaccines against the SARS-CoV-2 virus may take more time than expected, increasing the death toll of COVID-19.Krams et al. (2020) [7] suggested that the obesity paradox provides a possible explanation for the observed population differences in COVID-19 infection and mortality rates between Europe, the USA, and Southeast Asia. In the obesity paradox, obese patients may have better health outcomes than normal-weight patients despite greater risk of local and systemic inflammation in their fat tissue [8]. Lower mortality in individuals with larger fat reserves has been reported in both infectious and non-communicable diseases [9,10,11,12]. The beneficial impact of body fat in multiple diseases is considered to provide a buffer against disease-induced anorexia, which is known to impair immunity and increase susceptibility to diseases and infections [3,13,14]. According to the obesity paradox hypothesis as applied to COVID-19, populations differ in their COVID-19 incidence and mortality rates because of the differences in the distribution and condition of adipose tissue [15]. Although adipose tissue is generally considered as an energy reservoir [3,15], subcutaneous fat (superficial subcutaneous adipose tissue and deep subcutaneous adipose tissue) and visceral fat, two main types of adipose tissue in the body, have important diverging metabolic and endocrinological roles [15,16,17,18]. Most body fat is subcutaneous, and it initially grows by hyperplasia (cell number increase). In obesity, the maximum amount of subcutaneous fat seems evolutionarily constrained [3]. At the maximum number of fat cells in the subcutaneous fat depot, the fat tissue becomes locally inflamed [19] and adipocyte hyperplasia normally stops. Fat tissue shifts its growth from hyperplasia to hypertrophy (cell size increases), leading to even higher local inflammation in subcutaneous adipocytes [19]. An increase in subcutaneous adipocyte hypertrophy [20] determines the onset of fat-storing in the visceral tissue depots. Visceral fat has long been known to be associated with systemic inflammation [21,22], insulin resistance, and other metabolic syndromes [16,17,23]. Visceral fat is harmful because it produces pro-inflammatory cytokines released directly into the bloodstream and can lead to auto-amplifying cytokine production called “cytokine storms” [24]. Therefore, reducing visceral fat and/or increasing its metabolic health is traditionally considered a preventive measure for metabolic diseases [24].Surprisingly, visceral adiposity is considerably higher in Southeast Asian populations than in Europe or the USA [25], while COVID-19 incidence and mortality rates are higher in Western countries than in Southeast Asia (Figure 3A,B, respectively). In Southeast Asia, naturally occurring fat build-up involves building small subcutaneous fat depots and relatively large visceral adipose tissue depots [3,25,26]. Southeast Asians have greater body fat than Europeans and North Americans for the same BMI, meaning Southeast Asians cannot be treated as obese despite their larger visceral fat amounts [27]. It has been shown that the SARS-CoV-2 virus uses angiotensin-converting enzyme 2 (ACE2) as a gateway into the body [28,29,30]. ACE2 is a cell-surface exoenzyme that converts angiotensin II (Ang II) into vasodilatory angiotensin 1–7 (Ang 1–7). Importantly, adipose tissue, in general, has one of the highest expression levels of ACE2 of all tissues, being especially high in visceral adipose tissue [31]. Thus, ACE2 amount is expected to be highest in Southeast Asian populations that fare best in the COVID-19 pandemic [32,33]. However, it has been suggested that the ACE2 pool can be smaller in Southeast Asia than in Europe and North America [15]; otherwise, Southeast Asians would potentially host huge viral loads in their visceral adipose tissue, inducing more severe forms of the COVID-19 disease [34].ACE2 function is known to be dysregulated in certain metabolic pathologies when removed from cells’ surface via cleavage by the transmembrane disintegrin and metalloproteinase 17 (ADAM17) [35]. Such shedding of ACE2 is frequently seen in inflammatory states associated with adiposity and obesity [32,33]. Therefore, the loss of ACE2 might explain the paradoxical connection between adipose tissue and COVID-19 infection. While visceral adipose tissue is considered to harbor a large pool of functional ACE2 molecules, the actual ACE2 amount in visceral adipose tissue of Southeast Asians may be much lower than that of Europeans and North Americans because of higher systemic inflammation in visceral fat in Southeast Asians [15]. This specific way of the visceral fat build-up of Southeast Asians might be caused by previous encounters with coronaviruses [36]. High levels of visceral adiposity can prevent the virus from entering human cells, creating the observed association between visceral adiposity and the mildest COVID-19 severity observed in Southeast Asia. Since the disruption of the molecular processes required for normal lung homeostasis starts within hours of the virus’s entrance into the body [37], preventing the virus from entering the cells is of paramount importance.While city lockdowns and other restrictions are necessary measures to restrict the pandemics, the “Quarantine 15” (also known as the “Quarantine 19” in popular literature) has become a common way of referring to lockdown-induced weight gain during the COVID-19 pandemics [38,39]. Extra weight gained during lockdowns is caused by consuming more food than an individual can burn off through limited physical activity, especially when individuals consume energy-dense food which is poor in antioxidants and omega-3 fatty acids and rich in saturated fat and refined carbohydrates [40]. This so called “Westernized” diet and extra weight can affect the gut microbiome, resulting in changes in the host’s immune responses [41]. It has been also shown that consumption of energy-dense food poor in antioxidants and omega-3 fatty acids and rich in saturated fat and refined carbohydrates [40] may directly impair the host’s immune system [42]. Importantly, long-term lockdown and loneliness may cause people to eat foods known to cause obesity [43], which is associated with an increased risk of inflammation and mental health problems [44,45]. Therefore, phenomena like Quarantine 15 may have detrimental effects on people’s BMI, overall health, and, by extension, their ability to overcome COVID-19 [46,47,48].Besides promoting unhealthy lifestyles, lockdown and lockdown-induced weight gain may directly affect immune response to viral infections through alterations of the cellular immune system [49]. Adiposity has been shown to decrease the strength of immune response to influenza and hepatitis-B vaccination [50,51,52] via impaired activation and fewer functional markers of blood mononuclear cells [49].The effects caused by Quarantine 15 may differ between populations. We predict that a mechanism similar to the obesity paradox potentially alleviates the COVID-19 situation in Southeast Asia during the second wave [15,35,53]. A further increase in already high visceral fat levels in these populations will negatively affect the ACE2 pool not only in adipose tissue, but in all other parts of the body, including alveolar epithelial type 2 cells in the lungs due to visceral fat’s systemic effects operating via inflammation. Neuropilin-1 (NRP1) facilitates SARS-CoV-2 cell entry [54]; consequently, if there are no ACE2 receptors available on the surface of cells, as predicted by the reported high visceral fat levels in Southeast Asian populations, the potentiating factor effect of NRP1 is also expected to be absent, thus increasing protection against the SARS-CoV-2 virus. In contrast, in Europe and in the USA, where high visceral fat levels are less common than in Southeast Asia, people are expected to continue building all fat depots, including superficial and deep subcutaneous adipose tissues [53], causing mostly local adipocyte inflammation and not decreasing ACE2 levels throughout the organism. At the same time, weight gain causes deteriorating overall metabolic health and dysfunctional immune system, thereby decreasing an organism’s capability to withstand infections. Europeans, North Americans, and obese (because of their modern lifestyle) Southeast Asians are therefore expected to suffer from the second wave of the COVID-19 disease even more than during the first wave [55]. Thus, based on these hypothesized mechanisms, we predict that pandemic-induced weight gain has differential effects in different populations, generally beneficial in Southeast Asian populations and detrimental in Western populations [56,57].We hope that conclusions drawn from the obesity paradox in the COVID-19 context may help governments and individuals properly prepare for 2021 until vaccines against SARS-CoV-2 become widely available, and when their efficacy has been properly verified in real-world scenarios taking into account immune function heterogeneity and different strains of COVID-19 at the level of individuals and populations. Several pharmaceutical companies have started the production of vaccines, and many countries have already started vaccinations at the beginning of 2021. However, the large-scale effectiveness, population vaccination rate, and length of acquired immunity particularly in the face of different COVID-19 strains are currently unknown. Since many parts of the world are in the midst of an overwhelming COVID-19 wave, it is crucial to pay attention to this physiological hypothesis to prevent unnecessary fatalities before sufficient vaccination can be acquired at the population level. The research reviewed here indicates that adipose tissue heterogeneity requires more attention in the COVID-19 context. We highlight that according to the obesity paradox, distinctive types of adipose tissue may have different roles in affecting COVID-19 infection rates and outcomes.All authors equally contributed to writing the manuscript. Funding acquisition, I.A.K., T.K., and S.L. All authors have read and agreed to the published version of the manuscript.This work was funded by the Latvian Council of Science grants lzp-2018/1-0393 (I.A.K.), lzp-2018/2-0057 (T.K.), and lzp-2020/2-0271 (T.K.).Not applicable.Not applicable.The authors declare no conflict of interest.No relationship between population density and COVID-19 incidence (as of 10 January 2021) in three world regions combined (Panel A). Within three world regions, the relationships between population density and COVID-19 incidence go in the opposite directions (Panel B). Data points are scaled to reflect COVID-19 testing rates per 1 million population.No relationship between population density and COVID-19 mortality (as of 10 January 2021) in three world regions combined (Panel A). Within three world regions, the relationships between population density and COVID-19 mortality go in the opposite directions (Panel B). Data points are scaled to reflect COVID-19 testing rates per 1 million population.The peaks of (A) 7-day moving average of daily new cases per million population (DNC/1M); (B) 7-day moving average of daily mortality per million population (DM/1M) of East and South-east Asian countries (red dots), North American countries (blue dots), and European countries (green dots) on the log scale (as of 31 December 2020). Symbols above the diagonal (dashed line) indicate higher numbers in the second wave relative to the first wave, while symbols below the diagonal indicate the opposite. Framed labels indicate countries with zero mortality values on perpendicular axis.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01030.txt

@@ -0,0 +1 @@
+In the current coronavirus (COVID-19) pandemic, the definition of risk factors for susceptibility to adverse outcomes seems essential to support public and occupational health policies. Some specific issues need to be addressed to understand vulnerability in occupational settings. Among these, individual factors, e.g., age, sex, and preexisting comorbidities (hypertension, cardiovascular diseases, diabetes, obesity, cancer), that can predispose individuals to more severe outcomes and post-COVID-19 symptoms that may represent conditions of acquired susceptibility, possibly impacting the return to—and fitness for—work. Additionally, the risk of contracting COVID-19 through work should be addressed, considering the probability of being in contact with infected people, physical proximity to others, and social aggregation during work. Occupational health settings may represent appropriate scenarios for the early identification of vulnerable subjects, with the final aim to guide risk assessment and management procedures. These should include the systematic surveillance of work-related risk factors, collective preventive policies, stringent actions for specific groups of workers, decisions on occupational placement of employees, and health promotion activities. Concerted actions of general practitioners, hospital specialists, occupational physicians, and all the stakeholders involved in the occupational health and safety management should be focused on planning suitable preventive measures for susceptible subjects.In the current coronavirus (COVID-19) pandemic, the definition of risk factors for susceptibility to adverse and mortality outcomes represents an essential tool to support disease management issues. This can be useful to better define the clinical and epidemiological characteristics of the disease and to facilitate decision-making regarding the appropriate care for patients. From a preventive perspective, considering that countries are adapting to the longer-term challenges of COVID-19, a deeper knowledge of suscetibility risk factors may provide guidance for the development of suitable policies to protect the public and occupational health [1].To this aim, it is important to consider the huge toll paid by frail people in terms of severe clinical manifestations, access to intensive care units and mortality, in the early stage of the COVID-19 pandemic. Although frailty has its own clinical definition, as it is intended as “a medical syndrome with multiple causes and contributors, that is characterized by diminished strength, endurance and reduced physiologic function that increases an individual’s vulnerability for developing increased dependency and/or death” [2,3], data extrapolated from the experience of frail patients may be useful to identify conditions of susceptibility requiring specific preventive actions to possibly avoid the infection and achieve better management of the disease.Since its first recognition in China in December 2019, the disease has spread around the globe at an unprecedented pace. At the time this review was written, more than 76 million confirmed cases of COVID-19, including more than 1,700,000 deaths, have been reported worldwide [4]. The US has experienced more deaths from COVID-19 than any other country and has one of the highest cumulative per capita death rates [5]. Italy has been one of the first and most severely affected countries [6], with 1,977,370 confirmed cases of COVID-19 and 69,842 deaths from 3 January to 23 December 2020 [7].Additionally, worldwide reports have confirmed the high risk of severe illness from COVID-19, including the need for hospitalization, intensive care and death, in older adults, particularly in those with comorbidities (e.g., cardiovascular and respiratory diseases, morbid obesity, diabetes, chronic kidney disease, cancer, gastrointestinal, skin, muscle-skeletal, and immune diseases) [8,9,10,11,12,13].Acquiring this kind of information seems critically important for occupational physicians who are increasingly asked to advise on the fitness for work of patients who may be unusually vulnerable to the disease. It is essential to support employers in the adoption of suitable preventive and protective measures, including collective, individual, or organizational interventions, aimed to protect the health of workers, particularly for those at higher risk of unfavorable outcomes. In this scenario, additional efforts for occupational physicians are required to define the complex interplay between susceptibility conditions to COVID-19 with respect to specific occupational contexts and risk factors. Therefore, the aim of this narrative review is to provide an updated overview on possible individual risk factors predictive for unfavorable mortality and morbidity outcomes that may inform risk assessment and management in workplace scenarios, with the perspective to identify knowledge gaps and future research needs in this field.Advanced searches on PubMed, Scopus, and ISI Web of Science databases were conducted to identify studies published until the 20 December 2020, evaluating possible physiological or pathological conditions of susceptibility to SARS-CoV-2 infection and adverse outcomes, as well as relevant issues to be considered for the consequent risk assessment and management processes in different occupational sectors, taking into particular account the protection of vulnerable workers.We used two search lines that included the terms “COVID-19” or “SARS-CoV-2”, combined with “frailty” or “susceptibility” or “vulnerability”, which were further combined with additional individual terms referring to possible conditions influencing disease manifestation and course: “age”, “gender”, “cardiovascular disease*”, “diabetes”, “overweight and obesity”, “cancer”, “respiratory disease*”, “autoimmune disease*”. After removal of duplicates, the titles and abstracts of the articles retrieved through the computerized search were independently examined by two of the authors. All types of human peer-reviewed research articles (i.e., descriptive epidemiological surveys, cross-sectional, cohort, case-control studies, case series), as well as review papers, were included. Only studies published in languages other than English and publications not specifically focusing on the topics of the review were excluded. For the section on occupational risk factors for susceptibility to COVID-19, the first two search lines were combined with the terms “occupational risk assessment” and “occupational risk management”. Furthermore, some additional documents belonging to the gray literature from national and international governmental and research institutes involved in occupational health and safety management were considered for inclusion.As the review process was not intended to be systematic, and the idea of the authors was to provide inputs on relevant topics in the occupational health practice, articles considered better representative of relevant issues in the respective fields were discussed among the authors and finally included in such a critical revision. The following paragraphs will attempt to summarize important aspects regarding individual and occupational risk factors for COVID-19 that should be carefully considered for adequate protection of the health of workers, particularly for those with conditions of greater susceptibility to acquire and have a more severe course of the disease.A great series of epidemiological and observational studies demonstrated an increased number of cases and a greater risk of severe and fatal disease with increasing age [14,15,16,17,18,19]. Older peoples’ immune-senescence may be characterized by the disruption in both innate and adaptive arms of the immune system, which is not efficient enough to limit infection, as demonstrated by the higher peak viral load in the nasopharynx determined in older adults [20,21]. This results in an increased vulnerability and poor resolution of homeostasis following stress, which may lead to low resilience and frailty. In addition, it seems that very high pro-inflammatory cytokine release, which is described as a cytokine storm, is a pivotal pathophysiological mechanism in elderly COVID-19 patients [22]. Although the exact underlying mechanism of cytokine storm in elderly adults with severe COVID-19 infection is far from clear, it is likely that such dysregulation of the cytokine homeostasis may play a critical role in the development of the acute respiratory distress syndrome (ARDS) in some elderly patients. Many age-related pathophysiologic processes have been suggested as potentially related to such “inflame-aging phenomenon”, among those, the alteration of the expression of the angiotensin-converting enzyme 2 (ACE2) receptor [23], the excessive reactive oxygen species production [24], the autophagy alterations [25], the inflammatory phenotype of senescent cell activity, particularly adipose tissue [26], and immune-senescence [27], as well as lack of vitamin D content [28]. Additionally, aging-related worst outcomes may be dependent on the intrinsic pathophysiological changes that characterize the respiratory system, as well as by the increasing presence of chronic comorbidities with increasing age [29]. However, attention should be paid to consider chronological age per se without considering that individuals do not age the same and that the underlying mechanisms of aging may be different between subjects.First reports from China have pointed out a sex imbalance with regard to detected cases and case fatality rate of COVID-19 [30,31,32,33]. The largest body of publicly available sex-disaggregated data demonstrated no gender-related differences in the number of confirmed cases, with 45.7% female vs. 54.3% male cases in September 2020 [34,35], in line with previously published data [36,37]. However, the overall case-fatality ratio was approximately 2.4 times higher among men than among women [38,39]. Clinical and epidemiological evidence from China [37], the US [40], the UK [41] and European countries (Italy, Spain, France, Germany and Switzerland) [36] demonstrated a 60% to 80% increase in the risk of death for males compared to females. The diverse regulation of the immune system determined both by sex chromosome genes, and sex hormones can be responsible for gender-related differences in COVID-19 mortality [38,42,43]. Moreover, the higher tissue expression of the ACE2 receptors for SARS-CoV-2 in men can cause a greater susceptibility to the infection [44,45]. Psychosocial and behavioral factors can also play a role [46,47,48]. In fact, due to a different belief of the pandemic severity, men tend to engage in more high-risk behaviors, have lower rates of social distancing, wearing masks, and seeking medical help compared to women, thus enhancing the possibility for SARS-CoV-2 infection [49,50,51]. Smoking and drinking rates are higher among men than women, and such behaviors may be associated with the risk of developing comorbidities [36]. Possible differences in occupational risks between men and women may account for diverse COVID-19 outcomes. In the US, women are primarily employed in social and healthcare activities, while men are mainly involved in low-skilled or low-paid essential occupations, e.g., food processing, transportation, delivery, warehousing, construction, manufacturing, where men may experience a greater risk of mortality [52].Hypertension has been reported as the most common comorbidity among patients infected with SARS-CoV-2, with a prevalence ranging from 10–15% to more than 50% [53,54,55,56,57,58]. Hypertension has been proven to be associated with an increased risk of pneumonia, as well as acute respiratory disease and chronic low respiratory disorders independent of age, sex, smoking status and BMI [59,60]. The prevalence of hypertension was higher in patients with severe COVID-19 [61,62,63,64] and with lethal outcomes [65,66,67,68] with respect to non-severe cases. However, to achieve a correct interpretation of these data, it should be noted that hypertension is commonly diagnosed in aged populations, and it is accompanied by many comorbidities, i.e., obesity, diabetes, and cardiovascular diseases (CVDs), that may function as major determinants for COVID-19 severity. Concerning age, hypertension was reported in percentages ranging from 44.7% of the 55–64 years old Chinese population [69], in the 45.2% of the Italian population aged 60–69 years [70], up to the 63–77% of the US 55–64 years general population [71]. Overall, this underlines the relevance to further understand whether hypertension was a predictor of mortality independently of other concomitant risk factors [63].Additionally, also CVDs, i.e., heart failure and arrhythmia, appear as risk factors for severe complications of COVID-19 [53]. Patients with a previous history of CVDs were up to 4 and 6 times more likely to develop the severe and lethal disease, respectively [72]. In line with these findings, data from the Chinese Center for Disease Control and Prevention pointed to a 10.5% mortality rate in COVID-19 patients with CVD compared to a lower 2.3% rate in the overall infected population [73]. Moula et al. [74], in a meta-analysis, assessed the predictive role of different CVDs and coronary artery disease (CAD) separately with respect to COVID-19 mortality. The presence of other CVDs is a strong predictor of death. In fact, a 1.96- and 1.90-fold higher mortality risk was reported in patients with overall CVD and CAD alone, respectively, while these 2 conditions combined were found to increase the risk up to 2.03-fold. Cerebrovascular disease is a strong predictor of death as patients with the preexisting cerebrovascular disease tend to die 2-fold more than patients without cerebrovascular disease [74,75]. In addition to a history of CVD, being an aggravating factor in COVID-19 development, cardiovascular complications can also be the result of the infection, both through a direct SARS-CoV-2 induced myocardial injury as well as by an indirect action exerted by the acute respiratory distress syndrome and systemic illness [76,77,78].Most of the available information refers to patients with type 2 diabetes. This disease per se does not seem to increase the risk of contracting COVID-19 [79]. The prevalence of type 2 diabetes in patients with COVID-19 has been reported to be around 9–12% [80,81,82], raising up to 16–35% in hospitalized patients, including those who required intensive care for the severe disease [83,84,85]. Type 2 diabetes, in fact, has been reported to be more often associated with severe or critical disease, including ARDS, multiorgan failure and death, varying from 14 to 32% in different studies [59,86,87,88]. Additionally, in a summary report of 44,672 confirmed COVID-19 cases, the Chinese Center for Disease Control and Prevention reported an increased case fatality rate in diabetes (2.3% overall and 7.3% in diabetic patients) [89].As regards different types of diabetes, a UK survey found that among 23,804 COVID-19 patients died in hospital, 32% had type 2 diabetes, and 1.5% had type 1 diabetes, with 2.03 and 3.5 times the odds of dying compared with patients without diabetes, respectively. A multicenter observational French study, that was performed on 1317 diabetic patients hospitalized for COVID-19, showed that 3% had type 1 diabetes, while 88.5% were affected by type 2 disease [90].Concerning the impact of glycemic control on COVID-19 mortality, patients with an HbA1c of 86 mmol/mol (10.0%) or higher had increased COVID-19-related mortality compared to people with an HbA1c of 48–53 mmol/mol (6.5–7.0%), with a hazard ratio of 2.23 and 1.61 in type 1 and type 2 diabetes, respectively. In addition, in people with type 2 diabetes, those with an HbA1c of 59–74 mmol/mol (7.6–8.9%) and 75–85 mmol/mol (9.0–9.9%) showed a significantly higher COVID-19 mortality compared with those with an HbA1c of 48–53 mmol/mol, with an HR of 1.22 and 1.36, respectively [91].In summary, a worse prognosis is expected in patients with diabetes affected by COVID-19, most probably because of the concurring effect of multiple factors, i.e., age, comorbidities (hypertension, CVDs and obesity), and glycemic control before-, at the time of- and during hospital admission [79]. Many uncertainties still remain regarding the mechanisms linking diabetes with a more severe course of the disease. The high expression of ACE2 receptors used by SARS-CoV-2 to enter human cells, poor glucose control and high glycemic variability, preexisting diabetes-induced target organ damage, increased inflammatory factors, and hyper-coagulability can be responsible for higher susceptibility of diabetic patients to the adverse outcomes of the infection [82,92].Emerging evidence supports excessive body weight and obesity as risk factors for more serious COVID-19 disease and fatal outcomes [93,94,95,96]. A retrospective analysis of the body mass index (BMI) in USA SARS-CoV-2 infected patients demonstrated that subjects aged <60 years with a BMI between 30 and 34 and >35 were about two and four times more likely to be admitted to acute and critical care compared to individuals with a BMI < 30 [97], respectively. Other case studies confirmed obesity as a major risk factor for disease severity and intensive care unit requirements both in the US [53,83], China [98], the UK [99], France [100,101], Germany [102], and Italy [103,104]. COVID-19 deaths were also more frequently associated with obesity in several countries [105,106,107]. In line with these results, a recent meta-analysis including 30 studies for a total of 45,650 participants demonstrated that BMI defined obesity was significantly associated with a higher risk for severe COVID-19 including, hospitalization (OR 2.36), intensive care unit (ICU) admission (OR 2.32) and invasive mechanical ventilation (OR 2.63), as well as for mortality (OR 1.49) [108]. Further, excessive visceral adiposity appeared to be significantly associated with severe COVID-19 outcomes [108,109].The contribution of overweight/obesity to a more severe disease has biological and physiological plausibility. Obesity is characterized by abnormal secretion of several pro-inflammatory adipokines and cytokines from the adipose tissue, i.e., tumor necrosis factor-alfa, interleukin (IL)-1, IL-6, and monocyte chemoattractant protein, which are responsible for a low-grade, chronic pro-inflammatory state that may predispose to a greater COVID-19 severity [100,110,111]. Obesity may indirectly impact COVID-19 outcomes impairing lung function and respiratory system compliance to mechanical ventilation, thus placing these patients at high risk of severe illness and mortality [112]. In addition to the detrimental effects on lung function, obesity is a well-known risk factor for diabetes, hypertension and CVDs, all predictors of poor outcomes [93].Cancer patients are regarded as a highly vulnerable group due to weakened immune systems caused by both tumor growth and anticancer treatment [59,113,114]. In a nationwide cohort study on 2007 Chinese patients, Liang et al. [115] demonstrated that cancer patients had poorer outcomes from COVID-19. In fact, 39% of cancer patients experienced severe events, including ICU admission, need for invasive ventilation, or death, compared with only 8% of noncancer COVID-19 patients. Comparably, the analysis of data from the COVID-19 and Cancer Consortium (CCC19) cohort study on 928 cases also reported that 30 day-mortality and severe illness in COVID-19 cancer patients were significantly higher than the general population [116]. The most common malignancies in Kuderer et al. [116] were breast (21%) and prostate (16%) cancer; 43% had measurable cancer while 39% were on active cancer treatment.Among 3000 early COVID-19 cases from Italy, a previous (5 years) history of cancer was reported by 20% of non-survivor patients [9]. Higher rates of death, ICU admission, need for invasive mechanical ventilation, the manifestation of at least one severe or critical symptom was reported by Dai et al. [117] on 105 cancer patients. In this population, the most frequent cancer type was lung cancer (20.95%), followed by gastrointestinal (12.38%), breast (10.48%), thyroid (10.48%) and hematological cancers (8.57%). Moreover, the highest frequency of severe complications was reported in patients with hematological, lung, or metastatic cancer (stage IV). Concerning the role of treatment, patients who received surgery or an anticancer treatment within 14 days before COVID-19 diagnosis had higher risks of having severe events compared to patients without cancer, while such increase was not evident in those receiving only radiotherapy [118,119]. Patients who underwent chemotherapy or surgery in the prior month had a significantly higher risk (75%) of clinically severe events than those who did not receive chemotherapy or surgery (43%) [115]. Similarly, another study of 205 COVID-19 cancer patients showed that receiving chemotherapy within four weeks before symptoms onset and male sex were risk factors for death [120].The association between respiratory diseases and severe COVID-19 outcomes has not been definitively clarified. Concerning the prevalence of respiratory diseases among 13,184 patients with COVID-19, a recent metanalysis demonstrated that the chronic obstructive pulmonary disease (COPD) was the main respiratory disease documented [121], and only one study also reported pulmonary tuberculosis and asthma [122]. A significantly 4-fold higher odds of severe COVID-19 outcomes were determined in patients with underlying respiratory diseases, a result in line with those previously obtained by Yang et al. [123] in a smaller meta-analysis. A nearly four-to six-fold higher risk of developing severe COVID-19 was associated with the presence of COPD [121,124], although no significant association could be extrapolated for mortality association [124]. Few available studies have reported a high prevalence of asthma among COVID-19 patients, preventing definite conclusions [125,126].No definite conclusions can be extrapolated concerning the risk for COVID-19 infection and for more severe manifestations in patients with systemic autoimmune diseases in comparison with the general population with similar comorbidities [127,128,129,130,131]. A meta-analysis performed on 62 observational studies on autoimmune diseases that affected patients demonstrated that the prevalence of COVID-19 was 0.011 [128]. Among the seven case-controlled studies available, including psoriasis and rheumatic patients, the risk for COVID-19 was significantly higher than in controls (OR 2.19). This finding was also confirmed in an Italian series of systemic autoimmune diseases [129]. However, this was primarily attributed to the glucocorticoid use, while other therapies, including conventional (i.e., methotrexate, hydroxychloroquine, and sulfasalazine) or biological or target synthetic (i.e., infliximab, adalimumab, etanercept, tofacitinib) disease-modifying antirheumatic drugs (DMARDs) did not contribute to the risk [128]. No evidence for differences in hospitalization, ICU admission, mechanical/non-invasive ventilation, and death emerged from several investigations in comparison to the general population [128,132,133]. Only therapies with glucocorticoids, conventional DMARDs, combined conventional synthetic-biological/target synthetic DMARDs had a 2–3 times higher event rate of each clinical outcome when compared with biological-target synthetic DMARDs alone [128].In workplace settings, the story is not simply one of understanding and grading individual risk factors, but also to put individual conditions in relation to the risk of contracting COVID-19 through work. In fact, in a complementary manner, occupational analysis can address workplace environmental risk of exposure to and acquisition of infection, while personal features can determine the impact of infection once acquired [134]. From an occupational health perspective, it is clear that susceptibility should be considered as the result of the complex interplay between individual and the job or task features. This regards several types of working conditions, like healthcare and social works, where biological risks are intrinsically related to the job activities, but also other occupational sectors for which the risk of infection is comparable to that of the general population.Therefore, work has been recognized as a key determinant in the risk of infection [135]. Multiple outbreaks of COVID-19, in fact, have been reported in a variety of occupational fields, including food packaging and processing sectors, factories and manufacturing departments, and office settings [136]. As an ulterior confirmation, the analysis of the Italian compensation claims pointed out that the SARS-CoV-2 infection has been acquired at the workplace in a substantial number of cases (19.4% of the total amount) [137]. The probability of being in contact with infected people, the physical proximity to others during work activities, and the social aggregation connected to the job may function as key determinants in workplace SARS-CoV-2 transmission [135,138]. In the absence of control measures, a higher risk of infection may be experienced by workers involved in occupations in which it is difficult to maintain physical distancing from coworkers, customers, patients, as well as by employees performing work activities in indoor settings or with shared transport or accommodation [136].In this scenario, methodological approaches have been developed to estimate the levels of the risk of infection associated with various worksites. These first explore the exposure probability due to specific work tasks, including where, how, and to what sources workers may be exposed to SARS-CoV-2. Such sources may be represented by the general public, customers, and coworkers; sick individuals, like in the case of healthcare workers, or those at particularly high risk of infection, but also non-occupational risk factors at home and in community settings depending on the local infection prevalence [139]. Additionally, the extent to which the job entails either close proximity to people who may carry the infection or contact with materials that may be contaminated by the virus, aggregation factors as well as the effectiveness of any measure to reduce transmission, i.e., barriers or personal protective equipment, as well as how the individual commutes to work should be deeply assessed [138].According to all these aspects, job tasks may be categorized into different levels of exposure risk. From very high exposure risk jobs, like those in which healthcare or laboratory workers are exposed to aerosol-generating medical maneuvers or lab procedures requiring collecting or handling specimens from known or suspected COVID-19 patients, respectively, to high, medium and lower exposure risk jobs. This latter group includes jobs not requiring contact with people known or suspected to be infected with SARS-CoV-2 nor frequent close contact with the general public [139]. The importance of such classification relies on the corresponding preventive and protective protocols to be adopted to mitigate workplace transmission and the prompt identification and isolation of potentially infectious individuals within the company. According to the “hierarchy of controls”, such strategies may include (i) to remove or reduce the exposure at the source, e.g., through flexible work solutions, where possible, screening, testing, case investigating and contact tracing; (ii) to redesign the work environments with the adoption of barriers and protection elements to facilitate social distancing, and the implementation of areas for frequent hand washing and sanitizations; (iii) to support the adoption of organizational preventive measures in the workplace; (iv) to promote health and safety education of individual workers and the employment of personal protective equipment in order to minimize the exposure of receptors [140].However, suitable risk assessment and management strategies in the workplace should also integrate susceptibility evaluation and careful considerations of specific conditions of the individuals, e.g., older age, presence of chronic medical conditions, including immunocompromising ones. These issues appear extremely important considering the global aging of the workforce [141,142] and the growing population of immunosuppressed people who enter or reenter the workforce [143]. Overall, this perspective allows moving of the risk assessment from a population-based to an individually personalized approach. Moreover, it, inevitably, underlines the central role of occupational physicians in integrating personal information with job or task features as a unique opportunity to achieve suitable management of susceptible individuals at work.Risk assessment models should offer the possibility to quantify the vulnerabilities associated with demographic variables and comorbidities and their possible combined impact with occupational features in order to measure the equivalent increase in risk that may predict severe courses and case-fatality rate [1,144,145]. Age, gender, ethnicity and comorbidities are commonly included in occupational risk assessment tools, while measures of social circumstances are less frequently assessed [134]. This risk categorization should support individually tailored preventive and protective measures regarding workplace adaptations (e.g., unrestricted work or work-restricted to certain patient groups) or actions (e.g., staff redeployment or home working). However, caution should be paid in the application of such tools for workplace policymaking. Some limitations should be carefully considered, including the focus on the relative risk that is not able to assess whether the absolute risk is actually low, intermediate or high, and the possible lack of accuracy of such points score instruments where multiple risk factors co-exist that may provide an artificial classification of the risk. Finally, by addressing only “individual factors,” the scores do not include the interaction between these and the working environment or job task features. This points out the relevance of developing advanced models able to integrate all these aspects in order to have better representative risk assessment measures. Overall, clinical judgments, technical considerations, and prevailing advice from the government need to be carefully considered in managing occupational risks from COVID-19.The ongoing COVID-19 pandemic still requires a deeper understanding of the nature of the disease and its suitable treatment or preventive measures, as well as how to best safeguard public health in order to avoid the overwhelming of the health care system [146]. In this scenario, it also seems necessary to define appropriate strategies to identify and manage susceptible individuals with a higher risk of acquiring the disease and developing adverse outcomes. This means to translate data acquired on frail patients from clinical settings into a preventive perspective aimed to point out possible conditions of vulnerability that need specific attention and interventions in both public and occupational health contexts.Concerning possible conditions of susceptibility to COVID-19, literature evidence reported age and sex as relevant demographic risk factors for more severe outcomes. Additionally, a list of underlying health conditions, particularly preexisting CVDs, diabetes mellitus, obesity and cancer were estimated to increase the risk for severe COVID-19. However, additional investigations are warranted to acquire a deeper knowledge with respect to the role of individual conditions or the association of personal and/or medical illnesses with respect to both the risk of acquiring COVID-19 and developing severe manifestations. In this regard, better-designed studies should be aimed to overcome possible limitations related to the self-reporting of comorbidities but also to the possible under-reporting of diseases due to the lack of awareness and/or diagnostic testing. A suitable duration of follow-up should also support the extrapolation of unbiased associations between comorbidities and clinical outcomes. Additionally, the limitations due to the heterogeneity of some categories of comorbidity should be overcome, i.e., chronic pulmonary disease aggregates various disorders that are characterized by a wide range of severity. Future investigations should allow evidence-based risk assessments for more specific subcategories of diseases or rarer comorbidities, also considering the detailed history of each condition and specific pathological characteristics. This may include disease onset and course, current or recent activity/flare-ups, past histories of hospital admissions, and medications (past, current or recent) used.In the perspective of defining susceptibility conditions, a deep evaluation of possible consequences of SARS-CoV-2 infection should be included. In this view, the post-COVID-19 syndrome has been defined “as signs and symptoms that develop during or following an infection consistent with COVID-19, which continue for more than 12 weeks and are not explained by an alternative diagnosis. The condition usually presents as clusters of symptoms, often overlapping, which may change over time and can affect any system within the body” [147]. Notably, post-COVID-19 patients may exhibit extra-pulmonary manifestations, including fatigue, generalized pain, persisting high-temperature and psychiatric problems, neurological, cardiovascular, and musculoskeletal disorders, burdening their functional status [148,149]. Indeed, also physical, sensorial, cognitive, and emotional consequences in COVID-19 survivors should also deserve further investigation in order to better define their link with the previous COVID-19, considering the nonspecificity of most of these symptoms. Moreover, also their possible function as conditions of acquired vulnerability that may have implications for fitness for work also in the view of broader management of return to work in the face of COVID-19 health risks should be deeply addressed [150]. In this context, efforts for risk profiling based on a comprehensive assessment and testing markers of susceptibility in relation to outcomes, detection of early warning symptoms and atypical presentations among patients with multimorbidity should be strongly pursued. This may regard to define, i.e., biomarkers and immune function indicators able to predict how older adults’ immune responses help (or hinder) in fighting off the illness [151]. Finding possible biomarkers could be useful in identifying individuals at the highest risk for severe COVID-19 infection. Biomarkers may identify different risk management groups and yield important information on the mechanisms of severe disease. These may be associated with inflammation, endothelial function, mitochondria and apoptotic function, calcium homeostasis, fibrosis, neuromuscular function, sarcopenia, and bone/hormone metabolism and nutritional status [152,153].As pointed out in previous paragraphs, on one hand, the work can impact a person’s risk to contract the infection; on the other hand, demographic and pathological factors can influence COVID-19 severity. Both aspects are critical in determining the risk of a poor outcome and need to be carefully addressed. Therefore, an occupational risk assessment should include the evaluation of the level of work-based risk to become infected and its mitigation, and personal characteristics that may determine the different impact of the infection. In this perspective, the valuable participation of occupational physicians in risk assessment and management processes appears essential to define plans to protect vulnerable workers, to advise on the fitness for work and to develop a suitable epidemiological surveillance system, all priority measures in effective anti-COVID-19 workplace strategies [1,135].Additionally, it should be noted that the environmental and occupational risk, e.g., the prevalence of the disease at the community and hospital level, and the control strategies, e.g., measures to isolate infected patients, availability of personal protective equipment and vaccines, inevitably vary during the phases of the pandemic. This means that the risk assessment should not be viewed as a “one-off” process but as an iterative procedure to be updated when circumstances, whether environmental, occupational or individual, change. Overall, in actual conditions in which some uncertainties remain on several aspects, any risk assessment may be useful as a starting point for a discussion and the adoption of personalized preventive and protective plans [134]. To this aim, it seems important to underline that the complex interplay between the great variety of individual, physiological and pathological conditions and the huge spectrum of occupational realities requires a “case-by-case” approach to guide risk assessment and management measures. Any attempt to categorize personal or workplace features, in fact, can result in the failure of individually tailored preventive approaches that may result in an ineffective protection of susceptible workers.Occupational health settings may represent appropriate scenarios for the early identification of vulnerable subjects. However, this mission finds its strategic importance once it is able to guide risk assessment and management procedures. This latter may include the systematic surveillance of work-related risk factors, collective preventive policies, the adoption of more stringent actions specifically focused on particular groups of the population (e.g., stricter measures of personal hygiene; implemented personal protective equipment; prioritized access to the ongoing vaccine campaign), and decisions on occupational placement of workers (e.g., remote working; changes in job tasks) [151]. Suitable management strategies can also regard the need to implement the general preventive measures characterized by worker/patient information on occupational risk mitigations and education to recognize signs and symptoms of COVID-19, as well as to follow strict respiratory and hand hygiene. Health promotion activities tailored to support healthy behaviors, lifestyle, emotional health and wellbeing in vulnerable workers should be developed. Regular physical exercise and the use of functional food, smoking cessation programming and the importance of appropriate management of chronic conditions should be stressed in specifically focused populations [154,155,156]. The use of remote monitoring, teleconsultation, telemedicine and online apps and resources may be innovative means to achieve suitable risk communication, maintaining strong messaging to promote compliance with key protective behaviors. These tools can also assure vulnerable subjects regular follow-up, and constant provision of health, behavioral and psychological support, which may help in applying specific recommendations for facing COVID-19 related risks in the workplace.A suitable assessment of susceptibility conditions will provide guidance to develop effective surveillance regimens both through strict testing plans intended to reduce the asymptomatic spread and the rapid activation of contact tracing to reduce transmission and overall mitigate the impact on vulnerable individuals [157]. Additionally, a deeper knowledge of vulnerability may be useful to establish priorities in vaccination strategies [158] and to define policies for the correct use of public resources of insurance systems [138]. Overall, additional studies should be aimed to verify the impact that policy measures adopted to face the pandemic may have on vulnerable groups. Certainly, as the knowledge about COVID-19 risks further develops, risk assessments and management for workplace risk factors and individuals, particularly vulnerable subjects, will need to be reviewed. Concerted actions of general practitioners, hospital specialists, occupational physicians and all figures involved in the health and safety management in the workplace should be strongly encouraged to plan suitable preventive measures for vulnerable subjects [159].Conceptualization, I.I.; literature revision, V.L.; writing—original draft preparation, I.I., V.L., L.F.; writing—review and editing, I.I., V.L.; supervision, I.I. All authors have read and agreed to the published version of the manuscript.The research received no external funding.The authors thank Pietro Apostoli for providing input and comments on earlier versions of the manuscript.The authors declare no conflict of interest.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01031.txt

@@ -0,0 +1 @@
+This study was to observe smoking behaviours and infection control behaviours in smokers at outdoor smoking hotspots during the COVID-19 pandemic in Hong Kong. We conducted unobtrusive observations at nine hotspots during 1 July 2019–31 January 2020 (pre-outbreak, 39 observations), 1 February–30 April 2020 (outbreak, eight observations), and 1 May–11 June 2020 (since-outbreak, 20 observations). Sex, age group, type of tobacco products used, duration of stay, group smoking behaviours, face mask wearing and infection control behaviours of smokers, and mask wearing of non-smoking pedestrians were observed. Compared with pre-outbreak, lower volumes of smokers were observed during outbreak and since-outbreak. Smokers gathered more in a group (24.5% and 25.8% vs. 13.4%, respectively) and stayed longer (91.5% and 83.6% vs. 80.6% stayed ≥1 min) during outbreak and since-outbreak than pre-outbreak. Ninety-six percent smokers possessed a face mask. While smoking, 81.6% of smokers put the mask under the chin and 13.8% carried it in the hand, 32.4% did not wear a mask immediately after smoking, 98.0% did not sanitize hands, and 74.3% did not keep a distance of at least one metre. During the COVID-19 pandemic, smokers gathered closely and stayed longer at the hotspots, and few practised hand hygiene, all of which may increase the risk of infection.Smoking is likely associated with more severe outcomes of COVID-19 [1,2], but the controversy over whether smoking will increase the infection risk of COVID-19 is ongoing [3,4,5]. Smoking, with repetitive hand and mouth contacts, may put smokers at higher risk of infection, since the predominant transmission routes of COVID-19 are respiratory droplets, personal contact, and contaminated surfaces [6,7]. Although being the most developed and wealthiest city of China and among countries and regions with the lowest smoking prevalence, Hong Kong still has a large number of daily smokers (652,000; 10.5% of all persons aged over 15 years) [8]. Outdoor smoking hotspots have become the most common public places where smokers gather to smoke since the comprehensive indoor smoke-free legislation has taken effect on 1 January 2007 [9]. We have previously described such hotspots and our success in recruiting smokers for smoking cessation projects [10,11,12]. Smoking hotspots were selected if (1) group smokers were observed; (2) a rubbish bin with an ashtray for collecting cigarette butts was nearby; and (3) the observation area had a large number of pedestrians, including smokers and non-smokers [12]. They are open spaces near exits of mass transit railway stations, and entrances of shopping plazas and transport hubs, typically with a rubbish bin for collecting cigarette butts. Group smoking typically involves two or more smokers, standing closely or talking face-to-face in a circle. Some studies elsewhere suggested that group smoking at outdoor public areas restored the smokers’ social identity and increased their solidarity [13,14]. Talking and smoking at hotspots may help smokers fulfil their social needs of maintaining social connection. However, respiratory droplets produced while talking or exhaling smoke may endanger others if the smoker is infected by COVID-19. Therefore, hotspot smoking is potentially a high-risk behaviour for transmitting COVID-19. In July 2020 in Hong Kong, a COVID-19 case was reported to have close contacts with smokers at a smoking hotspot [15]. Our search of PubMed on 12 January 2021 using keywords of “COVID-19” or “SARS-CoV-2”, or “coronavirus” and “smoking”, showed no reports on smoking at hotspots and the infection control measures taken by smokers.The first confirmed case of COVID-19 was reported in Hong Kong on 23 January 2020, the same day lockdown began in Wuhan, China. The first wave in Hong Kong reached the peak of 10 daily cases on 9 February. Special work arrangement for government departments began on 29 January [16], allowing work from home except for emergency and essential services. Many private sector organisations followed suit. The second wave reached the peak of 65 daily cases on 27 March. In response, the government closed 11 types of businesses since 28 March and banned public gatherings including dining of more than four people since 29 March. With five days of zero local cases during 20 to 28 April, the government relaxed the maximum gathering group size to eight from 8 May (effective until 18 June) [17].The special work arrangements and social distancing measures have markedly reduced the volume of pedestrians in outdoor areas. However, the effects on the volume of smokers at hotspots and their smoking related behaviours were unclear. The World Health Organization (WHO) recommends 1 metre (3 feet) of social distance [18] and the US Centers for Disease Control (CDC) suggests 2 metres (6 feet) to avoid person-to-person spread of COVID-19 [19]. The Hong Kong government recommends a distance of at least one metre from others, but whether this was followed by smokers at smoking hotspots is unknown.In Hong Kong, mass masking had been advocated early before COVID-19 was declared a pandemic as a useful adjunct to social distancing and hand hygiene [20]. Face mask wearing by the general public in Hong Kong was 97.5% in February and 98.8% in March from a telephone survey, and was 96.6% in April 2020 from an outdoor observational study [21,22]. The prevalence of mask wearing in smokers and how they handle the mask during smoking are unknown. With no convenient hand washing facilities near the smoking hotspots, whether smokers use hand sanitizer before and after smoking is also unknown.This first unobtrusive observational study aimed to investigate: (1) the changes in volumes (persons per hour) of smokers from 1 July 2019–31 January 2020 (pre-outbreak period), to 1 February–30 April 2020 (outbreak period, i.e., 1st and 2nd wave), and 1 May–11 June 2020 (since-outbreak period, i.e., post 2nd wave), and whether similar changes occurred for pedestrians in general at the hotspots (including those who were not smoking), (2) smokers’ group smoking behaviours during the above three periods, and (3) smokers’ infection control behaviours from 20 May to 11 June 2020 (in the since-outbreak period). A total of 67 observations were conducted at noon (11 am to 2 pm) and in the afternoon (3 to 6 pm) on weekdays from 1 July 2019 and 11 June 2020. We started doing hotspot observations from July 2019 as part of a project on tobacco control policy commissioned by the government. Since the COVID-19 outbreak, the observations included COVID-19 related behaviours, although they were interrupted to protect our observers when case numbers were high.An observational study was conducted at smoking hotspots to measure the volumes (persons per hour) of smoking and non-smoking pedestrians, and to observe their behaviours. Ethical approval was granted by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (No. UW19-169, date of approval: 23/4/2019, revision approved on 20/5/2020).Based on our previous studies and pilot observations in different districts [10,11], we selected nine hotspots with the largest number of pedestrians and smokers in Hong Kong Island (four sites: Admiralty, Hong Kong Station, Causeway Bay, and Sheung Wan), Kowloon (three sites: Tsim Sha Tsui, Mong Kok, and Kwun Tong) and the New Territories (two sites: Tsuen Wan and Kwai Fong). The selected sites should be feasible to conduct multiple unobtrusive observations from a distance of at least 10 metres. The boundaries of each hotspot were clearly defined by fixed structures (e.g., poles, walls, and curbs) or markings in the environment to facilitate observations by two observers. The size of each hotspot was around 20 m2. People who were holding or using any tobacco products were recorded as smokers, and otherwise as non-smoking pedestrians. Each hotspot was observed 6–10 times, with a total of 39 observations during the pre-outbreak period, eight during outbreak, and 20 since-outbreak. Apart from the district of the hotspot location (Hong Kong Island, Kowloon, and the New Territories), the observation time of the day (noon: 11 a.m.–2 p.m., afternoon: 3 p.m.–6 p.m.), weather (rainy, sunny, or cloudy), and the average temperature of the day were also considered as possible determinants of the volume.A total of 35 observers were trained by the research team. Two observers were needed for each session to record smokers and non-smoking pedestrians unobtrusively. All observations were conducted in noon or afternoon sessions on weekdays when the volume was largest, each lasting for 3 h. During the observations, one observer recorded the smokers, while another observer recorded the non-smoking pedestrians. Smoker’s sex (male, female), age group (adolescents under 21, young adults 21–40, middle-aged 41–60, and elderly >60), types of tobacco products used (conventional cigarettes, e-cigarettes, heated tobacco products [HTPs], and others), duration of stay (passer-by, <1 min, ≥1 min), group smoking behaviours (if two or more smokers appeared to be smoking together as evidenced by staying close, facing each other or chatting), and the number of cigarettes consumed by each smoker (if they used two sticks or more during their stay at the hotspots) were recorded. In the training session, a set of photos of smokers taken on streets were provided to the observers for establishing a consistent standard on judging the age group. The total numbers of non-smoking pedestrians passing by (sex and age group) were recorded concurrently. The 3-h observations were separated into six sessions, with each session lasting for 30 min. Between the sessions, the observers could take short breaks or exchange the roles to avoid fatigue effect, e.g., change from observing smokers to pedestrians. It was common to record large number of pedestrians passing by in a short period. To save the time for each record and to reduce error, the observer only made a simple mark in the corresponding sex/age column (eight columns in total) during the observation and calculated the total number under each category after they had completed all the sessions. A new session always began on a new worksheet to avoid confusion with the previous records.The observation forms (smokers: sex, age group, types of tobacco used, and duration of stay; non-smoking pedestrians: sex, age group) were validated by two observers in June 2019. Two pilot 1-h observations of smokers and non-smoking pedestrians on two smoking hotspots (sites AD and UN) were conducted by two observers independently and simultaneously. Interrater reliability of the form was assessed by intra-class correlation coefficient (ICC) and consistency, for smokers and non-smoking pedestrians, respectively. We used ICC to assess the forms of smokers, since the data were individual-based and could be matched between the two observers. Matching was based on specific appearances (e.g., wearing a pink T-shirt, wearing a hat) and confirmed between the two observers immediately. We assessed the consistency of the forms of non-smoking pedestrians, because the data were aggregated numbers in eight sex-age subgroups (two categories of sex, four categories of age).Whether smokers possessed a face mask and non-smoking pedestrians wore a face mask were added to the form from 4 May to 11 June 2020 during the since-outbreak period. After several sessions of observation, the following infection control behaviours were also documented during 20 May to 11 June 2020: handling of the mask while smoking (put under the chin, carry in the hand, put in the pocket, etc.), wearing a mask immediately after smoking, sanitizing hands at least once during the stay, and maintaining social distance with other smokers (more than one metre).ICC estimates and their 95% confidence intervals were calculated based on a mean-rating (k = 2), consistency, two-way random effects model [23]. Values of ICC < 0.5, 0.5–0.75, 0.75–0.9, and > 0.9 indicate poor, moderate, good, and excellent reliability, respectively [23]. The consistency was calculated by summing up the consistent numbers of non-smoking pedestrians in the eight sex–age subgroups divided by the average total number of non-smoking pedestrians observed by the two observers (consistency of sex = [consistent number of female + male]/total number of participants; consistency of age = [consistent number of adolescents + young adults + middle-aged + elderly]/total number of participants).The volumes (hourly counts) of smokers were used as the outcome variable in the multilevel mixed effects model to evaluate the differences in three periods, after adjusting for observation time of the day, district, weather, temperature, and the clustering effect of location (nine sites). The mixed effects Poisson regression model was used to examine the change of the incidence rate of smokers, using the total number of pedestrians (sum of smokers and non-smoking pedestrians) per hour as offset (taking into account all the people of interest) after adjusting for the above factors. Rate ratio refers to the incidence rate ratio (IRR) [24]. An IRR with a p < 0.05 for two-tailed test indicates a significant difference in the rate of smokers. We fitted the models using Stata 13 (StataCorp. 2013. Stata Statistical Software: Release 13. College Station, TX, USA: StataCorp LP.). The daily mean temperature (°C) was not recorded during observations, but was obtained from Hong Kong Observatory [25]. The change of volumes of smokers and non-smoking pedestrians across the month was evaluated using a non-parametric test for trend as implemented in Stata [26]. The features of smokers and non-smoking pedestrians at smoking hotspots during the three periods were compared using logistic regression after adjusting for sex, age, type of tobacco products used, duration of stay, group smoking, and number of cigarettes consumed. The infection control behaviours of group smokers and single smokers were compared using Chi-square tests. All figures were generated using R version 4.0.2.The smokers (n = 166) and non-smoking pedestrians (n = 490) were observed by two observers simultaneously. The ICCs of sex (0.98, 95% CI 0.97 to 0.99), age groups (0.79, 95% CI 0.71 to 0.84), type of tobacco products used (0.90, 95% CI 0.86 to 0.93), and duration of stay (0.95, 95% CI 0.93 to 0.96) indicated moderate to excellent reliability of the observations on smokers. The consistency of sex and age group of non-smoking pedestrians were both above 0.90 (0.97 and 0.901, respectively).A total of 12,107 smokers were observed in 67 observations at nine hotspots during 1 July 2019 to 11 June 2020. Figure 1 shows decreasing trends in the unadjusted monthly volumes of smokers (nptrend z = −4.58, p < 0.001) and non-smoking pedestrians (nptrend z = −2.13, p = 0.03) (see volumes of smokers at each smoking hotspot in Supplementary Table S1).Table 1 shows the results from the Poisson regression model. The volumes (persons per hour) of smokers reduced significantly during outbreak and since-outbreak periods (outbreak vs. pre-outbreak: adjusted mean difference, −13.1, 95% CI −24.1 to −2.2; p = 0.02; since-outbreak vs. pre-outbreak: adjusted mean difference, −22.0, 95% CI −31.3 to −12.7; p < 0.001). The rate of smokers’ since-outbreak was significantly lower than pre-outbreak by 14% (adjusted IRR, 0.86, 95% CI 0.79 to 0.94; p = 0.001).Table 2 shows most smokers were male (72.2%), young adults (56.5%) or middle-aged (36.7%), and used conventional cigarettes (92.8%). Table 2 shows higher proportions of group smoking during outbreak (24.5%) and since-outbreak (25.8%) periods than pre-outbreak (13.4%). The smokers stayed longer in hotspots during outbreak and since-outbreak than pre-outbreak (duration of stay ≥1 min: 91.5% and 83.6% vs. 80.6%).Table 3 shows that from 4 May to 11 June 2020 during the since-outbreak period, 96.0% smokers possessed a mask while smoking and 98.8% non-smoking pedestrians wore a mask. Online Supplementary Figure S1 shows the day-to-day mask possession of smokers and mask wearing of non-smoking pedestrians during the same period. The proportions of mask wearing of non-smoking pedestrians were all above 97%. The proportion of mask possession in smokers was lower than 85% (84.7% and 83.0%, respectively) on 3 June and 9 June, and was 91.7% on 12 May, although the proportions on the other days were all above 95%.Table 3 shows 81.6% smokers put their masks under the chin, 13.8% carried in the hand, and only 2.0% sanitized their hands at least once during the stay. One-third (32.4%) did not wear a mask immediately after smoking. About three-quarters (74.3%) were within 1 metre from other smokers while smoking. The proportion of maintaining social distancing (at least 1 metre apart) in group smokers was significantly lower than that in single smokers (7.5% vs. 31.1%, p < 0.001). In the group smokers, 325 were with two smokers and 140 were with three or more smokers, of which 91.4% and 95.0% were within one metre from each other, respectively.This is the first unobtrusive observational study of hotspot smoking during the COVID-19 pandemic, which comprehensively assessed the volume of smokers at these spots and their infection control behaviours during smoking. The proportion of face mask possession in smokers and mask wearing in non-smoking pedestrians was, respectively, 96.0% and 98.8% in May 2020, which is similar to the proportions reported by two earlier Hong Kong studies [21,22]. The almost complete masking was voluntary until the third wave, when the government implemented mandatory masking in outdoor public places from 29 July.Our study showed that the volume of smokers decreased significantly in both the outbreak and since-outbreak periods. However, the rate of smokers among all pedestrians during outbreak was even higher than pre-outbreak, although the adjusted incidence rate ratio was not significant. This might be due to the insufficient number of observations in the outbreak period, as we cancelled some sessions to protect our observers. Nevertheless, our results showed that smokers’ presence decreased by a smaller extent than the pedestrians at smoking hotspots, which could be explained by smokers’ craving and higher priority to smoke over the need to prevent infection of COVID-19. Group smoking became more common in hotspot smokers during the outbreak and since-outbreak periods, just when public gatherings were limited to four or eight people from 29 March to 18 June. Furthermore, the smokers stayed longer at the hotspots during the outbreak than before. A recent study showed that the COVID-19 pandemic may enlarge the extent and severity of addictive disorders including smoking due to social distancing and poverty, poor mental health, and insecurity that were exacerbated by the pandemic [27]. With usual social activities such as lunch gatherings much affected by the outbreak, some smokers who still visited the hotspots might exploit such opportunities to socialise with other smokers. Our results may be relevant to other high-density urban settings where smoking is prohibited in public indoor places and people must smoke outdoors. Group smoking behaviours are also common in other regions, but may be in different outdoor settings [28]. In Japan, smoking on the street is also allowed if smokers could find an ashtray or remain at one place [29]. A study in the United States has reported multiple smokers gathered at outdoor public places such as parks, sidewalk cafes, restaurants, and pub patios [30].Apart from the physiological factors that render smokers more susceptible to COVID-19 infection [31], hotspot smoking may also put smokers at a higher risk of infection. We observed that more smokers did not wear masks during the since-outbreak period, while the non-smoking pedestrians were more adherent to mask wearing. Gathering and chatting without masks were common, and most smokers had unprotective behaviours in mask handling, hand hygiene and social distancing. About three quarters of smokers (74.3%) at smoking hotspots did not keep themselves one metre apart. It was unclear whether the smokers had washed their hands before coming out. Even if they had, hand contamination might occur when they used the lift or opened the doors. To reduce the infection risk, public education should emphasize the high risk for hotspot smoking and urge smokers to quit smoking.Our study had several limitations. First, our observation sessions were suspended for nine weeks due to safety concerns during the peak outbreak period (Mid-February to Mid-April 2020) of COVID-19 in Hong Kong. Second, the observation times and locations during the outbreak and since-outbreak were not fully comparable. This was because we had nine smoking hotspots in different districts but only eight observations covering six sites during the outbreak period. Although we adjusted for several factors, residual confounding could not be ruled out. Third, we tried to cover those hotspots with the largest number of pedestrians to observe more smokers. Therefore, the selected sites might not be representative of all the smoking hotspots, and the characteristics of smokers at hotspots could not represent all smokers in Hong Kong. Lastly, the characteristics of the smokers were different during three periods, thus the infection control behaviours of smokers during since-outbreak period might not represent the smokers observed in pre-outbreak and outbreak periods.We first reported that smokers gathered closely and stayed longer while smoking at outdoor smoking hotspots during the COVID-19 pandemic. Removal of face masks and poor hand hygiene were also common high risk behaviours that can increase the risk of infection. Group smoking not only puts the smokers at higher risk, but also may become a transmission route of COVID-19. Besides outdoor smoking hotspots, tobacco use is also common in bars in both Hong Kong and other regions. As the pandemic continues, and evidence of increasing risk of infection and more serious consequences for smokers emerging, smoking cessation campaigns and assistance are urgently needed. Warnings against unmasking for smoking and chatting in groups are needed at smoking hotspots and in mass media campaigns against COVID-19. Smokers should keep at least one metre apart to avoid spreading COVID-19. Due to the increasing severity of the third wave of COVID-19 outbreak in Hong Kong, the government further introduced mandatory masking in outdoor public places and clearly warned that unmasking for smoking was not exempted. Such measures are needed beyond Hong Kong, and could be an opportunity to motivate and support more smokers to quit in Hong Kong and other regions.The following are available online at https://www.mdpi.com/1660-4601/18/3/1031/s1: Figure S1: Mask carrying of smokers and mask wearing of non-smoking pedestrians during since-outbreak period, Table S1: Volumes of smokers and non-smoking pedestrians at 9 smoking hotspots during 1 July 2019 to 11 June 2020.Conceptualization: all authors; methodology: all authors; software: Y.S.; validation: Y.S. and X.Z.; formal analysis: Y.S., Y.T.D.C., and S.Y.H.; investigation: Y.S. and Y.W.; resources: Y.T.D.C., T.H.L., M.P.W., and S.Y.H.; data curation: Y.T.D.C. and S.Y.H.; writing—original draft preparation: Y.S.; writing—review and editing: all authors; visualization: Y.S.; supervision: T.H.L., M.P.W., Y.T.D.C., and S.Y.H.; project administration: Y.S.; funding acquisition: T.H.L., M.P.W., Y.T.D.C., J.C., W.H.C.L., and S.Y.H. All authors have read and agreed to the published version of the manuscript.This study was funded by Health and Medical Research Fund (Ref no.: TC-HKU).The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster (No. UW19-169, date of approval: 23/4/2019, revision approved on 20/5/2020).Consent was waived due to the nature of the study, which was an unobtrusive observational study.The data presented in this study are available on reasonable request from the corresponding author. The data are not publicly available, because the project is a part of a tobacco control policies study commissioned by the government and the data will only be available upon approval of the funder.The authors would like to thank all the research assistants for their work in data collection and observations.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.Monthly volumes of smokers and non-smoking pedestrians during July 2019 to June 2020.The volumes (persons per hour) of smokers at smoking hotspots in different periods.a Pre-outbreak period: 1 July 2019–31 January 2020. Outbreak period: 1 February–30 April 2020. Since-outbreak period: 1 May–11 June 2020. b Noon: 11 a.m.–2 p.m., Afternoon: 3–6 p.m. c Adjusted mean difference for observation time of the day, district, weather and temperature (continues variable), and clustering effect of location. d Adjusted incidence rate ratios for observation time of the day, district, weather and temperature (continuous variable, incidence rate ratios > 1 indicated higher temperature is associated with larger rate of smokers), clustering effect of location, and total number of pedestrians (sum of smokers and non-smoking pedestrians) per hour as offset (taking into account all the people of interest). * p < 0.05, ** p < 0.01, *** p < 0.001.Characteristics of smokers and smoking related behaviours at smoking hotspots in three periods.a Pre-outbreak period: 1 July 2019–31 January 2020. b Outbreak period: 1 February–30 April 2020. c Since-outbreak period: 1 May–11 June 2020. d 6 (0.05%) record was missing. e 18 (0.1%) records were missing. f 50 (0.4%) records were missing. * p < 0.05, ** p < 0.01, *** p < 0.001.Infection control behaviours of smokers and mask wearing of non-smoking pedestrians at smoking hotspots.a 2 (0.1%) were missing from 3061 cases recorded from 4 May 2020 to 11 June 2020. b 11 (0.5%) were missing from 2081 cases recorded from 20 May 2020 to 11 June 2020.c 104 (5.0%) were missing from 2095 cases recorded from 20 May 2020 to 11 June 2020. d 94 (4.4%) were missing from 2158 cases recorded from 20 May 2020 to 11 June 2020. e 117 (5.4%) were missing from 2158 cases recorded from 20 May 2020 to 11 June 2020.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01032.txt

@@ -0,0 +1 @@
+The two-point force-velocity model allows the assessment of the muscle mechanical capacities in fast, almost fatigue-free conditions. The aim of this study was to investigate the concurrent validity of the two-point parameters with directly measured force and power and to examine the generalization of the two-point parameters across the different functional movement tests of leg muscles. Twelve physically active participants were tested performing three functional lower limb maximal tests under two different magnitudes of loads: countermovement jumps, maximal cycling sprint, and maximal force under isokinetic conditions of the knee extensors. The results showed that all values from the two-point model were higher than the values from the standard tests (p < 0.05). We also found strong correlations between the same variables from different tests (r ≥ 0.84; p < 0.01), except for force in maximal cycling sprint, where it was low and negligible (r = −0.24). The results regarding our second aim showed that the correlation coefficients between the same two-point parameters of different lower limb tests ranged from moderate to strong (r −0.47 to 0.72). In particular, the relationships were stronger between power variables than between force variables and somewhat stronger between standard tests and two-point parameters. We can conclude that mechanical capacities of the leg muscles can be partially generalized between different functional tests.Muscle mechanical properties and their evaluation are known to be complex [1], as muscle strength depends on the current level of neural excitation, muscle contraction and time elapsed since the change in muscle excitation [1,2]. The slower a skeletal muscle shortens, the more force is generated during contraction (also force-velocity relationship), [3,4] is a basic principle of skeletal physiology [4]. Studies have been carried out first on isolated muscles and later on single and multi-joint movements [5]. Nevertheless, the expansion of scientific knowledge about the force-velocity relationship (F-V) began several years ago with the study of Jaric in 2015 [1], who proposed that F-V follows a linear form in multi-joint movements [5].The standard testing procedures applied for the assessment of leg muscle capacities often consider the performance of a single external testing load [6,7] and therefore assessment in a single mechanical condition. Consequently, the outcomes observed in this way do not allow differentiation among different muscle capacities, such as those for generating high force (F), velocity (V) and power (P) [1,8]. In addition, standard testing procedures often include movements that are not specific to sports or daily activities [9], or they may cause excessive strain on the musculoskeletal system. The outcomes of most routine testing procedures have been of limited informational value and therefore a number of issues in research have originated from arbitrarily interpreted experimental findings on specific muscle capacities. As opposed to standard testing procedures, the F-V relationship of multi-joint movements provides the possibility to selectively assess F, V and P generating capacity of the tested muscles [1]. Although it has been accepted for several decades that the F-V relationship has an approximately hyperbolic shape [4,10], recent studies indicate that the proposed relationship appears to be approximately linear and strong for multi-joint movements [11,12,13,14,15]. Moreover, although it is considered curvilinear, the F-V relationship also appears to be linear for single joint movements, as tested by isokinetic dynamometry [10,16]. Several authors [9,13,17,18,19] have already suggested that the linear F-V relationship could be developed into a routine test of mechanical muscle capacity in elite sports [9,16] or in older adults [20]. It has been shown that the application of different loads (regression F-V model) is time consuming, prolongs the procedure and tends to cause fatigue [9,10,21]. These findings dictate the use of a recently proposed two-point (i.e., two-load) method for testing various movement tasks that involve only two different external loads [8,9,15,21]. Specifically, this method provides the parameters representing F0 (i.e., the force intercept), V0 (velocity intercept) and P0 (calculated from the product of F and V) of the tested muscles [16]. The two-point model allows the assessment of the muscle mechanical capacities in fast, almost fatigue-free conditions. Therefore, it is suitable for testing more sensitive populations, such as young athletes, professional athletes recovering from injury or the elderly [21,22]. These two-point parameters correspond to the standard linear F-V relationship parameters obtained from several external load magnitudes [15,23]. Therefore, adding an additional load to the standard tests could allow the assessment of the mechanical muscle capacities (i.e., F, V and P), providing a deeper insight into the function of the tested muscles and resolve a number of questions questioned in the literature. In addition, such knowledge could also improve the outcomes of muscle testing in different environmental scenarios and physiological conditions to understand the human body’s adaptations and reactions to temperature [24,25,26], altitude [27] or dehydration [28,29,30].Although jumping on force platforms [31,32,33], cycling [32] and isokinetic dynamometry [23] are valid standard tests of leg muscle capacities and in assessing the F-V relationship [34], there is a lack of data regarding the relationship between outcomes of these tests. It should be kept in mind that the implicit assumption of any standard muscle capacity test is that the results typically observed in very few tests and muscles can be partially generalized to other muscle systems that perform different functional movements [33,35]. Accordingly, only one study examined the generalizability of the linear parameters of the F-V relationship for leg muscle capacities [15]. The authors concluded that the linear F-V relationship parameters could only partially be generalized to different muscle groups [15]. However, to our knowledge, the relationship and generalization between parameters obtained from a two-point model for leg muscle capacities have not yet been evaluated.To address the issues discussed, we designed a study to investigate the two-point model parameters based on the linear F-V relationship. The first aim of this study was to assess concurrent validity by comparing the parameters of the two-point model with directly measured F and P obtained using standard testing procedures. The second aim was to investigate whether the two-point model parameters could be generalized across the different functional movement tests that assess leg muscle capacity. Possible results could lead to a practical application of the simple two-point model as well as contribute to a better understanding of mechanical muscle capacities and the function of our muscular system.Twelve physically active participants (female physical education students; age 21 ± 2 years, body mass 67.4 ± 6.2 kg, height 172 ± 7 cm) were recruited for the study. The sample sizes ranging from 3 to 12 appeared to be necessary to detect differences between dependent variables obtained from different loading conditions [13,18]. Participants reported no recent injuries or chronic diseases that could affect the performance tested. All participants were physically active during their academic curriculum, which typically included about 10 h per week of moderate physical activity, and none of them were active athletes. They did however, have experience working out in the gym. The study was conducted in accordance with the Declaration of Helsinki and all participants signed an informed consent form approved by the University of Belgrade, Faculty of Sports Review Board (ID 02-35-1).Body height and body mass were measured with a standard anthropometer (Martin Anthropometer GPM 101, Duebendorf, Switzerland) and a digital scale (SECA 888 Digital Scale, SECA, Hamburg, Germany). The main part of testing procedure consisted of three functional tests for maximum performance of the leg muscles, which were carried out under different loads: countermovement jumps (JUMP), maximal cycling sprint (CYCLING) and maximal F under isokinetic conditions of the knee extensors (ISOKINETIC).The experimental procedure used for both groups of participants was performed during the 4 sessions separated by at least three days of rest. The first test session consisted of anthropometric measurements, followed by a familiarization with JUMP, CYCLING and ISOKINETIC tests. In the second, third and fourth testing sessions, each test was performed separately. Note that the order of the tests was randomized for each participant. Moreover, the loads within each test were randomized. The sessions usually lasted about 90 min. For all tests except CYCLING, the first trial served as a practical test, while the second trial was used for further analysis. Prior to each session, each participant was given a 5-min warm-up period on a stationary bicycle, followed by 5 min of active and passive stretching exercises. Afterwards, participants had a specific warm-up consisting of several trials of jumping and isokinetic extension, but not cycling (because they already had it in the general warm-up). All measurements were performed in the university research laboratory. The process of data collection is shown in Figure 1.The test JUMP with weighted vest and belt (MiR Vest Inc; San Jose, CA, USA; weight approx. 1 kg) was performed on a force plate (AMTI, BP600400; Watertown, MA, USA). Participants were instructed to perform unconstrained maximum vertical jumps “as high as they can” from an upright, standing position with hands on hips [15]. No specific instructions were given regarding the depth of counter-movement. The CYCLING test included the evaluation of the maximum power output of the 6-s maximum wheel sprint [15,36,37] performed on a Monark 894E leg bike ergometer (Monark, Varberg, Sweden). Participants were instructed to perform an “all-out” effort from the beginning of the test and to remain seated throughout the sprint [15,37]. The test started with the preferred leg in the crank position at 45° forward. The seat height was adjusted for each participant based on the height of the greater trochanter while standing parallel to the seat and following the instructions of the bike ergometer [38]. The ISOKINETIC test was performed on the isokinetic dynamometer Kin-Com AP125 (Chatex Corp., Chattanooga, TN, USA). The participants sat in an upright position and were fastened to the test device with the straps around the pelvis, thigh and ankle. The axis of rotation of the dynamometer was aligned with the lateral femoral condyle. For the ISOKINETIC knee extension tests, the range of motion of the knee extension was set from 90° to 170° [39].For the evaluation of the maximum F, V and P (Fmax, Vmax and Pmax, respectively) in various functional tests, the external load condition that is usually used in standard test procedures was selected. The test JUMP was carried out with unloaded vest and belt. For the test CYCLING the external load of 6 kg was used, which corresponded to approximately 8.9% of the participant’s body weight. For the assessment of force in the ISOKINETIC test, the angular velocity was 60°/s, while the angular velocity for maximum power was 180°/s [40].The two-point model consisted of two loads or two velocities, depending on the test, to obtain the parameters of maximum F, V and P (F0, V0 and P0 respectively). Magnitudes corresponded to the lowest and highest loads/velocities that were used in our previous studies [8,10,15]. For the JUMP test, the participants performed 4 countermovement jumps (2 loads × 2 tests). The first load was performed with empty vest and belt, while the second was performed with a load of 24 kg. The trial with the highest peak P was used for further analysis. The familiarization procedure showed that all participants were able to jump with the heaviest load (24 kg). The rest period between two consecutive jumps was 1 min and 3 min between different loading magnitudes [15].For the test CYCLING the participants performed two sprints with the lowest external load of 2 kg and with the heaviest load of 10 kg (2 loads × 1 trial). The rest period between the consecutive sprints was 4 min [15].For ISOKINETIC, the two-point model was not applied to the lowest and highest V that the participants could perform, but rather on the most frequently used test V—60 and 180°/s (2 × 2 trials). Each trial consisted of a single contraction performed as hard as possible and the trial with the highest peak F was used for further analysis. The rests were 30 s between the trials and 1 min between 2 consecutive velocities. A real time visual feedback of the F-time curve was available during the strength assessment [22,41].The same experienced examiner supervised all the tests. Before each test, a detailed explanation and qualified demonstration was given and a standardized verbal stimulus was given. Participants were asked to complete two to three submaximal exercise repetitions before each test series.With regard to JUMP, a specially developed LabVIEW program (National Instruments 2013; Austin, TX, USA) was used to record and process the vertical component of the reaction force. The signals were sampled at 1000 Hz and low-pass filtered with a second-order recursive 10 Hz low- pass Butterworth filter. Integration of the acceleration signal obtained from F was conducted to calculate V [13,42]. The analyzed motion phase covered the time interval from the lowest position of the body center of gravity to the beginning of the flight phase. Thereafter, the maximum value of F, V and P, were obtained from the jumps’ concentric phase.Regarding CYCLING, device software (Monark anaerobic test) was used to acquire P and the frequency data. To obtain the corresponding linear measures, V was calculated from the frequency and the crank length, while F was calculated as P divided by V [15]. The maximal values were obtained for further analysis.With regard to ISOKINETIC, a customer-specific LabVIEW program was used for data acquisition and processing. The force-time curves were recorded at 500 Hz and low-pass filtered (5 Hz) with a second-order Butterworth filter (zero phase delay). Since F was recorded directly, the angular V (rad/s) was transformed into a linear V (m/s) by multiplication with the length of individual lever arms, so the results could be comparable with other tests. The maximal values of F and V were obtained for further analyses.Descriptive statistics were calculated and the data were presented as mean and standard deviation. Prior to the statistical analyzes, initial tests showed that none of the dependent variables deviated significantly from their normal distribution (Shapiro-Wilk test). The variables: Fmax, Vmax and Pmax were assessed using standard test procedures. The two-point parameters F0, V0 and P0 were calculated by fitting a linear regression through the maximum values of the F and V data obtained from 2 loads, i.e., angular velocities, depending on a test. The F-V relationships were extrapolated to determine the maximum F (F0; F-intercept) and the maximum V (V0; V-intercept) and the slope of the relationship (a = F0/V0). Finally, the maximum P was calculated from the product of F0 and V0 (P0 = F0 × V0/4). The relationship between two-point parameters and maximum values from standard testing was tested using Pearson correlations. The Student’s t test for dependent samples was used to test the differences between the two-point parameters and the maximum values obtained from standard tests. The Pearson correlations and the corresponding 95% confidence intervals (95% CI) were calculated to test the relationships between the same variables between different tests. The data were analyzed using SPSS 20.0 software (SPSS Inc. Chicago, IL, USA). Alpha was set at 0.05.Figure 2 shows a two-point model of three different functional tests for the leg muscles. Two-point parameters were determined from the 1 kg and 24 kg for JUMP, 2 kg and 10 kg for CYCLING and 60 and 180°/s for ISOCINETIC. The F0 and V0 were highest for JUMP and lowest for ISOKINETIC. The steepness of the slope, which represents the ratio of F and V, was again highest at JUMP, while it was lowest at CYCLING.Figure 3 shows the differences between the magnitudes of the same variables observed with the two-point model and standard testing procedures. The results in Figure 2 showed significant differences (p < 0.05) for all three tests. The highest values of F and P were evaluated in the JUMP test, while the lowest values were obtained in ISOKINETIC. Note that all values from the two-point model were higher than the values from the standard tests. Figure 2 also shows the relationship between the same variables from different tests. All correlation coefficients were found to be strong (r ≥ 0.84; p < 0.01), except F in CYCLING where it was low and negligible (r = −0.24).Table 1 shows the generalizability of two-point parameters and maximum values from standard tests by correlating the same variables obtained from three different tests. In general, the correlation coefficients ranged from moderate to strong. In particular, the relationships were stronger between P (0.68 on average) than between F variables (0.47 on average) and somewhat stronger between standard tests (0.64 on average) and two-point parameters (0.51 on average).In this study, we investigated the parameters obtained from the two-point model in various functional tests of the lower limb tests. As for our first aim, we compared the parameters with directly measured mechanical muscle capacities assessed by standard testing procedures. The results showed that the two-point parameters were higher than the directly measured variables in all tests. The correlation between them was strong, except for the F parameter in CYCLING, which was low and insignificant. Our second aim was to determine to what extent the parameters of the two-point model can be generalized across the different tests of the lower limb tests. The results showed that the correlation coefficients between the same two-point parameters of different lower limb extremities tests ranged from moderate to strong. In particular, the relationships between P-variables were stronger than between F-variables and somewhat stronger between standard tests and two-point parameters.Although recent studies suggested that F-V relationships could be used in routine testing [15,23], only a few of them investigated the two-point model (i.e., the load). The results of the studies mentioned above showed that the parameters obtained from the two-point model were very similar to those obtained from the linear F-V relationship. Furthermore, the investigation of [34] showed that reliability and validity were highest when the most distant pair of loads (i.e., 20% and 70% of 1 RM) was used among all two-point methods evaluated. Based on this fact, and in line with our previous study [15], we have applied the specific magnitudes of load (described under Methods) in the present study. The results showed that the two-point parameters for JUMP, CYCLING and ISOKINETIC were higher than the directly measured variables force and power. Moreover, the correlation between them was strong, except for the correlation between the parameters F0 and Fmax in CYCLING, which was low and negligible. The possible reason for this result could be that a standard test procedure involving a load of 6 kg (corresponding to 8.9% of the body mass of the participants) was the optimal load for the development of maximum power [17,37] rather than the maximum force. Note that velocity variable was excluded as it could be considered constant in ISOKINETICS. It should be noted that the distance between the applied loads was the furthest in the ISOKINETIC compared to the other two tests (see Figure 1). This could explain the highest correlation between two-point parameters and directly measured variables of F (r = 0.99) and P (r = 0.98) in this test.In routine testing there is assumed that results obtained with only a few muscles can be partially generalized to the entire muscle system [33,35]. Our results regarding correlations between standard leg tests support these findings. In particular, P could be generalized between different leg tests, while F could only be partially generalized. To our knowledge, the relationships between muscle capacities determined by linear F-V parameters (F0, V0, P0) from different tests have so far only been presented by Zivkovic with colleagues [15]. The authors have shown that the generalization of parameters obtained from the standard regression model was inconsistent for arm and leg muscle tests. In general, the results showed that the correlation between the P-variables was higher than between the F-variables. It was concluded that the observed parameters can only be partially generalized. Similar results concerning the leg muscle tests were obtained in the current study, only between two-point parameters. In particular, a moderate correlation was observed between JUMP and CYCLING for P0, while a high degree of agreement was found between ISOKINETIC and the other two tests. These results could be explained by the fact that ISOKINETIC is considered a routine test for assessing muscle capacity [10,22,43]. Similar to the F in standard tests, the parameter F0 could only be partially generalized. These findings represent an advance in the assessment of the mechanical properties of muscles. Thus, with new methods, the mechanical properties of muscles can be assessed by only one test and thus generalized to the entire muscular system. In our case, the possibility of such generalization is much higher when it comes to P, not F, which can only be partially generalized and must be verified with multiple tests.We recognize possible limitations in this article: (i) during testing procedure, the two most commonly used angular velocities (i.e., 60°/s and 180°/s) that are far from the velocity section, therefore it is possible that the accuracy of the F-V relationship could be improved by velocities closer to the velocity section by reducing the extrapolation required to achieve V0 [44]; (ii) there is some information due to the limitations of isokinetic devices for testing very fast movements [5]; (iii) only women were included in the present study, so we cannot generalize the results to both sexes; (iv) only physically active population was included in the study, so we cannot generalize the results of the study to the non-active population; (v) the sample size is rather small, but in line with previous studies in the same field [10,13]; (vi) we did not control for the depth of the squat within the countermovement jumps, however, majority studies from this specific field did not control for the depth of the squat when determining F-V parameters with novel two-point method [13,15,42,45]; furthermore, to minimize the possible effect of the depth of the squat we gave participants instruction “jump as high as you can”; (vii) the nature of the countermovement jump provides an approximate small distance between loads; nevertheless, the F-V parameters in the countermovement jump test have been shown to be reliable and valid [13,14,22,45,46].In summary, the present study showed a high degree of agreement between standard tests and the novel two-point model in general. Furthermore, the results showed that the mechanical capacities of the leg muscles can be partially generalized between different functional tests. The addition of only one additional load or velocity to the standard functional tests of muscle capacities could distinguish the basic mechanical capacities of the tested muscles. A fairly consistent data set observed when comparing maximum power from standard tests and P0 from the two-point model suggests that it could be used for routine testing. The two-point model could further improve test protocols by allowing easier and faster assessment of maximum F, V and P. Although the correlations for the same variables obtained from different standard tests were moderate to high, further investigation is needed. Further investigation should include more different functional tests performed on different types of subject samples to assess the validity and sensitivity of two-point parameters in the future. In addition, for the application of the two-point model in practice, the methodology must first be standardized, which includes the selection of the type and magnitudes of load and velocity included in testing procedures.Conceptualization, S.Đ. and M.Ž.; methodology, S.Đ., M.Ž. and V.S.; software, S.Đ.; validation, S.Đ., V.G. and N.M.; formal analysis, S.Đ., M.Ž., N.M. and V.S.; investigation, S.Đ., V.G. and M.Ž.; resources, V.G.; data curation, S.Đ., M.Ž. and V.G.; writing—S.Đ., V.G., M.Ž., N.M. and V.S.; original draft preparation, S.Đ. and M.Ž.; editing, S.Đ. and V.S.; visualization, S.Đ., M.Ž. and V.S.; supervision, S.Đ. and V.S.; project administration, V.S.; funding acquisition, S.Đ. All authors have read and agreed to the published version of the manuscript.This work was supported by the Ministry of Education, Science and Technological Development of the Republic of Serbia under Grant [number 175037 and 175012] and Slovenian Research Agency within the Research programme Bio-psycho-social context of kinesiology No P5−0142.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Faculty of sports and Physical Education Committee (protocol code 02-35-1,14.01.2016).Informed consent was obtained from all subjects involved in the study.Data generated and analyzed during this study are included in this article. Additional data are available from the corresponding author on request. The authors declare no conflict of interest.Flow chart of data collection.Two-point model presented for three different functional tests of lower limb muscles averaged across participants.The averaged across the participant values of F (top panel), and P (bottom panel) obtained from the two-point model (open bars) and standard tests (filled bars) for all lower limb tests (means with SD error bars). The correlation coefficients are presented in parentheses above bars (* p < 0.05—significance of correlation). Significant differences between the two-point parameters and maximal values obtained from standard tests are marked (* p < 0.05).Pearson’s correlation coefficients observed among the same two-point parameters and maximal values obtained from standard tests between three different leg tests.In parentheses are shown 95% CI for corresponding correlation coefficient (* p < 0.05; ** p < 0.01—significance of correlations).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01033.txt

@@ -0,0 +1,8 @@
+Potentially toxic elements (PTEs) pollution in the agricultural soil of China, especially in developed regions such as the Yangtze River Delta (YRD) in eastern China, has received increasing attention. However, there are few studies on the long-term assessment of soil pollution by PTEs over large regions. Therefore, in this study, a meta-analysis was conducted to evaluate the current state and temporal trend of PTEs pollution in the agricultural land of the Yangtze River Delta. Based on a review of 118 studies published between 1993 and 2020, the average concentrations of Cd, Hg, As, Pb, Cr, Cu, Zn, and Ni were found to be 0.25 mg kg−1, 0.14 mg kg−1, 8.14 mg kg−1, 32.32 mg kg−1, 68.84 mg kg−1, 32.58 mg kg−1, 92.35 mg kg−1, and 29.30 mg kg−1, respectively. Among these elements, only Cd and Hg showed significant accumulation compared with their background values. The eastern Yangtze River Delta showed a relatively high ecological risk due to intensive industrial activities. The contents of Cd, Pb, and Zn in soil showed an increasing trend from 1993 to 2000 and then showed a decreasing trend. The results obtained from this study will provide guidance for the prevention and control of soil pollution in the Yangtze River Delta.Agricultural soil is the basis of human production and social development [1]. Compared with the world average, China faces severe pressure on per capita farmland resources [2]. As of late, however, the shortage of land resources in China has been exacerbated as a result of pollution by potentially toxic elements (PTEs), which is caused primarily by industrial production and rapid urbanisation [3,4,5]. Therefore, it is of great significance to assess the soil environmental quality for the rational use of land resources and the protection of public health.The results of a national survey conducted between 2005 and 2013 showed that approximately 16.1% of the collected soil samples exceeded the safety limits countrywide, thus facing severe problems of PTEs pollution, especially in developed areas, such as the Yangtze River Delta (YRD) region [6,7]. Although the general state of national soil pollution can be ascertained from this survey, it does not provide a comprehensive understanding of the elemental distribution and concentrations. In addition to this national investigation, various studies have been conducted regarding specific contamination situations in many fields of interest [8,9,10,11,12]. However, most of this research focuses only on local regions, which provide an insufficient representation of the overall status of large regions. Thus, it would be valuable to establish a method to evaluate large-scale PTEs pollution in soil.The temporal trend of soil pollution research is significant for regional soil environmental risk management [13,14]. However, currently, there is a lack of large-scale and long-term soil pollution field monitoring networks. Pollution situations at varying times can be obtained via a retrospective analysis of previously published studies [15,16,17,18]. Using this long-term trend information, researchers can reasonably predict the development trend of regional soil PTEs pollution and provide a decision basis for soil pollution prevention and control [19,20,21]A meta-analysis is an effective strategy used to synthesise the results of multiple studies in order to obtain the overall trend of the target subject [22]. This method has been widely applied in evidence-based medicine and has been certified to be useful with ecological data. For example, researchers [23] compared 193 studies to evaluate the temporal yield stability of different cropping systems using a meta-analysis. Guo et al. [24] discovered the effect of land-use changes on soil carbon stocks by reviewing 74 publications. Nevertheless, few studies have conducted a large-scale and long-term analysis of PTEs pollution in agricultural soil using a meta-analysis [25,26].As one of the most economically active regions in China, the YRD creates nearly a quarter of China’s Gross economic product with less than 4% of its land area. The YRD is densely populated (with a population of 227 million), and a large number of chemical, pharmaceutical, printing and dyeing plants, etc. located here. Human activities have a huge impact on the soil environment, therefore, soil PTEs pollution in the YRD has attracted increasing attention [27,28,29]. Therefore, this study conducted a meta-analysis based on 118 published papers from 1993 to 2020 to evaluate the status and temporal trends of PTEs pollution in its agricultural soil. The main objectives of this study were to (1) assess the overall pollution status of eight PTEs (Cr, Pb, Hg, As, Cd, Zn, Cu, and Ni) in the agricultural soil of the YRD; (2) explore the spatial pattern of PTEs pollution in the agricultural soil of the YRD; and (3) investigate the temporal trends of metal contents and identify potential drivers.Peer-reviewed 118 publications between 1993 and 2020 were collected using the keywords ‘potentially toxic elements’ OR individual elements (Cd, Cr, Hg, Pb, As, Cu, Zn, and Ni) AND ‘farmland soil’ OR ‘agricultural soil’ AND ‘Yangtze River Delta’ OR individual provinces (Zhejiang, Shanghai, Jiangsu, and Anhui) in the Web of Science and China National Knowledge Infrastructure databases. These primary studies were further screened according to the following criteria: (1) only field experiments monitoring surface (0–20 cm) soil in the YRD farmland region were collected; (2) selected studies should record the number of sampling sites and the size of the research area; (3) to ensure the quality of documents, the preparation, and analysis of soil samples should refer to the standards of China environmental protection industry [30,31]; and (4) the mean, standard deviation, and range could be extracted directly from the graphs, tables, and text, or could be calculated from the primary studies.The extraction values of each study contained (1) the basic information (title, published year, author, keyword, and journal); (2) the location (name and administrative code of province, city, and country); (3) emission source (according to the description of the environment around the study area and the main anthropogenic emissions mentioned in the article), wherein the areas were divided into ‘normal group’ (rural farmland) and “HS group” (mining and smelting areas or industrial areas) for further subgroup analysis; and (4) the summary statistics of the contaminations (mean, standard deviation, maximum and minimum element contents for Cd, Cr, Hg, Pb, As, Cu, Zn, and Ni). The extracted data are presented in Supplementary 2.Due to the varying experimental regions and analysis methods, there was bias and heterogeneity in the primary extracted data. To ensure that the results were statistically significant, the data were tested and classified before analysis and calculation. The I2 statistic was calculated to assess heterogeneity [32] because it represents the percentage of individual heterogeneity in the total heterogeneity according to the values of Q and df. df represents the degree of freedom (k–1), and Q follows the chi-square test. An I2 less than 50% showed that multiple similar studies had homogeneity, and the fixed effect model was chosen to calculate the fitting effect value [33]. An I2 more than 50% showed heterogeneity, and thus, further sensitivity analyses, removal of outliers, or subgroup analyses were required. In this study, Cook’s distance method was chosen to find and remove outliers [34].The literatures selected in this study were monitored according to the national standard method of China for soil heavy metal content [30]. According to the standard, the total concentration of Cd, Zn, Ni, Cu, and Pb of the soil samples should be acid-digested with HCl-HNO3-HF-HClO4 and then analysed by atomic absorption spectrometer. Different studies may use different proportions of these acids and few studies used HCl-HNO3-H2O2 in the digestion procedure. The total Hg and As determination recommended for using cold atomic fluorescence spectrophotometry with digesting by H2SO4-HNO3-KMnO4 in the standard, or digested by HCl-HNO3 with bathing in the hot water can also be recognised [31].Weight is a very important indicator in the process of meta-analysis, as it is used to calculate the average value. All the studies selected for analysis herein were conducted with field experiments, and several important indicators could be used as reference values for weight calculations, including research area and the number of sampling sites, which determine the degree of representativeness of the study for the whole region, and the variation of the measured value, which determines the reliability of the study. Therefore, the weight value calculation in this study can be obtained using the following:(1)Wi=Ai×NiSdi,
+        where Wi is the weight related to each individual observation, and Ai, Ni, and Sdi are the size of the research area, the number of soil samples, and the standard deviation of the PTEs in each study, respectively.The fitting effect value referred to is
+        (2)C=Ci×Wi∑i=1nWi,
+        where C is the weighted mean, and Ci and Wi represent the calculated mean concentration of the PTEs reported and weight in each study, respectively.To solve the problem of several studies having high weights that affect the fitting mean, the natural logarithm of the weight in every study was calculated using Equation (3), and the weighted mean was recalculated with Wi* as follows:(3)Wi*=lgAi×Ni/Sdi,
+        (4)C*=Ci×Wi*/∑i=1nWi*,
+        where Wi* and C* are the logarithmically transformed weight and recalculated weighted mean, respectively, and Ai, Ni, Sdi, and Ci have the same meaning as in Equations (1) and (2).The effect of PTEs pollution on the biological population and potential ecological risk is described using the potential ecological risk index (RI) [35], which can be calculated using the following:(5)Eri=∑i=1nTri×CiSi,
+        (6)RI=∑i=1nEri,
+        where Eri  represents the potential ecological index for a single element, Tri is a fixed value that represents the toxin response (Cd = 30, Cr = 2, Hg = 40, Pb = 5, As =10, Cu = 5, Zn = 1, and Ni = 5), and Ci and Si are the calculated mean concentrations of PTEs reported and the threshold, respectively.Data extraction and conversion were conducted using Microsoft Excel 2019 (v.2019, Microsoft Corporation, Redmond, WA, USA). The spatial distribution of the PTEs contents in the YRD was determined using ArcGIS (v10.3, ESRI Inc., Redlands, CA, USA). A meta-analysis was performed using R (v.3.5.3, AT&T, Murray, NJ, USA) with the meta and metafor packages [36].In the data analysis influenced by extreme values, the distribution of the eight elements’ concentrations represented various degrees of skewness (Figure S1, Supplementary 1). Thus, Cook’s distance method was used for outlier diagnosis. Meanwhile, we noticed that agricultural areas at higher risk of pollution are more likely to attract researcher attention, such as mining and smelting areas and industrial production areas, which might introduce extreme values. Therefore, a subgroup analysis was performed to identify publication bias, wherein the data were divided into ‘HS’ and ‘Normal’ groups according to their emission source. Figure 1 shows the comparison of the weighted average calculated by the four datasets, which were ‘all data’ (all data collected from the literature), ‘removed outliers’ (all data excluding outliers), ‘normal group’ (rural farmland), and ‘HS group’ (mining and smelting or industrial areas). We found that the mean values of the HS group were obviously higher than those of the normal group, especially for Cd, Hg, and Cu, which proved the existence of publication bias. However, there were no remarkable differences between the contents of removed outliers and normal group, suggesting that publication bias had no significant impact on the overall results. Therefore, the data with the outliers removed were used in the subsequent analyses.The overall regional PTEs contents in the YRD agricultural soil are summarised in Table 1. The sampling number showed that more attention was paid to Pb (175) and Cd (153), and less to Zn (104) and Ni (49). The regional mean concentrations of Cd, Hg, As, Pb, Cr, Cu, Zn, and Ni were 0.25 mg kg−1, 0.14 mg kg−1, 8.14 mg kg−1, 32.32 mg kg−1, 68.84 mg kg−1, 32.58 mg kg−1, 92.35 mg kg−1, and 29.30 mg kg−1, respectively. Compared with the national standard, the proportion of samples that exceeded the screening value (GB 15618-2018) [6] was the highest for Cd, wherein approximately 26.14% of the samples were contaminated by Cd, followed by Zn (6.94%) and Pb (3.43%). There were no samples with excessive Cr or Ni concentrations. Nevertheless, in general, the weighted mean concentration of the eight PTEs was within the safe range, exhibiting varying degrees of accumulation compared with the background values.The article numbers were counted and illustrated by points that differ in size based on county-level division (Figure 2). We found that more researchers were concerned about the soil environment in regions along the eastern seaboard, such as Shanghai, Ningbo (Zhejiang), and Suzhou (Jiangsu), and what they have in common are frequent industrial activities and high population density. In addition to the overall pollution situation analysis of the YRD, the spatial distribution was analysed by the administrative division. The specific calculated mean concentrations of eight elements in the soil were performed in Table S1. Figure 3 shows the results of the average concentration calculated by provinces in the YRD, including Shanghai, Jiangsu, Zhejiang, and Anhui, as compared with the background values. There was no obvious PTEs accumulation in any province’s agricultural soil, with the exception of Cd and Hg. The Cd concentration in Anhui was significantly higher than its background value, as was the Hg content in Jiangsu. Moreover, a subgroup analysis based on the city level was conducted, wherein darker colours represent higher concentrations (Figure 4).The Cd concentrations in Chizhou (Anhui), Nanjing (Jiangsu), and Ningbo (Zhejiang) were higher than those of other cities, while the Hg and Pb concentrations were higher in Nanjing (Jiangsu) and Wuhu (Anhui). Note that the Pb, Cr, Cu, and Zn contents in Tongling (Anhui) were the highest in the entire YRD. The main reason contributing to this is that Tongling is famous as an important mining region in China which may lead to clear accumulation of soil potentially toxic elements.The ecological risks of each province and the entire YRD were assessed using Hakanson’s ecological risk index [32] (Table 2). Hakanson gave the corresponding risk index classification standard (RI < 150 is low, 150 ≤ RI < 300 is moderate, 300 ≤ RI < 600 is considerable, and RI ≥ 600 is very high). The ecological risk level of PTEs in the agricultural soil of the entire YRD was moderate, wherein the risk was predominantly contributed by Hg and Cd, while other PTEs s did not pose significant ecological risks. The PTEs pollution status of the four provinces showed a trend of Jiangsu > Anhui > Zhejiang > Shanghai, and the risk levels of them were considerable, considerable, moderate, and low. The major pollutant PTEs was Hg in the agricultural soil of Jiangsu, and Cd in that of Anhui. Ecological risk assessments and mapping were conducted for 25 cities in the YRD (Table 2, Figure 5). In terms of cities, no city among them reached a ‘very high’ ecological risk level, 10 cities were of moderate risk, 10 were of low risk, and five were of considerable risk, and the main contaminations in considerable risk cities were Cd and Hg. The results revealed that Tongling (Anhui), Nanjing (Jiangsu), Taizhou (Jiangsu), Suzhou (Jiangsu), and Chizhou (Anhui) are facing a serious pollution risk, wherein Tongling, Chizhou, and Taizhou were mainly contaminated by Cd, while Nanjing, and Suzhou were primarily contaminated by Hg.To further understand the change mechanism and trend of PTEs pollution in the agricultural soil of the YRD, the literature databases were grouped and calculated for a single year from 1993 to 2020. All the data extracted from the literature and weighted mean values of a single year are shown in Figure 6, wherein the temporal variation is represented by a polynomial regression curve. With the change of time, the contents of eight PTEs in the soil have different trends. Compared with other elements, the fluctuations of Cu and Cr contents were slight. The contents of Cd, Pb, and Zn in the agricultural soil showed an overall increasing trend from 1993 to 2000 and then showed a decreasing trend. The Hg, As, and Ni concentrations in the agricultural soil of the YRD showed continuous decreasing trends of different rates for different periods.The weighted mean value calculated by the meta-analysis showed that soil PTEs pollution in the YRD was generally slight. However, Cd and Hg presented a relatively stronger risk of pollution due to their higher accumulation and toxicity levels, especially in some cities in eastern YRD, such as Suzhou (Jiangsu), Nanjing (Jiangsu), and Ningbo (Zhejiang). Previous studies found that human activities, such as industry, agriculture, and transportation, have significant impacts on the soil environment, playing a critical role in the processes of accumulation, spatial distribution, and migration of PTEs in the soil [42,43]. In industrial production processes, wastewater, waste gas, and waste residue directly or indirectly pollute the soil environment [44,45]. In Suzhou, the atmospheric deposition was proved the most important source of soil Hg [46]. The main anthropogenic source of atmospheric mercury involves coal-fired power plants and industrial furnaces in the YRD [47]. As one of the world’s largest industrial cities, Suzhou has numerous industrial boiler equipment and large coal-fired power plants, which could be the cause of its serious soil Hg pollution. Researchers analysed the source in an abnormally high soil heavy metals agricultural area in Nanjing and found that the annual input flux of Cd in the soil through atmospheric deposition reached 7.00 g hm−2, which may be mainly related to chemical activities such as coking in industrial parks near the study area. At the same time, agricultural fertilisation is also an important source; the annual input flux of Cd in the soil via agricultural fertilisation reached as high as 8.94 g hm−2 [48]. In addition, PTEs with high availability are released into the environment directly during the process of metal mining, especially for copper mines [49,50]. For this reason, strong risk of PTE pollution was present in Tongling, which was engaged in frequent copper mining for decades.As previously mentioned, the concentration of PTEs, such as Cd, Pb, and Zn, first increased and then decreased with an inflection point around 2000. According to the mass balance theory, when the input is greater than the output, PTEs accumulate in the soil; otherwise, the opposite occurs [51]. Numerous studies have shown that fertiliser and pesticide applications are directly related to PTEs accumulation, such as Cd, Pb, and Zn, in agricultural soil [52,53,54,55]. To relieve environmental pressure, regulations have been implemented to limit the use of fertilisers and pesticides, including prohibiting pesticides containing Hg, As, and Pb in Chinese agriculture since 2002 [56]. Meanwhile, industrial emissions were also an important source of soil PTEs pollution. Since the 1990s, with the Chinese economy reforming and opening-up, the industry in the YRD region entered a stage of accelerated development. Therefore, the PTEs concentration in agricultural soil in the YRD was higher in the early 21st century. Recently, to establish a sustainable economic development model, the Chinese government adopted a series of policy reforms and control measures to alleviate soil PTEs pollution [57].The output of PTEs in agricultural soil main through crop removal, leaching, and surface runoff [58], and the proportion of their output contribution varies in different regions. In Zhejiang, researchers proved that the annual flux of Cd output from farmland through crop harvest and leaching were 1.26 g hm-2 and 1.80 g hm-2, contributing 34.52% and 49.32% of the total output flux, respectively [59]. However, researches usually proved the leaching losses and crop uptake were usually relatively small compared with the total fluxes of PTEs input into the agricultural soil [60,61]. Therefore, the decrease of PTEs contents in the soil is more likely be the result of the anthropogenic intervention that is the remediation and treatment of contaminated soil. Phytoextraction of Zn and Cd contaminated soil by hyperaccumulator, e.g., Sedum plumbizincicola, has been shown to be effective in the YRD [62]. However, as soil PTEs pollution is highly hazardous, long-term, and irreversible, soil pollution prevention and control should be an ongoing effort.Comparisons between the agricultural soil PTEs content in the YRD and other regions in China, including the Pearl River Delta (PRD), Huabei Plain, and other provinces [63,64,65,66,67,68,69], are summarised in Table 3. The results of this study were highly consistent with the field monitoring results of the YRD farmland soil by Shao et al. [26], especially for Cd, Pb, and Ni. However, the concentrations of Cu and Zn in this study were slightly higher than those found by Shao et al. This discrepancy is probably due to the existence of publication bias, as the concentrations in the ‘normal group’ of Cu (28.69 mg kg−1) and Zn (90.61 mg kg−1) were closer to the field measured results in the YRD. Meanwhile, our results were consistent with the national field monitoring results of Song et al. [70], indicating that PTE pollution in the YRD was at the national average level. The results of this study were also compared with those of the Hunan Province, whose PTE pollution situation has attracted widespread attention as a result of public events, including ‘cadmium rice’ [71,72]. Apparently, the concentrations of PTEs in the soil of the YRD were lower than those in the Hunan Province, especially for Cd, As, and Zn (Table 3). In addition, as an economically developed region, the PRD showed more serious soil PTE contamination than the YRD. However, compared with southern China (YRD, PRD, and Fujian), the soil in northern China (Heilongjiang and Hebei) had less PTE pollution [53]. In summary, the results of the meta-analysis were confirmed to be reliable, revealing that the PTE pollution level of agricultural soil in the YRD was close to the national average level. However, PTE pollution in the agricultural soil of the YRD should not be underestimated, especially for Cd and Hg, even though the pollution levels in the YRD were relatively low compared with other regions such as Hunan Province and the PRD.Average contents of soil potentially toxic elements (PTEs) in the soil of various Chinese regions (mg kg−1).This paper reviewed 118 studies about agricultural soil PTEs contamination published between 1993 and 2020 in China. Overall, the concentration of PTEs did not exceed the national standard but was close to the national average level. With the exceptions of Cd and Hg, the PTEs did not show significant accumulation in the soil as compared with the background value. The eastern YRD showed a higher risk of pollution due to the industrial agglomeration effect, wherein some cities, such as Tongling (Anhui), had a higher risk of PTEs pollution due to robust mining activities. At the beginning of the 21st century, the PTEs content in soil was relatively higher and then experienced a decreasing trend. The available data provide a reference for the prevention, treatment, and remediation of soil pollution by PTEs in the YRD, by the environmental agencies.The following are available online at https://www.mdpi.com/1660-4601/18/3/1033/s1. Figure S1: Probability distribution histogram of eight elements extracted from all studies. Table S1: The calculated concentration of eight elements in agricultural soil in the Yangtze River Delta (YRD) (mg kg-1).Conceptualisation, S.S. (Shufeng She), B.H., L.Z., and Z.S.; data curation, S.S. (Shufeng She), B.H., X.Z., and Y.J.; methodology, S.S. (Shufeng She) and B.H.; resources, S.S. (Shufeng She), X.Z., and Z.S.; software, S.S. (Shufeng She) and S.S. (Shuai Shao); supervision, L.Z. and Z.S.; writing—review and editing, S.S. (Shufeng She), B.H., X.Z., S.S. (Shuai Shao), and Z.S. All authors have read and agreed to the published version of the manuscript.This research was funded by the National Key Research and Development Program (2018YFC1800201), the consulting research project of the Chinese Academy of Engineering (2019-XZ-24), the National Natural Science Foundation of China (41771244), and Key Program of the Natural Science Foundation of Zhejiang Province (LZ21D010002).Not applicable.Not applicable.The authors declare no conflict of interest.Weighted mean values of (a) Cd; (b) Hg; (c) As; (d) Pb; (e) Cr; (f) Cu; (g) Zn; and (h) Ni in different groups.Literature quantity statistics of agricultural soil potentially toxic elements (PTEs) pollution in the Yangtze River Delta (YRD).Average concentrations of (a) Cd; (b) Hg; (c) As; (d) Pb; (e) Cr; (f) Cu; (g) Zn; and (h) Ni in four provinces and the entire Yangtze River Delta (YRD) (background value reference: Shanghai [38], Jiangsu [39], Zhejiang [40], and Anhui [41]).Spatial distribution of (a) Cd; (b) Hg; (c) As; (d) Pb; (e) Cr; (f) Cu; (g) Zn; and (h) Ni concentrations in the cities of the Yangtze River Delta (YRD).Ecological risk levels of potentially toxic elements (PTEs) in the cities of the Yangtze River Delta (YRD).Temporal trend of (a) Cd; (b) Hg; (c) As; (d) Pb; (e) Cr; (f) Cu; (g) Zn; and (h) Ni contents in agricultural soil in the Yangtze River Delta (YRD); (red points represent all data extracted from articles except outliers; black triangles represent the weighted mean data grouped by years; and red curve is a nonlinear fitting for the extracted data).Regional mean values (mg kg−1) of eight elements in agricultural soil in the Yangtze River Delta (YRD).a Regional weighted mean values were calculated using the database without outliers. b Background values [37]. c Standard values were derived from national specification (GB 15618-2018) elements screening value (paddy fields with pH ≤ 5.5). d Percentage of sites exceeding the standard values (GB 15618-2018).The regional ecological risk index (RI) of eight elements in agricultural soil in the Yangtze River Delta (YRD).The national eight elements background values were adopted in the RI calculation of the entire YRD, and the regional background values were used for the provinces and cities.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01034.txt

@@ -0,0 +1 @@
+The COVID-19 pandemic continues to wreak havoc across the globe. According to the Centers for Disease Control and Prevention, limiting face-to-face interaction is the best strategy for reducing the spread of COVID-19. We investigate the impact of social distancing on social connection and well-being, while also considering the moderating influence of smartphone use. In a survey of 400 students, the study presented herein finds that smartphone use attenuates the negative impact of social distancing on social connection and well-being. Contrary to popular sentiments regarding the influence of smartphone use on well-being, increased smartphone use during the pandemic may foster social connection and well-being. Overall, the research presented provides evidence that the perceived loss of social connection with others is not a de facto outcome of social distancing. The study’s findings have important implications for public policymakers, government officials, and others, including consumer researchers. These implications include stressing the important role technology can play in staying socially connected during the current pandemic and the importance of reframing “social distancing” as “physical distancing with social connectedness”.The current COVID-19 pandemic continues to wreak havoc across the globe. As of 29 December 2020, there have been 19,055,869 COVID-19 cases diagnosed in the US and 82,086,503 worldwide. To date, 332,246 have died in the US and 1790,597 worldwide as a result of the Coronavirus pandemic [1,2]. According to the US Centers for Disease Control and Prevention (CDC), limiting face-to-face contact with others is the best strategy for reducing the spread of the Coronavirus disease [3]. Social distancing, providing a safe buffer between ourselves and others outside of our homes, can take on many forms including: (1) staying at least six feet (about 1.8 m) apart from other people, (2) working from home, (3) limiting the size of social gatherings, (4) not using public transportation, (5) avoiding crowded places and large gatherings, (6) sheltering in place, and (7) self-quarantining if one suspects they may have the disease or have been exposed to someone who does. Schools have transitioned to online learning, businesses have closed, many cities and states at one point have issued shelter-in-place orders, and airline travel has declined significantly [4].Strict social distancing may be the best way to avoid contracting or spreading the COVID-19 disease, but it only addresses our physical well-being. Our overall well-being must consider our mental health as well. The negative impact of the COVID-19 pandemic on our mental health is beginning to emerge. US adults are reporting significant increases in symptoms of psychological distress. One survey found that 13.6% of adults (compared to 3.9% in 2018) reported serious psychological problems. It appears that adults between the ages of 18 to 29 are suffering the most from the pandemic with 25 percent reporting significant psychological distress [5]. Many of those surveyed reported feeling socially isolated, lonely, stressed, anxious, and depressed because of the pandemic [6]. Social connectedness is best understood as the extent to which an individual feels an emotional bond with others [7]. Feeling connected to others is a reward for people who foster and maintain social connections and often leads to a heightened sense of well-being [8].A lack of a sense of social connection to others undermines our mental well-being. Jamil Zaki [9] and other health-care professionals [6,10] have argued that we must reframe how we approach social distancing. Zaki argues that we are practicing “physical distancing” rather than “social distancing” to highlight the need for people to remain socially connected even while keeping their physical distance from others.The primary objective of the present study is to investigate the impact of social distancing on well-being. Given the importance of social connection to human well-being [11,12,13], social distancing would likely undermine an individual’s sense of being socially connected to others and ultimately reduce one’s psychological well-being. This line of thinking, however, may not hold during the current pandemic for two possible reasons. First, technology may allow individuals to remain in touch with those from who they are currently physically distancing. Staying at home may provide individuals with the time to connect or reconnect with meaningful others who were not part of a person’s everyday routine prior to the pandemic. Stories abound about family dinners, dates, gatherings on Zoom, calling friends and family to talk rather than sending an impersonal text, and getting the chance to interact with one’s neighbors since the pandemic began [14]. The pandemic may have reminded us of the importance of relationships [15]. The smartphone, often derided as something that has alienated us from others [16,17], might very well be the tool that allows us to connect or reconnect with family, friends, coworkers, and others [9]. The present study investigates the role smartphone use plays regarding the relationship between social distancing and social connection. Maintaining social distance from others runs counter to human well-being [6]. Humans have a fundamental need to connect with others. In fact, the lack of social connections is tantamount to such health risks as smoking cigarettes or not exercising [18]. Van Orden et al. [19] argue that social changes at the macro level (e.g., the current pandemic) can have a negative impact on individual well-being. The authors use the hypothesis, Is society “pulling together” or “pulling apart”? to explain how societal-level changes can impact individual well-being. For example, the terrorist attacks on 9/11 pulled together US citizens as they came together to fight a common foe. These types of events may bring people closer together. On the other hand, tragedies or events that create social distancing (such as the current pandemic) may be associated with “pulling-apart” communities and individuals. Such “pulling apart” has been found to increase mental health problems [19].Using a survey of college students, Bian and Leung [20] found that loneliness was the most powerful predictor of both bonding and bridging capital. Loneliness has been consistently found to be related to decrements in both physical and mental well-being [21]. When we feel a lack of connection with others, we are more likely to exhibit signs of depression, stress, and anxiety [19].Reframing social distancing, however, may change the direction of its relationship with social connection and well-being. As called for by Zaki [9] and others [22], reframing “social distancing” to “physical distancing” could potentially change its influence on an individual’s sense of social connection and well-being. Social distancing suggests breaking off our connection with others while “physical distancing” connotes that we must keep our physical distance from others, but we need not become socially distant. Ironically, the smartphone and other technologies that have been criticized for separating us from our fellow humans [16,20,23] may now be the best tools available to maintain or grow our social connections. As argued above, social distancing can undermine an individual’s sense of social connection and have a negative effect on one’s psychological well-being. During the current pandemic, however, technology may enable individuals to maintain their social connections despite physical distancing. Smartphones allow us to call, text, direct message, shop, surf the internet, and scroll (or post, comment, and like) our social media feeds. A small body of research has found that social media use can help individuals build bridging, bonding, and maintained social capital [24,25,26]. In a national survey of Chinese adults, Chan [24] found that mobile phone use was positively associated with both bridging and bonding capital as well as psychological and emotional well-being. An experiment conducted by Deters and Mehl [25] asked subjects in the treatment group to make more posts than usual to Facebook. Results of the experiment found that this increased posting led to lower levels of reported loneliness. These same subjects reported that they felt more connected to their friends because of their increased posting. In a large (n = 2708) sample of South Korean adults, Cho [27] found that the use of communication apps on smartphones was associated with lower levels of reported loneliness and enhanced feelings of social capital. Such social capital can reduce loneliness and enhance psychological well-being [28]. Research by Primack et al. [21] and Verduyn et al. [29] found that social media use may foster social connection by providing avenues for social support. Wei and Lo [30] p. 53 argue that smartphone use can enhance social connection providing “instant membership in a community”.Given the current circumstances surrounding the COVID-19 crisis and call for social distancing, the smartphone and other technologies may moderate the otherwise negative influence of social distancing on our sense of social connection and ultimately, our psychological well-being. Next, we present a study designed to test the proposed moderating effect of smartphone use on the social distancing-social connection relationship.We designed and conducted a survey to assess the relationship between social distancing and social connection, as well as well-being. The study also examines the role of smartphone use, in the relationship between social distancing and feelings of social connection, to explore whether the predicted negative association between distancing and feelings of connection is attenuated by smartphone use. The sample consisted of 400 undergraduate students from a large U.S. university (52% female, Mage = 20). Students completed the study in exchange for course credit. The majority of participants were sophomores (i.e., 62% were in their second year of university studies), followed by juniors (28%) and then seniors (9%) and freshman (1%). The ethnicities of the sample were as follows: 75% white/Caucasian, 11% Hispanic/Latino, 8% Asian/Pacific Islander, 4% black/African American, and 2% other. The survey was completed online through Qualtrics in exchange for course credit. The independent variable, social distancing (α = 0.92), was measured using two items developed for the current study in which respondents indicated their agreement on a seven-point Likert scale to the following statements concerning their response to the Coronavirus outbreak: “I have strictly practiced social distancing” and “I have done my very best to quarantine myself from other people”. The mediator, social connection (α = 0.89), was assessed using the Lee and Robbins [31] measure; example items include “I feel distant from people” and “I have no sense of togetherness with my peers”. A seven-point Likert scale was used to assess the items. The moderator, phone usage, was assessed using an objective self-report directly from participants’ iPhones. Specifically, participants who have an iPhone were asked to follow several steps on their iPhone which involved going to Settings and Battery information to find and report the exact amount of “Screen On” time. Several participants reported not having an iPhone, resulting in a final sample size of 378 for the study analyses.The dependent variable, well-being (α = 0.87), was assessed using two items (e.g., “Overall, I am satisfied with my life”) in which participants indicated their agreement on a seven-point scale. The study also included a measure of negative well-being (α = 0.94) which was assessed using the four-item measure (PHQ-4) of depression and stress developed by Kroenke and colleagues [32] and used in related work by David and Roberts [33]. Descriptive statistics and correlations between the study variables are provided in Table 1.The Preacher and Hayes [34] PROCESS macro for SPSS was used to run our empirical analyses. This method uses an ordinary-least-squares (OLS) path analysis to estimate model coefficients and to assess the indirect and/or direct effects of an independent variable [35]. The PROCESS models use a bootstrapping procedure (n = 5000), which does not rely on any assumptions about the normality of the sampling distribution, to calculate the bias-corrected 95% confidence intervals associated with the statistical significance of the indirect effects [34,35,36]. The Preacher and Hayes [34] PROCESS bootstrapping Model 7 was used to test whether social distancing is negatively associated with feelings of connection, which leads to lower well-being, and whether one’s use of his/her smartphone may attenuate the negative impact of social distancing. The model first tests the effects of social distancing, phone use, and the interaction of these variables on perceived social connection (F3, 374) = 9.47, p < 0.10, R2 = 0.07). (Of note, social connection was measured such that higher scores corresponded to greater feelings of being disconnected). The results show that social distancing (b = 0.33, p < 0.05) and phone use (b = 0.27, p < 0.05) are associated with social connection; in addition, the interaction between social distancing and phone usage is significant (b = −0.03, p < 0.05). Specifically, the impact of social distancing on social connection is significant when phone use is low (b = 0.23, p < 0.05), but becomes non-significant when usage is high (p > 0.05). Next, the model tests the impact of social distancing and social connection on subjective well-being (See Table 2 for full results). The main effect of social distancing is non-significant, but the effect of social connection is significant (b = −0.29, p < 0.05). Importantly, the results show support for moderated mediation such that the indirect effect of social distancing on well-being (through social connection) is significant, but only among individuals who had lower phone use (b = −0.06, p < 0.05). These results indicate that social distancing during the current pandemic is harmful to one’s subjective well-being unless individuals actively use their phones, likely to engage with and interact with others virtually. Similar results were found for the depression and stress measures of well-being.Overall, the results show that phone use mitigates the negative impact of social distancing on feelings of connection, and ultimately well-being. In addition, the results provide evidence that could suggest that the negative effects of social distancing may be explained by the way in which individuals experience social distancing, albeit either focusing on being socially distanced such as not being able to see friends or focusing on being physically distanced but yet being able to socially connect or reconnect with meaningful others.As humans, we are social animals and have a primordial need to socially connect with others. Social distancing during the current COVID-19 pandemic on its surface appears to interfere with our basic need to connect with others. The current study, however, provides evidence that the perceived loss of social connection with others is not a de facto outcome of social distancing. Our study found that social distancing is negatively associated with social connection. The more we socially distance ourselves from others the more socially disconnected we feel. Given the importance of social connection, this lack of social connection was found to be associated with higher reported levels of stress and depression and lower subjective well-being.Interestingly, we found that smartphone use moderates the relationship between social distancing and social connection. Higher use of one’s smartphone improves an individual’s perceived social connection which is associated with better psychological well-being. This finding is particularly insightful given the often-maligned impact of smartphone use on relationships and social connections prior to the pandemic [16]. We believe that the use of smartphones to connect with others is still not as effective as face-to-face interaction [8] but appears to be a suitable alternative during a public health crisis like the current COVID-19 pandemic. According to the societal “pulling together” or “pulling apart” hypothesis, societal events like the current pandemic tend to separate individuals from one another and increase mental health problems during and after the crisis [19]. People around the world have found a myriad of ways to leverage technology, including smartphones, to maintain some semblance of connection to others.The current findings have several important implications for public policymakers, government officials, and others, including consumer researchers. First, the importance of technology in keeping socially connected must be stressed. Prior to the pandemic, smartphones were a separating force between people and others in their social network. Since the pandemic, this may no longer be the case. People need to be encouraged to use whatever technology is available to them to stay socially connected. This may be particularly difficult for the elderly and the economically disadvantaged segments of society. Computer-mediated communications are no substitute for face-to-face interactions but appear to be an adequate stopgap measure during the current crisis.A second important implication of the current research is that something as simple as a positive reframing of social distancing can have a salubrious impact on both a sense of social connection and subjective well-being, as well as stress and depression. People must learn that even though they are being asked/require to socially distance themselves from others this does not mean they cannot stay socially connected with others. Although it may be difficult to get people to reframe in a positive manner the idea of staying physically apart from others, the undetermined length of this “new normal” may justify efforts to shift how people think about the difficult task of separating ourselves from others. Psychology professor, Jamil Zaki, suggests we should encourage people to practice “distant socializing”—remaining physically distant but maintaining social connection through technology [9]. Again, the technology that has been blamed for weakening our connections to others may, at least for the time being, be the best means of preserving our imperiled social connections.Although this study is the first to examine the impact of social distancing on social connection and well-being and the moderating role of smartphone use, certain limitations do exist. Larger random samples would help in generalizing the results to the larger population or specific segments that might be particularly vulnerable to the effects of social distancing including adolescents and the elderly.Future research into the role of technology during tragedies that necessitate social distancing like the current COVID-19 pandemic would be enlightening. With the likelihood of the return to a “new normal,” it is critical that we improve our understanding of technology’s role in decreasing the perceived psychological distance between people being asked or required to practice social distancing. Media usage during the current pandemic has risen sharply. Television viewing, internet traffic, and social media usage have all shown sizeable gains [37]. Previous research has shown that some social media platforms are more conducive to fostering connections than others [38]. Additionally, it will be helpful to investigate which individual activities on one’s smartphone are the best at enhancing social connection. Research in this area is essential because many people are spending a lot of time online and are getting their “COVID news” from online sources as well.Lastly, more experimental and longitudinal research is needed. Although the immediate danger to our physical health of not social distancing is clear, little is known about both the short and long-term psychological effects of such pandemics. These short and long-term mental health consequences are of significant importance to individual and societal-level well-being, [39]. Wiederhold [37] p. 275 states that “…mental health issues in the surviving population can have far greater and longer lasting impacts.”Utilizing a large-scale survey of US college students, the present study finds that social distancing leads to lower levels of reported social connection and well-being. This relationship, however, is moderated by smartphone use. Prior to the current pandemic, the smartphone was seen by many as hampering social connection and undermining well-being. Current results suggest that smartphones can foster social connection and ultimately, overall well-being. Important implications of the study include stressing the importance of technology like smartphones in bridging the perceived social distance between people during the COVID-19 pandemic (or other situations that require physical distancing) and reframing “social distancing” to stress that, although we must physically distance ourselves from others, this does not mean that we cannot stay socially connected with the help of modern technology.Conceptualization, M.E.D. and J.A.R.; methodology, M.E.D. and J.A.R.; writing, M.E.D. and J.A.R. 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 guidance of the Baylor University Institutional Review Board (IRB), and was determined April 16, 2020 to be EXEMPT from review by the Baylor University Institutional Review Board (IRB) according to federal regulation 45 CFR 46.104(d)(2).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 issues included in the informed consent.The authors declare no conflict of interest.Descriptive Statistics and Correlations between Social Distancing, Iphone.Use, Social Connection, and Well-beingNote: Correlation coefficients shown in cells; ** p < 0.01; M = mean; SD = standard deviaton.Study Results from PROCESS Model 7 Analyses.Note: Results based on the Preacher and Hayes (2008) PROCESS Model 7. SE = standard error; 95% CI = 95% confidence interval; LL = lower limit; UL = upper limit. a Social connection results (F(3,374) = 9.47, p < 0.01 R2 = 0.07). b Well-being results (F(2,375) = 26.06, p < 0.01 R2 = 0.12).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01035.txt

@@ -0,0 +1 @@
+Background: People who are homeless experience poor health. Reflective of overall health and factors such as acquired injuries, physical ability or functioning is often low among people who are homeless, but there is a lack of consistency of measures used to evaluate this construct. The aim of this study was to evaluate the feasibility of a broad test battery to evaluate limitations in physical functioning among people who are homeless. Methods: This cross-sectional, observational study occurred in a hospital in Dublin, Ireland. We evaluated lower extremity physical function (Short Physical Performance Battery), falls risk (timed up and go), functional capacity (six-minute walk test), stair-climbing ability (stair climb test), frailty (Clinical Frailty Scale), grip strength (handgrip dynamometer) and muscular mass (calf circumference measurement) in a population of people experiencing homelessness admitted for acute medical care. The test completion rate was evaluated for feasibility. Results: The completion rate varied: 65% (Short Physical Performance Battery), 55.4% (timed up and go), 38% (six-minute walk test), 31% (stair climb test), 97% (Clinical Frailty Scale), 75% (handgrip dynamometer), 74% (calf circumference measurement)). Collectively, the most common reasons for test non-participation were pain (24.1%, n = 40), not feeling well or able enough (20.1%, n = 33), and declined (11%, n = 18). Conclusion: The feasibility of the test battery was mixed as test participation rates varied from 31% to 97%. Physical functioning tests need to be carefully chosen for people who are homeless as many standard tests are unsuitable due to pain and poor physical ability.Homelessness is a significant societal and global problem. A person experiencing homelessness is someone without stable housing who may live on the streets, in a shelter, in temporary accommodation, or in some other unstable or non-permanent situation [1]. It is estimated that there are 307,000 people experiencing homelessness in the U.K. [2], 550,000 in the U.S. [3], and 235,000 in Canada [4] at any one point. Homelessness has increased rapidly in Ireland since 2015 [5].Homelessness profoundly affects health [2], and the mortality rate for homeless adults is almost four times higher than in the general population [6]. Recent data from the U.K. report a mean age of death among people who were homeless as 45 years for men and 43 years for women, in comparison with 76 and 81 years, respectively, in the general population [7]. People experiencing homelessness suffer a high burden of ill health and frequently have multiple chronic medical conditions [8,9] as well as mental illness and addiction issues [8]. Common chronic diseases such as chronic obstructive pulmonary disease, asthma, epilepsy, heart disease, and stroke are considerably more prevalent among people experiencing homelessness compared to housed individuals [10]. It has been reported that people who are homeless are 60 times more likely to present for unscheduled health care in the emergency department [11] and there is a higher rate of acute medical admissions [12]. Doran et al. showed that 70.3% of all hospitalisations of homeless people result in either an emergency department visit or readmission within 1 month of discharge, with the majority of hospital readmissions occurring within the first 2 weeks after hospital discharge [13].People who are homeless and admitted for acute inpatient care may represent an especially vulnerable group with distinct care needs due to the complexity of their medical and social problems [14]. It is therefore recommended that clinicians screen this population for physical deficits so appropriate rehabilitation services can be initiated [15]. Due to the earlier onset of geriatric conditions [16] such as falls, poor balance, frailty, and poor mobility, outcomes commonly used to assess people who are homeless are often extrapolated from the geriatric setting [15], but with a much lower mean age [15], the usefulness of these measures is not fully known.Identification and application of a suitable test battery may indicate a person’s ability to perform everyday tasks, their ability to live independently [8,16] and may help identify at risk individuals who need input that is more intensive. This would also target resource use and appropriate discharge planning [16] as well as providing insight into early signs of disability, poor health, and increased death risk [16,17].The objectives of this study were to assess the feasibility of a comprehensive test battery to assess physical functioning limitations in adults experiencing homelessness.This single-centre observational cross-sectional study occurred in St. James’s Hospital, which is a large university teaching hospital serving adults resident in the south inner city in Dublin, Ireland. It is estimated that approximately 1000 people are sleeping rough or in emergency accommodation within the catchment area of St. James’s Hospital [12]. The institutional review board of Tallaght University Hospital/St. James’s Hospital approved this study. All participants provided full written informed consent.A team of expert physiotherapy and medical clinicians/academics devised the test battery by consensus. Considerations were to include tests that: (i) evaluated constructs of impairments, physical functioning and performance, (ii) were not burdensome in terms of time, (iii) required minimal resources in terms of cost and equipment, (iv) could be easy applied to the clinical setting, (v) would be applicable across a spectrum of functional levels, and (vi) displayed sound psychometric properties. The test battery chosen is summarised in Table 1.The clinical lead of the Inclusion Health Service, a consultant general physician (CNC), performed an initial eligibility screen of all inpatients registered as homeless using live daily updates of the Power BI software system from November 2018 to May 2019 in St. James’s Hospital. The European Typology on Homelessness and Housing Exclusion (ETHOS) definition of homeless was employed to register patients [33], which included those who were sleeping rough (those sleeping outside without cover); those living in emergency accommodation such as a hostel, night shelter, or bed and breakfast (B&B) accommodation; those living with family and friends (where possible, this was ascertained); or in a squat.The route to admission for participants was unplanned self-presentations to the Accident and Emergency/Emergency Department, which in cases of medical necessity, unscheduled medical admission to an inpatient ward setting for acute care followed. In the inpatient ward setting, potential participants were flagged to SK and then definitively screened against the following criteria: (i) hospital inpatient, (ii) homeless, and (iii) >18 years. Exclusion criteria were: (i) insufficient level of English to follow instructions required for study participation (unless translator present); (ii) cognitive impairment, delirium, agitated state, or other reasons to a degree that precluded assessment; (iii) medical or orthopaedic reasons that would preclude ability to complete test battery; and (iv) confirmed pregnancy.Suitable patients were given a participant information leaflet and verbal information about the study. Study information was read aloud and worded appropriately to accommodate participants with literacy issues. All participants provided written informed consent prior to participation in the study, and following a process of rolling consent, the participant could quit the assessment at any point. Participants voluntarily participated in this study and no remuneration was provided. Each test was explained briefly and demonstrated to the participant. If they were willing to proceed, each test was carried out in turn.Descriptive analysis was performed with the percentage compliance with each element of the test battery reported. The reasons for non-completion were recorded. The test feasibility index or rate was assessed as a percentage. This was calculated from the number of participants who were able to participate in the test battery divided by the total number of participants who completed the test. The feasibility rates were interpreted based on pre-specified feasibility rates identified by Wouters et al. [34]: <50%, not feasible; 50–75%, quite feasible; and >75%, feasible.The flow of participants through the study is shown in Figure 1. Out of 122 patients assessed for eligibility, 57 were excluded for various reasons. The most prevalent reasons were that the patient was off the ward (n = 23) at the time or patient refusal (n = 17). In some cases, potential participants who were off the ward were recruited at a later time and thus included in the study numbers (n = 65). The results of the test battery are reported elsewhere (paper under review).Participant demographics are presented in Table 2. The majority of participants (n = 44, 67.7%) were men and the median (IQR) age was 45 (38, 56) years with a range of 23 to 80 years. The majority of participants (n = 57, 87.7%) were born in Ireland. Most participants (n = 41, 64%) used hostel accommodation or were rough sleepers (n = 11, 17%). Eleven participants (16.9%) were re-admitted during the data collection period. More than half of the participants (n = 34, 52%) admitted to consuming excess alcohol. A smaller percent (n = 23, 35%) admitted to actively using heroin/intravenous drugs. Many participants suffered from pre-existing health conditions, with hepatitis (n = 27), liver disease (alcohol related) (n = 13), epilepsy/seizure disorders (n = 11), and mental health conditions (n = 17) being among the most common.Participants completed some or all of the outcome measures outlined in the test battery, as shown in Table 3. The entire test battery took approximately 30 min to complete and was completed in the same day where possible or within the same inpatient stay.The quite feasible [34] performance-based tests were the Short Physical Performance Battery (SPPB) (completion rate 65%, n = 42) and Timed Up and Go (TUG) (completion rate 55%, n = 36). The Six-Minute Walk Test (6MWT) and Stair Climb Test (SCT) were deemed not feasible” tests [34] with completion rates of 38% (n = 25) and 31% (n = 20), respectively.The completion rate of tests performed by the assessor was higher than performance-based tests. The Clinical Frailty Scale (CFS) was a highly feasible test [34], which was completed by the majority of participants (97%, n = 63). Measurement of calf circumference was quite feasible [34] with a completion rate of 74% (n = 48).The specific reasons for non-completion of each test are outlined in Table 3. Collectively, the main reasons for non-completion were pain (24.4%, n = 40), not feeling well enough (20.1%, n = 33), and declined (11%, n = 18).This appears to be the first study to evaluate a broad physically focused test battery in hospital in-patients who were registered as homeless. In our sample, the median age was 45 years and 67.7% were men. This is broadly comparable to other cohorts of inpatients who were homeless with an average 46 years of age and 76.1% men [35] from a U.S. study, and 48 years and 76.5% men [36] from a Spanish study. A striking finding was the inability of many participants to conduct simple standard physical tests due to pain and poor physical ability.No physical-focused outcomes [15] have been validated and no core outcome set exists [37] specific to people who are homeless. In this study, we piloted a comprehensive experimental test battery using standard clinical and psychometrically sound tests, mainly extrapolated from the geriatric setting. We chose tests that could be easily applied in the ward-based setting, require no specialist equipment, and be easily interpreted. Despite this careful planning at the outset of this study, not all evaluation tools were feasible for use.We found that pain was the most common reason (24.4%) that precluded participation, highlighting the possible under-treatment of pain in this cohort. We also found that many standard geriatric tests were too challenging for this group to perform, despite a low median age of 45 years. Performance-based tests such as the SCT (31%) and 6MWT (38%) were not feasible [34]. Tests that were completed by the study assessor, such as the CFS (data generated for 97% of participants) and measurement of calf circumference (data generated for 74% of participants) were much more feasible [34]. There was also a higher level of feasibility [34] in low-threshold tests such as handgrip dynamometry (75%). Due to the process of rolling consent, participants could decline to participate in a test at any time, yet declining to complete the test only applied in 11% of cases.Positive aspects of the test battery were the duration of testing (20–30 min), which did not appear to be overly burdensome to participants. Tests were easily conducted in the clinical environment and were safe as evidenced by the lack of adverse effects, but a qualified physiotherapist conducted all tests and assessed whether participants were suitable for test participation. This indicates the need to optimise pain and refine physical evaluation tools in this cohort.Based on the results of this study, to optimise the test battery for further use, we propose a quick standardised test battery outlined in Table 4 below that could be applied to the clinical setting and future research studies. Tests might also be useful in a primary care setting where intervention programmes could be implemented that might improve physical status. To the best of our knowledge, this is the first time that a battery of tests has been proposed for the evaluation of impairments, physical functioning, and performance among people experiencing homelessness. In recommending these tests going forward, we concede there may be a ceiling effect for a small number with high functional capacity [38,39]; therefore, close evaluation of whether tests are applicable across the spectrum of functioning [15] would be required. The next step would be to assess the feasibility, validity, reliability, and sensitivity to detect change of this test battery and establish cut-off points to identify high-risk patients. In addition, as only 38% were able to walk for 6 min, the feasibility of other tests should be investigated such as the 1 min sit-to-stand test, which has been used in Chronic Obstructive Pulmonary Disease (COPD) populations [40], and gait speed over a 6 m course [41]. As pain was the main reason for not participating in tests, we recommend that pain is screened and optimised before the conduction of any physical tests in clinical or research settings. Although the proposed test battery would not be onerous in terms of time, we recommend that if testing could not be completed in its entirety in one session due to reasons such as fatigue or difficulties maintaining focus on physical tasks, that it could be conducted over a number of sessions within a meaningful time period (e.g., single hospital admission).Contextual factors may also be important to consider in the interpretation of this study. Firstly, homeless populations in the U.S. include a high proportion of veterans and ethnic minorities, while the study cohort in Dublin, Ireland, reflecting previous work, includes predominantly white Irish participants and very few war veterans [12]. For instance, in our study >90% were white, which is higher than the U.S. study of hospitalised homeless, which included 62% white people [35]. A second major difference is that publicly-funded free primary and secondary healthcare is available to those falling into the lowest one-third income bracket in Ireland, so insurance status is not a factor limiting inpatient hospital care. Thirdly, in Dublin, homelessness is closely linked to drug use: up to 70% of homeless individuals report having used illegal drugs, many with poly drug use, and >50% report injecting drugs [42]. Approximately 70% of homeless individuals in Dublin consume alcohol at dangerous levels [42]. This reflects that diseases related to alcohol and drug use (abscesses, hepatic failure and haematemesis) as well as seizures are also more common among homeless inpatients in Ireland, which may result from the increased rate of traumatic brain injury and substance use in this population [12]. It is not known whether this profile limits the interpretation of results in other settings.Other limitations to consider were that participants experienced a burden of physical and medical conditions that may have interacted with testing, which is an inherent limitation of evaluating a physical test battery in an acute hospital population. The exact reason for unscheduled medical admission was not recorded for the purposes of the present study. A previous detailed analysis of homeless inpatients (n = 459) within our centre [12] revealed that 94.9% of unscheduled medical admissions were due to physical health needs including pneumonia/bronchitis (11.8%), seizures (8.5%), syncope and collapse (5.7%), acute exacerbation of COPD/asthma (5.3%), abscess (5.0%), cellulitis (4.8%), venous thromboembolism (3.5%), haematemesis (3.27%), hepatic failure (2.18%), and alcohol withdrawal (2.18%). It is likely that a similar pattern pertained to the present study. Acute or chronic presentation of one or more of these conditions or acute trauma may have influenced mobility levels at the time of testing and therefore the ability to participate in the test battery, the extent to which is difficult to elucidate. A further limitation was that this study was subject to selection bias, as all participants were recruited as hospital inpatients. Data were collected from one urban hospital setting, but as deficits in physical functioning ability are prevalent [15] among a range of homeless settings, results may be cautiously applicable to homeless shelters and hostels, but this requires further evaluation. We concede that comparing other non-homeless hospitalised patients matched by factors such as age, sex, and co-morbidities would provide a useful objective comparison, but this was beyond the scope of the present study. This should be a focus for a follow-up study. Finally, we excluded participants with severe impairment who were unable to complete the majority of the test battery due to orthopaedic or medical reasons. Therefore, this study may unwittingly be a snapshot of participation levels in those with less-severe physical ability.The strengths of this study were the application of psychometrically sound measures to evaluate constructs of impairments, physical functioning and performance. A qualified physiotherapist familiar with the testing battery performed all tests and followed standardised methodology. In addition, a reasonable sample size to gather perspectives on the feasibility of these measures was generated in this study.To assess physical ability, it is necessary to use appropriate tools that are context-specific to the population under evaluation. We found the most feasible tests were the CFS, handgrip dynamometry, and calf circumference measurement. The 6MWT and SCT were not feasible tests for use in this cohort. Based on results of this study, we propose a test battery that may be feasible for use in this population. This requires further evaluation, but may be useful for research and clinical studies to more closely investigate physical functioning limitations in people who are homeless.Original conception and design of study: C.N.C., N.M., J.D., and J.B. Participant recruitment and screening: C.N.C. Participant testing, data collection and analysis of results: S.K. and J.B. Study Supervision: C.N.C., J.B., and N.M., Drafting of paper: J.B. and C.N.C. Editing of paper: J.B., C.N.C., N.M., J.D., and S.K. All authors have read and agreed to the published version of the manuscript.This study did not receive grants from any funding agencies in the public commercial, or not-for-profit sectors.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Tallaght University Hospital/St. James’s Hospital Research Ethics Committee (Reference: 2018-10 List 35(3)) which was granted on 30 October 2018.Informed consent was obtained from all subjects involved in the study.Not applicable.The authors have no conflict of interest to declare.Flow diagram of participants through the study.Summary of test battery.Demographic characteristics of participants.HIV: human immunodeficiency virus.Numbers of participants who completed physical test battery.Proposed test battery to evaluate physical functioning limitations in people who are homeless.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01036.txt

@@ -0,0 +1 @@
+Big Data approaches raise hope for a paradigm shift towards illness prevention, while others are concerned about discrimination resulting from these approaches. This will become particularly important for people with mental disorders, as research on medical risk profiles and early detection progresses rapidly. This study aimed to explore views and attitudes towards risk prediction in people who, for the first time, sought help at one of three early detection centers for mental disorders in Germany (Cologne, Munich, Dresden). A total of 269 help-seekers answered an open-ended question on the potential use of risk prediction. Attitudes towards risk prediction and motives for its approval or rejection were categorized inductively and analyzed using qualitative content analysis. The anticipated impact on self-determination was a driving decision component, regardless of whether a person would decide for or against risk prediction. Results revealed diverse, sometimes contrasting, motives for both approval and rejection (e.g., the desire to control of one’s life as a reason for and against risk prediction). Knowledge about a higher risk as a potential psychological burden was one of the major reasons against risk prediction. The decision to make use of risk prediction is expected to have far-reaching effects on the quality of life and self-perception of potential users. Healthcare providers should empower those seeking help by carefully considering individual expectations and perceptions of risk prediction. The technical and scientific progress in the field of personalized medicine is paving the way for new insights into the development, treatment and prevention of diseases. Particularly in the field of psychiatry, recent advances in risk prediction are associated with high hopes that a paradigm shift towards disease prevention will be promoted and that a significant reduction in the burden of disease can be achieved in the long term [1,2,3]. Considerable progress has been made, for example, in psychosis research [4,5,6]. Early detection and prevention centers for mental illness have been established in recent decades, following the modern program of predictive, preventive and personalized medicine [7]. Key goals of prediction in the context of mental disorders are to identify valid indicators for an increased risk, to predict the onset of illness and to offer risk-adapted prevention measures in order to improve overall health outcomes. This could pave new pathways in disease prevention and thereby lower the burden of mental disorders for those at risk for developing a mental disorder [8,9]. However, prevention methods and goals in public health are critically discussed—especially when it comes to mental health [10]. Mental disorders require specific approaches in distinction to physical illnesses, which must also be reflected in the implementation and design of prevention measures [11]. Furthermore, the handling of Big Data is controversially discussed with regard to sensitive patient and health data [12] and its potential impact on privacy and discrimination risks [13].The application of such predictive methods based on Big Data approaches in routine clinical psychiatric practice has yet to be tested [14] and requires careful consideration of the expectations and attitudes of those affected. People with mental health problems will be faced not only with the decision of whether or not to make use of risk prediction, but also the need to understand and carefully evaluate the consequences related to the predicted risk information in order to derive self-determined decisions for their health. An individual’s health literacy, defined as the knowledge, motivation and competencies to access, understand, appraise and apply health information [15], is widely regarded as one of the key factors for conscious and self-determined health decisions with regard to disease management, prevention and health promotion. The concept of health literacy serves as a theoretical framework for dealing with health-relevant information and making decisions concerning one’s own health. Promoting or improving health literacy is not only about conveying information and developing skills to process health information in order to apply medical measures. It is also argued that improved health literacy is crucial for empowerment [16].Thus far, the perception and expectations of the general population have been the main focus in research of risk prediction for mental illness, and only a few empirical studies have been conducted on attitudes towards risk prediction of potential user groups in the field of mental health [17]. Little is known about how people deal with probability-based risk knowledge, what expectations they have with regard to risk prediction and what consequences individuals draw in terms of the use of preventive measures. The present study aimed to explore the views and attitudes towards risk prediction in a population of people who, for the first time, sought help at early detection centers for mental disorders in Germany. We thereby focused on their individual motives for a hypothetical approval or a rejection of risk prediction to shed light on the perspective of those who are and will potentially be confronted with the decision about using risk prediction. This study was part of a larger survey on health literacy in cooperation with three German university hospitals [18]. The initial survey addressed people who, for the first time, sought help in an early detection center for mental disorders in Cologne, Dresden or Munich between September 2014 and September 2017. Inclusion criteria were initial contact, a minimum age of 15 years, sufficient knowledge of German to complete the question independently and the ability of the interviewees to give their informed consent. Informed consent was obtained from all subjects involved in the study. The research project was ethically approved by the institutional review board of the participating health care institutions (Ethics Committee Medical Faculty, University of Cologne: 14-165; Ethics Committee Medical Faculty, LMU Munich: 93-15; Ethics Committee at the Technical University Dresden: EK 141042016).Of 310 participants who took part in the health literacy study, N = 269 persons responded to the subsequent question about potential use of risk prediction. Sociodemographic and clinical characteristics as well as general level of health literacy are listed in Table 1.The study included a paper–pencil questionnaire on self-perceived health literacy and two open questions on the views and attitudes towards risk assessment as well as underlying motives for the potential approval or rejection of risk prediction. Health literacy was assessed by the European Health Literacy Questionnaire (HLS-EU-Q47). The HLS-EU-Q is a validated assessment tool including 47 question on accessing, understanding, appraising and applying health information [15]. Methods and results of the health literacy assessment are described in detail elsewhere [18]. In deliberate distinction to this quantitative research approach, we used a qualitative design to explore the broad range of subjective risk perception and the views, attitudes and underlying motives for the potential use, i.e., the approval of future risk prediction measures or its rejection without anticipating answers and/or drawing attention to (initially) unconscious aspects [20]. The approach of asking about general risk prediction was deliberately chosen to inquire about attitudes towards the researched possibilities of risk prediction in general, including measures to predict physical as well as mental illness. The qualitative data were collected by means of additional open response fields that were embedded in the parent survey, referring to the following question on risk prediction: “Your personal appraisal: Imagine it would be possible to predict which diseases you will get in your life and how existing diseases will develop. This would require a lot of information, such as your diet, exercise, lifestyle, environmental influences, genetic information (DNA) and images of your organs. If you could use this method to predict the likelihood of contracting a particular disease, would you do so? Please give reasons for your answer”.The analysis was carried out in a two-stage procedure using qualitative content analysis (see Figure 1). At the core of the qualitative content analysis stands the step-by-step inductive development of categories alongside the data [21,22]. To ensure content validity, two researchers (PM and AB) independently screened the whole data set and inductively developed a category system. The first set of categories were discussed between the authors and refined throughout the research process, resulting in one common category system. Disagreements were resolved by discussion.In the first step of the analysis, four overarching themes emerged with respect to the potential use of measures of risk prediction: approval—conditional approval—rejection—indifference. In this analysis step, the respective statements were assigned to only one of the above categories, which allowed a percentage distribution of all respondents.In the second step, we sub-coded the participants’ motives in each of these four categories to understand the reasons for the potential approval, conditional approval, rejection or indifference on whether to make use of prediction or not. Due to the thematic density of the statements, multiple assignment to the corresponding motive categories was possible in this step of analysis. Quality criteria of the analysis were achieved in terms of exhaustion and saturation [23,24]. The analysis was carried out using MAXQDA 2018 software [25,26]. Overall, half of the respondents (49%) were in favor of using risk prediction, while 35% would reject such an assessment. In total, 10% indicated that they would make use of risk prediction under certain conditions. Only 6% of the respondents stated that they had no conclusive opinion on the potential use of risk prediction and expressed themselves indifferently (see Figure 1).Most participants took a very clear and unequivocal position on whether they would approve or reject risk prediction. The different perspectives that underlie this overall attitude were reflected in the respective answers. Proponents who were in favor of using risk prediction, for example, replied: “Of course, it is very important to get knowledge about your current and future state of health”, “Yes, it would be stupid not to do it” or “100%, because I would rather disclose private data than risk a possible illness”. Equally strong were the responses related to a rejection of predictive risk assessment. For example, the answer of one interviewee was “No, because knowing about the risk would drive me crazy”. Another participant replied: “No, because then I’ll be just waiting to become ill”.The participants perceived the anticipated benefit of a risk prediction differently. The handling of risk knowledge as well as the consequences of action resulting from this knowledge seemed to be very individual. The question of the potential use showed a strong positioning of the participants, which was partly based on normative arguments, revealing different beliefs and health or life expectations of the interviewees. That highlights the different views and attitudes on risk prediction, which affect the anticipated appraisal of the risk information resulting from it and, in turn, the decision on its potential use.The following section shows the central motives that were reflected in the decisions to approve, conditionally approve or reject the potential use of predictive measures. The respective motive category is printed in bold; a full overview of all categories is shown in Figure 1.The most frequent reason given by participants in favor of risk prediction related to the possibility to make use of preventive measures in order to reduce risk and thereby prevent the onset of a disease. This category was labelled option for preventive measures. Participants in this category largely formulated their responses impersonally in the sense of “one can” instead of “I can”, e.g., “Yes, because in case of a positive result you can initiate preventive measures at an early stage”. It often remained unclear whether the person concerned played an active or passive role in the implementation of preventive measures and behavior. The ambition to make use of risk prediction seemed to lie in the subsequent option to apply the information by making use of preventive measures. The statements indicate that it remains to be seen whether the predicted risk would inevitably result in the actual use of these measures. Following the statements of the interviewees, participants would (re)consider the use of preventive measures on the basis of the predicted risk and the available methods for illness prevention.In contrast to this, the category active health behavior summarized statements of persons who indicated that they would make use of risk prediction in order to use the risk knowledge to actively engage in their own health, e.g., by changing their lifestyle to reduce the risk of developing the disease. One participant answered, “Since I am a person who is very health conscious, I would use this procedure. That way I could adapt my lifestyle to it”. Proactively drawing consequences of action from the risk probability is already part of the answers in this motive category.Other interviewees indicated that they are in favor of using risk prediction because it allows them to adapt their (life) planning. In this context, some of them explicitly expressed the desire to maintain control over their lives. Examples of statements assigned to this category were the following: “I would use this procedure because it allows you to adjust your life plan accordingly” and “[…] so I can prepare myself for any illness that may arise”.Thus, the desire for control and self-control of one´s own actions turned out to be a central common feature of the statements on active health behavior and (life) planning.Answers assigned to the category health consciousness were characterized by the necessity of being informed about one’s own health status. One participant stated that “it is of course very important to get knowledge about your state of health”. Some statements in this category reflected the inherent hope that knowing a certain risk would contribute to an increased (self-)awareness of one’s own body and better health consciousness in general. A further motive for making use of prediction in the future was the participants’ interest in such measures as well as a certain curiosity about the method of risk assessment and the new pathways resulting from it in general. The majority of respondents who indicated that they would make use of risk prediction solely under certain conditions stated that their decision depended on the predicted diseases and therapy options available. Participants answered, for instance, “For diseases with treatment options that would limit the quality of life severely without treatment, I would accept the offer. However, if the quality of life is only slightly limited, or if the disease is untreatable, the information would probably be a much greater burden […]”.Interestingly, further categories related to the main category conditional approval included motives which were partly also mentioned as reasons for the rejection of prediction but were, in contrast, formulated as conditions in this context. These included data security (“I would do this only if the data is collected with my consent and remains absolutely private.”), skepticism towards the method (“I would use it only if this approach has reached the necessary maturity.”) and the certainty of cost coverage of the predictive measure by their health insurance (“Yes, if it’s free of charge.”). A majority of those respondents who were opposed to risk prediction stated that they were concerned that knowing about a certain risk could have negative effects on their mental well-being. Statements referring to mental stress based on the information of a predicted risk were summarized in the category emotional burden. Statements by respondents assigned to this motive category included the following: “No, I would rather not use it. I always worry too much about my body and a prediction of a possible disease would be an extreme psychological stressor for me. For me, personally, the disadvantages outweigh here.”; “No, the probabilities would scare me.”; “No, it would unsettle me too much and make me panic constantly”. Within this category, respondents frequently gave subjective answers including narratives and biographical references. Some interviewees explicitly referred to their mental state, e.g., they spoke of an already “strained mental state” or an “unstable psyche”, which would be further strained by an outcome of a higher risk. Feelings of stress and fear of the future were other personal reasons given in this context. In contrast to other categories, e.g., options for preventive measures or predicted disease and therapy options, statements within this category included personal reference and the individual life situation and were cited as the main motives for rejection of risk prediction.Furthermore, the knowledge about a risk was perceived as a possible interference in life course and life planning. The interviewees indicated that they fear the knowledge about an increased risk, because it would influence their life planning in a negative way. One participant stated, for instance, “No, predictions could influence my life plan to such a degree that it would proceed differently, if this prediction had not taken place.”, while others replied, “No, I wouldn’t use that, because I’m convinced that life shouldn’t be planned”.The category self-fulfilling prophecy included answers in which respondents expressed a suspicion or concern that knowledge of a risk might lead to its fulfilment and thus increase the actual risk. Furthermore, respondents feared that knowledge of a risk alone could make them ill: “I wouldn’t do it because I think that fear of a possible or probable disease could make you sick, even if you don’t end up with the disease in question”.Overall, both categories, interference in life course and life planning as well as self-fulfilling prophecy, represent a clear contrast to the categories active health behavior and (life) planning, which were presented as motives for approval of risk prediction. Here, the fear of losing control seems to be an essential decision component for rejection.Another motive for rejection was based on mistrust in the concept and methodology of risk assessment. The motive category included statements related to skepticism towards the method of risk prediction. Respondents indicated that they did not believe that reliable statements on risk could be made with the aid of the latest technologies. Some interviewees substantiated their rejection with concerns about data security. In this respect, the respondents mentioned on the one hand, a lack of confidence in the technology/IT, with which the patient data would be aggregated and analyzed (e.g., “No, I’m very skeptical about personal data retention and the technology behind it.”). On the other hand, they claim a lack of confidence in the people (doctors, researchers, etc.), who would ultimately handle the data (e.g., “No, I’m a dedicated data protector and do not want to entrust this data to anyone. The best data are no data collected at all. […] The industry standards of privacy are ridiculous.”).Other respondents rejected risk prediction because they did not wish to undertake the effort that a prediction or a potentially increased risk would require and/or simply did not see the need for such a measure. A further motive, which made the interviewees reject risk prediction measures, was the prioritization of their own body awareness, which, according to their own statements, they preferred to rely on, rather than on the analysis of data.Acceptance of predictive measures is an essential requirement for the successful implementation of measures to prevent mental illness. This study shows the variety of attitudes towards risk prediction in a population of people with perceived mental health problems. The consideration of hypothetical individual perspectives and motives in favor of or against risk prediction are particularly important in this population, since recent research demonstrates the feasibility of developing risk prediction models for psychiatric disorders and clinical practice is moving towards embracing prediction and prevention [8]. Thus, people with mental health problems may be increasingly confronted with the decision to choose for or against risk prediction in the near future. The statements of the interviewees show that the expected benefit of risk prediction is perceived very differently by individuals. Reasons and motives behind the attitude for potential use (approval or rejection) illustrate the diverse views and expectations towards risk prediction. Above all, an emotional burden was stated as a central reason against the use of risk prediction, which was thought to be caused by the risk knowledge rather than the potential illness itself. With regard to the study population, we expected that people with current mental health problems may have a particularly high level of attention in terms of possible negative emotional implications. For future research it would be of interest to know whether those who use risk prediction also include the potential harm of emotional burden in their personal risk–benefit analysis. With regard to the population of help seekers, it must be considered that these people are generally not healthy or free of complaints. In moments of help seeking, most already have first symptoms that can show the onset of a mental illness. They may turn to an institution that provides diagnostic and risk clarification. It is therefore all the more surprising that some respondents were vehemently opposed to the risk prediction. At the same time, it should be borne in mind that pre-existing symptomatology, such as depressive symptoms, which affected more than 70 percent of this sample (see Table 1), can have a negative impact on the current future prospects and plans of those affected. However, the justifications were, for the most part, written in a differentiated manner and their content partly indicated a personal risk–benefit analysis that forms the basis of the decision-making process for or against the use of risk prediction. The given motives for the potential use of risk prediction point to far-reaching consequences for users—both for one’s own self-image and for shaping one’s own life. Overall, the differentiated responses reflected different attitudes toward risk prediction, which did not seem to be positively or negatively influenced by the presence of current mental health problems per se. In this respect, it should be noted that participants were, on average, of a relatively young age and had a high level of education (see Table 1).Furthermore, it became clear that the desire for self-determination and personal responsibility is of great importance, both for the motives of rejection and for the approval of risk prediction. This is reflected, for example, in the motive of (life) planning, which was emphasized in favor of risk prediction. Statements by these respondents indicated that they would make use of the predicted risk and all the associated trade-offs in order to plan for their future in a self-determined manner. In contrast to this, the desire for self-determination became equally clear in the category intervention in the life course and life plan, which was a motive for rejecting risk prediction. Both positions convey different values with regard to personal life plans, beliefs and health expectations. Another key example that illustrates these fundamentally different beliefs and approaches to risk prediction is the desire for control of one’s own life. On the one hand, there seems to be a wish for risk knowledge in order to specifically counteract an existing risk through active health behavior and (life) planning. On the other hand, the fear of loss of control was mentioned as a motive against the use of risk prediction as it was formulated, for example, in the motive of self-fulfilling prophecy. Fears of loss of control, as reflected in several categories, could be indicative of internal processes on the part of respondents who reflect concerns of social stigma when deciding for or against risk prediction. In the clinical setting, these concerns may be influenced by the social desirability to risk prediction, namely unexpressed (hidden or even unconscious) motives on the part of clinicians or researchers in favor of risk prediction. From a clinical point of view, giving consent to risk prediction may seem desirable to prevent the undesirable occurrence of the mental disorders. These pre-assumptions could create subtle pressure and force the fear of stigmatization for people concerned in the decision-making situation, which eventually is a major ethical problem.The results show that the respondents who sought help at an early detection center for mental disorders were driven by very different motives when dealing with risk knowledge. The decision for approval or rejection of risk prediction seems to be strongly dependent on individual values and concepts. Thus, the evaluation basis of a risk prediction seems to go far beyond an understanding of objective medical facts. From an ethical perspective, health care providers should take these individual values into account and support a personal risk–benefit analysis when it comes to decision-making about the use of risk prediction in the context of mental health and beyond. Concurrently, this process may be particularly challenging for people with mental health problems. The parent study on self-assessed health literacy has shown that those affected find it particularly difficult to critically appraise health-relevant information [18].In line with these findings, the latest evidence on health literacy showed that people with mental illness experience particular challenges in dealing with health information [27,28,29].Following up on the statistical analysis of the parent study on health literacy among people with mental health problems [14], the intention to make use of future risk prediction or not was not significantly associated with the health literacy levels of the respondents This finding indicates that the decision whether to make use of such future measures is highly individual and relies on personal values, preferences and feelings regarding Big Data approaches on the one hand and the perceived benefit of risk prediction on the other hand. These very personal reasons should be considered separately from individual skills and abilities to process health information.Studies from other medical fields such as breast cancer [30], heart failure [31] and diabetes [32] also impressively showed that risk prediction and risk knowledge are perceived very differently and can have manifold consequences for the individual. Blakeslee et al. (2017) stress that “accepting patients’ experiences and beliefs in their own right and letting them guide the discussion may be important for a satisfying decision-making process” [30] (p. 2353). In contrast to somatic diseases, the risk prediction of mental illnesses is accompanied by disease-specific challenges. While somatic diseases are communicated relatively openly in society, mental illnesses have long been tabooed. While there has been increasing public attention to mental health in recent years, people with mental illness continue to face stigmatization and exclusion, which can make their participation in society considerably more difficult or even prevent it altogether [33]. From an ethical perspective, the question arises how risk knowledge can influence affected persons and their environment. The interviewees anticipated far-reaching consequences of risk prediction for their personal well-being and their life paths. Negative social consequences that could result from this risk knowledge, such as insecurities and prejudices in the social environment or possible consequences for family members that would result from a genetically increased risk, were not explicitly addressed in this population. According to labelling theories of psychiatric stigmatization, a positive test result on an increased risk could lead to early stigmatization of those affected [34]. Persons who are at an increased risk of mental illness could thereby experience similar stigmatization and discrimination as those who are already ill. However, not only stigmatization of third parties but also self-stigmatization is significant in this context. Initial study results show that knowledge of an increased risk can also trigger increased stigmatization stress and self-labelling as “mentally ill” in those affected, which leads to negative effects on their well-being [35,36]. At the same time, clinical practice has shown that risk labelling can also lead to psychologic relief of the affected persons, since the previously irritating, uncontrollable and inexplicable complaints can then be assigned to a disease model [33].The protection of privacy and the sensitive handling of personal health information are rightly emphasized in this context; published risk information (e.g., an increased risk of schizophrenia) can negatively impact individual circumstances, such as insurance relationships or decisions in hiring an individual at increased risk [37]. The challenge is to find a viable way to protect individuals’ health data while enabling the flow of information necessary to promote quality health care [38].As the results of this study show, dealing with and deciding upon risk prediction cannot be broken down to clinical parameters from an objective perspective. On the contrary, individual expectations and perceptions of risk prediction must be considered to maintain or improve the quality of life of an individual. Health literacy in people with mental health problems is critical, especially with regard to appraising health information [18]. Meanwhile, the individual preference as a result of weighing up the advantages and disadvantages of risk prediction is becoming increasingly important, as knowledge of a risk can already have far-reaching consequences for an individual’s lifestyle and quality of life. Interventions to support the decision-making process and dealing with risk knowledge should focus even more on the promotion of competencies to appraise health information against the background of personal values and life concepts. On the basis of the available results, quantitative assessment approaches should be developed to deepen knowledge on subjective risk perception and risk expectations in the context of mental health to facilitate decisions on risk prediction in clinical practice.Decisions on risk prediction require all the more successful communication between the healthcare professional and the person at risk, including comprehensive information on the procedure of risk prediction and its potential consequences for the individual. This is of particular importance for people with low health literacy, who need to be appropriately supported by healthcare professionals in the process of decision-making on the basis of all relevant information. In light of other findings on the lack of understanding probability values even among clinicians [39], training and education of medical personnel should be considered to ensure adequate communication of options and results of risk prediction. The increasing possibilities of risk prediction and the resulting options for action raise hope for the development of sustainable prevention opportunities for people with mental illness. Dealing with risk knowledge, however, must be implemented sustainably by decision-makers, thereby considering individual risk attitudes in order to ensure a high individual benefit. These results can be seen as a call for action for health care policy makers to include, beyond clinical parameters, the personal resources and values of patients for or against the use of risk prediction in medical decisions. Study participants were recruited from three early detection centers, where young people and people with a high level of education were significantly over-represented compared to other populations with mental disorders. Results in the investigated group should, therefore, be regarded as explorative and a bias of the selection of “help seekers” should be considered when interpreting the results. It is of note that of the 310 participants who took part in the parent study on health literacy, 41 did not respond to the question on risk prediction. Possible reasons could be that the question was asked at the end of the survey and, unlike the previous questions, was open-ended and required written responses. In addition, the participants surveyed may have found the question complex and difficult to answer. This may also indicate that the risk prediction decision is not trivial and may overwhelm some potential users. Limitations of this study are that a hypothetical use of risk prediction in general but not specifically related to mental disorders was queried. We surveyed a population at increased risk (in this case, people with increased risk for mental disorders) potentially facing risk prediction options within their initial contact with an early detection center. However, asking about general risk prediction might have influenced the results in two ways. Firstly, studies show that the intention of health-related behavior and theoretical preference may differ from actual behavior. Nevertheless, regardless of actual utilization, there is clear evidence about how different individual preferences and anticipated consequences are perceived with respect to risk prediction. Secondly, it is not clear whether the participants related their answers to certain physical or mental illnesses, which might have influenced their answers. Against the background of societal discrimination and stigmatization tendencies towards mental disorders, it cannot be excluded that a risk prediction focused on mental illnesses would have been met with even greater rejection among the respondents. However, the main purpose of this study was to investigate the manifold views and attitudes towards risk prediction approaches in general. Attitudes towards predicting physical illness are of no less interest in this population, as physical comorbidities are very common in people with mental illness. This study aimed to draw an initial picture of the views and attitudes of potential users with mental health problems. Future studies should also include other consumer groups to get a deeper understanding of the needs and preferences of those concerned. Qualitative content analysis is based on interpretative analyses that do not aim to draw quantifiable conclusions on statistical relationships. However, qualitative research enables access to reality through subjective opinions, perspectives and underlying interpretative processes [20]. In the context of risk prediction, the subjective meaning is given a particularly great significance to determine naturalistically what attitudes participants have towards risk prediction. The results can be seen as an important starting point for further investigations of affected persons’ perspectives, which can contribute to a differentiated implementation of risk prediction in clinical practice.The new possibilities of Big Data-based risk prediction promise a rapid upswing in research of new predictive, preventive, diagnostic and therapeutic approaches to care. In the field of mental health, this approach is also associated with high hopes for a paradigm shift from treatment of illness to illness-prevention and health-promotion. Thus far, little research has been conducted on potentially affected individuals’ values and preferences regarding risk prediction.This study showed that persons with mental health problems tend to have clear preferences for the use or rejection of risk prediction. Preferences were accompanied with various and sometimes contrary motives. The anticipated impact on self-determination was, in particular, a driving decision component, regardless of whether the person in question would decide for or against risk-profiling. At the same time, the results of this study show that decisions on risk prediction are expected to have far-reaching effects on the quality of life and self-perception of potential users. A differentiated view of possible advantages and disadvantages of risk prediction, considering individual expectations and values, seems indispensable. This is especially true in the context of mental disorders, where public and self-stigma is still alarmingly present. The results of such risk prediction approaches could reinforce such tendencies including discrimination. These implications should be taken into account when considering the development and implementation of risk prediction in the psychiatric context.Conceptualization, P.K.M. and C.W.; methodology, P.K.M, C.W. and A.B.; formal analysis, P.K.M., A.B. and A.J.; resources, P.K.M., S.R. and C.W.; writing—original draft preparation, P.K.M.; writing—review and editing, A.B., S.R., A.J. and C.W.; supervision, S.R. and C.W.; project administration, P.K.M.; funding acquisition, C.W. All authors have read and agreed to the published version of the manuscript.This study was funded by the Federal Ministry of Education and Research (BMBF), Berlin, Germany. It was carried out in the context of “SysKomp”, a collaborative project of ethics, law, health system research and psychiatry, which aimed to analyze societal and individual competences for dealing with innovative approaches of risk prediction in medicine. The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board (Ethics Committee Medical Faculty, University of Cologne: 14-165; Ethics Committee Medical Faculty, LMU Munich: 93-15; Ethics Committee at the Technical University Dresden: EK 141042016). 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.The authors would like to thank Nikolaos Koutsouleris and team, Andrea Pfennig and team, as well as Stephan Ruhrmann’s team who assisted in data collection for the study. We also thank Caren Bertram for assistance in data preparation.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.Overview of the two-stage analysis procedure resulting in overall attitudes (approval, conditional approval, rejection, indifference) and the inherent motive categories.Characteristics of study participants.1 ICD-10: International Statistical Classification of Diseases and Related Health Problems. 2 BDI-II: Beck Depression Inventory II [19]. 3 European Health Literacy Questionnaire [15]: Thresholds were set to compare levels of HL, divided in excellent (>42–50), sufficient (>33–42), problematic (>25–33) and inadequate (0–25). The bold words show the category of the following subcategories in this column.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01037.txt

@@ -0,0 +1,4 @@
+Microalgae growth inhibition assays are candidates for referent ecotoxicology as a fundamental part of the strategy to reduce the use of fish and other animal models in aquatic toxicology. In the present work, the performance of Chlorella vulgaris exposed to heavy metals following standardized growth and photosynthesis inhibition assays was assessed in two different scenarios: (1) dilutions of single heavy metals and (2) an artificial mixture of heavy metals at similar levels as those found in natural rivers. Chemical speciation of heavy metals was estimated with Visual MINTEQ software; free heavy metal ion concentrations were used as input data, together with microalgae growth and photosynthesis inhibition, to compare different effects and explain possible toxicity mechanisms. The final goal was to assess the suitability of the ecotoxicological test based on the growth and photosynthesis inhibition of microalgae cultures, supported by mathematic models for regulatory and decision-making purposes. The C. vulgaris algae growth inhibition test was more sensitive for As, Zn, and Pb exposure whereas the photosynthesis inhibition test was more sensitive for Cu and Ni exposure. The effects on growth and photosynthesis were not related. C. vulgaris evidenced the formation of mucilaginous aggregations at lower copper concentrations. We found that the toxicity of a given heavy metal is not only determined by its chemical speciation; other chemical compounds (as nutrient loads) and biological interactions play an important role in the final toxicity. Predictive mixture effect models tend to overestimate the effects of metal mixtures in C. vulgaris for both growth and photosynthesis inhibition tests. Growth and photosynthesis inhibition tests give complementary information, and both are a fast, cheap, and sensitive alternative to animal testing. More research is needed to solve the challenge of complex pollutant mixtures as they are present in natural environments, where microalgae-based assays can be suitable monitoring tools for pollution management and regulatory purposes.Several environmental regulations rely on ecotoxicological data for the assessment and management of chemical contaminants. In Europe, the Water Framework Directive (Directive 2000/60/EC [1]) defines the protection of European waters (surface continental waters, transition waters, coastal waters, and subterranean waters) and regulation, management, monitoring, assessment, etc. rely, among others, on the ecotoxicological test. Ecotoxicology has a holistic approach to the study of a toxic effect on the ecosystem. It has gained a fundamental role in the assessment and management of environmental risks [2]. In aquatic ecotoxicology, the tested compartments typically belong to three basic trophic levels: microalgae (at the bottom of the trophic chain), zooplankton (the intermediate link), and fish (top trophic level) [3]. Depending on the concentration of the compound and the exposure time, the tests are able to assess chronic or acute toxicity. The OECD’s (Organisation for Economic Co-operation and Development) revised version of acute toxicity tests [3,4] states that fish-based tests are only to be performed at one concentration of toxic compound: the lowest toxic concentration for microalgae or zooplankton. It is based on results showing a higher sensitivity of microalgae and zooplankton [5,6,7] and the European requirement of reducing the use of animals in laboratory assays [8,9]. Consequently, zooplankton- and microalgae-based tests (growth and photosynthetic activity the parameters mostly measured) are a fundamental part of the strategy of reducing the use of fish, including fish larvae and other animal models, in aquatic toxicology, and several standardized protocols on microalgae growth inhibition assays have been published [3,10,11,12]. However, for other toxicity assays involving sublethal effects, such as neurobehavioral disorder (affecting reproductivity capacity), the use of zebrafish as the organism model is more feasible [13]. Likewise, for studies of excretion kinetics and preferential pathways of metal nanoparticles with possible applications in new treatment for neoplasms or neurological disorders, not only is a zebrafish model suitable [14,15], but also murine animal models [16]. In microalgae growth inhibition assays, the half-maximal effective concentration (EC50) of a compound is a toxicity endpoint based on algal population parameters such as abundance, growth rate, and biomass. Standardized toxicity tests usually last between 2 to 120 h and examine the effects upon multiple generations of an algal population, thus they should not be considered acute toxicity tests as in animal tests [17].Heavy metals are considered essential or toxic depending on their role in microalgae physiology and their concentration. Generally, Cu and Zn are a priori essential or micro-nutrients for many microalgae species, Ni is essential for diatoms, and As and Pb are non-essential [18,19]. Some studies have shown greater metal effects on microalgae growth at 48 h in microalgae cultures and species showing high growth rates [20]; for this reason, the length of the assays can be reduced in order to save costs, get faster assessments, and avoid misleading results about toxicity with the development of microalgae detoxification mechanisms [17,21].Several studies showed that heavy metals inhibit microalgae photosynthesis (e.g., [22,23]). Photosynthesis efficiency, also called photosystem II (PSII) quantum yield or photosynthetic yield, is the fraction of the calories of radiation absorbed, which are stored as calories of chemical potential. Photosynthesis inhibition measurement by pulse-amplitude-modulation (PAM) fluorometry and the saturation pulse method is based on the fact that light energy captured by photosynthetic pigments can be either (i) used to drive photosynthesis, (ii) be dissipated as heat, or (iii) emitted as light (fluorescence). An increase in the yield of one process is directly linked to a decrease in the other two. It is therefore possible to measure a change in the efficiency of photochemical processes by measuring the yield of chlorophyll (or other pigments) fluorescence [24,25,26]. The PAM fluorimeter operates with a specific modulation of the measuring light signal that allows for the detection only of fluorescence excited by the measuring light. This method has been proven to be very efficient for the detection of effects of metals in microalgae as well as a very wide range of other toxicants (for a review, see [27]). PAM fluorometry applied to assessing short- and long-term toxic effects on microalgae provides several advantages: (1) a combined and rapid evaluation of several functional parameters in parallel, (2) screening of trends over time, (3) observing effects in replication, and (4) being non-destructive.The sensitivity to a toxic compound in microalgae is species-dependent [28]. Therefore, few freshwater microalgae species are used as model species, such Chlorella vulgaris. This species fulfils three basic requirements: It is a single cell that does not form aggregates in culture and test conditions, it is easy to maintain in laboratory conditions, and it is very sensitive to contaminants. Additionally, the combined toxicity of a mixture of compounds depends on complex interactions between the test organism, the contaminants, and other compounds (i.e., nutrients) present in the sample. Many mathematical models have been developed for the analysis of toxicological data and chemical speciation, and eventually for understanding the mechanisms of action and supporting pollution management, regulation, and related decision-making. The most common toxicological endpoints used in algal growth and photosynthesis inhibition tests are cell density and cell photosynthetic yield at the end of the exposure period. A mathematical description of the relationship between the pollutant concentration (dose) and the final cell density and cell photosynthetic yield (response) is useful to interpolate, extrapolate, or derive metrics such as EC50, other effective concentrations (ECx), or maximum effects of toxicity. In most cases, a single inflection is observed, which can be modeled using a classical Hill model [29]. In other cases, the laboratory results show an initial stimulatory effect (hormesis), a full inhibition with two inflection points (bi-phasic curve), or a combination of these features, which requires other mathematic descriptions [30]. The toxicity of heavy metals depends on their bioavailability, which is influenced by their chemical speciation. Heavy metal speciation in aquatic environments depends on the conductivity and pH of the medium, as well as the other dissolved compounds (i.e., their chemical speciation and their concentration). The Eh-pH diagram of The National Institute of Advance Industrial Science and Technology of Japan [31], based on the pH range of the culture medium, is an extensively used model for determining the chemical speciation of heavy metal. More elaborated models such as Visual MINTEQ [32] also take into account the temperature, the concentration, and the chemical speciation of other chemicals in the medium. Heavy metals are present in nature rarely as a single compound; instead, a variety of mixtures is commonly present [33]. How to deal with the toxicity of mixtures is still an open challenge [34]. This work also analyzes the effects of some selected metal mixtures, based on some of the most detected heavy metals in the environment. The aim of the present work is to assess the suitability of regulatory and decision-making resolutions obtained with the ecotoxicological tests based on the growth and photosynthesis inhibition in microalgae cultures exposed to heavy metal mixtures at environmentally relevant concentrations. A combination of calculated EC50 plus metal bioavailability based on mathematical models was used. The measured effects were then compared to the prediction of the concentration addition (CA) [35]) and independent action (IA) [36,37] mixture models. Moreover, the obtained data gave the opportunity to study the differences in toxicity of the selected heavy metals in microalgae regarding the specific mode of action (MoA) and bioavailability.Chlorella vulgaris (Chlorophyta, strain SAG 211_11b) was cultured in nutrient-enriched modified Bold’s Basal Medium (BBM) non-autoclaved medium [38], pH 6.8 ± 0.1, natural light (maximum irradiance 4180 μE m−2 s−1, average day length 14L:10D), and a natural temperature range from 23.7 to 39.8 °C.The growth rate was calculated as follows:(1)U=(ln(N1N2))/t
+        where U is the growth rate, N1 is the cell density at the beginning of the test (cells mL−1), N2 is the cell density at the end of the test (cells mL−1), and t is the time lapse (in days). Microalgae cells were counted on a Neubauer chamber at the light microscope.All reagents were from Sigma AldrichTM (Saint Louis, MO, USA) and high purity. Single dilutions of metals were prepared in distilled water in previously cleaned glass material using previously calibrated automatic pipettes (Table 1). The concentration range of single dilutions of heavy metals in the test tubes included concentrations detected in natural samples from Catalan rivers [39,40] and the US Environmental Protection Agency screening benchmark for chronic effects [41]. The artificial mixture of heavy metals was prepared with single dilutions of each heavy metal in distilled water.Ecotoxicological assays of growth inhibition followed the conditions used for R. subcapitata in the paper by Expósito et al. [42]. Chlorella vulgaris cultures at exponential growth phase and a climatic chamber (KBWF 240, Binder GmbH, Tuttlingen, Germany) at constant 174.6 μE m−2 s−1 light and 32.4 °C were used. Initially, 190 cells µL−1 were distributed to glass test tubes containing BBM culture medium, plus 100 µL of single heavy metal dilution, an artificial mixture of heavy metals, or distilled water (control groups); the final volume was 2.1 mL per test tube. Five replicates were prepared per single heavy metal dilution and artificial mixture of heavy metals. At the end of the assay, 48 h later, iodine lugol was added in order to stop cell growth. Cell density was checked using a Neubauer chamber under a light microscope. Pictures of cell morphology were taken with a Nikon TE2000E microscope equipped with a camera and managed with NIS-Elements light software (Nikon Corporation, Tokyo, Japan).The growth rate of controls was the quality control of the results: If it was equal to or higher than the specific growth rate of the microalgae culture in exponential phase, the results were accepted.The performance of C. vulgaris was assessed in two different scenarios: (1) single dilutions of heavy metals and (2) an artificial mixture of heavy metals.Photosynthesis inhibition was studied in acute exposure and after 24 h exposure and incubation in the darkness in a climatic chamber (KBWF 240, Binder GmbH, Tuttlingen, Germany) at 25 °C and 75% humidity. Microalgae cultures were used when they reached the exponential growth stage. The test was performed on black 96-well flat-bottom microplates; each well contained 400 µL of culture medium (initial density of 3.3–3.8 × 106 cells mL−1) plus 4 μL of single heavy metal dilution or distilled water (control groups). The effect of 12 concentrations of single heavy metals (except 6 in the case of Pb) and 7 concentrations of metal mixtures (Table 1) ranging from 0.1 µM to 10 mM was measured with a MAXI Imaging-PAM instrument (Heinz Walz GmbH, Germany) equipped with LED lights with a wavelength of 470 nm.After 5 min dark adaption, the minimum fluorescence yield F0 was measured on a non-actinic measuring light (ML) with low intensity (3 µmol quanta m−2s−1 PAR (frequency 8 Hz, modulated pulses of 100 µs). The gain was adjusted in order to get a fluorescence signal in the range of 0.150–0.200 units. The saturation pulse was 2700 µmol quanta m−2 s−1 Photosynthetic Active Radiation (PAR) and lasted for 800 ms. The maximal level of fluorescence when all PS II centers were closed (Fm) was measured after dark adaptation. The unquenched variable fluorescence (Fv) was calculated as the difference between Fm and F0. Maximal PS II quantum yield (Max YII) was determined after dark adaptation and was calculated as Max YII = Fv/Fm. Four saturation pulses were applied for 100 s each, maintaining the sample in the dark phase to check homogeneity of the replicates. The test substances were added during a short break after pulse 4, maintaining the dark phase and starting from lower concentrations; the time delay between the first and last addition was about 200 s. The dark phase was maintained during 3 saturation pulses after the addition of the test substances in order to check the early effects on PS II not related to photosynthetic processes. Afterwards, actinic light was applied at two PAR intensities: 83 and (after 1000 s) 283 µmol quanta m−2 s−1, and additional saturating pulses were given at intervals of 100 sec to measure the maximal level of fluorescence under illuminated conditions (Fm′). The transient fluorescence (Ft) was monitored during the entire duration of the test (2000 s). The effective PS II quantum yield (YII) was calculated by the formula YII = (Fm’ − F)/Fm’, where F is the level of fluorescence immediately before the saturation pulse (3 s average). After the acute test the samples were incubated 24 h in the dark and afterwards the same measurement protocol was repeated.The heavy metal concentration and percentage of growth inhibition of microalgae cultures, obtained from tests with single heavy metal dilutions, were the input data for Matlab R2017a software (KTH Royal Institute of Technology, Stockholm, Sweden). All data were log-transformed prior to data analysis to ensure the normality of the data and homogeneity of variances. The results were fitted using the hormetic concentration-response model by Yoshimasu et al. [30] implemented in Matlab R2017a. Concentration-response graphs were also plotted in Matlab R2017a. Matlab software suggests the best model that fits the obtained results with single dilutions of heavy metals. Values of EC50 and expected growth inhibition from single and artificial metal mixtures of contaminants were calculated using this software, interpolating or extrapolating the missing information. Mixture models were based in the concentration addition (CA) [35] and independent action (IA) hypothesis [36,37]. In the case of CA, the following equation is applied:(2)ECxMix=(∑i=1npiECxi)−1
+        where n denotes the number of mixture components, pi is the relative fraction of chemical I in the mixture, and x is a common effect level, which is provoked by an exposure to a single substance or mixture concentration ECxMix resp. ECxi.In the case of IA, the following equation is applied:(3)E(cMix)=1−∏i=1n(1−E(ci))The individual effects of mixture constituents E (ci) can be calculated from the concentration response functions Fi determined for single substances: E(ci) = Fi (ci). Again, the individual concentrations ci can be expressed as relative proportions pi of the total concentration cmix, and under the condition that the total effect E(cmix) equals x%, cmix is defined as ECxmix. Thus, by substitution we can transform Equation (3) into:(4)x%=1−∏i=1n(Fi(pi(ECxmix)))
+        which implicitly provides a prediction of effect concentrations of a mixture under the hypothesis of IA.Metal speciation was simulated by Visual MINTEQ 3.1 (KTH Royal Institute of Technology, Stockholm, Sweden) [32]. This model is a chemical equilibrium model for the calculation of metal speciation, solubility equilibria, and sorption in water. The heavy metal concentration can be total (i.e., assuming 100% bioavailability, data in Table 1) or corrected with the percentage of bioavailable chemical species, estimated with Visual MINTEQ.Data on and heavy metal concentrations in single dilutions and in artificial mixtures were the input data for determining the chemical speciation of each heavy metal in the test tubes, based on mathematic models (Supplementary Information Tables S1–S3). The chemical speciation of each heavy metal obtained with the Eh-pH diagram [31] was compared to the chemical speciation obtained with Visual MINTEQ [32].The statistical analysis was performed with XLSTAT software (Addinsoft, New York, NY, USA). To avoid the bias of a small number of samples and verify data normality of the growth variable before ANOVA analysis, three normality tests were applied: Lilliefors, Shapiro–Wilk, Jarque–Bera and Anderson–Darling. Significant differences were determined with ANOVA and Tukey and Dunnett’s multiple comparisons test (α = 0.05) followed by a verification of the standardized residues at 95%. The specific growth rate in the culture for C. vulgaris averaged 0.55 divisions per day, the average growth rate in control groups exposed to single dilutions of As, Pb, and Ni was 1.30 divisions per day, in control groups exposed to single dilutions of Cu and Zn it was 0.93 divisions per day, and the average growth rate of the controls exposed at combined metals test was 1.19 divisions per day. These growth rates were within the range of the growth of C. vulgaris in the exponential stage for cultures kept at the established culturing conditions; consequently, the results obtained were suitable for ecotoxicological assessment according to the standard procedures.The use of a culture medium as the exposure solution rather than natural water provides an unnatural level of nutrients that is not found in natural waters. Nevertheless, the culture medium was selected for three reasons: first, in order to be sure that the detected toxicity was due to the metals or the natural sample and not to other unexpected or undetected compounds in the natural water. Second, the composition of the exposure solution was needed for the speciation/bioavailability models, the more information the better the speciation prediction would be, and culture media have a known composition. Third, the test targeted a clear toxic effect, i.e., an effect that can be detected in an optimum growth scenario such as the exponential growth stage with enough nutrients in the solution to keep on growing.A colorless, mucilaginous matrix embedding pigmented C. vulgaris cells was observed (Figure 1) when exposed to low Cu (2.5–5.0 µg L−1) concentrations. It did not appear at higher concentrations of Cu, nor in any of the concentrations tested for the other heavy metals or the artificial mixture of heavy metals. A similar mucus was observed in Raphidocelis subcapitata exposed to Cu at similar concentrations, and in Zn at low concentrations [42]. The mucilage impaired the quantification of C. vulgaris by direct count and biases indirect quantifications based on the assumption that this species is always single suspended cells in culturing and test conditions. The external mucilage could be a defense mechanism against Cu toxicity by adsorption of the heavy metal ion to the mucilage and preventing its entrance to the cell. Secretion of metal-binding organic compounds (exopolysaccharides and exoproteins) to the surrounding environment has been described in other microalgae species and suggested as a mechanism for reducing the heavy metal toxicity by avoiding any interaction between metal cations and the microalgae cell [21]. The metal cation might form complexes with negatively charged residues of the organic compounds (pyruvate, succinate, acetate, phosphate) with a reduced bioavailability [43]. Heavy metal cations are considered the most toxic chemical species; nevertheless, the bioavailability of heavy metal-EDTA (Ethylenediaminetetraacetic acid) complexes should not be overlooked. Inorganic arsenic compounds, which are anions, are more toxic than organic compounds, and trivalent species (As III) are more toxic than pentavalent species (As V) [19]. As III act as a cross-linking agent by binding up to three monothiol molecules, such as the antioxidant GSH (glutathione), and this arsenic–protein binding often triggers cellular responses such as oxidative stress [44,45,46]. Copper, as ionic form Cu2+ disrupts many microalgae metabolic pathways, such as photosynthesis, respiration, ATP (Adenosine triphosphate) production, and pigment synthesis, as well as inhibits cell division [47]. According to Stauber and Florence [48], copper binds rapidly to many non-specific sites on the cell membrane, including carboxylic, sulfhydryl, and phosphate groups, and specifically to copper transport sites. Once internalized, it can oxidize thiol groups in the cytoplasm, leading to a lowering of the ratio of reduced to oxidized glutathione, which in turn affects spindle formation and cell division. Nickel, as ionic form Ni2+, is an essential metal in the cellular physiology of some eukaryotes and prokaryotes, including microalgae [49,50]. However, at higher concentrations, this metal can diminish growth [51], decrease the lipid content of autotrophic systems [52], and directly affect their photosynthetic system [53,54]. Nickel toxicity can also be related to its ability to replace essential metals in the metalloenzymes, which results in the disruption of metabolic pathways [18].Lead (Pb) metal ions are able to replace other bivalent cations such as calcium, magnesium, and iron, and monovalent cations such as sodium, which ultimately disturbs the biological metabolism of the cell [55]. Zinc, as ionic form Zn2+, is a cofactor for enzymes participating in CO2 fixation (i.e., carbonic anhydrase), DNA transcription (i.e., RNA polymerase), and phosphorus acquisition (i.e., alkaline phosphatase) [56]. According to Ouyang et al. [23], Zn promotes the quantum yield of PSII, one of the components of chloroplasts and a clue to photosynthesis. The effect of ethylenediaminetetraacetic acid (EDTA) on heavy metal toxicity is heavy metal- and EDTA concentration-dependent. The general assumption that heavy metals are less bioavailable when they form chemical complexes with ligands such as humic acids or synthetic compounds (i.e., EDTA) was contradicted long ago. For instance, Tubbing et al. [57] evidenced that Cu is biologically available for Selenastrum capricornutum (synonym of R. subcapitata) when complexed with EDTA, and Vassil et al. [58] found a 78-fold higher Pb concentration when exposed to EDTA in tissues of Brassica juncea. Opposite results were found by Franklin et al. [59] and Ma et al. [60]: The addition of 34 µM EDTA into a Scenedesmus subspicatus culture enabled a 55% reduction in growth inhibition exerted by 40 µM Cu. In aquatic environments, free metal ions are in equilibrium with metals bound to organic and inorganic compounds [21]; both can be bioavailable and cause toxicity, and therefore they should be studied individually.According to the Eh-pH diagram, the chemical species of As, Cu, Ni, Pb, and Zn were inorganic arsenate (HAsO42−), copper (II) (Cu2+), nickel (II) (Ni2+), lead (II) hydroxide (PbOH+), and zinc (II) (Zn2+), respectively, which are assumed to be the most toxic chemical species because of their bioavailability. Based on these results, 100% bioavailability could be assumed. According to the Visual MINTEQ model (Supplementary Materials Tables S1–S3), the chemical species of As, Cu, Ni, Pb, and Zn were more varied, with a smaller percentage of the most toxic species considered a priori (HAsO42−, Cu2+, Ni2+, PbOH+ and Zn2+) and showing a strong interaction with the EDTA present in the BBM culture medium. Heavy metals interact with other chemicals present in a BBM culture medium. Visual MINTEQ takes these interactions into account when estimating the chemical speciation. In these conditions, all metals formed complexes with EDTA (Supplementary Materials Tables S1–S3). In the growth inhibition tests with single heavy metal dilutions, EDTA-heavy metal complexes ranged from 16.8% (for ZnEDTA2–) to 99.9% (for CuEDTA2– and NiEDTA2–); in these cases, the free ion concentrations were dramatically reduced compared to the assumption based on the Eh-pH diagram. The percentage of free ions also depended on the concentration of the source compound in the tested dilutions, i.e., the percentage increased with higher concentrations of the compound.As the medium contains EDTA (8.5 mg L−1) and other components that can chelate metal ions, only metal species, which were bioavailable, as well as effective concentrations and free ions were considered for the EC50 calculations. The range of concentrations used for photosynthesis inhibition tests was wider (Table 1, Supplementary Materials Tables S1–S3). For Cu all species were considered except CuEDTA3− and CuHPO4 aq. Serra et al. [61] found that Cu toxicity in fluvial periphyton was reduced 1.6 times in a test performed with a high soluble reactive phosphorus (SRP) concentration (~50 µM). Since it was not expected that the microalgae would suffer from phosphorus (P) limitation during exposure under laboratory conditions (low SRP test ~5 µM), the study supported the hypothesis that the P-Cu interaction in the media leads to a reduction in Cu bioavailability. For this reason, we excluded from our calculation of EC50s the fraction of P-Cu complexes. Dissolved Zn2+ was the reactive species considered for Zn, and Ni2+ was the reactive species considered for Ni. Cu, Zn, and Ni showed the most complex behaviors: (a) at lower concentration, they had a high sequestration rate by EDTA in common and decreasing bioavailability, (b) whereas at higher metal concentration, the EDTA was saturated and the metal became almost totally bioavailable. Arsenic was dissolved and bioavailable as anions HAsO42− (46.30%) and H2AsO4− (53.70%), and the two species had about the same concentration in the medium. Lead (Pb) was complexed by EDTA as PbEDTA2− with a very high percentage (>99%), and only a small fraction was bioavailable as Pb2+.The calculated EC50 for growth inhibition tests using 100% bioavailability and taking into account the chemical species estimated by Visual MINTEQ (Table 2) was based on Matlab concentration-response curves (Supplementary Materials Figures S1 and S2). Parameters and goodness-of-fit coefficients are indicated in Table S4 of the Supplementary Materials. Chlorella vulgaris was less sensitive than R. subcapitata exposed to the same heavy metal concentrations in the same conditions [42]. The 48 h EC50 indicated that for this type of test C. vulgaris was more sensitive to As (36, 22 µg L−1 for both speciation models), followed by Zn (437.6 µg L−1 based on the chemical speciation estimated with Eh-pH diagram and 2.984 µg L−1 based on chemical speciation estimated with Visual MINTEQ).Yan [44] and Guo et al. [45] hypothesized that HAsO42−, which is a molecular analogue of phosphate (HPO42−), can compete for phosphate anion transporters (transporter proteins). Once in the cell, As (V) can be readily converted to As (III), the more toxic of the two forms. Trivalent arsenicals bind to thiols that are contained in numerous intracellular and cell-surface functional proteins.Zinc essentiality can explain the higher concentrations needed for achieving growth inhibition; however, the zinc toxicity is related to the disruption of calcium uptake, another essential cation necessary for the Ca-ATPase activity in cell division [62]. The calculated 48 h EC50 by Eh-pH diagram for As and Zn of 36.2 and 437.6 µg L−1, respectively (Table 2 and Supplementary Materials Figure S1), were close to the US EPA screening benchmark [41] (Table 1), but the values estimated by Visual MINTEQ for Zn were far below the US EPA benchmark (Table 1 and Table 2). C. vulgaris exposed to Cu, Ni, and Pb gave a tendency to negative impact on growth but it was not possible to obtain a clear concentration-response curve with the concentrations used. In Table 2 some selected lower-level effective concentrations are also reported; this information was useful when comparing to photosynthesis inhibition tests (Table 3). The calculated 48 h EC50 for Cu, Pb, and Ni based on the chemical speciation estimated with the Eh-pH diagram and hormetic model were very high and out of the range of tested concentration (Table 2 and Supplementary Materials Figures S1 and S2); nevertheless, the calculated 48 h EC50 for Pb estimated by Visual MINTEQ was slightly below the US EPA screening benchmark (Table 1 and Table 2). The toxicity of heavy metals has been studied and determined by many authors for different microalgae species, including R. subcapitata, and 100% bioavailability was assumed, i.e., all dissolved heavy metal becomes the cation estimated by the Eh-pH diagram. Their findings are publicly available in scientific publications and toxicological databases. Single metallic species rarely exist in the natural environment; instead, the metal ions present in the natural environment might enter a variety of interactions with physical and chemical water parameters and microorganisms. The co-existence of a multiplicity of metal ions is considered to give rise to distinct interactive effects [54], namely, synergism, antagonism, and non-interaction. In experimental assays for an artificial mixture test using natural dissolve metals and metalloids concentration from the Puig River (Catalonia region) shown in Table 4 and Table 5, no inhibitory effect was observed because no significant differences were observed between this treatment and control (Figure 2). The experimental results suggest antagonist effects among them with possible inhibition of their individual toxic effect.Expected growth inhibition concentrations were obtained from single metal concentration-response curves using the concentration addition (CA) and independent action (IA) models (Table 4 and Table 5). The EC50 calculated with the CA model was 160.84 µg L−1 (fitting parameters available in Supplementary Materials Table S6), using dose response curves obtained with nominal concentrations assuming 100% bioavailability. Meanwhile the predicted EC50 calculated by the CA and IA models, assuming only free ions bioavailable in the mixture, were 30.610 µg L−1 and 34.17 µg L−1, respectively (Table 4, model fitting parameters available in Supplementary Materials Table S6). The growth inhibition measured after the exposure to the experimental mixture was lower with respect to the expected growth inhibition by the CA and IA models (%). This result indicates possible antagonistic effects in the metal mixture (Table 5). The calculated EC50 for photosynthesis inhibition tests, taking into account the chemical species estimated by Visual MINTEQ (Table 5), was based on Matlab concentration-response curves (Supplementary materials Figure S3). The parameters and goodness-of-fit coefficients are indicated in Tables S5 and S6 of the Supplementary Materials. The effects on photosynthetic yield YII in C. vulgaris were measured immediately after the addition of the metal (2000 s test) at 83 PAR (photosynthetically active radiation) and after an incubation period of 24 h at 83 and 283 PAR intensities. The results indicated that moderate and low concentrations of metals had a small stimulating effect (Supplementary Materials Figure S3). Effective nominal concentrations were quite high in most of the cases (>1000 µg L−1) and only copper, nickel, and zinc showed negative effects. Copper was the most effective metal (Supplementary Materials Figure S3), and in the dark phase a small amount of inhibition (10–20%) was already observed at the highest concentrations (results not shown). Light intensity affected the magnitude of the response: When more processes were activated in the cells, the effects of the metals appeared more quickly (Supplementary Materials Figure S3). Lower concentrations could affect the algae, but a quick recovery was shown. After 24 h exposure, a very steep concentration-response curve was observed. According to Visual MINTEQ, divalent metals showed interactions with EDTA present in the culture medium, with the free ion concentrations in some cases being dramatically reduced compared to the nominal concentration of metal added in each test. Copper has a vital function in the regulation of PSII-mediated electron transport either as a part of a polypeptide involved in electron transport, or as a stabilizer of the lipid environment close to the electron carriers of the PSII complex [63]. Nevertheless, Cu is a potent inhibitor of photosynthesis [27]. High Cu concentrations inhibit the photosynthetic electron transport rate, especially the PSII, and both Cu deficiency and Cu toxicity interfere with pigment and lipid biosynthesis and, consequently, with the chloroplast ultrastructure, thus negatively influencing the photosynthetic efficiency. Chen et al. [64] found a significant decline of Y (II) in C. vulgaris exposed to a concentration of ~190 µg L−1 Cu (II), to almost zero at a concentration more than ~254 µg L−1 (acute test). Xia and Tian [65] found that Chlorella pyrenoidosa exposed for a 12 h period to 130, 320, 640, 1.270, and 2.540 µg L−1 Cu exhibited a significant decrease in the amount of active PSII reaction centers per excited cross section (39.5%, 41.9%, 59.8%, 55.1%, and 61.2%, respectively). Ouyang et al. [23] reported a reduction in photosynthesis by C. vulgaris in the presence of 320 µg L−1 Cu. These values are about one order of magnitude lower respective to the effective concentration of the tests conducted in this study (Table 6). This result can be explained due to specific algae strain resistance and taking into account our specific culture condition of using a eutrophic or nutrient-enriched medium. Apart from affecting the photochemistry of PSII, copper also slows down the synthesis of PSII D1 protein, thereby inhibiting the recovery from photo inhibition in C. pyrenoidosa [66]. Zn+2 acts as a substitute of Mg, which is the central atom of chlorophyll; this mechanism leads to the breakdown of photosynthesis [67]. Early toxic effects of Zn on photosynthetic activity of the green alga Chlorella pyrenoidosa were previously studied by Plekhanov and Chemeris [68]; the early effect of heavy metals was manifested as a rapid (within 0.5–2 h) reduction of photo-induced oxygen release by the algal cells with concentration of 10 µM (~653.8 µg L−1). Zn at a concentration of 100 µM (~6538 µg L−1) reduced Fv/Fm in C. pyrenoidosa cells almost to zero after a 30 min incubation. These results are in accordance with our findings (Table 6). An analysis of the induction curve of the delayed chlorophyll fluorescence in C. vulgaris cells suggested that the early toxic effects of Zn at the above concentrations manifested itself not only in inhibited electron transport in PS II, but also in the reduced energization of photosynthetic membranes. The impact of sub-lethal concentrations of the heavy metals copper (Cu), zinc (Zn), and lead (Pb) on the photosynthesis of C. vulgaris was studied during 96 h exposure experiments by OuYang et al. [23]. It was found that under a concentration of 5 μM (~320 µg L−1), the quantum yield of PS II was reduced by Cu, promoted by Zn, and not affected by Pb. These results are also in accordance with our findings and can be explained by taking into account the maximum exposure concentrations for Zn. In fact, we measured an EC50 of 12.96 mg L−1 in the 24 h exposure test and we detected a hormetic effect at lower concentrations. For Pb the lack of effects is explained by the speciation as not bioavailable PbEDTA2− (Table 5).The green microalgae A. falcatus was highly sensitive to Ni concentrations as low as 1 μg L−1: Photosynthetic pigments were reduced and Ni2+ increased antioxidant enzyme responses [69]. In our study the EC50 of Ni (~10 mg L−1) was higher than the values found in the literature for single metal species. Acute toxicity effects of As were previously observed on the photosynthetic efficiency of diatoms and cyanobacteria, whereas green algae were less affected [70,71]. The toxicity measured in this study for as in photosynthesis inhibition assays was not significant with respect to the other heavy metals (except for Pb). Taking into account the listed previous studies, the justification of our results is probably linked to the specific species sensitivities in our algae strain and also to specific eutrophic medium conditions.The results of the mixture experiments are shown in Supplementary Materials Figure S4 and Table 3. In Supplementary Materials Figure S4 the experimental data were plotted together with model AC and IA prediction. The EC50 values were comparable to the EC50 of single metal exposures but the concentration response curves were less steep and effective concentrations of EC1, 10, and 20 showed lower values with respect to both model predictions for 24 h tests. The CA and IA models underestimated the mixture effects for lower effective concentrations and slightly overestimated the mixture effects at higher concentrations (Supplementary Materials Figure S4). This indicates possible synergism in the metal effects, as the effects are more than additive (Table 5). For acute effects, as only copper showed measurable inhibition, the mixture curve was compared to the Cu concentration-response curve. Some synergistic effects are also shown in this case: The mixture was more toxic with respect to what could be predicted based on single metal exposure results (Supplementary Materials Figure S4).The input data on heavy metal concentrations varied depending on the percentage of bioavailable chemical species, estimated with Visual MINTEQ speciation based on each culture medium. For acute tests only Cu showed measurable effects; the AC and IA models were not run for acute exposure.The variability of the EC50 values between the growth and the photosynthesis inhibition tests and literature values can be explained by specific toxicity mechanisms and the difference in resistance of the specific algae strain and culture conditions. For example, our tests were performed with an excess of nutrient availability, with chelators resulting in higher EC50 values compared to if the tests had been conducted with lower nutrient concentrations with lower EC50 values (as Ware et al. [72] reported).The methods proposed in this study have several advantages over other species models, such as a shorter time lapse, the possibility of miniaturizing and reducing the amount of chemicals used, and a high-throughput (more replicates, more data) considered as a good alternative to animal testing. Based on our results, the Chlorella algae growth inhibition test was most sensitive to As, Zn, and Pb exposure whereas the photosynthesis inhibition test was most sensitive to Cu and Ni exposure. The effects for growth and photosynthesis were not related. Chlorella evidenced the formation of mucilaginous aggregations at lower copper concentrations that could be a defense mechanism against Cu toxicity. The toxicity of a given heavy metal is not only determined by its chemical speciation; other chemicals (such as nutrients or chelators) and biological interaction play an important role in the final toxicity. Regarding nutrient concentrations, eutrophic media could protect against metal pollution. In our study, the metal sequestration rate by EDTA was important for metal bioavailability, however, the effects of the lack of EDTA in modified oligotrophic mediums have to be also considered to avoid biased results. The toxicity of metal mixtures cannot reliably be predicted based on the toxicity of individual metals using AC or IA models, because antagonistic or, more likely, synergic interactions may occur. These interactions are complex and affect the test in different way for different endpoints. More research is needed to face the challenge posed by pollutant mixtures as they are present in natural environments, and make microalgae-based assays suitable for pollution management and regulatory purposes.The following are available online at https://www.mdpi.com/1660-4601/18/3/1037/s1, Figure S1: Concentration-response curves (lines) and experimental data (o) for C. vulgaris based on the growth inhibition results of single dilutions of heavy metals, taking into account the chemical species estimated by Eh-pH diagram (100% bioavailability); Figure S2: Concentration-response curves (lines) and experimental data (o) for C. vulgaris based on the growth inhibition results of single dilutions of heavy metals, taking into account the chemical species estimated by Visual MINTEQ; Figure S3: Concentration-response curves (lines) and experimental data (o) of YII for C. vulgaris based on the photo-synthesis inhibition results of single dilutions of heavy metals, taking into account the chemical species estimated by Visual MINTEQ; Figure S4: Concentration-response curves (lines) and experimental data (o) of YII for C. vulgaris based on the photo-synthesis inhibition results of mixture dilutions of heavy metals, taking into account the chemical species estimated by Visual MINTEQ. Table S1: Metal speciation at the different treatment concentrations in single metal exposure expressed as % of total metal concentration for growth inhibition tests; Table S2: Metal speciation at the different treatment concentrations in single metal exposure expressed as % of total metal concentration for photosynthesis inhibition tests; Table S3: Metal speciation at the different treatment concentrations in mixture metal exposure expressed as % of total metal concentration for photosynthesis inhibition tests; Table S4: Parameters and fit goodness coefficients of growth inhibition concentration-response curves for single heavy metal exposed samples; Table S5: Parameters and fit goodness coefficients of photosynthesis inhibition concentration-response curves for single heavy metal exposed samples; Table S6: Parameters and fit goodness coefficients of photosynthesis and growth inhibition concentration-response curves for mixture heavy metals exposed samples and AC and IA models. N.E., R.C. and G.G.P.: Conceptualization, Methodology, Formal analysis, Project administration, Funding acquisition, Writing—original draft, Writing—review & editing, Visualization, Data curation, Investigation, Validation, Methodology, Resources, Formal analysis. J.S.: Software, Supervision. V.K.: Software, Supervision. M.S.: Supervision, Resources. All authors have read and agreed to the published version of the manuscript.This work was supported by the Beatriu de Pinós program (2010 BP_A2 00018) co-funded by the 7th Framework Programme of the European Union (Marie Curie PEOPLE action) and Agencia de Gestión de Ayudas Universitarias y de Investigación AGAUR co-funded by European Regional Development Fund FEDER (code project: 2019 PROD 00113)., as well as the Tecniospring program (TECSPR14-2-0012) co-funded by the 7th Framework Programme of the European Union (Marie Curie PEOPLE action) and ACCIÓ (Generalitat de Catalunya). The authors thank the Tecniospring Plus programme (TECSPR17-1-0012) and ACCIÓ for funding this project. This project received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 712949. Not applicable.Not applicable.The data presented in this study are available on request from the corresponding author.Universitat Rovira i Virgili’s Technical Services assisted with the pictures of the C. vulgaris cells. Chlorella vulgaris (SAG) was purchased from the Culture Collection of Algae at Göttingen University (Göttingen, Germany). The authors declare no conflict of interest.Comparison of a normal culture of C. vulgaris, with single, isolate, suspended cells (upper picture) and a culture exposed to low concentrations of Cu (lower picture) showing the mucilaginous matrix with cells embedded. In both cases, C. vulgaris cells keep their characteristic “O” shape.Growth of C. vulgaris exposed to artificial mixture of heavy metals (Artificial Mix.) (*) No significant difference to control.Tested single heavy metal dilutions: United States Environmental Protection Agency (US EPA) benchmarks in water hardness of 100 mg L−1 [41], and heavy metal levels in the artificial mixture in growth (selection based on data from [39,40]) and in photosynthesis inhibition tests.Note that these were the nominal concentrations of heavy metals. All concentrations referred to the single metal.Selected effective concentrations (ECx) for 48 h exposure to single dilutions heavy metals for growth inhibition calculated with Matlab concentration-response curves.Input data on heavy metal concentrations varies depending on the percentage of bioavailable chemical species, estimated with the Eh-pH diagram or Visual MINTEQ speciation based on each culture medium. Bioavailable species are underlined.Maximum stimulation concentration (hEC) and selected effective concentrations (ECx, ppb) of YII in artificial heavy metal mixtures and modeled with the AC model and calculated with Matlab concentration-response curves.Selected expected effective concentrations (ECx) for 48 h exposure to heavy metals mixture for growth inhibition calculated with concentration-response curves using concentration addition (CA) and independent action (IA) models.Input data on heavy metal concentrations varies depending on the percentage of bioavailable chemical species, and nominal or Visual MINTEQ speciation based on each culture medium.Expected growth inhibitions for 48 h exposure to the artificial mixture calculated with concentration-response curves using concentration addition (CA) and independent action (IA) models.Input data on heavy metal concentrations varies depending on the percentage of bioavailable chemical species, and nominal or Visual MINTEQ speciation based on each culture medium.Maximum stimulation concentration (hEC) and selected effective concentrations (ECx, ppb) of YII for Cu, Ni, and Zn, calculated with Matlab concentration-response curves for single metal tests.Input data on heavy metal concentrations varies depending on the percentage of bioavailable chemical species, estimated with Visual MINTEQ speciation based on each culture medium.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01038.txt

@@ -0,0 +1,28 @@
+Sexually transmitted infections (STIs) continue to exert a considerable public health and social burden globally, particularly for developing countries. Due to the high prevalence of asymptomatic infections and the limitations of symptom-based (syndromic) diagnosis, confirmation of infection using laboratory tools is essential to choose the most appropriate course of treatment and to screen at-risk groups. Numerous laboratory tests and platforms have been developed for gonorrhea, chlamydia, syphilis, trichomoniasis, genital mycoplasmas, herpesviruses, and human papillomavirus. Point-of-care testing is now a possibility, and microfluidic and high-throughput omics technologies promise to revolutionize the diagnosis of STIs. The scope of this paper is to provide an updated overview of the current laboratory diagnostic tools for these infections, highlighting their advantages, limitations, and point-of-care adaptability. The diagnostic applicability of the latest molecular and biochemical approaches is also discussed.Sexually transmitted infections (STIs) are some of the most frequent communicable conditions globally. They represent a major health and social issue and, in particular, four treatable STIs contributed to up to 376.4 million new infections in 2016 [1], corresponding to approximately one million new cases per day. This number included infections due to Chlamydia trachomatis (CT) (127.2 million), Neisseria gonorrhoeae (NG) (86.9 million), Trichomonas vaginalis (TV) (156 million) and Treponema pallidum subspecies pallidum (TP) (6.3 million). Additionally, the World Health Organization (WHO) [2] estimated up to 417 million infections of herpes simplex virus type 2 (HSV-2) and 291 million women infected with human papillomavirus (HPV). Although frequently underdiagnosed, genital mycoplasmal infections are also among the most common STIs, with a prevalence of 1–3% in the sexually active general population for Mycoplasma genitalium (MG), and 20–50% and 40–80% in asymptomatic sexually active women for Mycoplasma hominis (MH) and Ureoplasma urealyticum (UU), respectively [2].STIs represent a considerable burden for public health, which is difficult to assess because asymptomatic cases are very common. Undetected infections which are not treated can have serious complications, which disproportionately affect women and their newborn babies. Congenital syphilis can cause spontaneous abortion and premature delivery [3], and the WHO estimates that this condition is responsible for more than 300,000 fetal and neonatal deaths, and for an increased risk of premature death for 215,000 infants each year [4]. Furthermore, up to 530,000 cases of cervical cancer and 264,000 associated deaths are caused each year by HPV infections [4]. The two most common bacterial STIs, CT and NG, not only cause pelvic inflammatory disease and chronic pelvic pain in women, but also result in ectopic pregnancies and miscarriages, preterm labor, increased risk of mother-to-child HIV transmission, conjunctivitis, and pneumonia in neonates [5]. Although the evidence is inconclusive, MG and TV infection in pregnancy have also been associated with preterm birth [6]. Neonates can become infected with HSV at birth from an infected mother, with potential involvement of the central nervous system and fatal consequences [7]. Both CT and NG can cause irreparable damage to the fallopian tubes, leading to infertility in women [8], and a possible role of these organisms in the reduction of sperm quality and fertility in males has been suggested [9]. Furthermore, STIs increase the susceptibility to human immunodeficiency virus (HIV) infection and its transmission risk due to augmented viral shedding in genital secretions [10].In 2016, the WHO launched its global strategy for tackling STIs [4]. One of the main cornerstones includes improved surveillance through the development and implementation of better diagnostic algorithms and tests. In particular, the early diagnosis and identification of asymptomatic carriers (screening) is a prerequisite to efficiently guide treatment and prevention interventions for STIs [4,11]. In recent years, the advent of molecular and mass spectrometry tools revolutionized the clinical microbiology laboratory [12]. In particular, the latest technological advancements paved the way for the development and implementation of point-of-care (POC) testing [13].The purpose of this review is to discuss the latest advancements in the laboratory diagnosis of the main STIs, in particular: (1) the importance of POC and molecular assays in complementing the traditional syndromic approach for the early diagnosis and the identification of asymptomatic carriers; and (2) the current and future technological developments for STI diagnostics (including omics approaches).The syndromic management of STIs is based on obtaining a presumptive diagnosis via the observation of a set of signs and symptoms (“syndrome”) in the patient [14], without the need for laboratory confirmation of the underlying pathogen. This approach is cheap, it can be used in resource-poor and remote areas, and it allows immediate treatment during the first visit to the clinic or health center without any laboratory trained staff or technically demanding procedures [15]. However, the inconspicuous clinical presentation of STIs affects the applicability of the syndromic approach. These infections are often asymptomatic and the same set of signs and symptoms may correspond to different underlying organisms, leading to subjective judgement. Furthermore, even physiological discharge may be misinterpreted as pathological [15].The syndromic approach is efficient and cost-effective when most cases are symptomatic, as in the case of CT and NG infections causing urethral discharge in men [16,17]. The opposite situation is seen for vaginal discharge in women, as shown by various studies. For instance, based on the results of a study that collected samples from sex workers in India, Desai and colleagues [18] estimated that a syndromic approach based on vaginal discharge had a low specificity (50–55%) and a very low positive predictive value (PPV) (from 11.5% for CT to 24.6% for TV), missing 30–40% of cases of CT and NG and leading to up to 90% of cases being treated without infection. Similar results have been reported in Delhi [19], where the reported specificity of the syndromic approach using vaginal discharge as a symptom was only 37.5%. Vaginal discharge showed an even lower PPV (14.1%) in pregnant women in Sudan [20].Hence, the syndromic approach has poor predictive value and should not be used for screening, as it can lead to either overtreatment or undertreatment. In the first case, antibiotics are prescribed to healthy patients, contributing to the development of antimicrobial resistance, unnecessary costs, and emotional upset. In the latter case, no antibiotic is given to patients whose test is positive, potentially increasing the risk of reproductive sequelae and other health complications, in addition to the further spread of STIs among the population [21,22].The clinical presentation per se does not always allow a specific STI to be identified. In these cases, etiological diagnosis via laboratory confirmation of the underlying organisms is essential. Other than supporting a more rational use of antimicrobials, laboratory confirmation is pivotal in surveillance (to determine the true scale of the spread of STIs in communities) and in screening (i.e., testing of at-risk people without recognized signs or symptoms) [11]. By identifying infected people regardless of their symptoms, both the complications and further transmission of STIs can be reduced [4]. Approximately 30 different pathogens can be transmitted sexually, ranging from different bacterial species, viruses, fungi, and parasites [23], so the choice of the appropriate diagnostic approach can be difficult for physicians.To target the most relevant pathogens, various tools have been developed for the laboratory diagnosis of STIs throughout the years [11]. The choice of the most appropriate diagnostic test (Box 1) relies not only on the performance of the tool itself in terms of sensitivity, specificity, and predictive values, but also on the logistics (technical requirements, cost, throughput) and its purpose. 
+          
+            
+              Sensitivity: ability to correctly identify individuals with the infection (true positive rate)
+            
+            
+              Specificity: ability to correctly identify individuals without the infection (true negative rate)
+            
+            
+              Positive predictive value (PPV): probability that positive individuals truly have the infection
+            
+            
+              Negative predictive value (NPV): probability that negative individuals truly do not have the infection
+            
+            
+              Complexity: includes all the technical requirements (equipment, reagents, personnel) needed for optimal test performance
+            
+            
+              Cost: both materials- and labour-related
+            
+            
+              Throughput: number of tests completed in a given amount of time
+            
+            
+              Time to result: time needed to get a response from the test
+            
+          
+        Sensitivity: ability to correctly identify individuals with the infection (true positive rate)Specificity: ability to correctly identify individuals without the infection (true negative rate)Positive predictive value (PPV): probability that positive individuals truly have the infectionNegative predictive value (NPV): probability that negative individuals truly do not have the infectionComplexity: includes all the technical requirements (equipment, reagents, personnel) needed for optimal test performanceCost: both materials- and labour-relatedThroughput: number of tests completed in a given amount of timeTime to result: time needed to get a response from the testThe use of the most sensitive and specific tests is often impractical in resource-poor or remote areas due to their high cost and technical requirements. In addition, considering the clinical presentation and the severe complications associated with some STIs if left untreated, the “ideal” diagnostic test should be quick so that the patient is treated on the spot. In these settings, it is a common for patients to not return to the clinic for treatment while waiting for a laboratory result [24], and loss to follow-up can also affect the sensitivity of laboratory diagnosis. This is well exemplified by a study undertaken on female sex workers in Benin [25]; the authors found that up to 57.6% of infected women did not return to the clinic within 10 days of the visit, resulting in a decrease in the laboratory test sensitivity from 74.6% to 28.2%.Point-of-care (POC) tests represent an answer to the problem of needing to reach a diagnosis quickly outside of a standard laboratory. By definition, a POC test can be performed at the patient’s hospital bedside or own house, the physician’s office, or in the field [26]. Ideally, a POC test should meet the WHO affordable, sensitive, specific, user-friendly, rapid and robust, equipment-free and deliverable to end-users (ASSURED) criteria [27]. POC tests based on the detection of pathogen nucleic acids (NAs), antigens, or even antibodies have been developed for a variety of infectious diseases [13,28]. Regarding STIs, POC tests are available for CT [29], NG [30], and TV [31], for example. Some of these POCs, such as the OSOM® antigen test for TV, can be executed directly by patients with comparable sensitivity and specificity to that of a test performed by a clinician in a hospital setting [32], and patients can easily collect the specimen and interpret the result at home [33]. The development and implementation of POCs for STI management is considered a priority by WHO [4], and target product profiles (TPPs) for POCs have been developed [34,35].Recent systematic reviews have shown that, in many instances, the use of POCs can result in a substantial reduction in both missed treatments and overtreatment [28,36]. Treatment uptake is also improved, and patients show appreciation for receiving a quick and specific diagnosis, which also facilitates the treatment of the partners of patients [37].POC testing is so widespread that it also has an important role in the management of bacterial vaginosis (BV), even though it is a dysbiotic condition. However, BV is the most common vaginal disorder [38], affecting approximately 30% of women of reproductive age and up to 50% in sub-Saharan Africa [39,40]. Notwithstanding, BV results from a dysbiosis, and the transmissibility of BV implies that it may also result from external contamination in addition to internal imbalance, resulting in a true STI [41,42]. The major issue with BV is that its diagnosis remains problematic and subjective; hence, innovations in this field are urgently needed, and reliable and low-complexity tests are required.A POC test detecting sialidase activity in vaginal discharge has been implemented, with an excellent sensitivity and specificity in comparison with standard Gram staining [43]. Metabolomics-based POC tests based on the identification of characteristic organic compounds (i.e., biogenic amines) are being developed, such as the VGTest, which measures the presence of these molecules using a portable desktop ion mobility spectrometer (IMS) [44,45]. Given the potentially relevant adverse effect on women’s reproductive health, promotion of the implementation of more rapid and accurate POC tests to improve the routine diagnosis of BV should be a primary issue.Currently available laboratory diagnostic approaches for the detection of active STIs are summarized in Table 1 and discussed in more detail below. The following sections focus on POC formats, molecular diagnostics, and the latest technological advancements.Most of the etiological agents of STIs can be visualized microscopically or grown under laboratory conditions following isolation from clinical samples. These approaches are still in use (particularly for bacteria) in resource-poor settings and in peripheral or intermediate-level laboratories, but their performance varies. Both microscopy and culture are ideal for confirming the patient diagnosis and identifying a proper course of treatment.Microscopy is cheap and does not require special equipment, and can be performed at the point-of-care if needed [11]. However, sensitivity is affected by the microscopist’s skills and tends to be lower in asymptomatic patients, as seen for example in gonorrhea [30] and trichomoniasis [46]. Another drawback is that microscopic analysis should be carried out within 10 min of sample collection, making the examination unsuitable for high throughput laboratories [47]. Culturing the organisms using specific media shows good sensitivity and specificity, and is a highly standardized method. Despite being a slow process (often taking several days depending on the organism), it remains the only reliable method to test for antimicrobial resistance. The efficiency and precision of culture has been recently increased due to the use of matrix-assisted laser desorption/ionization–time-of-flight (MALDI-TOF) mass spectrometry [48]. This technique can precisely identify not only the species, but also the subspecies or serotype of the organism within minutes. Despite its ability to identify some viral species, spectrometry still does not enable the routine identification of viruses due to a lack of commercially available databases [48].Some drawbacks of the culture approach include the fact that organisms such as TP [49] or HPV [50] are not culturable, and the growth of CT [29] or HSVs [51] requires specialized laboratories and is both time-consuming and expensive. In addition, NG is a fragile microorganism, and hence culture assays may exhibit low sensitivity due to inappropriate transportation conditions [52].Furthermore, cultural approaches could miss some fastidious microorganisms with particular growth requirements that need the organism to be viable, making the sample collection step and storage extremely important to ensure the reliability of the result.Immunoassays detect either the pathogen’s antigens or the presence of antibodies produced by the immunological response to infection. These tools can be adapted into POC formats, particularly in the form of lateral flow immunochromatographic (ICT) assays. However, they are not readily available for all STIs (for example, genital mycoplasmas and HPV), and only a few of these tests have been thoroughly validated in clinical settings.Antigen detection-based assays allow the diagnosis of current infections. A few of these tests have been developed and tested for the detection of NG, showing relatively high specificity and returning results in just 25–40 min [53]. However, their sensitivity is highly variable and, in many instances, too low [30]. The same issues apply to antigen tests for the detection of CT [29], and the WHO does not recommend their use if other methods are available [11]. The reduced sensitivity represents a problem, particularly in swabs collected from asymptomatic cases [11,29,51]. Furthermore, because the interpretation is visual, the result is operator-dependent and this fact may impair diagnosis, particularly in weakly positive cases. However, the development of automated strip readers and the use of artificial intelligence can help to manage this disadvantage [54].A few antigen rapid assays have been developed for trichomoniasis [46]. The OSOM® Trichomonas Rapid Test is a good example of an efficient POC test [31]. Featuring good sensitivity and specificity, this POC test can be conducted at home and was found to be highly acceptable by women [31]. Despite their sub-optimal sensitivity, antigen-based POC tests can improve patient management (leading to more cases being treated) in high infection prevalence contexts due to their rapidity. This has been observed for both CT [24] and TV [55].Antibody tests have the drawback of detecting past exposure to the pathogen, and further tests are needed to confirm the presence of a currently active infection [11,29,51]. However, serological assays for the presence of serum antibodies represent a sensitive and specific tool for the diagnosis of syphilis [49]. A variety of serological assays have been developed for this infection and they play a crucial role in screening pregnant women to prevent adverse pregnancy outcomes. These tests show a good performance in both sensitivity (90% or more depending on the test) and specificity (>95%), and their ICT format makes them ideal for resource-poor settings [56]. Tests are now also available to detect both treponemal and non-treponemal antibodies, allowing discrimination between past infection and active syphilis [56].Although the multiplexing potential of immunoassay tests is still in its infancy, some examples are available. Latest technological developments regarding microfluidics and biosensors have sparked the development of a test for the parallel detection of CT and NG [57]. The assay is sensitive and rapid, and can be used successfully on crude urine [57]. Simultaneous antibody detection of both TP and HIV can be now accomplished through various ICTs with excellent sensitivity and specificity for both pathogens [56]. Interestingly, a duplex enzyme-linked immunosorbent assay (ELISA) for the detection of HIV/syphilis antibodies in fingerprick blood has been adapted in a device which is operated by a smartphone, facilitating the testing and reading of the results directly in the field [58]. The combination of nanotechnology and immunoassays has proved to be a promising approach, with the introduction of quantum dots (fluorescent labels of protein size) for labelling secondary antibodies, replacing conventional fluorescent molecules [59]. The use of quantum dots can also increase the potential for multiplexing in immunoassays [60].Methods based on the amplification and detection of an organisms’ nucleic acids (DNA or RNA) in clinical samples now represent the gold standard for the diagnosis of many STIs. Nucleic acid amplification tests (NAATs) are particularly important for organisms for which traditional tools have always been either unavailable or too expensive, such as C. trachomatis, genital mycoplasmas, and viruses. Overall, these tools overcome some important limitations of microscopy and culture-based approaches.NAATs have a significantly lower turnaround time (they can be easily automated, and the time to result is significantly lower than that of traditional approaches), they allow the detection of organisms which cannot be grown in culture, and they have a significantly higher sensitivity (a dramatically relevant advantage for asymptomatic infections where the organisms’ load is low) [61,62]. Due to their high sensitivity, NAATs are an invaluable tool for detecting extragenital infections, as shown in the case of rectal or pharyngeal gonorrhea in men [63]. Furthermore, NAATs do not require viable organisms, allowing the collection and utilization of more diverse and less invasive biological samples (such as urine or vaginal swabs) which can be easily self-collected; they can be automated and standardized, and they can be performed by minimally trained people; and they can be multiplexed, allowing the detection of multiple organisms from the same sample and in a single test run [64,65]. Due to the latest technological advances, NAATs also have significant potential to become the POC tests of the future for STIs due to a reduction in their cost [36].Polymerase chain reaction (PCR) has revolutionized clinical microbiology. Tiny amounts of DNA or RNA can be amplified from samples which are unsuitable for analysis by traditional methods; thus, it is not surprising that PCR-based tools now play a central role in the laboratory diagnosis of infectious diseases. Various PCR-based assays (single-step, nested, real-time, singleplex, or multiplex) have been developed for STIs, thereby simplifying the diagnostic workflow and reducing the turnaround time. Many of these tests have been perfected and are now commercially available; multiplex assays are available for the detection of up to eighteen different organisms (e.g., the CLART® STI A&B, Genomica S.A.U., Madrid, Spain), and their sensitivity and specificity are comparable with their respective singleplex assays. Multiplexing allows for the cost-effective analysis of one sample through the exploitation of a syndrome-associated specific panel, hence with an augmented probability of identifying the responsible pathogen.For instance, the BD Max™ CT/GC/TV (MAX) assay (BD Diagnostics, Franklin Lakes, NJ, USA) is a multiplex real-time assay capable of detecting CT, NG, and TV. The test shows an excellent performance using both vaginal and endocervical swabs and male urine, with sensitivities in the range of 91.1–99.1%, and consistently high specificity (≥98.6%) for all the organisms tested [66]. Another example is the Anyplex™ II STI-7 (Seegene Inc, Seoul, South Korea), which is based on the same principle, but it can detect up to seven STIs including CT, GC, TV, MG, MH, and two Ureaplasma sp. (urealyticum and parvum). In various studies, including both symptomatic patients and asymptomatic volunteers, the assay performed extremely well in terms of both sensitivity and specificity (≥99%), and showed good concordance and reduced cost compared to similar (but only duplex) platforms [67,68]. The assay was used successfully in Ghana to test asymptomatic women, revealing a high prevalence of MG and co-infections with up to three organisms [69], and confirming the applicability of this assay for the extensive screening of genital mycoplasmas as reported elsewhere [70]. Another multiplex assay, the FilmArray® (BioFire Diagnostics, Salt Lake City, UT, USA), can detect up to nine STIs (CT, GC, TV, MG, and UU, plus TP, Haemophilus ducreyi, and HSV 1 and 2) [71].Several PCR-based assays have been developed specifically for the detection of T. pallidum, not only from genital swabs and lesions, but also from blood and cerebrospinal fluid [74], and even semen [75]. Although they show high sensitivity and specificity, these tests are not commercially available and their use is limited to some laboratories [49]. There are various PCR-based assays for the detection of viral STIs, such as HSV1,2 [51] and HPV [76], many of which are commercially available and approved by the relevant regulatory bodies, and which constitute the tests of choice for these infections.There are some PCR-based platforms which have potential as POCs. The GeneXpert® (Cepheid, Sunnyvale, CA, USA) platform is a real-time PCR benchtop instrument adaptable for the detection of a variety of infections. In particular, the CT/NG assay running on this platform showed excellent sensitivity (100%) and specificity (≥99.8%) in detecting these organisms in the urine of both men and women [77]. However, its high price and the need for electricity do not allow this platform to be considered as a fully realized POC test [26].Current PCR-based assays have reached extremely high levels of sensitivity and specificity of detection, and they are powerful tools for screening samples for multiple organisms. However, their ability to be implemented as POC tests in resource-poor settings is still hindered by their relatively high cost and the need for specialized instrumentation (thermocycler) and laboratories.Isothermal amplification (IA) assays can offer a rapid and cost-effective solution for the diagnosis of STIs compared to PCR-based assays. In IA assays, the amplification of the target DNA/RNA does not require thermal cycling (as in PCR) and happens at a significantly lower and constant temperature; the time-to-result is generally reduced (<1 h), and the amplification results can be easily assessed visually (by observing turbidity or a change of color in the reaction tube) [78]. There are many IA methods which have been exploited in molecular diagnostics involving different chemistries. IA approaches used for STIs include loop-mediated isothermal amplification (LAMP), transcription-mediated amplification (TMA), strand displacement amplification (SDA), helicase-dependent amplification (HDA), and recombinase polymerase amplification (RPA).The enzymes used in IA assays are less affected by the inhibitory substances present in biological samples, opening the possibility of reducing both the time and cost associated with sample preparation and extraction. A LAMP assay developed for the detection of NG showed the same sensitivity of a reference PCR test when used on crude urine [79]. The applicability of using samples without performing traditional DNA extraction was also proven for a LAMP assay used to detect CT in endocervical swabs, with a lysis step of just five minutes before amplification [80]. Target amplification can be assessed using a lateral flow approach in the form of paperfluidic devices, as shown in a similar semi-quantitative LAMP-based approach for CT [81]. Such a format is used in rapid diagnostic tests, such as the HDA-based AmpliVue Trichomonas assay (Quidel, San Diego, CA, USA) for TV [82].IA approaches, and in particular RPA, can produce results very rapidly. Notable examples are the assay for the detection of CT directly from urine, giving a result within 20 min [83], or the 15-min time-to-result of a prototype for the simultaneous detection of CT and NG [84]. In addition, the TMA approach (which targets pathogen RNA rather than DNA) has been exploited for STI testing, most notably in the form of the Aptima® platform (Hologic, Marlborough, MA, USA), which has been used to detect NG/CT and TV [46].Multiplexing is also possible with IA approaches, and recent advances in microfluidics technology are pushing the boundaries of the “lab-on-a-chip” idea, in which sample extraction, amplification, and detection occur sequentially in the same instrument. Considerable promise has been shown by microfluidics. This technology is characterized by a precisely engineered manipulation of small volumes of fluids in channels on the micrometer scale. Microfluidics use ranges from nucleic acid purification to immunoassays, and its use suggests a revolution of medical diagnostics to overcome existing limitations. It allows for faster assay time-to-result, lower technical requirements for laboratories, smaller volumes of reagents, and increased automation and integration levels [85]. Furthermore, complex technologies can be adapted in POC tests through miniaturization. For instance, magnetic bead-based mechanisms allow for the analysis of complex and crude biological samples [86,87]. Real-time amplification is then carried out through miniaturized modules combining thermal cycling and detection. Melting profiles can be used to differentiate PCR products, in addition to melting temperatures, thereby allowing the detection of multiple pathogens simultaneously.The detection of major causes of STIs is one of the most promising applications of microfluidics [88]. Magnetofluidic cartridges may be assembled, combining reagents for both nucleic acid extraction and amplification with complete process integration from sample to detection, as shown in the case of CT detection operated by a mobile phone interface [89].A LAMP assay was recently developed using magnetic beads and a capillary-based architecture to simultaneously detect NG and CT [90], and another has been tested for identifying genital mycoplasmas (MG, MH, UU) [91], showing concordance with the results of PCR on clinical samples but with a higher sensitivity at lower DNA concentrations. Another assay exploiting LAMP applied to the concept of “lab-on-a-chip” successfully detected NG, CT, MH, and UU directly from genitourinary secretions [92]. The Illumigene (Meridian Bioscience, Cincinnati, OH, USA) HSV-1,2 multiplex LAMP assay has been thoroughly evaluated in a multicenter study, and detects and differentiates the two viruses with a higher sensitivity compared to other molecular assays and even higher than that of culture [93]. A very rapid detection (less than one hour) has also been successfully achieved simultaneously for CT and NG using a strand invasion-based (SIBA) approach [94].Regardless of the approach (PCR-based or IA), NAATs require a strategy to amplify DNA/RNA from clinical samples. Novel methods have been described for the detection of DNA or RNA directly from the sample without amplification. These technologies highlight the possibility of skipping the amplification steps, potentially resulting in a significant reduction in sample handling and processing times, and making POC testing possible.One of these approaches, called microwave-accelerated metal-enhanced fluorescence (MAMEF), combines low-power microwave heating to release DNA directly from a sample and metal-enhanced fluorescence for its detection. This method was first described for the rapid detection (less than a minute) of cultured CT [95]. In another study, MAMEF was used to detect CT and plasmid DNA from genital swabs within nine minutes [96]; despite the suboptimal sensitivity (<90%), the method showed a good agreement with PCR and showed promise to be adapted into a POC test following further improvements.Much attention is now being given to the use of nucleic acid-binding proteins as biosensors for diagnostics [97]. These include zinc-finger (ZF) and transcription activator-like effector (TALE) protein domains, and the clustered regularly interspaced short palindromic repeat (CRISPR)-derived RNA-guided engineered Cas proteins. These programmable proteins can be used to detect very specific DNA/RNA sequences previously amplified via PCR or IA, but their nucleic acid binding properties also allow detection without amplification. Although the use of ZF and TALE proteins has not been closely explored in molecular diagnostics [97], Cas proteins (in particular the Cas12 class of molecules) have attracted increased attention for the POC detection of many pathogens [98].The use of Cas12 proteins has been recently described for viral STIs. Extremely low amounts of HPV DNA (in the attomolar range) were detected using a Cas12 system following preliminary RPA amplification [99]. Furthermore, the assay can discriminate two HPV types (16 and 18) from both anal swabs and cell cultures. A similar strategy was implemented to detect HPV16/18 from swabs and human plasma using a 3D-printed microfluidic device [100]. Electrochemical biosensors for the detection of HPV DNA without amplification via Cas12 have also been reported [101]. Interestingly, a method for the detection of Mycoplasma DNA (to detect culture contamination) using RPA and Cas12 with a naked eye in-tube visual readout has also been described [102]. It is expected that more assays for the diagnosis of STIs will appear in the future using programmable nucleic acid binding proteins and exploiting POC platforms.In addition to the detection of the organism per se, it is important to identify the presence of potential antimicrobial resistance (AMR). This information is pivotal, not only to guide the most appropriate pharmacological treatment course for the patient, but also to prevent the further selection and spread of resistance determinants. AMR resistance is a well-known issue for NG and MG globally, but it is also emerging in CT, TP, and TV [103]. Currently validated methods for the detection of AMR in clinical isolates are time-consuming and involve the cultivation of the organisms followed by in vitro antimicrobial susceptibility testing [104].If the molecular mechanisms of resistance are known, NAATs can be used for the rapid (and potentially POC) detection of resistance genes and mutations. PCR-based molecular assays for this purpose have been developed for STIs, including tests for the detection of macrolide resistance in MG [105] and fluoroquinolone resistance in NG [106]. The absence of known AMR determinants for some antibiotic classes (e.g., fluoroquinolones) has high positive predictive value for phenotypic susceptibility. Conversely, other antimicrobial classes, such as cephalosporins, penicillins, and macrolides, are so complicated that an accurate prediction of resistance cannot be performed due to multiple underlying mechanisms of resistance [107]. Multiplex molecular tests can provide a rapid tool to detect and monitor the prevalence of resistance genes and mutation in at-risk populations [108], and their adaptation to POC formats would benefit patient treatment and management. However, AMR is a complex phenomenon involving several molecular mechanisms, and novel resistance genes and mutations constantly emerge, so the clinical relevance of nucleic acid-based tools remains to be verified against traditional phenotypic tools [107].In recent years, DNA and RNA sequencing technologies have been extensively used to characterize the microbial communities—usually indicated as the microbiome—inhabiting the human body. Metagenomics allows the characterization of the collective genome of microbial communities in a sample, leading to the identification of novel organisms without the need for culture. Furthermore, changes in microbial communities which are potentially associated with disease can be studied. A dysbiotic vaginal environment, like the one observed in bacterial vaginosis (BV), is interlinked with an augmented risk of contracting STIs [109] and adverse pregnancy outcomes [6].Metagenomic approaches allow for the identification of specific bacterial species, such as Gardnerella vaginalis and Atopobium vaginae, which are predictive for BV [110]. The presence of these two bacteria has also been associated with preterm birth [111]. Next generation sequencing has also been used to detect AMR genes directly from urine samples in complicated urinary tract infections [112], potentially paving the way for rapid POC nanopore-based whole genome sequencing technologies. Potentially, these approaches could assist in identifying pathogens in culture-negative clinical samples, as recently shown in the case of synovial fluid [113] and urine [114]. However, before metagenomic approaches can become part of the routine clinical diagnosis of STIs, additional aspects must be considered. Among others, these include developing appropriate laboratory workflows and standards to avoid contamination and difficulty in the interpretation of the results, reducing the cost and turnaround time of the instrumentation, and establishing appropriate (and possibly more user-friendly) bioinformatic analysis pipelines [115].Culturomics is another promising approach for its potential applications to STIs. It consists of the high-throughput analysis of microbes cultured under different conditions using MALDI-TOF mass spectrometry [116]. Culturomics often integrates the information obtained through metagenomics, and allows the functional analysis of enzymes and the metabolic activity of the cultured microbes to further refine the culture conditions. Despite still being in its infancy for STIs, culturomics can aid in the characterization of bacterial species putatively associated with BV [117].Due to the limitations of a syndromic approach, the laboratory diagnosis of STIs is essential to ensure timely and appropriate patient treatment. Furthermore, the actual burden and spread of these infections can only be quantified by including the asymptomatic carriers. Tools based on antigen and DNA/RNA detection have revolutionized the field, allowing for a more rapid and sensitive diagnosis compared to traditional microscopy or culture, and highlighting the possibility of STI screening in at-risk groups. Advancements in the development of novel materials, chemistries, and portable devices has created the potential for POC testing.Some challenges remain for STI diagnostics. NAATs are a powerful tool, but further improvements are needed to make them less technologically demanding so that they may be more affordable in resource-poor settings. Isothermal tools and their future refinements show great promise in this respect, particularly regarding pathogen detection with minimal sample handling and without the need for amplification due to protein-based biosensors. Rapid testing for the presence of antimicrobial resistance remains an important issue for STIs. Long-term goals include thorough research to improve the development of microfluidic technologies and their definitive establishment in diagnostic platforms. Thus, independence from cold chain transport and storage, in addition to a further reduction in the workflow, can be achieved, encouraging the spread of laboratory diagnostics in resource-limited settings and welcoming a new era in healthcare.Furthermore, the extensive use of next generation sequencing, metagenomics, and culturomics on clinical samples could provide assistance, not only in the characterization and detection of AMR mechanisms, but also in developing strategies for previously unculturable organisms. Whole-genome sequencing can be used to track the origin and evolution of hospital outbreaks [115], and allows for the high-resolution typing of microorganisms such as NG to unravel the sexual networks behind STI transmission [118]. Untargeted metagenomic sequencing from a clinical sample can, in principle, identify any microorganism present without previous knowledge of its genome. Appropriate databases and bioinformatic pipelines allow for the rapid identification of pathogens [119], and metagenomics can be used to discover novel viruses in animal and human samples [120]. The use of these approaches on genito-urinary samples could lead to a better understanding of the microbial communities and their impact on the physiology and pathology of the genito-urinary tract, and to discover potentially new pathogens. The widespread and routine use of -omics technologies in a clinical context remains to be seen, particularly in laboratories with limited economic and technical resources, but the future nevertheless appears promising.Overall, novel technological solutions should be focused on improving the sensitivity, specificity, and cost of current POC tests. Cheap and user-friendly tests for STIs could be routinely used on a much larger scale, resulting in a significant reduction in long-term morbidity and also in costs for the healthcare system. Significant advances have been made in POC testing for STIs, and more tests are in the pipeline [121]. However, the need remains for better integration of STI POC testing into healthcare systems [122]. Due to the hidden nature of STIs, ensuring the extensive and rapid screening of at-risk people and their partners is pivotal to successfully controlling these infections.G.C. and T.F. were responsible for ideation, G.C. and R.V. performed literature research, G.C. and A.G. drafted and revised the work. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.Data sharing not applicable.The authors declare no conflict of interest.Comparison of current laboratory diagnostic tools for sexually transmitted infections (STIs) based on performance, applicability, and adaptability to a point-of-care (POC) format.HSV = herpesvirus; HPV = human papillomavirus; DFA = direct immunofluorescence assay; DF = dark field; OI = optical immunoassay; ELISA = enzyme-linked immunosorbent assay; ICT = immunochromatographic test; NAAT = nucleic acid amplification test; PCR = polymerase chain reaction; IA = isothermal amplification; AMR = antimicrobial resistance; POC = point-of-care; EC = endocervical; VA = vaginal; UR = urethral; CO = conjunctival; OP = oropharyngeal; RE = rectal; ASYM = asymptomatic; SYM = symptomatic.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01039.txt

@@ -0,0 +1 @@
+Physical activity has been suggested to be beneficial in preventing disease and improving body function in older people. Older people’s leisure-time physical activity (LTPA) is affected by various factors, especially environmental factors. However, the differences in the association between older people’s LTPA and the built environment in different sex groups remain unclear. Perceived built environment scores and older people’s LTPA were collected for 240 older people in Jinhua using the Neighborhood Environment Walkability Scale and International Physical Activity Questionnaire, respectively. A linear regression method was used to analyze the associations between older people’s LTPA and the built environment in men, women, and all participants. The results showed that land use mix diversity was associated with LTPA in older people for both sexes. In men, LTPA was also associated with access to services. However, in women, LTPA was associated with residential density, street connectivity, and crime safety. The relationship varied when demographic variables were incorporated into the regression analysis. Those results indicated that a shorter perceived distance from home to destination would motivate older people to engage more in LTPA. Older people’s LTPA was affected by various built environment factors according to different sex groups. Women’s LTPA was generally more sensitive to the built environment. More studies are needed to confirm the association between LTPA in older people and the built environment in men and women in mid- or small-sized Chinese cities in the future.Population aging is becoming a global social problem. According to the 2019 Revision of World Population Prospects, 16% of people in the world will be over 65 years old, and the number of people at or over 80 is expected to triple by 2050 [1]. The population aging problem is even more serious in China. By the end of 2040, the percentage of older people over 60 years old will be 24% [2]. The prevalence of diseases, chronic health issues and functional losses greatly increase with aging [3]. Physical activity has been suggested to play an important role in preventing diseases and improving body function by previous studies [4,5,6,7,8,9,10]. Also, as the World Health Organization (WHO) has stated, the benefits of physical activity include lower risks of coronary heart disease, hypertension, stroke, diabetes, colon and breast cancer, enhanced cardiorespiratory and muscular fitness, and so on [11]. Additionally, older people obtain more health benefits from leisure-time physical activity (LTPA) due to greater energy consumption [12]. LTPA refers to body movements that involve energy expenditure caused by skeletal muscle contraction in leisure time [13], and it is influenced by various factors [14]. Socio-ecological models indicate that physical activity is affected by internal factors (e.g., lack of interest, fear of falling), social support (by peers or family) and environmental factors (e.g., weather and access to infrastructure) [14]. Meanwhile, as the person-environment-occupation model suggests, environment should be seen from a broad perspective including the cultural, socio-economic, institutional, physical and social considerations of the environment from the perspectives of the person, household, neighborhood or community [15]. For older people, fears about safety, health problems, and lack of guidance from healthcare professionals have been suggested as barriers to physical activity [16,17]. In a systematic review, Baert et al. concluded that barriers to physical activity at the intrapersonal level are quite similar for younger and older adults except for fear, which seems to be a more specific barrier for older people [18]. Motivators for older people’s participation in physical activity include health benefits, lifelong activity and social support [19]. Because older people are often single or widowed (socially isolated), exercise programs involving social interaction may be especially effective in engaging them in physical activity [20].There is plenty of scientific evidence showing that the built environment is related to the physical activity of residents [21,22,23,24,25,26]. However, few studies have focused on whether this association differs by gender. In a review study, Smith et al. found that older females used parks and greenspaces at a lower frequency than older males did, and parks and greenspaces only promoted older people’s LTPA in males [27]. Sun et al. found that men’s LTPA was associated with street connectivity, walking/cycling facilities and aesthetics, but no significant association was found in females in China [28]. Additionally, Jefferis et al. found a positive association between access to social and leisure actives for males and a nil association for females [29]. While there are a vast number of studies aimed at diverse ages and countries, more studies are needed to investigate the gender differences in the association between built environment and LTPA. A recent report published by the World Bank in 2020 suggested that with women occupying just 10% of the highest-ranking jobs in the world’s leading architectural firms, cities are always built for men and neglect the needs of women. However, people have different needs and routines when it comes to access to the city. Six elements of the built environment combined with gender inequity were mentioned in the World Bank report including access, mobility, safety and freedom from violence, health and hygiene, climate resilience and security of tenure [30]. On the other hand, the model used in the regression analysis would also affect the associations. In Wu et al.’s study, the geographical location of the destination was significantly related with older people’s LTPA in Nanjing, but no significant relationship was found after adjustment by demographics variables [31]. The relative Chinese studies have usually been conducted in super large-scale cities like Shanghai [32], Shenzhen [33] and Hongkong [34,35] and large-scale cities like Hangzhou [36], Nanjing [31], and Xi’ an [28]. All the above cities are in the top 20 cities in terms of economic development and population size. Only a few studies have been aimed at cities of middle scale [37]. The participants’ demographic characteristics and social-economic status are definitely different between cities of diverse scale. This would affect the relationship between physical activity and the built environment as well. Yu et al. found that the relationships were different between a first-tier city, Hangzhou and a second-tier city, Wenzhou. All factors varied between cities except for crime safety and walking/cycling facilities [37]. To our knowledge, only one study has targeted middle-size cities in China [37]. Hence, more studies are needed to provide information on mid- or even small-scale cities in China. In summary, the objective of this study was to investigate the association between older people’s LTPA and the built environment in a mid-scale city, Jinhua in China. We especially focused on the differences in the associations in different sex groups. Additionally, we also wanted to examine whether the linear regression model results would differ when demographic variables were added as covariables. The findings of this study could help to quantify the influence of the built environment on older people’s LTPA in a more accurate and reliable way and would provide support for decision-making in the urban construction process. This study was conducted in the city of Jinhua, which is located in Zhejiang province in the southeast coastal area of China. It is the fourth largest city in Zhejiang province and ranked 49th in 2019 in terms of economic development in China. The official resident population of Jinhua was 5,624,000 in 2019. Both economic development and the resident population in Jinhua are smaller than those cities used in previous Chinese studies [28,31,32,33,34,35,36,37]. Jinhua is a national-level historic and cultural city and a famous tourism city. The largest Chinese film and television studio, the Hengdian World Studio is located here and is open for tourism. Jinhua is also one of the top ten habitable cities in China. Moreover, it has been awarded with a national title as a sanitary city and it has been set as an example of a law-based governed city. For public transportation, the bus is the main traffic mode for residents in Jinhua and two subway lines are under construction. The study was conducted in Wucheng district in Jinhua, which is the downtown area of Jinhua according to its geographical location. Older participants from nine diverse communities were recruited with the assistance of community resident committees. A cross-sectional survey of random samples of older people was conducted. The inclusion criteria for participants in this study were (a) older people over the age of 60 years, (b) residents of the selected communities and have lived there for at least 6 months, and (c) have normal communication abilities and without cognitive impairment. All the questionaries were collected from July to December in 2019, and collected in a one-to-one interview by three team members to ensure the quality. All of them were trained to get familiar with the interview procedure and to understand all the interview questions precisely before they carried out the interview. Demographic data of participants were collected using an individual characteristic questionnaire including gender, age, education level, income situation, travel mode choice and motion sickness. LTPA data for the past seven days were obtained by using the International Physical Activity Questionnaire-Short version (IPAQ-S) [38]. Metabolic equivalent scores (MET) were computed according to the IPAQ scoring procedure. Perceived scores of built environments were collected using the Chinese version of the Neighborhood Environment Walkability Scale-Abbreviated (NEWS-A). The reliability and validity of NEW-A has been shown by a research team from Hongkong [39] and has been widely used in previous studies [28,32,34,35,36,37].Eight built environment elements were evaluated using the NEW-A. A detailed description of this questionnaire can be found in our previous paper [37]. In brief, the diversity of the land use mix evaluated the distances from home to 20 destinations. Thirty-one questions were included in respect to the other seven built environment elements in total. The average score represents the older person’s perceived score in respect to each built environment element. Descriptive statistical analyses were conducted to describe the demographic characteristics in different sex groups. The Chi-square test was conducted to locate the differences in demographic characteristics between sex groups. The differences in older people’s LTPA and perceived scores of built environments in sex groups were examined using the independent t-test. A multivariate linear regression method was used to analyze the relationship between older people’s LTPA and built environment components in males and females separately, and then in all participants together. We also analyzed the association between older peoples LTPA and built environment with demographic variables as covariables. Statistical significance was set at p < 0.05. SPSS 19.0 software was used to conduct all the analyses (SPSS Inc., Chicago, IL, USA).Table 1 shows the demographic characteristics of the participants in different sex groups. Significant differences were found in education level (X2 = 21.17, p < 0.001), income (X2 = 53.41, p < 0.001), and travel mode choice (X2 = 11.07, p = 0.004) between men and women. The education level of most participants was primary or below in both male and female groups. However, the percentage of lower secondary in men was much larger than that in women (42.4% vs. 18.5%), which means that the education level of men was higher than that of women. This result is in line with a previous study in which Zhang found that the education level of older men was higher than that of older women on a national level [40]. With regard to income situation, the income of men was significantly higher than that of women, 33.3% of female participants’ income was 1501–2500 RMB with only 3% of male participants at this income level, and 30.3% of male participants’ income was 3501–4500 RMB, with only 18.5% of female participants at this income level. A previous study also found that the income of older men was higher than that of older women on the national level [41]. The main travel mode was car or bus in both sex groups, but more women chose bicycle (22.2% vs. 12.1%) or walking (14.8% vs. 6.1%) as their travel mode than men. No significant differences were found in age and motion sickness between different sex groups.Table 2 shows the results of a comparison of older people’s LTPA and perceived built environment scores in different sex groups. Men’s LTPA level was relatively lower than women’s LTPA level (2789.27 vs. 3137.22), without significant difference. Regarding perceived built environment scores, significant differences were found in residential density, access to services, and land use mix diversity. Except for access to services, both the perceived scores for residential density and land use mix diversity in the female group were higher than those in the male group. Table 3 shows the association between the built environment and older people’s LTPA in different sex groups. Our results showed that older women’s LTPA was more easily affected by the built environment, and older people’s LTPA was affected by various built elements in different sex groups. In the female group, four built environment elements were significantly associated with older people’s LTPA, including residential density, street connectivity, crime safety, and land use mix diversity. Except for land use mix diversity, all the other three built environment elements were positively associated with older females’ LTPA. Land use mix diversity was negatively associated with older females’ LTPA, which means that a closer distance from home to destination would result in a higher level LTPA in older people. However, in the male group, only two built environment elements were significantly associated with older people’s LTPA. Access to services was positively related with older males’ LTPA. Land use mix diversity was negatively related with older males’ LTPA, which is the same as in the female group. Table 4 shows the association between the built environment and older people’s LTPA with the demographic variables included in the linear regression model. For men, access to services and land use mix diversity were still associated with older people’s LTPA. Additionally, crime safety became a positive influence on older people’s LTPA. No significant associations between demographic variables and older males’ LTPA were found. The association changed a little more in the female group compared with the male group. For women, residential density and land use mix diversity were still significantly associated with older people’s LTPA, but the association between street connectivity, crime safety and LTPA became insignificant. Moreover, aesthetics was positively related with older women’s LTPA as a new influence. A significantly negative association was also found for income. Table 5 shows the association between the built environment and older people’s LTPA in all participants, without considering sex differences. The results showed that four built environment elements were significantly associated with older people’s LTPA, which were the same as the results for the female group shown in Table 3. The results with the demographic variables included in the linear regression analysis showed that the four significant built environment elements in model 1 were still related with older people’s LTPA. Additionally, aesthetics became a new influence on older people’s LTPA, which was the same as in the female group shown in Table 4. Significant associations were also found in two demographic variables, including education level and income. Both were negatively associated with older people’s LTPA. The purpose of this study was to find the association between perceived scores of built environment elements and older people’s LTPA in the city of Jinhua in the eastern region of China. Furthermore, we wanted to find the sex differences in the LTPA and-built environment association. The finding of this study could help to quantify the effect of the built environment on older people’s LTPA in a more accurate and reliable way and could provide support for decision-making in the urban construction process.Our results showed that sex differences existed in the built environment and LTPA association. Older females’ LTPA was more sensitive to the built environment (Table 3). LTPA was affected by only two built environment elements in males, while it was affected by four elements in females. Except for land use mix diversity as a common factor in older people’s LTPA in both sex groups, all the other significant built environment factors were different. Older females’ LTPA was positively affected by residential density, street connectivity, and crime safety. However, older males’ LTPA were not affected by those built environment elements. On the other hand, access to services affected LTPA in males, but did not affect LTPA in females. The association differences that varied by gender were also found in previous studies [28,36,42,43]. Those results suggest that future studies should consider sex differences when investigating the association between older people’s LTPA and built environment elements.The association between older people’s LTPA and the built environment was also affected by the model used. In the linear regression analysis without demographic variables, four built environment elements were related with older people’s LTPA in all participants. However, when demographic variables were included in the linear regression analysis, one more built environment element was correlated with older people’s LTPA (Table 5). This might be due to the fact that two demographic variables were found to be significantly related with older people’s LTPA, and the linear regression model was adjusted by demographic variables. Similar findings could also be observed in Table 4 in the current study and in previous studies [31,33]. Wu et al. found that the geographical location of the destination was significantly related with older people’s LTPA in Nanjing, but no significant association relationship was found after adjustment by demographic variables [31]. Indeed, older people’s LTPA can be affected by various factors. Built environment elements and demographic variables were included in this study, but psychosocial factors like self-efficacy, social support, and perceived benefits and barriers can also have an influence on older people’ LTPA [44,45,46,47]. However, it seems impossible to collect all the factors relative to older people’s LTPA in one empirical study. The association relationship without being adjusted by other factors, or with part of factors adjusted, could also explain how the older people’s LTPA were affected by the built environment to some extent. Land use mix diversity was a strongly influenced older people’s LTPA in this study. It was negatively associated with older people’s LTPA, and did not vary by gender and model used. The land use mix diversity element was evaluated on a 5-point Likert Scale by IPAQ-S, and a higher score means a longer distance from home to the destination. Therefore, our result indicates that a shorter distance from home to destination would motivate older people to engage more in LTPA. The average scores of land use mix diversity were 3.02 and 3.20 in the male and female group, respectively, which means that older people need 11–15 min to walk from home to destination. The results above indicate that a 11–15 min walking distance is acceptable for older people, and a shorter walking distance such as 1–5 min or 6–10 min would encourage older people to take part in more LTPA. On the other hand, walking distances such as 20–30 min and more than 30 min are not very friendly for older people. Older people might abandon their outdoor physical activity plan or select another travel mode instead of walking if the destination is not within an acceptable distance from home. A similar finding was also found in previous studies [42,48,49]. Van et al. suggested that older people’s walking/cycling behaviors were significantly positively related with perceived short distances to services in Belgium [42]. A significant difference was found in land use mix diversity between different sex groups (Table 2). Older females perceived a longer distance from home to destination than older males. Access to services was only positively associated with older males’ LTPA, independent of the model used. This finding is consistent with previous studies. Cerin et al. suggested that better access to shops, park/recreational facilities, public transport, and health-related destinations was positively associated with older people who walk for transport [48]. Good access to recreational facilities and parks was also found to be associated with older people’s LTPA in Barnett et al.’s study [49]. No significant association between access to services and female LTPA was found. This result might imply that access to services is not an important consideration when females take part in outdoor physical activity. The perceived score of access to services in the male group was significantly higher than that in the female group (Table 2), which suggested that older males perceived that they had better access to services. A recent report published by the World Bank suggested that because public spaces are often designed to cater primarily to men and are less accessible to women, women were 15% less likely to use public spaces and this indirectly decreased women’s physical activity [30]. Moreover, crime safety was positively associated with male LTPA after the model was adjusted with demographic variables. This positive association is consistent with previous studies [34,37,50,51]. Yu et al. found that crime safety was positively related with older people’s LTPA in the cities of Hangzhou and Wenzhou, which are also located in Zhejiang province, China [37]. However, they also supposed that because no association was found in other Chinese studies [31,36] and based on the high ranking of China in the Law-and-Order Index report published recently, the relationship between crime safety and older people LTPA in China needs more studies. Residential density was positively associated with older females’ LTPA, no matter what model was used (Table 3 and Table 4). No significant association was found in the older male group. The perceived score of residential density in the female group was significantly higher than that in the male group (Table 2). These results indicate that a higher residential density would motivate older females to engage more in LTPA, however, it does not affect older males. The positive association between older people’s LTPA and residential density was consistent with previous studies in China [32,37] and other countries [22,43,50,51,52,53,54]. To our knowledge, square dancing has become more and more popular in China. Residents like to take part in square dancing in crowds after dinner, and most of the participants are women. A higher residential density might help women to more easily find square dancing companions, and thus increase their LTPA. However, our results were contradicted by other Chinese studies in Xi’ an and Hangzhou [28,36]. The potential reason for the differences might be the diverse participants in these studies. Compared with adults in these two studies, older people in our study are e more able to take part in LTPA like square dancing because of more free time and less perceived pressure from work and home. Moreover, street connectivity, crime safety (Table 3), and aesthetics (Table 4) were positively related with older females’ LTPA. A convenient, safe and pleasing environment is helpful for older people to feel at ease and to overcome fear in their outdoor physical actives. Finally, it will increase older people’s LTPA and nudge them in a health-preventive direction as plenty of previous studies have suggested [51,55,56]. Finally, four built environment elements were found to be significantly associated with older people’s LTPA in all participants of this study, and these were also found to be significantly related with older females’ LTPA. Except for land use mix diversity, which was a common factor in both sex groups, residential density, street connectivity, and crime safety were factors only for older females LTPA. Those results mean that the associations between LTPA in older people’ and the three above built environment elements in all participants were mainly caused by the significant associations found in the female group. On the other hand, those results also indicated that LTPA in older females was more sensitive to the built environment in Jinhua. Compared with older males, older females consider more built environment elements when they take outdoor physical activity. Additionally, income situation was also found to be negatively associated with older females’ LTPA (Table 4), which suggests that older females with a lower income would engage more in LTPA. Hence, examining the association between older people’s LTPA and the built environment by gender is a better way to investigate how the built environment affects LTPA in older people. The strength of this study is that we investigated the association between older people’s LTPA and the built environment, especially by focusing on the association differences in different sex groups. We found that sex differences exist in the association between older people’s LTPA and the built environment. These sex differences cannot be neglected and they provide a more accurate explanation of the effect of built environment on older people’s LTPA. Secondly, most previous studies in China have targeted been at cities of super large-scale or large-scale that rank in the top 20 in terms of population size and economic development [28,31,32,33,34,35,36,37]. Few of them have focused on mid-sized cities [37]. The social-economic status of the population undoubtedly has an effect on the association between older people’s LTPA and the built environment. The current study expanded the scientific base about this association in cities of midscale. However, the present study has several limitations. Firstly, although we conducted a one-to-one interview to avoid deviations as much as possible in the data collection process, some source bias was inevitably included in the data and might influence the association relationship between older people’s LTPA and built environment; this is mainly because self-reporting tools and a cross-sectional survey of random samples of older people were used in this study. Secondly, only demographic variables were included into the linear regression analysis in this study, and social and psychological variables were not combined to explain the effect of the built environment on LTPA in older people. Those variables have been suggested to be related with older people’s LTPA in previous studies [44,46,47,56]. Thirdly, the sample size in this study was a little smaller than those in previous Chinese studies [28,31,32,33,36,37], and this might influence the association between older people’s LTPA and the built environment. Whether the associations found in Jinhua would be changed after enlarging the sample size needs further study. Additionally, the results of this paper should be extended to other populations and cultures with caution because of differences in the demographic profile and social-economic status of participants.Older people’s LTPA was affected by various built environment elements in different sex groups, and older females’ LTPA level was more sensitive to the built environment. Older females consider more built environment factors when they take outdoor physical activity. Higher residential density, better street connectivity and higher crime safety would only encourage older females to take part in LTPA. A shorter perceived distance from home to destination would also motivate older people to engage in LTPA, independent of gender. Otherwise, better access to services would encourage older males to take part in LTPA. The finding of this study helps to quantify the effect of the built environment on older people’s LTPA in a more accurate and reliable way.In order to develop an age-friendly environment in the city of Jinhua, the policy makers should take several vital built environment elements into consideration in the urban construction process as follows. Land use mix diversity is an important and common factor for encouraging older people to engage in LTPA in both sex groups. Additionally, the differences in the effect of the built environment on older people’s LTPA should also be considered by policy makers in order to increase the level of LTPA in both older males and females. A better built environment with access to service for older males and better street connectivity, crime safety and higher residential density for older females would encourage them to take part in outdoor activities. More studies are needed to investigate the association between older people’s LTPA and the built environment in mid-scale or even smaller-scale cities in China. Conceptualization, J.Y. and J.L.; Funding acquisition, J.Y.; Methodology, J.Y., S.Z., D.Z. and A.W.; Writing—original draft, J.Y.; Writing—review & editing, C.Y. All authors have read and agreed to the published version of the manuscript. This study was supported by the MOE (Ministry of Education in China) Project of Humanities and Social Sciences (Project No. 17YJC890040).The study was conducted according to the guidelines of the Declaration of Helsinki, and was approved by the ethics committee of the Research Academy of Grand Health, Ningbo University (RAGH20190600021; June 2019).Informed consent was obtained from all subjects involved in the study.Data available on request due to restrictions privacy. The data presented in this study may be available on request from the corresponding author and with authorization of funding origination. The authors thank the other investigators, the staff, and the participants of the study for their valuable contributions.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.Demographic characteristics of the study participants in Jinhua (n = 240).Note: * indicates significant differences between sexes in Jinhua.Comparison of older people’s leisure-time physical activity (LTPA) and perceived built environment scores for men and women in Jinhua.Note: * indicates significant differences between sexes. MET represents metabolic equivalent score.Association between the built environment and older people’s LTPA in different sex groups.Note: Depend variable: total score of older people LTPA, B stands for regression coefficient, SE represents stand error, * represents significant difference (p < 0.05).Association between the built environment and older people’s LTPA with demographic variables as a covariable in different sex groups.Note: Depend variable: total score of older people LTPA, B represents the regression coefficient, SE indicates the stand error, * represents significance (p < 0.05).Association between the built environment and older people’s LTPA in all participants.Note: Model 1 only includes the built environment variables, Model 2 incorporates the built environment variables and demographic variables. Dependent variable: total score of older people LTPA, B represents the regression coefficient, SE stands for the stand error, * represents significance (p < 0.05).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01040.txt

@@ -0,0 +1,2 @@
+Interventions to discourage sugary beverages and encourage water consumption have produced modest and unsustainable behavioral changes to reduce obesity and noncommunicable disease risks. This systematic scoping review examined media campaigns to develop a typology to support healthy hydration nonalcoholic beverage behaviors. Our three-step methodology included the following: (1) review and summarize expert-recommended healthy beverage guidelines; (2) review six English-language electronic databases guided by PRISMA to describe existing campaign types by issue, goal and underlying theory; and (3) develop a media campaign typology to support policies, systems and environments to encourage healthy hydration behaviors. Results showed no international consensus for healthy beverage guidelines, though we describe expert-recommended healthy beverage guidelines for the United States. Of 909 records identified, we included 24 articles describing distinct media campaigns and nine sources that defined models, schemes or taxonomies. The final media campaign typology included: (1) corporate advertising, marketing or entertainment; (2) corporate social responsibility, public relations/cause marketing; (3) social marketing; (4) public information, awareness, education/ health promotion; (5) media advocacy/countermarketing; and (6) political or public policy. This proof-of-concept media campaign typology can be used to evaluate their collective impact and support for a social change movement to reduce sugary beverage health risks and to encourage healthy hydration behaviors. A robust evidence base shows that adequate and healthy hydration behaviors are essential to promote optimal human health outcomes [1]. The frequent and excessive consumption of sugary beverage products by individuals and populations is associated with adverse diet and health outcomes, including dental caries, obesity and type 2 diabetes [2,3,4,5]. Sugary beverages include different types and brands of soft drinks, sports and energy drinks, fruit drinks and sweetened teas, coffees and milks [3].Temporal trends in sugary beverage intake by children and adolescents in 15 countries (1990 to 2020) documented that consumption was highest before 2000, and only slight reductions were observed for these populations over 30 years [6]. Evidence is inconsistent on whether the long-term consumption of sugary and artificially sweetened beverages is linked to a higher cardiovascular disease mortality risk in adults [7,8]. Prospective cohort studies suggest that replacing a daily serving of a sugary beverage with water, coffee or tea, but not an artificially sweetened beverage, is associated with a 2 percent to 10 percent lower type 2 diabetes risk [9].Governments in countries worldwide have enacted policies and programs to define sugary beverage categories and products to limit or discourage their purchase and use. Examples of policies include enacting sugary beverage taxes, requiring manufacturers to place warning labels on sugary beverages to increase awareness about their contribution to obesity and diet-related health risks and establishing nutrition standards for beverage sales in institutions such as schools, while also encouraging healthier beverages to reduce obesity and noncommunicable disease (NCD) risks [10,11,12,13,14,15,16]. However, well-funded corporate campaigns have opposed these government and advocacy public policy efforts and used the media to influence the views of decisionmakers and the public in many countries to challenge the need for these types of policies [12,14,15]. Additionally, many interventions designed to reduce sugary beverage or increase water purchase and intake have had limited to modest effects to sustain behavioral changes to significantly reduce population health risks [17,18].Media campaigns have delivered corporate advertising slogans to encourage sugary beverage purchases and intake by individuals since the late 1880–1890s by The Coca-Cola Company [19] and PepsiCo, Inc. [20], two companies that currently dominate beverage markets and sales in over 200 countries worldwide [21]. In 2019, Coca Cola ranked first and PepsiCo ranked sixth as the most chosen brands by consumers worldwide [22].The Coca-Cola Company and PepsiCo, Inc. together market beverage brands across five categories: sparkling soft drinks or sugary beverages (e.g., Coke, Sprite, Pepsi, Gatorade and Mountain Dew); water (e.g., Dasani, Smartwater, Aquafina, Bubbly and LIFEWTR); juices (e.g., Minute Maid, Fanta, Sunkist and Tropicana); dairy and plant-based drinks (e.g., Fairlife, Simply Almond and Muscle Milk); and coffees or teas (e.g., Costa, Lipton, Pure Leaf and Starbucks ready drinks) [23,24]. Nestlé is the third largest beverage manufacturer that markets many brands in 187 countries including: ready-to-drink chocolate, tea and coffee (e.g., Milo, Nesquik, Ovaltine and Nestea); water (e.g., Nestlé Pure Life, Perrier, Poland Spring and Pellegrino); and specialized toddler milks [25].The global beverage market is projected to generate $1.86 trillion by 2024 [26]. Transnational beverage firms have partnered with chain restaurants in primarily industry self-regulated markets where national governments have allowed sugary beverages to be widely marketed and socially normalized. These firms have built and sustained consumer demand through advertising delivered across media to encourage sugary beverage consumption, which is linked to obesity and diet-related NCDs in populations worldwide [2,3,4,5].Mass media campaigns have been used by public health practitioners and policymakers in countries worldwide to promote healthy dietary and physical activity messages to prevent obesity and NCDs among children, teens and adults. Evaluations have shown that these media campaigns have focused primarily on changing an individual’s awareness, knowledge or attitudes about obesity as a problem rather than influencing the upstream environmental factors that perpetuate obesogenic environments [27]. Studies reveal that media campaign effectiveness depends on the issue, message content and audience targeted [27,28]. Many public awareness campaigns have not changed parental attitudes about the causes and consequences of childhood obesity, or have not effectively mobilized public support for broader policies, systems and environmental strategies to address obesity [29]. Campaign planners must determine the specific behaviors to target and how to frame communications persuasively to increase the perceived message effectiveness among target audiences to adopt and sustain behaviors to improve health outcomes [27].Stead et al. 2019 conducted an exhaustive review of media campaigns to change behaviors to reduce obesity and NCD risks [30] and concluded that evidence was limited for the effectiveness of diet-related media campaigns, while results were mixed for media campaigns to effectively reduce sedentary behaviors and promote physical activity. These investigators suggested that future evaluations should examine how campaigns contribute to multicomponent interventions and explore how local, regional and national campaigns work synergistically to achieve a clear goal [30].Media campaigns have been used to influence beverage behaviors [14,17,18] but many have not been evaluated. A 2019 Cochrane systematic review of environmental interventions to reduce sugary beverage sales and consumption included only multicomponent media campaigns with a control community as part of the evaluation design [31]. An in-depth evaluation of the design, outputs, outcomes and impact of beverage-related media campaigns has not yet been published.Reviews of mass media campaigns to promote public health outcomes suggest that they are more likely to be effective when planners apply recommended principles, test and develop persuasive messages for targeted populations, adequately fund and execute campaigns as planned to disseminate behaviorally focused messages, update messages as evidence is revised, set realistic milestones and expectations for the media campaign outcomes and expect small to modest behavioral effects for targeted populations [30,32,33,34]. Best-practice recommendations encourage campaign planners to clarify the situation and priorities linked to a vision and mission [35] and to convey actionable, credible, relevant, trusted and understandable messages [36]. Campaign planners must also understand the target audiences’ awareness about the health risks of behaviors, their knowledge about policies to mitigate the risks and their confidence to take recommended actions [36].Social change is a long-term process that fosters collective action to transform social norms, attitudes and behaviors of populations for a specific issue over years or decades [37,38]. Figueroa et al. 2002 [38] identified mass media campaigns as one of many factors used to catalyze community dialogue to influence individual and collective actions within a social change process. Social change movements have harnessed print, broadcast and digital media and used social networking platforms to deliver persuasive messages through media campaigns to influence social norms, attitudes, beliefs, values, behaviors and actions to benefit populations [37,38].Advocacy campaigns implemented in the United States (U.S.) during the 19th and 20th centuries evolved into effective social movements by harnessing the collective social and political actions of people to improve urban housing conditions in cities; demand higher wages and reduce working hours; enhance children’s health and social welfare; and discourage substance use [39]. Social change movements have often been led by people most affected by health inequities or racial or gender biases to challenge existing narratives, assumptions and paradigms about power and privilege in societies and disrupt the status quo [37,38].Examples of effective media campaigns used to support social change movements include those that enacted tobacco marketing policies to restrict availability in public settings; promoted environmental conservation and recycling behaviors; required transportation safety with seat belt laws for automobiles and helmet use for bicycles and motorcycles; and protected civil and gender rights and racial and health equity for historically disenfranchised populations [37,38,39,40,41].The tobacco prevention and control movement in the U.S. that expanded globally is one of the most effective social change movements that used mass media campaigns with a broader policy, systems and environmental (PSE) approach to transform cultural values and behaviors to reduce tobacco use and smoking rates from the 1950s to 2010 [42,43]. The U.S. Food and Drug Administration’s Real Cost public information campaign [44], American Legacy Foundation’s Truth social marketing campaigns [45], the Tobacco-Free Generation proposal [46] and the Tobacco Endgame strategy [47] have used media to build on previous public policy and social change achievements to prevent the sale and supply of tobacco products and socially de-normalize and discourage tobacco and vaping products among young people. The concept of creating a Sugary Beverage-Free Generation has been described [48], but it has not been fully explored to understand how media campaigns can be used as part of a broader PSE approach to discourage sugary beverage products and encourage and socially normalize drinking water and other healthy hydration behaviors among populations.The purpose of this review is to develop a coherent media campaign typology to examine how different types of campaigns can support broader PSE approaches to reduce sugary beverage health risks and encourage healthy hydration behaviors. There is limited published literature on how different types of media campaigns can support a social change movement to discourage sugary beverages and normalize healthy beverage intake as part of a broader PSE approach that includes sugary beverage taxes and warning labels to benefit population health. We also sought to examine expert-recommended healthy beverage guidelines for children, adolescents and adults to have a reference point for analyzing the campaign messages when the typology is tested in the U.S. context.This study addresses these issues with three objectives. First, to summarize expert-recommended guidelines issued by the World Health Organization (WHO) and other international and U.S. expert and authoritative bodies to encourage healthy hydration nonalcoholic beverage behaviors for populations. Second, to conduct a systematic scoping review of relevant evidence to identify types of media campaigns used to influence awareness, preferences, purchase and consumption behaviors of alcohol, tobacco, food and/or beverage products; and to identify conceptual models, taxonomies, typologies and categorization schemes used to delineate or categorize media campaigns. Third, to use these campaign definitions, models, taxonomies and typologies to develop a unique media campaign typology that can be used to describe the goals and underlying paradigms for each campaign type. Thereafter, we plan to test this typology in the U.S. context to examine the evidence to evaluate the collective impact of beverage-related media campaigns to encourage healthy beverage behaviors. The typology may support a Sugary Beverage-Free Generation social movement, similar to the Tobacco-Free Generation proposal, to encourage healthy hydration behaviors to reduce sugary beverage health risks.This study used a systematic scoping review process guided by three research questions (RQ).
+      RQ1:What are existing international and U.S. recommended guidelines or targets to encourage healthy hydration beverage behaviors and targets for individuals and populations?RQ2:What types of conceptual models, taxonomies, typologies or categorization schemes have been used to identify and categorize media campaigns aimed to influence awareness, preferences, purchase and/or consumption behaviors for alcohol, tobacco, food and/or beverage products?RQ3:How can existing models, taxonomies, typologies or categorization schemes be adapted into a media campaign typology to evaluate the collective impact of media campaigns on policies, systems and environments to support a social change movement to establish a Sugary Beverage-Free Generation?What are existing international and U.S. recommended guidelines or targets to encourage healthy hydration beverage behaviors and targets for individuals and populations?What types of conceptual models, taxonomies, typologies or categorization schemes have been used to identify and categorize media campaigns aimed to influence awareness, preferences, purchase and/or consumption behaviors for alcohol, tobacco, food and/or beverage products?How can existing models, taxonomies, typologies or categorization schemes be adapted into a media campaign typology to evaluate the collective impact of media campaigns on policies, systems and environments to support a social change movement to establish a Sugary Beverage-Free Generation?RQ1 provides relevant background information for media campaigns to promote healthy beverage behaviors. To address RQ1, the principal investigator (VIK) used the Google search engine (up to 100 search hits) to conduct a rapid scoping review of healthy beverage recommendations issued by authoritative or expert bodies between 1 January 2000 and 1 December 2020. Websites were searched for grey-literature reports published by and accessed through the World Health Organization (WHO), the six WHO Regional Offices, the Food and Agriculture Organization (FAO) of the UN that describes countries that have translated their national dietary guidelines into food-based dietary guidelines, and the World Cancer Research Fund’s Continuous Update Project and NOURISHING database.VIK also searched U.S. government agencies responsible for developing healthy beverage guidelines, including Health and Human Services (HHS), U.S. Department of Agriculture (USDA) and the Centers for Disease Control and Prevention (CDC); private foundations, including the Robert Wood Johnson Foundation’s Healthy Eating Research National Program; nongovernmental organizations (i.e., National Academy of Medicine and American Heart Association); and professional health societies and organizations (i.e., American Academy of Pediatrics, American Dental Association, Academy of Nutrition and Dietetics, American Society for Nutrition, American Public Health Association and the American Medical Association). The results were independently reviewed by the co-investigator (KCS) to identify expert-recommended guidelines and targets for healthy beverages that could be used in media campaigns to prevent dental caries and reduce obesity and NCD risks.To address RQ2, the co-investigators (VIK and KCS) identified six published reviews of media campaigns used to discourage sugary beverages or encourage heathy beverages to identify search terms and develop a “working media campaign typology” that included: (1) corporate or commercial advertising and marketing, corporate social responsibility or cause marketing campaigns; (2) social marketing campaigns; (3) public information, awareness, health education or health promotion campaigns; (4) counteradvertising or media advocacy campaigns; and (5) political or public policy campaigns. We also reviewed persuasion theories, models and frameworks used to evaluate persuasive communications [49] that served as a foundation for the campaign designs.KCS worked with university research librarians to design a search strategy for a systematic scoping review of evidence that described the purpose, characteristics and underlying paradigms and theories of media campaigns that have been used to influence the awareness, preferences, purchase and/or consumption behaviors of alcohol, tobacco, food and beverage products. We selected a systematic scoping review because of the broad nature of the research questions and to more comprehensively examine the evidence gap for this topic. Many expert recommendations for sugar beverages are unlikely to be initially published in peer-reviewed literature, and we were aware of multiple stakeholder groups that produce relevant reports and other grey literature for campaigns. Likewise, evidence quality was not an initial priority as we were solely focused on identifying distinct categorizations, models, typology or taxonomies and we anticipated a qualitative evidence synthesis [50].We followed five steps for the scoping review described by Arksey and O’Malley 2005 [51] to examine the peer-reviewed literature and grey-literature sources. We identified articles that described a conceptual model, typology, taxonomy or categorization scheme either within one type of media campaign or across different types of media campaigns. The search strategy used the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocol (PRISMA-P) [52] and the PRISMA Extension for Scoping Reviews (PRISMA-ScR) checklist [53]. We did not plan to assess study quality or risk of bias given the exploratory and conceptual nature of the research questions.The search period established was the database inception to 1 September 2020. We reviewed six English-language electronic databases (i.e., PubMed, Web of Science, PsycInfo, Political Science Complete, Academic Search Complete and Health Source Complete: Consumer and Nursing/Academic editions); and conducted a separate review of Google Scholar (first 100 search hits) for evidence relevant to RQ2. Supplemental Table S1 summarizes the pre-defined search terms used for the title and abstract searches.RQ2 inclusion criteria were: (1) English-language peer-reviewed review articles, book chapters or grey-literature reports; (2) evidence sources that proposed a unique conceptual model, taxonomy, typology or categorization scheme to understand or categorize a campaign that aimed to influence awareness, preference, purchase and/or consumption behaviors for alcohol, tobacco, foods and/or beverages; and (3) evidence sources that described a model, category, framework, taxonomy or typology for health media campaigns more broadly and use examples of food, beverage, alcohol and tobacco campaigns. RQ2 exclusion criteria were: (1) non-English-language evidence sources; (2) evidence sources that presented primary research findings or evaluations for specific media campaigns; and (3) evidence sources that did not describe or analyze health-related media campaigns for alcohol, tobacco, food or beverages (i.e., safety or HIV/AIDS prevention) or did not discuss a distinct conceptual model, taxonomy, typology or categorization scheme to inform the development of the media campaign typology.Step one involved KCS reviewing the full-text articles and creating an evidence table to summarize the first author’s last name and year published; the campaign issue(s) examined; type(s) of media campaigns described; any typology or scheme and the basis for its development; target audiences of the media campaign type; underlying theory or theories of change; and documentation of the effectiveness or impact of the media campaigns. VIK also conducted backward searches on inclusion articles to identify any additional relevant evidence sources and separate hand searches using our proposed campaign typology themes to identify any other potentially relevant sources. Step two involved the two co-investigators independently and then collectively analyzing the results to refine our working media campaign typology based on the evidence sources analyzed. The co-investigators discussed and resolved any disagreements with interpreting the article or report for inclusion or exclusion. We address RQ3 in the Discussion section by synthesizing insights from RQ1 and RQ2 to refine our working media campaign. We discuss how it can be used to evaluate the collective impact of media campaigns on policies, systems and environments to support a social change movement to establish a Sugary Beverage-Free Generation.This section describes the evidence synthesized for the three research questions as a narrative review. The first section describes evidence for international and U.S. expert-recommended beverage guidelines and targets to encourage healthy beverage behaviors. The second section presents evidence used to develop and refine the categories of our working media campaign typology. The third section synthesizes the evidence to describe how existing models, typologies, taxonomies or categorization schemes can be adapted and operationalized into a media campaign typology to evaluate the collective impact of media campaigns on policies, systems and environments to support a social change movement to establish a Sugary Beverage-Free Generation.The WHO recommends that children and adults should limit their consumption of foods and drinks that contain high amounts of free sugars, including all types of sugary beverages, throughout the life course to less than 5–10 percent of total energy intake [54]. The WHO also recommends that national governments implement the WHO’s 2010 recommendations to restrict the marketing of unhealthy food and nonalcoholic beverage products to children and adolescents through age 18 years [55] that includes targets for beverages. Between 2015 and 2019, six WHO Regional Offices released nutrient-profile models for national governments to develop policies and laws to restrict the marketing of unhealthy food and nonalcoholic sugary beverage products to children [56,57].Between 2000 and 2019, the WHO had updated quality standards for governments to ensure safe drinking water for populations [58]. However, we were unable to identify any WHO Regional Office reports that provided technical guidelines to Member States for healthy hydration beverage categories, targets or behaviors that could be used in media campaign messages and applied across countries, regions or globally as part of a comprehensive obesity and NCD prevention and management plan. The FAO has identified 100 countries that have translated their national dietary guidelines into food-based dietary guidelines, and a recent review documented varied recommendations for nonalcoholic beverages (i.e., sugary, fruit juices, milk and nondairy substitutes) across countries [59]. In 2018, the World Cancer Research Fund International’s evidence update, published by scientific researchers, recommended that people limit their consumption of sugary beverages and drink mostly water or unsweetened beverages but found insufficient evidence to recommend artificially sweetened beverages containing non-nutritive sweeteners [60].The Dietary Guidelines for Americans (DGA) 2000, 2005 and 2010 [61,62,63] recommended that people moderate their sugar and calorie intake from beverages to maintain a healthy body weight but did not specify daily beverage targets for water, milk or 100 percent juice. In 2013, USDA issued the Smart Snacks in School Standards that recommended school administrators provide children with portion sizes based on their age and established beverage targets, including plain water (carbonated or uncarbonated); unflavored low-fat milk and milk alternatives; 100 percent fruit and vegetable juices; and full-strength juices diluted with water (carbonated or noncarbonated) and with no added sweeteners [64].The U.S. National Academy of Medicine (formerly called the Institute of Medicine) recommended quantitative Dietary Reference Intake targets in 2005 for total water, beverage water (20 percent from food intake), milk and 100 percent juice for healthy individuals and populations that varied by age and sex across the lifespan [65]. These targets were considered when the DGA developed the sugary beverage recommendations. In 2006, an Expert Beverage Panel comprised of U.S. academic researchers published a seminal paper that proposed a new beverage guidance system that offered quantitative daily targets for various beverage categories that prioritized beverages with few or no calories, including water, tea, coffee, low-fat or non-fat milk and non-calorically-sweetened beverages that may provide some nutritional benefits, to replace sugary beverages [66].In 2013 and 2019, the U.S. Healthy Eating Research National Program, supported by the Robert Wood Johnson Foundation, released principles and recommended beverage targets for water, milk and 100 percent juice (but not artificially sweetened beverages) throughout the life course (Table 1) [67,68]. The guidelines for infants, toddlers and children (birth to five years old) were updated in 2019 and endorsed by four U.S. health organizations and professional societies. These are the most comprehensive targets for different beverage categories for infants, toddlers and preschoolers, older children, youth and adolescents and adults. These recommendations have not been widely disseminated, and it is not known whether U.S. media campaign messages have incorporated these recommendations.The DGA 2015–2020 recommended that individuals aged two years and older consume water, fat-free or low-fat milk and/or 100% juice [69]; and USDA’s ChooseMyPlate—the pictorial version of the DGA—showed milk as the default beverage with meals (though not explicitly low-fat or non-fat milk) rather than water [70]. The DGA 2015–2020 also recommended that Americans consume ≤10 percent of their daily calories from added sugars (i.e., 200 calories for a 2000-calorie diet) [69], which was translated by the American Heart Association (AHA) into a recommendation of ≤6 teaspoons or 25 g of added sugars daily for children 2–18 years and adult women, and ≤9 teaspoons or 37.5 g of added sugars daily for adult men [71].The DGA 2015–2020 acknowledged that high-intensity sweeteners (including artificial sweeteners approved by the Food and Drug Administration, such as aspartame and stevia) could be used but noted that there is insufficient evidence to support their long-term use for weight management [69]. An AHA expert advisory committee encouraged Americans to consume primarily water (plain, carbonated or unsweetened); and issued a qualified recommendation that children with diabetes could substitute low-calorie sweetened beverages for sugary beverages and adults could replace sugary beverages with low-calorie sweetened beverages to reduce overall sugar intake [72].The 2020 Dietary Guidelines Advisory Committee report [73] recommended that Americans reduce their added sugars to ≤ six percent of daily energy intake. However, the DGA 2020–2025 report did not incorporate the advisory committee’s recommendation and upheld the <10 percent of calories from added sugars included in the DGA 2015–2020 report. The new DGAs also recommended that parents establish a healthy beverage pattern for infants and toddlers under age two years and that adults consume water or other calorie-free choices as their primary beverage—or other nutrient-dense choices, including fat-free and low-fat milk and 100 percent juice [74]. The DGA 2020–2025 lacked specific guidelines for healthy beverages for adults and suggested that individuals “meet their food groups needs with nutrient-dense foods and beverages; limit foods and beverages higher in added sugars, saturated fat and sodium; and limit alcoholic beverages” [74]. No recommendation was made regarding the use of low- or no-calorie high intensity sweeteners in beverages due to questions about their long-term effectiveness as a weight management strategy [74]. Figure 1 shows the PRISMA flow diagram for the systematic scoping review used to address RQ2. We identified a total of 909 records: 809 records from the six electronic databases and 100 records identified through the Google Scholar search engine. After removal of duplicate records, we screened 666 records, of which 49 full-text records were assessed for inclusion between August and September 2020. We excluded 34 records and selected 15 articles that met the inclusion criteria. An additional 16 articles were identified through a supplemental Google search and by hand searching the reference lists of the included articles between October and November 2020. We included 31 evidence sources in the final review, including 24 articles that defined or described a specific type of media campaign [75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98]. Two of these articles [80,95] defined one or more types of campaigns that also qualified as a unique categorization scheme. Seven additional articles, grey-literature reports and media websites described a distinct media campaign model, categorization scheme, typology or taxonomy [79,99,100,101,102,103,104,105].The systematic scoping review identified 31 evidence sources published between 1990 and 2020 described below and summarized in Table 2 (n = 24 articles) and Table 3 (n = 9 evidence sources), with two articles providing relevant information for inclusion in both tables. We organized the campaigns into the five categories of our “working” media campaign typology that we refined after our analysis. We also noted underlying theories for the media campaigns when reported.Four articles described various types of media campaigns used by tobacco, alcohol, food and/or sugary beverage companies and industry trade associations to promote brands and products to populations to maximize revenues and commercial profits [80,82,96,98]. These included using traditional and social media for advertising and marketing campaigns, corporate social responsibility campaigns, cause marketing campaigns and public relations campaigns.Cruz et al. 2019 [80] examined pro-tobacco advertising and marketing campaigns that used media to tailor package design and messages to reach and appeal to various socio-demographic groups. These investigators defined tobacco marketing as a broad term including paid advertising and promotions in movies and television; sponsorship or loyalty programs; and product design or pricing used by tobacco manufacturers and distributors to reach racially and ethnically diverse populations [80].Dorfman et al. 2012 [82] described corporate social responsibility (CSR) campaigns and social media cause-marketing campaigns, noting that the latter are a variation of CSR campaigns, whereby a company associates its corporate brand and/or product(s) to a social benefit, such as supporting the development of sports centers at schools. CSR campaigns of soda companies were launched in response to increased public concerns about the health risks of sugary beverage products, and also to increase beverage sales among young people. In contrast, tobacco CSR campaigns that were launched to defend their corporate reputation but not to increase the sales of tobacco products [82].Weishaar et al. 2016 [96] analyzed how the corporate campaigns of tobacco, alcohol, soft drink and processed food companies are portrayed by the media. These investigators found that commercial actors frame media messages as individual responsibility for obesity and NCDs, and unhealthy consumption or lifestyle behaviors as a personal choice. Moreover, industry actors also portray government regulation as producing negative economic implications based on a “market justice” frame; whereas public health advocates frame messages with regard to how these corporate actors’ production and marketing practices harm human health and serve as systematic causes of NCDs and poor health outcomes. Public health actors use a social justice frame to advocate for government regulation and population-based interventions to improve health outcomes. None of the articles reported underlying theories to explain the design or influence of these campaigns. Wood et al. 2019 [98] examined two Coca-Cola Company advertising campaigns to highlight how beverage companies use public relations to strategically communicate with customers and use brand marketing to target children, adolescents and their mothers through social media platforms. None of the articles reported theories to explain the design or influence of these campaigns.Nine articles provided definitions for social marketing programs or campaigns used by public health and nongovernmental organizations [75,78,79,81,84,85,88,93,94]. Overall, social marketing involves adapting commercial marketing principles, strategies and techniques to analyze, plan and evaluate programs or campaigns to influence the voluntary behaviors of target audiences. Stead et al. 2007 [93] defined social marketing as adapting commercial marketing principles, strategies and techniques to analyze, plan and evaluate programs or campaigns to influence the voluntary behaviors of target audiences to improve their personal welfare and that of society. Luca and Suggs 2013 [88] described features that distinguish a social marketing campaign from other campaigns and interventions that use the commercial marketing-mix framework to identify a product, price, place and promotion.Several articles described social marketing campaigns used to raise awareness and encourage the adoption of healthy eating behaviors, such as increasing fruit and vegetable intake. Evans et al. 2008 [84] and Evans et al. 2015 [85] described branded health campaigns more broadly and noted that branding can be used in both health communication and social marketing programs. The definition provided for health branding that emphasized using marketing principles to encourage behavior change most closely aligned with social marketing campaigns.Shawky et al. 2019 [92] examined whether social marketing campaigns used social media, defined as tools and platforms for social interaction including digital and web-based and mobile technologies, to augment the messages of social marketing campaigns. These investigators found that Facebook was the primary social media platform used by only a small proportion of campaigns to share content to reach target audiences in order to raise awareness, influence health behaviors and encourage advocacy for campaign activities.Only three articles described specific theories or frameworks for the social marketing campaigns. Cugelman et al. 2011 [81] and Luca and Suggs 2013 [88] described social marketing as drawing on many theories, including the transtheoretical theory or stages of change model, social cognitive theory, cognitive behavioral therapy, health belief model, diffusion theory and the theory of reasoned action. Stead et al. 2007 [93] indicated that a social marketing framework is based on many disciplines and theories to explain human behavior and described six steps to benchmark a social marketing campaign summarized in Table 2. These authors noted that using the term “social marketing” is inconsistent among health interventions and found many interventions that used it but did not meet the six benchmark criteria for social marketing campaigns, while other interventions were not called social marketing but met the criteria [93].Six articles described media campaigns used to share public information to educate and raise awareness among target individuals and populations about the hazards of using alcohol and tobacco products [80], raise awareness and educate about reducing sugary beverages [94] and dietary salt [95] or encourage individuals and populations to engage in healthy lifestyle behaviors [77,95,97]. Bouman and Brown 2010 [77] described that lifestyle campaigns could promote health by building awareness and influencing the public’s attitudes, beliefs, values or behaviors. Three articles noted at least one kind of theory or framework used in the design of health education [95]; public information [97]; and lifestyle campaigns [77].Five articles described either countermarketing campaigns used by public health advocacy groups to counteract the commercial advertising and marketing messages of tobacco, alcohol and sugary beverages corporations and reduce demand for products [76,89,90]; or media advocacy campaigns used to challenge or influence corporate industry practices that harm health [86] and to target policymakers and citizens to drive policy change to address the social determinants of health [83].Bellew et al. 2017 [76] described how social countermarketing is a social change process that starts at the community level and is focused on public good. These authors noted that while social countermarketing campaigns are often created to oppose certain policy positions or commercial marketing tactics that may be harmful to the health or well-being of people and society as a whole, they can also be used to counteract potentially harmful sociocultural norms that have been created through persistent corporate influence [76]. Dorfman et al. 2014 [83] emphasized that with media advocacy campaigns, the focus is not on delivering public health messages to the general public to change behaviors for better health outcomes but is instead more narrowly focused on using media to elevate dialogue and mobilize individuals to put pressure on policymakers, with the ultimate goal of eliciting policy change that will improve the health and well-being of the wider population Only three articles reported on theory: one article reported on agenda setting and media framing theories [83], one on an integrative social countermarketing framework to explain the influence of these campaigns [76] and one mentioned how social cognitive theory helped to understand the influence of peer engagement in tobacco campaigns on individuals’ behavior [90].Only one article, Iyengar and Simon 2000 [87], emphasized the importance of using conceptual models for political campaigns in order to consider the relationship between message content and the predispositions of the targeted populations as well as the interactions between competing campaign messages. These authors noted that political and public policy campaigns may be used to either support or undermine social change and described three theoretical models that can be used to identify and understand the objectives of political campaigns: the resonance model, the strategic model and the traditional model (summarized in Table 2) [87].Table 3 summarizes nine distinct models, categorization schemes, typologies or taxonomies across one or more types of media campaigns. Two of the sources were included in both Table 2 and Table 3. Three typologies were relevant to corporate actors that described the purpose, motives or views about corporate branding [101]; used different types of corporate social media marketing campaigns to reach customers, improve a company’s reputation and maximize sales [103]; and used corporate misinformation or disinformation campaigns to influence decisionmakers or deceive the public about scientific research that produced consequences for public health and safety [104].Five models, schemes or typologies described various public information or health promotion media campaigns that could be implemented by nonprofit organizations and health organizations: a public relations communication campaign typology [99]; a model that described pro-tobacco marketing media campaigns in order to develop more effective anti-tobacco education campaigns [80]; a multicomponent campaign characterization scheme that combined public awareness and health education to influence the behaviors of targeted populations [95]; a campaign typology that influenced target audiences through three different behavioral changes [102]; and a taxonomy for a public communication campaigns defined by purpose, scope and maturity [100]. One typology described a multicomponent media advocacy and public policy campaign to encourage government regulation of industry marketing practices targeted to children [106].Schroeder 2017 [101] described a corporate branding typology with four branding motives to understand how businesses “articulate, embody and embrace cultural contradictions and corporate strategy.” The typology includes a (1) corporate view to develop a firm’s brand strategy to build awareness, recognition, engagement and loyalty; (2) consumer view focused on the role of brands in customers’ lives and in consumer culture exemplified by brand relationships and brand tribes; (3) cultural view where brands are embedded in culture and have meaning, history and a legacy; and a (4) critical view that reveals how brands function as ethical, ideological and political objects [102].ThriveHive 2017 [103] also offered a typology for corporate campaigns, describing four kinds of social media marketing campaigns that individuals or corporations can use to reach large groups of people with a specific product or message. The typology consists of (1) partnership campaigns, (2) holiday campaigns, (3) milestone campaigns and (4) charity campaigns (described in more detail in Table 3) [103].Union of Concerned Scientists 2018 [104] provided a categorization scheme for understanding five corporate campaign tactics that some companies use to undermine or distort science and public health and to protect their own interests. These tactics—given the nicknames the fake, the blitz, the diversion, the screen and the fix—have been used by corporations to cause scientific confusion and ambiguity and blur the lines between science, policy and corporate interests [104].Bünzli and Eppler 2019 [99] defined public relations as a strategic communications process used by nonprofit organizations to build trust and mutually beneficial and sustainable relationships with the public to achieve goals. Public relations may take many forms, including organizational communication, employee communications, media relations, community relations, corporate and social responsibility, public affairs, crisis management and social media. The authors describe four categories of public relations matched to different stages of change (based on the trans-theoretical model) to systematize different types of communication approaches that organizations use when interacting with the public to encourage target audiences to adopt a specific behavior: directing, platforming, involving and mobilizing (see Table 3 for category descriptions) [79].Snyder et al. 2004 [102] conducted a meta-analysis of the effect of U.S. health communication campaigns categories based on their desired effect on behavior to: (1) prevent an undesirable behavior, (2) start a new behavior or (3) stop an existing behavior. Using the diffusion theory, these authors maintain that it is easier to convince people to begin a new behavior rather than to prevent an undesirable behavior or stop a current behavior (especially if it is an addictive behavior), and that campaigns that include enforcement messages are more successful at influencing behavior [102]. As a new behavior becomes socially normalized and more people engage in it, positive role models and public health campaigns may reinforce behavioral changes. However, campaign planners should establish realistic goals because health campaigns have small measurable short-term effects on population behaviors [102].Watson and Martin 2019 [105] describe a multicomponent media campaign typology that combined media advocacy with encouraging citizen protest and engagement and developing relationships with decisionmakers to build political support for government regulation of industry practices that undermine population health. These investigators describe the important role of civil society organizations to mobilize the public and the media to support health policies, and recommend sustained campaigns given the long-term nature of driving health policy change [105].This systematic scoping review examined the available evidence to develop a media campaign typology to support healthy hydration nonalcoholic beverage behaviors. We used three steps to identify relevant evidence summarized in Table 1, Table 2 and Table 3. Thereafter, we synthesized the evidence into a unique media campaign typology. Our analysis for RQ1 revealed a lack of universally accepted, expert guidance for healthy hydration behaviors and targets for nonalcoholic beverage categories across the life course. We discussed several recommended guidelines issued by U.S. expert and government bodies, and found that the Healthy Eating Research provides the most comprehensive healthy beverage guidelines for different age groups (Table 1). Given the lack of comprehensive guidance or a clear definition for healthy hydration behaviors by U.S. and international bodies, we propose a working definition for healthy hydration behaviors. Healthy hydration behaviors prioritize the purchase and/or consumption of water, low- or non-fat milk and other healthier beverage choices outlined in the Healthy Eating Research recommendations (Table 1). We also propose the promotion of water as the default healthy beverage choice. We recommend that the WHO harmonize national and international guidelines for Member States to normalize healthy beverage behaviors for populations, delivered through food-based dietary guidelines and promoted through media campaigns in geographically diverse countries, regions and globally.Figure 2 defines the goal and underlying paradigm for each campaign in our proposed typology based on the evidence analysis and synthesis. The final typology included: (1) corporate advertising, marketing or entertainment campaigns; (2) corporate social responsibility, public relations or cause marketing campaigns; (3) social marketing campaigns; (4) public information, awareness, education or health promotion campaigns; (5) media advocacy or countermarketing campaigns; and (6) political or public policy campaigns.Media—in digital, broadcast and print forms—is a major component of our typology as the key vehicle through which these campaigns share their messages and aim to influence large audiences in order to reach their broader campaign goal. Mass media frames public health policy debates to influence the public’s views by selecting which issues to report and how they are discussed [106]. Mass media also frames the nature and drivers of public health problems and potential policy solutions [106]. Weishaar et al. 2016 [96] and Henderson and Hilton 2018 [106] discuss how corporate actors use a “market justice” media frame versus public health advocacy actors who use a “social justice” media frame to justify their actions in the marketplace. These contrasting frames reflect different values and underlying paradigms associated with media. It is important to analyze how the media deliver campaign messages, as well as analyze the institutions that produce and use the media, and the impact on actor networks, practices, technologies and political ideologies [106].We provide illustrative examples of U.S. beverage campaigns below that fit into the proposed campaign typology based on different goals and underlying paradigms. These examples will require further analysis based on available evidence for campaign evaluations. The first category for corporate media campaigns combines advertising, marketing and entertainment to promote a specific brand or product with the intent to increase its commercial success through sales, purchasing, use and consumption. U.S. media advertising campaign examples include: The Coca-Cola Company’s Open Happiness (2009–2015), Polar Bears Catch (2012) and Share a Coke and a Song (2015) [19]. Popular PepsiCo, Inc. campaigns include The Choice of a New Generation (1984), Pepsi for Every Generation (2018) and For the Love of It (2019) [20].The second category includes public information, awareness, education or health promotion campaigns designed to educate or inform individuals or populations about a health-related issue, such as the harms associated with the purchase, use or consumption of specific products or with practicing certain behaviors. Examples of U.S. beverage campaigns include: Rethink Your Drink (2011–2020), Life’s Sweeter Without Sugary Drinks (2011), The Bigger Picture (2013) and The Sour Side of Sweet (2017) [92,96,107,108,109]. Future evaluations should examine whether the framing of health messages in these media campaigns have led to meaningful changes in cognitive or behavioral outcomes for populations. A Cochrane review of health-message framing suggests that the message goal and attributes have had limited or no consistent effect on consumers’ behaviors [110].The third category includes social marketing campaigns where planners have used commercial marketing principles to encourage individuals and populations to voluntarily begin, reject, modify or stop a behavior to improve their diet or health. U.S. beverage campaign examples include: 1% or Less to encourage low-fat milk (1990s); Got Milk and Milk Mustache? (1995–2015); and Drink Up (2013) to promote water [111]. Given the inconsistent use of “social marketing” within existing campaigns and interventions, we recommend that campaign planners follow the guidelines and benchmark criteria set out by existing social marketing researchers [79,93].The fourth category includes corporate cause marketing, CSR and public relations campaigns. Dorfman et al. 2012 [82] and Wood et al. 2019 [98] describe the purpose of these campaigns as associating a company’s brands or products with a social benefit and communicating with target populations to defend a corporate reputation in response to concerns about their products. Schroeder 2017 [101] and ThriveHive 2019 [103] describe many ways that corporations use branding and social media marketing campaigns to achieve their goals. This category is different from the corporate advertising, marketing and entertainment campaigns because CSR, cause marketing and public relations campaigns are used to defend a company’s reputation, align a company with a cause to positively influence the public’s perceptions of the company and/or maintain “business as usual” to encourage moderate or continued consumption of branded products or introduce new products.Examples of U.S. cause marketing, CSR and public relations media campaigns are the Coca-Cola Company’s Live Positively (2010), Movement is Happiness (2013), The Great Meal and Together Tastes Better (2020) and Together We Must (2020) racial justice dialogue campaigns to support the #BlackLivesMatter movement [82,100,112,113]. PepsiCo, Inc. has also used CSR and public relations campaigns such as the Pepsi Refresh Project (2010) that provided grants to local communities using social media [83]. In 2020, PepsiCo, Inc. launched a combined CSR, cause marketing and public relations campaign that linked social justice and environmental recycling commitments through Black Art Rising, which sponsored visual Black artists connected with a new branded water product called LIFEWTR packaged in 100% recyclable plastic containers and promoted through a hashtag social media campaign [114]. PepsiCo, Inc. also launched the Food for Good “purpose-driven” initiative to address food insecurity and hunger among U.S. children in response to the coronavirus (COVID-19) in 2020 by partnering with local communities to distribute PepsiCo products [115]. Both Coca-Cola Company and PepsiCo have also used CSR and public relations in partnership with the American Beverage Association to promote the Mixify campaign (2014) that was part of the U.S. Balance Calories Initiative (2014) designed to reduce the calories consumed by Americans from sugary beverage products nationwide by 20 percent by 2025 [116].The fifth category includes media advocacy or countermarketing campaigns that share the goal to organize and mobilize communities to challenge corporate marketing practices to advance public health outcomes. U.S. beverage campaign examples include: Kick the Can (2012), The Real Bears (2012), Happiness Stand and Coming Together: Translated (2013), Open Truth (2015), Share a Coke with Obesity and Change the Tune (2015) campaigns [90,94,117].The sixth category includes political or public policy campaigns that use media to engage the public and policymakers to support legislation or laws to regulate products associated with poor diet and health. Watson and Martin 2019 [105] identified the need for a multicomponent campaign to support political and social change that uses media advocacy to frame an issue and potential solutions, encourage citizen protest and engagement and develop relationships with policymakers and decisionmakers to build political support for government regulation. Kennedy et al. 2018 use the term “upstream social marketing” to influence structural and systems change through policymakers [118]. Examples of U.S. beverage campaigns that fit into this category are ones that successfully enacted legislation and local ordinances to support sugary beverage taxes in seven U.S. cities or municipalities, including Berkeley and Oakland, California and Boulder, Colorado [10,12,119].Foster et al. 2012 [35] suggest that planners use a theory of change framework or conceptual model to identify the long-term goal and objectives of media campaigns and the pathways needed to achieve them and recommends clarifying the evaluation processes before implementation. Planners should also identify the assumptions and external factors that may undermine campaign effectiveness; and monitor and evaluate indicators of success for inputs, outputs and outcomes [35]. Figure 3 provides a conceptual model to plan and evaluate media campaigns to support a social change movement to reduce sugary beverage health risks. Our proposed typology is an important tool for planners because a media campaign strategy must examine the competing messages from corporate advertising and marketing campaigns that may undermine public health campaigns prior to implementing and when evaluating a media campaign.Future research should examine the content and perceived effectiveness of messages reported by media campaigns, and the alignment of the campaign messages with government- and expert-recommended beverage targets that promote healthy eating patterns. No published study has examined how media campaigns have been used to support policies, systems and environmental strategies to discourage sugary beverage sales and consumption, while promoting water and low-fat milk choices and other healthy beverage behaviors among populations.To address RQ3, we discuss how this typology may be used to evaluate the collective impact of campaigns on policies, systems and environments to support a social change movement to establish a Sugary Beverage-Free Generation. A collective impact approach involves developing a common agenda and shared measurement as well as providing mutually reinforcing activities, continuous communication and organizational support for the entire initiative [120]. Effective collective impact must consider “who is engaged, how they work together and how progress happens” [121]. A collective impact evaluation approach is better positioned to address large-scale complex societal challenges, such as sugary beverage overconsumption that contributes to dental caries, obesity and diet-related NCDs. We believe that this proof-of-concept media campaign typology may support the strategic communications for a Sugary Beverage Free Generation, which could draw from the Tobacco-Free Generation social change movement that has built support across several Asian countries to phase out the sale and use of tobacco products among young people [122,123,124].Several topics emerged during the evidence screening and analysis that represent cross-cutting issues relevant to the design, implementation and evaluation of different media campaigns to achieve public health and social change outcomes. These issues include: (1) planning for unintended effects or consequences of media campaigns, (2) harnessing social media used by all campaign types and (3) anticipating that media campaigns may spread misinformation or disinformation and produce both intended and unintended effects for audiences.Dillard et al. 2018 suggest that using fear or scare tactics to influence beverage-related health behaviors may lead audiences to resist or reject media campaign messages that may require counter-persuasion to encourage healthy behaviors [125]. Cho and Salmon 2007 developed a typology of 11 unintended side effects of public health communication campaigns including: obfuscation (creating confusion or misunderstanding of a health risk or prevention behavior), dissonance (psychological distress based on the difference between health recommendations and an individual’s experience), boomerang (audience reaction is opposite to the intended response), apprehension (unnecessary concern about health in response to the health-risk message), desensitization (public may not respond after repeated exposure to the health messages), culpability (messages are perceived as blaming individuals rather than the social or environmental circumstances), opportunity cost (selection of certain health issues over others to improve the population’s health), social reproduction (campaigns that reinforce existing knowledge and behaviors among different audiences), social norming (marginalization of certain populations disproportionately impacted by health risks), enabling (promote images and finances of industries or improve the power of individuals and institutions), and system activation (the campaign may influence unintended sectors, and may mediate or moderate the campaign effects on the intended audiences) [126].Social media is used across different campaign types [83,88,94,104], which is a major social innovation that has changed how humans communicate, with important implications for public relations, advertising, marketing and business [127] as well as public health advocacy and policies [41,84,106,128]. Recent political campaigns have used micro-targeting through social media platforms that are tailored to the demographic and psychographic profiles of voters to influence them to support or reject certain political candidates or policy issues [129]. Personalization and microtargeting strategies are used with virtual social media hashtag campaigns such as Share a Coke and a Song, whereby the company offered customers personalized bottles of Coca Cola to increase sales while building an emotional relationship and brand loyalty [130]. Social media platforms can be used to launch either small- or large-scale campaigns to motivate people to take steps toward health-related behavior changes [131]. Social media indicators should be measured as part of the overall campaign evaluation [129].Various institutional actors may use social media campaigns to spread misinformation and disinformation. One typology that we identified described five corporate strategies used to promote misinformation (the unintentional sharing of information) or disinformation (the intentional spreading of misleading political, reputational or financial information) to prioritize business interests over public health and safety [104]. Krishna and Thompson 2019 [132] and the WHO [133] state that misinformation campaigns spread false or scientifically inaccurate information and impede people’s health literacy to make appropriate decisions. These types of campaigns have been used extensively by tobacco firms [134]. Future research is needed to examine how beverage companies use these types of campaigns to influence the public and policymakers about the link between beverage products and diet-related health outcomes, and how these campaigns may influence their views about corporate legitimacy and accountability for marketing practices.This is the first study to develop a unique typology to understand how different media campaigns are used to influence awareness, preference, purchase and/or consumption behaviors for alcohol, tobacco, foods and/or beverages. The typology describes each media campaign based on its unique goal and underlying paradigms. A strength is the use of both peer-reviewed articles and grey-literature sources to develop the typology that allowed us to identify corporate strategies and media campaign tactics not readily available in the published literature. Another strength is the compilation of existing healthy hydration guidance from U.S. and international expert bodies that may inform our analysis of media campaign messages used in the U.S. context to promote healthy hydration behaviors. However, future research should examine whether international consensus exists for healthy beverages since we did not conduct a comprehensive review of all existing international guidelines for this topic. There is also a need to differentiate between corporate media campaigns used to promote sugary beverage brands and products linked to obesity and NCD risks from branded functional beverages emerging in the marketplace that may provide phytonutrients that support healthy dietary patterns [135]. Study limitations are the use of solely English-language articles and the inability to include evidence sources that did not clearly define or describe interventions used in various types of campaigns. An additional limitation was the difficulty to accurately describe distinct media campaigns given the evolving media and marketing landscape as the boundaries between the various campaign types have become blurred.Interventions to discourage sugary beverages and encourage water consumption have produced modest and unsustainable behavioral changes to reduce obesity and NCD risks. This systematic scoping review examined evidence for expert-recommended beverage guidelines. The WHO should harmonize national and international guidelines that clearly define sugary beverage products and help countries to establish national guidelines and targets for healthy beverage behaviors for various populations that can be promoted through media campaigns in geographically diverse countries, regions and globally. We also reviewed the evidence for existing campaigns to develop a unique media campaign typology that has six distinct categories: (1) corporate advertising, marketing or entertainment campaigns; (2) corporate social responsibility, public relations or cause marketing campaigns; (3) social marketing campaigns; (4) public information, awareness, education or health promotion campaigns; (5) media advocacy or countermarketing campaigns; and (6) political or public policy campaigns. This proof-of-concept campaign typology can be used to evaluate the collective impact of media campaigns on policies, systems and environments in a country, region or globally for a specific social, environmental, diet or health issue. This typology should be tested to evaluate whether it may promote healthy hydration behaviors and support a social change movement to establish a Sugary Beverage-Free Generation.The following are available online at https://www.mdpi.com/1660-4601/18/3/1040/s1, Supplemental Table S1: Search strategy used for the systematic scoping review to develop a media campaign typology to promote healthy hydration beverage behaviors to reduce obesity and noncommunicable disease risks.V.I.K. and K.C.S. conceptualized the research questions and study design; K.C.S. designed the search strategy in consultation with V.I.K., and conducted the systematic scoping review; V.I.K. wrote the first draft manuscript and led the submission process. K.C.S. compiled the evidence tables and provided input into subsequent drafts. Both authors read and approved the final submission. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable. Not applicable.V.I.K. received partial funding from the Department of Human Nutrition, Foods, and Exercise at Virginia Tech to support staff salary to complete this paper. V.I.K. and K.C.S. did not receive any funding from government or nongovernmental agencies or commercial or private-sector entities to support the research included in this manuscript. We thank Juan Quirarte for designing the figures.The authors declare no conflict of interest related to the content of this paper.PRISMA flow diagram for the systematic scoping review of evidence for media campaign models, taxonomies, typologies and categorization schemes.A typology of six different media campaigns defined by the goal and paradigm.A conceptual model to plan and evaluate media campaigns to support a social change movement to reduce sugary beverage health risks. Adapted from: Foster, B.; Horton, B.; DeFrancesco, L.; Wedeles, J. Evaluating Social Change. Washington, DC: Vanguard Communications Purple Paper; 2012 (reference [35]).U.S. Healthy Eating Research healthy beverage recommendations by beverage category and age group.Sources: Healthy Eating Research. Recommendations for Healthier Beverages. March 2013. https://healthyeatingresearch.org/wp-content/uploads/2013/12/HER-Healthier-Bev-Rec-FINAL-3-25-13.pdf; and Healthy Eating Research. Healthy Beverage Consumption in Early Childhood. Key Recommendations from Key National Health and Nutrition Organizations. September 2019. https://healthyeatingresearch.org/wp-content/uploads/2019/09/HER-HealthyBeverage-ConsensusStatement.pdf (references [67,68]).Evidence from the scoping review of specific types of media campaigns for alcohol, tobacco, food and/or beverages defined by goal, aim or objective and underlying theory or conceptual framework, 1990–2020.Evidence from the scoping review of unique media campaign models, schemes, typologies or taxonomies, 1990–2020.Directing: Educate the audience by sharing their own point of view, not in receiving audience feedback or input.Platforming: Educate the audience using active listening to hear audience concerns and ideas and acts as a platform for exchangeInvolving: aim Empower audiences by encouraging dialogue on issuesMobilizing: Empower audiences by suggesting specific actionsPro-tobacco marketing campaigns: Use media to market to individuals and populations to increase tobacco sales and use.Anti-tobacco public education campaigns: Use media to educate individuals and populations to discourage and reduce tobacco purchases and use.Purpose: what the campaign is trying to accomplish; purpose is a continuum with personal behavior at one end and policy change at the other.Scope: the size and extent of the campaign; continuum with small, targeted on one side and large, broad on the other; andMaturity: of the campaign and issue based on a continuum with younger (informal) versus older (formal) campaigns.Corporate view: Focuses on developing a firm’s brand strategy to build brand awareness, recognition and loyalty;Consumer view: Focused on the role of brands and branding in the lives of customers and in consumer culture exemplified by brand relationships, brand community and brand tribes;Cultural view: Focused on brands embedded in the culture that have meaning, heritage, history and a legacy rather than being used exclusively as a management tool; andCritical view that reveals ways that brands function as ethical, ideological, and political objects.Campaigns that promote a new behavior;Campaigns that discourage an old behavior; andCampaigns that prevent a new undesirable behavior.Partnership campaigns: Associate a brand with a more successful or better-known brand, raise awareness about both brands, reach an audience currently not targeted by a company’s products or services, and create strong business to business relationships.Holiday campaigns: Design an advertising campaign around many holidays.Milestone campaigns: Celebrate a company’s anniversary, opening a new business location, launching a new product or service, customer loyalty targets, or reaching a certain number of social media followers.Charity campaigns: Support charitable causes and ask staff to volunteer for the company’s charity of choice and document it using social media, such as through Facebook Live streaming.Public awareness campaigns: Change population behavior on a large scale and are characterized by short messages delivered through mass media, print and digital media.Health education programs/campaigns: Provide information about salt-reduction delivered directly to groups of people.The fake: Companies conduct counterfeit science and try to pass it off as legitimate research by ghosting writing articles, selectively publishing positive results or commissioning scientific studies biased toward predetermined results.The blitz: Companies or industry trade associations that oppose or attack scientists who speak out with results or views that challenge the industry.The diversion: Companies that may create scientific uncertainty about evidence that challenges a product’s adverse effects (e.g., tobacco, dietary salt or sugary beverages) by deceiving the public and undermining government regulatory agencies that have a mission to protect the public’s health.The screen: Companies may purchase credibility and legitimacy by developing alliances with academic universities or professional societies, or by sponsoring academic positions, students, or fund research.The fix: Influence government officials or the policy process.Use media advocacy to frame an issue and potential solutions, and boost the discussion of an issue;Encourage citizen protest and engagement; andEngage with decisionmakers to build political support for government regulation.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01041.txt

@@ -0,0 +1 @@
+The literature shows that social pressure promotes non-suicidal self-injury (NSSI) Eating disorders, along with self-injury, are also favored by underregulated social media. Tik Tok is one of the most used social media platforms among adolescents. It has been shown that the time young children spend on this platform doubled during the lockdown. The theme of anorexia is very common on this platform. While most “pro-ana” (pro-anorexia) videos, where users exchanged advice on how to pathologically lose weight, have been censored by the application, other “anti-pro-ana” (anti-pro-anorexia) videos, officially aimed at raising awareness of the consequences of anorexia, have become increasingly popular. However, our case shows how even these safer videos paradoxically lead the users to emulate these “guilty” behaviors.Anorexia is the most prevalent eating disorder, primarily involving female adolescents, with a mortality rate 5 to 10 times higher than in the general population [1,2] and a high prevalence of non-suicidal self-injury (NSSI) [3]. A study also proved how the models of beauty displayed on social media can lead adolescents to develop an eating disorder [4]. The social restrictions due to Covid-19 have led adolescents to spend more and more time on social networks [5]. One of them is Tik Tok, a mobile application created in 2016 that in a few years reached more than 800 million users, at least 20% of which turned out to be teenagers [5,6,7]. This platform has been criticized on several occasions for poor content filtering and mediocre privacy protection [8]. The theme of eating disorders and self-harming behaviors is still present on the application, although disguised.We describe the case of a 14-year-old patient with a diagnosis of anorexia nervosa. Starting from April 2020 (a month after the Covid-19 related lockdown in Italy), the girl reported progressive restriction of food intake and increased physical activity (5 h per day of treadmill), introduction of low-calorie meals, and intake of laxative (8 pills per day) to accelerate the process of weight loss.When presenting to the emergency room, the patient was alert and oriented, she had sinus bradycardia (39–40 bpm), her affectivity was flattened and her mood deflected, with marked alexithymia. She had lost 16 kg in 6 months (BMI 14.2) and had been amenorrheic for 9 months. She reported self-injury (nail scratches on the forearms) but denied suicidal ideation (total C-SSRS score 0).In the initial interview with the child neuropsychiatrist, stubborn refusal of food intake with the declared intent to be hospitalized and desire to lose weight emerged, allegedly not for poor acceptance of her body image, but to try an “extreme experience”, claiming that her life had always been “too simple and free of suffering”. She defined this thought as “perverse” while strongly denying the possibility of death and perceiving herself with “excessive physical strength”. She reported being inspired on Tik Tok, where she saw many teens sharing experiences of deep suffering, often centered on non-suicidal self-injury or eating disorders. She claimed to engage in such conduct with the intent of being hospitalized to demonstrate to herself and others the difficulties of her newly acquired condition.The patient was admitted to the Department of General Pediatrics of the Bambino Gesù Children’s Hospital, in Rome, where she had weekly interviews with psychologists and neuropsychiatrists specialized in eating disorders. She was fed with enteral nutrition, and then switched over for a period of 4 weeks to feeding entirely by mouth. She was allowed to use the mobile only for education purposes and to go out a couple of hours a day to perform group activities with other patients hospitalized with the same diagnosis. She was initially treated with aripiprazole, as indicated [9,10]. Although a diagnosis of major depression was never made, due to the important mood deflection, it was necessary to treat her with fluoxetine. She demonstrated a strong perseverance in refusing food, even though she was aware of her poor health and the reasons that had led her to that condition. She partially recovered her weight (BMI 15.8) and was no longer bradycardic, thus she was discharged with a specific diet and neuropsychiatric follow-up.The intent of this article is to show how social media can promote self-injury and eating disorder behaviors. The social media platform referred to in this case is Tik Tok, a platform whose public is primarily made of young adolescents. Unlike other social networks where written content (i.e., Facebook) or images (i.e., Instagram) are favored, Tik Tok provides 15- to 60-second long videos where users usually perform dances to a musical basis [11]. Over time, it has encompassed a wide range of topics, from cooking to performances by professional dancers to political discussions.The application algorithm records data from the single users and proposes videos that catch the kid’s attention specifically, by creating a personalized “For You” page [12]. This feed will suggest videos from anyone on the platform, not just from the followed accounts [11]. Therefore, if a user accidentally views a video dealing with anorexia on the homepage and, intrigued, searches for other videos alike, the algorithm will keep suggesting such videos, contributing to the development of obsessive behaviors, as in the examined case [12]. These algorithms, with the aim of increasing the diffusion of user-sensitive virtual content, are based on solid models whose development is still ongoing [13]. However, these algorithms are not yet able to discriminate harmless and harmful content, as this, rather than the videos they like, is related to the user’s own life experience, which cannot be embedded in the “formula”.The theme of anorexia entered Tik Tok through “pro-ana” videos (pro-anorexia), where users supported and encouraged each other to lose weight, exchanging advice and challenging one another [14]. Luckily, most of these videos were blocked by the platform control over content (15.6% of removed videos supported suicide, self-harm, and other dangerous acts) [15], although some still easily circumvent the controls [14,16]. Currently, “anti-pro-ana” (anti-anorexia) videos, officially aimed at raising awareness on anorexia, are much more prevalent [17]. Unlike explicit pro-ana videos, these videos are not subject to banning, with increased content visibility. The same process has also been observed for videos on self-injury [15].In our case, the patient attended high school with excellent performance and a tendency to perfectionism, she practiced contemporary dance and played cello with a strong commitment. Her family environment was not disrupted, as often occurs in these patients. During the lockdown, with much more free time on hand, she turned to Tik Tok for entertainment and came across videos about eating disorders, non-suicidal self-injury (NSSI) and depression, officially aimed at showing the darker sides of these conditions and scare those who romanticized or promoted them. However, she reported that the video makers are often competitive with one another, as they are tempted to prove they are in the worst condition by showing the numbers, calories, and parameters of “being really sick”. We believe that this kind of content, if shown to fragile teenagers, insecure about their body image, may have a paradoxical effect. In fact, when our patient decided to live such experiences “to grow-up”, she learnt to hide food and vomit, while deceiving others, through these same “good” videos. She then gradually began to lose weight and to self-inflict cuts across her forearms and wrists, until she was admitted to hospital, finally showing herself and others her “true state of suffering”.Our study has the main limitation of being a case report. As such, it cannot assess a causality relationship between two given events, in our case the exposition to “anti-pro-ana” videos and the development of eating disorders. We therefore hope that further prospective studies may be carried out to better address this putative relationship.Nevertheless, this is the first article that explores the paradox behind “anti-pro-ana” content on social networks. “Anti-pro-ana communities” have already been mentioned on previous studies, yet always with the intent of demonstrating their positive effect on adolescents [17]. To our knowledge, only one other work approached the same topic, in a different way [18]. In this recently published study, virtual content under the hashtag #Edrecovery (eating disorder recovery) were analyzed, highlighting the link between these and pro-eating disorder videos on Tik Tok [19].With this article, we intend to raise awareness on the influence that social media exerts on young adolescents, whose parents are often unaware [16] and whose teachers and pediatricians need to consider more closely. With the restrictions due to the Covid-19 pandemics, the lives of adolescents have radically changed, losing daily habits and sources of leisure and entertainment, and turning to social networks to fight the sense of isolation and boredom [20]. We showed how easy it is to come across messages that are harmful both to the mental and physical health of adolescents on such platforms. In fact, what seems like harmless videos about anorexia or self-injury may instead paradoxically trigger the emulation of such destructive behaviors.Conceptualization, M.R.M.; writing—original draft preparation, G.L., M.R.; writing—review and editing, all of the authors. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.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.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.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01042.txt

@@ -0,0 +1 @@
+The aim of this study was to analyse the association of the use of the mobile phone with physical fitness (PF) and academic performance in secondary school students and its gender-related differences. A total of 501 high school students participated in the study (236 girls and 265 boys; 12–18 years). Use of the mobile phone and sample distributions were done with the Mobile-Related Experience Questionnaire (CERM): low use of mobile phone (LMP = 10–15 points), medium use of mobile phone (MMP = 16–23 points) and high use of mobile phone (HMP = 24–40 points). PF via Eurofit test battery and academic performance were recorded, and gender was used as a differentiating factor. The HMP group registered lower values than the LMP group for academic performance (Spanish: 4.78 ± 2.26 vs. 3.90 ± 1.96 points; p = 0.007, Mathematics: 4.91 ± 2.23 vs. 4.00 ± 1.84 points; p = 0.007) and PF (Abdominals: 6.83 ± 2.40 vs. 5.41 ± 2.46 points; p < 0.001, Broad jump: 6.24 ± 3.02 vs. 4.94 ± 2.28 points; p = 0.013). The boy students showed greater values than girl students for PF in the LMP (medicine-ball-throw: 6.34 ± 2.24 vs. 5.28 ± 1.86 points, p = 0.007) and MMP (medicine-ball-throw: 6.49 ± 2.52 vs. 5.02 ± 1.68 points; p < 0.001) groups, but no gender-related differences were found in the HMP group. In conclusion, high use of the mobile phone was related to worse results in the PF tests and academic performance. Gender-related differences were found for academic performance regardless of the use of the mobile, but for physical fitness no gender differences were found in HMP group.The advent of new technologies, above all the smartphone, has meant a considerable change in the lives of people in general and especially in adolescents. The latter deserve special attention in terms of their relationship with these technologies that are present in their daily lives [1]. Currently, more than 60% of adolescents have a mobile, and it is commonly used in their activity time [2]. Thus, they tend to be more vulnerable to suffering from problematic situations such as negative feelings or low self-esteem when they do not have a mobile phone or do not receive messages or calls [3]. In addition, the reduction in time spent on physical activity and studying in favour of mobile phone use could be related to poor development in the adolescent’s academic and physical area [4]. Therefore, adolescence is considered an important risk factor due to the high frequency of Internet connection and mobile phone use [5]. Physical fitness is defined as a set of attributes related to health [6] and the level of physical fitness is considered a powerful indicator of health in adolescence [7]. Physical fitness values in Spain are somewhat lower compared to other countries [8]. This may be due to the decrease in the practice of physical activity that occurs especially in adolescence as shown in previous research [9]. Physical inactivity or the decrease of physical activity during adolescence is related to lower values of physical fitness, to the detriment of health and quality of life, as well as the development of cardiovascular diseases, obesity and osteoporosis during adulthood [10]. In addition, the use of a smartphone could contribute to the decrease in physical activity and, therefore, the reduction of physical fitness levels [11]. Most of the research has been focused on the use of the Internet, apps and smartphones as a solution for combatting sedentarism [12], or other factors such as academic performance [13], although the relationship between the use of the mobile phone and the risk of sedentarism has been less reported [14]. High school may become the only real option for performing physical activity for adolescents [15], and therefore it is also a favourable environment for evaluating physical fitness and adopting possible measures that may prevent diseases in adulthood.Academic performance could be defined as the students’ performance in each year, which is rated on a point scale. In relation to the physical fitness of high school students, the best values of academic performance during adolescence used to be associated with good health and low body fat in adulthood [16]. Furthermore, other studies point out that on the one hand adolescents who have physical activity habits or practice sports at a competitive level obtain better academic results than sedentary students [17,18], and on the other hand, spending more than two hours per day in front of a screen could be associated with lower academic achievement among school-aged children [19]. Some research showed differences between boy and girl students in physical fitness and academic performance; while boys recorded better physical fitness performance [20], academic performance was higher for girls [21]. In addition, the relationship between leisure screen time and the probability of getting high academic performance has been reported in boy students but not in girl students [22].Interest in the possible addiction of high school students to mobile phones has grown in recent years. Some studies have noted that adolescents with high mobile-phone dependence showed a greater risk of suffering psychological and social problems [23]. The Mobile-Related Experience Questionnaire (CERM) is a questionnaire based on 10 criteria and was designed to examine the degree of “addiction” to the mobile phone [24] showing higher values in female than male university students [25]. The consequences of the increase in mobile phone use in recent years and its gender-related differences should be studied in high school students. Therefore, the objective of this study was to analyse the association of the use of the mobile phone on physical fitness and academic performance in secondary school students and gender-related differences.A total of 501 secondary school students volunteered to participate in this cross-sectional and correlative study. Participants were divided into three groups according to their results in the CERM: low use of mobile phone (LMP), (N = 133, Age: 14.7 ± 1.4 years, Height: 163.9 ± 9.9 cm, Body Mass: 60.1 ± 14.3 kg, Body fat: 22.4 ± 10.8 %, mobile phone use: 13.5 ± 1.7 CERM points), medium use of mobile phone (MMP) (N = 305, Age: 15.2 ± 1.6 years, Height: 164.4 ± 11.9 cm, Body Mass: 59.9 ± 13.3 kg, Body fat: 21.7 ± 9.8%, mobile phone use: 18.7 ± 2.1 CERM points) and high use of mobile phone (HMP) (N = 63, Age: 15.3 ± 1.6 years, Height: 163.4 ± 11.8 cm, Body Mass: 59.4 ± 17.0 kg, Body fat: 24.4 ± 11.2%, mobile phone use: 25.7 ± 2.0 CERM points). The students came from state schools in the region of Castilla–La Mancha (Spain) (Figure 1).All participants and their parents were informed in writing and verbally of the purpose and procedures of the investigation, and the parents of the participants provided a signed informed consent before the start of the study. The participants were free to leave the activity without the need to give any kind of explanation and without their departure implying any sanction. The study was approved by the Ethics Committee of Clinical Research at the Hospital Complex in Toledo (Spain) (number 487, dated 25 February 2020) according to the principles of the latest version of the Declaration of Helsinki.The CERM was employed to divide the sample into 3 groups according to the points obtained [26]: LMP (total points were between 10 and 15 points, which means no problems derived from mobile phone use), MMP (total points were between 16 and 23 points, which means occasional problems derived from mobile phone use) and HMP (total points were between 24 and 40 points, which means problematic use of mobile phone). The questionnaire consists of 10 Likert items with four responses, classified from ‘1’ to ‘4’, in increasing order of intensity. It is used to analyse two mobile-related factors: problems due to emotional and communicational use, and conflicts related to mobile use. The CERM scale has good validity and reliability, and also showed a positive high correlation with the use of the mobile phone [24]. The CERM questionnaires were completed by each of the students at the beginning of the measurement process and the total score was used for further analysis. For the evaluation of physical fitness, we included several tests included in the Eurofit [27] which are validated and standardised by the Council of Europe and showed very good test-retest reliability and good criterion validity for adolescent evaluation [28]. The tests were applied by a researcher (JMG) in collaboration with the physical education teacher during a class after a familiarization session carried out on a different day. The marks obtained by the participants were evaluated according to their age and gender, and rated from 0 to 10 points, 10 being the highest qualification [29,30]. All tests were performed twice, and the best performance was chosen and expressed as a 0–10 mark. The test battery was applied in the following order:Abdominals in 1 min: it consisted of performing the highest number of trunk crunches in 1 min from a supine position on a mat and with the feet attached to a trellis. Repetitions to complete 1 min of exercise were recorded.2 kg overhead medicine-ball throw: it is based on throwing a medicine ball over the head as far as possible by extension–flexion of the trunk and upper limbs. The medicine-ball throws were performed using a 2 kg rubber medicine ball. Each throw was measured for distance.Standing Long Jump Test (Broad jump). A tape measure was employed to measure the maximum horizontal distance jumped. The student stood behind a line marked on the ground with feet together. A 2-foot take-off and landing was performed, with a swing of the arms and bend of the knees to provide forward drive. The subject attempted to jump as far as possible, landing on both feet without falling backwards. Each jump was measured for distance.50 m sprint: students performed 50 m runs at a maximum pace. Time to complete the 50 m run was recorded. The instructions given to start were a countdown from “ready, 3, 2, 1, go.”Trunk flexion test: it was based on performing a flexion of the body by bringing the arms back between the legs without bouncing. The objective was to analyse the flexibility of the trunk and lower extremities. Each attempt was measured for distance.10 × 5 m shuttle run test (Agility test): Subject was required to run back and forth as fast as possible 10 times, along a 5 m course. Time to complete the agility test was recorded.Abdominals in 1 min: it consisted of performing the highest number of trunk crunches in 1 min from a supine position on a mat and with the feet attached to a trellis. Repetitions to complete 1 min of exercise were recorded.2 kg overhead medicine-ball throw: it is based on throwing a medicine ball over the head as far as possible by extension–flexion of the trunk and upper limbs. The medicine-ball throws were performed using a 2 kg rubber medicine ball. Each throw was measured for distance.Standing Long Jump Test (Broad jump). A tape measure was employed to measure the maximum horizontal distance jumped. The student stood behind a line marked on the ground with feet together. A 2-foot take-off and landing was performed, with a swing of the arms and bend of the knees to provide forward drive. The subject attempted to jump as far as possible, landing on both feet without falling backwards. Each jump was measured for distance.50 m sprint: students performed 50 m runs at a maximum pace. Time to complete the 50 m run was recorded. The instructions given to start were a countdown from “ready, 3, 2, 1, go.”Trunk flexion test: it was based on performing a flexion of the body by bringing the arms back between the legs without bouncing. The objective was to analyse the flexibility of the trunk and lower extremities. Each attempt was measured for distance.10 × 5 m shuttle run test (Agility test): Subject was required to run back and forth as fast as possible 10 times, along a 5 m course. Time to complete the agility test was recorded.To assess student achievement, we used their grade point average from the first trimester, provided by the participants and verified by the academic secretary of the corresponding centre. These grades summarise the student’s performance in the following common subjects: Spanish language and literature (Spanish), mathematics, English and physical education; they range from 0 (=very poor) to 10 (=excellent) and are standardised by Spanish law.The statistical analysis was performed with IBM SPSS Statistics 23.0 (SPSS, Chicago, IL, USA). All data are expressed as mean ± standard deviation. The data were tested for normality with a Kolmogorov–Smirnov test. Since the assumption of normality (all variables p > 0.05) was verified, a two-way ANOVA (2 × 3) was used to establish the differences in the CERM score, physical fitness and academic performance variables between the two gender groups (boys and girls) and among the three mobile-phone-use groups (LMP, MMP and HMP) and subsequently post-hoc Bonferroni’s tests were used when a significant main effect or an interaction between factors was found. The relationship between physical fitness tests and academic performance variables was analysed with a simple linear regression, from which the Pearson correlation coefficient was calculated. The effect size (ES) was calculated for all pairwise comparisons according to the formula proposed by Cohen [31]. The magnitude of the ES was interpreted using the scale of Cohen [31]: small (<0.2), medium (0.5) and large (>0.8). A probability level of p < 0.05 was defined as statistically significant.After the CERM questionnaire had been answered, the results of our study showed that 12.8% of the high school students had addictive behaviour in relation to mobile phone use. The number and the percentage of all groups of boy students assigned to risk groups (MMP and HMP) was higher than for girl students (68.8% boys and 64.2% girls). We did not find differences between boys’ and girls’ average results in the CERM questionnaire (boy vs. girl): 17.89 ± 4.01 vs. 18.61 ± 4.25 points; p = 0.055. No differences were described between boys’ and girls’ results in each group: LMP (13.32 ± 1.68 vs. 13.72 ± 1.64 points; p = 0.167); MMP (18.56 ± 2.02 vs. 18.95 ± 2.18 points; p = 0.110) and HMP (17.89 ± 4.01 vs. 18.61 ± 4.25 points; p = 0.055).The physical variables results are presented in Table 1. Significant interactions between use of mobile phone groups and gender were found for the abdominals test, medicine-ball throw, broad jump, 50 m sprint, trunk flexion test and agility test. The boys’ group showed higher values than the girls’ group in the LMP and MMP groups for the medicine ball (p < 0.008), broad jump (p < 0.003), 50 m sprint (p < 0.014) and agility (p < 0.025) tests. Boys also achieved higher scores than girls on abdominal tests for the MMP group (p < 0.001). No differences were found between boys and girls in the deep trunk-flexion test (p > 0.100). Finally, we did not find differences between the genders in HMP for any test (p > 0.129). Without the assumption of the gender effect, significant differences were described for the abdominals test (p = 0.001), medicine-ball throw (p = 0.001), broad jump (p = 0.011), and trunk-flexion test (p = 0.005) with higher values in the LMP and MMP than the HMP group (Figure 2).Within the boys’ group, intermobile phone-use differences were observed in the PF performance being higher in the values of LMP group than HMP group for the abdominals test (p = 0.005), medicine-ball throw (p = 0.047), broad jump (p = 0.009) and trunk-flexion test (p = 0.047). In addition, higher values were also described for MMP than HMP in the abdominal test (p = 0.007), medicine-ball throw (p = 0.009) and trunk-flexion test (p = 0.004). No intergroup differences were found in the 50 m sprint and agility test (p > 0.05). For the girls’ group, no differences among mobile-use groups were found (p > 0.05).Significant interactions between mobile-phone-use groups and gender were found for Spanish, English and Physical Education subjects. Spanish results for girls were 32.3% greater than for the boys’ group in the LMP group (p < 0.001). Girls showed 17.7% higher values in Spanish than boys in the MMP group (p = 0.001) and 46.22% higher in the HMP group (p = 0.006). English marks for girls were 10.7% higher than for boys in the MMP group (p = 0.019) and 25.5% greater in the HMP group (p < 0.001). No differences were described between boys and girls in mathematics and physical education marks. The group means are presented in Table 2. Without the assumption of the gender effect, significant differences were described for Spanish (p = 0.006), mathematics (p = 0.036) and physical education (p = 0.036), with higher values in the LMP and MMP than in the HMP group (Figure 3).For boys, differences among mobile-phone-use groups were observed in academic performance. The Spanish marks in the MMP group were 30.6% greater than in the HMP group (p = 0.004). The physical education marks in the LMP group were 16.5% greater than in the HMP group (p = 0.009). No intergroup differences were found in mathematics and English. For the girls, no differences among mobile-phone-use groups were described (p > 0.05).A positive correlation was found between physical fitness and academic performance (r = 0.359; p < 0.001). Physical education marks also showed a positive correlation with academic performance (r = 0.607; p < 0.001).This study describes the association between use of the mobile phone and PF and academic performance in high school students and the differences between both genders. The main findings of our study were: (i) The students with the highest use of mobile phon, without the effect of gender, showed worse results in PF and academic performance than in the LMP and MMP groups; (ii) Boy students presented higher values in PF than girl students when the use of the mobile phone was low or medium (LMP and MMP groups); (iii) English and Spanish academic results were better for girls than boys independently of mobile phone use; (iv) A positive correlation was described between physical fitness and academic performance. All of these results suggest that higher use of the mobile phone was related to worse results in physical fitness and academic performance in high school students—although longitudinal studies are warranted to confirm it—and also that the association with other variables such us socioeconomic status and physical activity should be investigated.The results of our research show higher marks for boy students in most of the physical tests evaluated compared to girl students. These differences could be attributed to the action of testosterone, which has an important role in physical performance, especially in relation to strength [32]. Based on the premise that the maturation process is responsible for the gender differences in physical fitness, boy students should show better marks than girls in all the mobile-phone-use groups examined. Previous studies have reported greater than 55% gender-related differences in a similar abdominals test and upper-body-strength test [33]. However, contrary to this hypothesis, no differences were observed in the physical fitness test between boy and girl students in the HMP group. Therefore, maturity is not enough to provoke the gender-related differences in physical fitness and even less when, according with the EUROFIT guide [27], the marks are rated in a specific way for gender; therefore, other factors must be associated with the differences, such as daily time devoted to physical activity [20]. Previous research showed the relevance of physical-activity time per day in young athletes and its relationship with coordinative aspects [34]. Therefore, the greater use of the mobile phone for HMP students could provoke a decrease in physical-activity time and lead to a reduction of the gender-related differences. Finally, mental fatigue associated with mobile-phone use could also contribute to a decrease in physical fitness [35].In our research we found intergroup differences in physical fitness for boy students but not for girls; thus, use of the mobile phone could affect them in a different way depending on gender. Boys in the HMP group showed lower values in physical fitness tests than those in the LMP and MMP groups. Currently, more than 85% of students have a mobile phone and described smartphones as an integral part of their daily lives [2]. This phenomenon could lead adolescents to have less free time for necessary physical activity. On the other hand, Stephens and Allen [36] found that appropriate use of the mobile phone supported by education was an effective tool to reduce sedentarism and obesity; the problem arises when previous studies support that the bulk of time students spend on their mobile phone is for leisure (~70%) and when they are using their mobile phone they are typically sitting (~87%) [37]. Previous research suggested gender-related different responses to the decrease in physical-activity time [20]. Lower values of physical-activity time regardless of exercise intensity reduced the boys’ physical fitness, while for girls the differences in physical fitness were in relation to the time spent in physical activity of high intensity [20]. Higher values of CERM (>20 points) showed a relationship with a greater level of addiction to the mobile phone [25]. In our study, although we did not find differences in CERM points between boy and girl students, the percentage of boys in the MMP and HMP groups was higher than that of girls (~69% vs. ~64%), contrary to previous research [23,25]. Therefore, the high percentage of boy students in the groups with a greater addiction to mobile-phone use could result in less physical-activity time and also could be the reason for the physical-fitness differences described among the boys’ groups.Academic performance in students was evaluated based on criteria established by the curriculum for secondary education. From our results and according to previous research [4], there was an association between mobile-phone use level and academic performance, and those groups with lower levels of mobile-phone use, LMP and MMP, achieved higher academic results. In our study, girls showed better results than boys in language subjects (Spanish and English), and these differences did not change when the student level of mobile-phone use was factored in. The higher use of mobile phones for the boys could be the reason for the lower marks in academic performance. In addition, contrary to physical fitness, the gender-related differences did not disappear for the HMP group following the line described in a previous meta-analysis that showed better school-grade marks for girl students [21]. Finally, a significant correlation was found between physical fitness and academic performance, therefore preserving high levels of physical fitness could help to reduce the negative impact of mobile-phone use. Coinciding with our results, previous studies confirmed that those students with lower physical fitness and higher body fat obtained poor academic achievements [38,39].This study has several limitations. First, the use of subjective methods, such as questionnaires, presents important limitations, because it is well known that self-perceptions in questionnaires tend to overestimate usage. Future research should provide specific and accurate analysis of mobile-phone use incorporating objective methodologies. Second, due to the differences in academic curriculums and assessments in secondary schools all around the world, it is difficult to establish comparisons in academic performance with other investigations. In addition, only common subjects were studied; therefore, new investigations are warranted to analyse the association between the use of mobile phones with performance in other subjects such as economics or history. Third, the individual physical-activity time was not recorded, although previous studies showed high correlations between physical fitness and physical-activity daily time. It is reasonable to think that a more in-depth analysis of physical activity, sports practice and socioeconomic status would have to be implemented, as well as the kind of apps used by the students, to gain insight into their roles as contributors to the problematic use of the mobile phone. Finally, the study design (cross-sectional study) did not allow establishment of causal effects; therefore, the results should be corroborated by longitudinal research to confirm the association between mobile-phone use and PF and academic performance. In conclusion, the results of the present study indicate that the groups of secondary school students with lower values of mobile-phone use outperform students with greater values both physically and academically. Boys showed better results for physical fitness than girls when mobile-phone use was low. The gender-related differences in physical fitness disappeared when mobile-phone use was high. Academic performance was higher for the girls’ group and did not appear to be conditioned by mobile-phone use. Finally, a positive correlation was described between physical fitness and academic performance.Conceptualization, A.B.-S., J.M.-G., P.A. and J.A.-V.; data curation, A.B.-S., J.M.-G. and P.A.; formal analysis, A.B.-S., P.A. and J.A.-V.; investigation, A.B.-S., J.M.-G. and P.A.; methodology, J.M.-G. and J.A.-V.; supervision, J.A.-V.; writing—original draft, A.B.-S. and J.M.-G.; writing—review and editing, P.A. and J.A.-V. All authors have read and agreed to the published version of the manuscript.This research received no external funding.The study was approved by the Ethics Committee of Clinical Research at the Hospital Complex in Toledo (Spain) (number 487, dated 25 February 2020) according to the principles of the latest version of the Declaration of Helsinki.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 restrictions. The authors would like to thank the study participants for their involvement in the study as well as the participating centres for their help in this study.The authors declare no conflict of interest.Flow chart of the study subjects. Mobile-Related Experience Questionnaire (CERM) = questionnaire about experiences related to the mobile phone; LMP = low use of mobile phone; MMP = medium use of the mobile phone; HMP = high use of the mobile phone.Results of physical fitness test. LMP = low use of the mobile phone; MMP = medium use of the mobile phone; HMP = high use of the mobile phone. * significant differences between groups; significant criteria p < 0.05.Results of academic performance. LMP = low use of the mobile phone; MMP = medium use of the mobile phone; HMP = high use of the mobile phone. * significant differences between groups; significant criteria p < 0.05.Results of physical variables.Data are expressed as means ± SD; LMP = low use of the mobile phone; MMP = medium use of the mobile phone; HMP = high use of the mobile phone; 95% CI = 95% confidence intervals for the difference; ES = effect size; a significant differences from LMP group; b significant differences from MMP group; † significant differences between boys and girls; significant criteria p < 0.05.Results of academic variables.Data are expressed as means ± SD; LMP = low use of the mobile phone; MMP = medium use of the mobile phone; HMP = high use of the mobile phone; 95%-CI = 95% confidence intervals for the difference; ES = effect size; a significant differences from LMP group; b significant differences from MMP group; † significant differences between boys and girls; significant criteria p < 0.05.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01043.txt

@@ -0,0 +1,7 @@
+Acquired Immune Deficiency Syndrome (AIDS) has become one of the most severe public health issues and nowadays around 38 million people are living with the human immunodeficiency virus (HIV). Ensuring healthy lives and promoting well-being is one of 17 United Nations Sustainable Development Goals. Here, we used the Markov chain matrix and geospatial clustering to comprehensively quantify the trends of the AIDS epidemic at the provincial administrate level in the mainland of China from 2005 to 2017. The Geographically Weighted Regression (GWR) model was further adopted to explore four groups of potential influencing factors (i.e., economy, traffic and transportation, medical care, and education) of the AIDS incidence rate in 2017 and their spatially distributed patterns. Results showed that the AIDS prevalence in southeastern China had been dominant and become prevalent in the past decade. The AIDS intensity level had been increasing between 2008 and 2011 but been gradually decreasing afterward. The analysis of the Markov chain matrix indicated that the AIDS epidemic has been generally in control on the Chinese mainland. The economic development was closely related to the rate of AIDS incidence on the Chinese mainland. The GWR result further suggested that medical care and the education effects on AIDS incidence rate can vary with different regions, but significant conclusions cannot be directly demonstrated. Our findings contribute an analytical framework of understanding AIDS epidemic trends and spatial variability of potential underlying factors throughout a complex extent to customize scientific prevention.Human immunodeficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) has become one of the most severe public health issues since the first cases of AIDS were identified in 1981 [1]. At the end of 2019, approximately 38.0 million (31.6 million–44.5 million) people are living with HIV [2]. Due to gaps in HIV services, about 32.7 million (24.8 million–42.2 million) people have died from AIDS-related illnesses [2]. Ensuring healthy lives and promoting well-being for all ages were addressed by the United Nations as one of 17 Sustainable Development Goals in the 2030 Agenda [3]. Currently, the research on AIDS is mostly focused on epidemiology, and comparative analysis between different social characteristics and geographical populations has been done by epidemiologists [4]. Since the generation, transmission, and distribution of AIDS are closely related to geospatial information [5], tabular records could lose critical information for vaccination tracking and surveillance due to the neglect of the spatial and geographic attributes related to AIDS cases.As a branch of epidemiology, spatial epidemiology has developed rapidly in recent years. With spatial statistical methods, spatial epidemiology improves the understanding of disease from qualitative level to quantitative level, by analyzing geographically indexed health data related to demography, environment, behavior, socio-economy, genetics, etc. [6,7]. Supported by geographic information science (GIS) and spatial analysis technology, spatial epidemiology describes and analyzes the spatial distribution characteristics and development pattern of epidemiological health events. With a vivid summary and precise pattern of areas at apparently high risk [8], the related disease maps can be used for various descriptive purposes, including aid policy formation and resource allocation [6,9].Spatial analysis has become increasingly common in HIV/AIDS-related research, specifically for the geographic distribution assessment [10]. For instance, Bautista et al. [11] analyzed the spatial distribution of HIV infection among U.S. civilian applicants for US military service from 1985 to 2003, and found the difference in HIV spatial aggregation between white people and African Americans. Focusing on HIV infection rate in rural areas of South Africa, Tanser et al. provided clear empirical evidence for the localized clustering of HIV infection [12]. The study detected considerable geographical variation in local HIV prevalence within relatively homogenous population, i.e., population do not have much geographical socioeconomic variation. In a recent study of the prevalence of HIV in sub-Saharan Africa, Dwyer-Lindgren et al. [13] revealed considerable within-country variation and local differences in the change direction and rate in HIV prevalence from 2000 to 2017, which brought attention to the subnational variation in HIV prevalence and the related fine-scale estimation and interventions. In China, trend surface analysis and spatial autocorrelation analysis on AIDS have been conducted for a single province or city, including Yunnan [14] and Lanzhou [15]. However, most publications have merely addressed characteristics of the HIV infection at the regional level.For truly controlling HIV and directly reducing the number of new infections, an understanding of the spatial distribution needs to be combined with an investigation of potential influencing factors. Smith et al. [16] found that education level is an important factor in AIDS risk in rural Uganda. Liu et al. [17] studied the spatial distribution and influencing factors of HIV-infected people in Henan Province, and found that economic level and medical level will have a certain impact on the incidence of AIDS. Most recently, Yang and Li [18] found that the economy, medical treatment, and transportation (i.e., population flow) have an impact on the AIDS incidence in China. Conventional statistics (e.g., ordinary least squares (OLS)) are the primary tools to examine the impacts of underlying factors. However, the prominent limitation of conventional statistics in geoscience is the incapability in capturing the spatial non-stationarity, i.e., spatially varying relationships between dependent and independent variables [19,20]. Moreover, OLS has been shown to be of limited utility when spatial data are coupled with highly correlated independent variables [21,22,23]. An alternative to conventional statistics is Geographically Weighted Regression (GWR).The discipline of the spatial and temporal analysis of the AIDS incidence and its influencing factors are not uniform. So far, the research mainly focuses on the analysis of AIDS spatial distribution, especially in South Africa [9,12,13,24,25], At present, the related research on AIDS in China mainly focuses on the analysis of epidemiological characteristics [26], prevention and control [27,28], the epidemic situation of the specific population [29,30], etiology and sociology [18,31], however, there are few studies on the spatial-temporal change analysis and influencing factors analysis. The long time series study on AIDS incidence, the future trend, and factors affecting AIDS spread in China have not yet explored.To fill the gap, this study collected the AIDS incidence data on Chinese mainland from 2005 to 2017 to study the spatial distribution characteristics and influencing factors of the AIDS epidemic, so as to provide a theoretical basis for policymakers to customize scientific prevention and control policies. The study focused on AIDS incidence, i.e., the body’s immune system is badly damaged because of the HIV virus. The paper is organized as follows. After the description of the research data and sources, Section 2 presents the mean research methods, including AIDS intensity level mapping (Section 2.1), Markov chain matrix analysis (Section 2.2), Spatial clustering character (Section 2.3), and GWR analysis (Section 2.4). The related results are followed in Section 3. Section 4 discusses key issues of the AIDS intensity on the Chinese mainland, and the findings based on the analytical method are also summarized. Finally, short conclusions are provided in Section 5.AIDS incidence data was acquired from the Data Center of China Public Health Science (www.phsciencedata.cn). The AIDS prevalence in China for the period 2005–2017 was investigated, considering the data resource consistency and the fact that AIDS was drawn attention by the Chinese government around the year 2005 [32]. The personal information of the individual participant had already been taken away before the release for privacy protection. The study area includes 31 provincial administrative units on the Chinese mainland.The selection of the potential underlying factors was guided by expert knowledge from wide studies and publications on the epidemic and infectious diseases. Considering the variables type and their relationship with the AIDS infection, the explanatory factors were classified into four categories: economy, traffic and transportation, medical care, and education. First, in most rural areas of China, the economic level is widely regarded as an important factor affecting people’s medical treatment and medical examination affordability. In addition, population flow [33,34] and drug trafficking [35] are important factors in the spread of AIDS and HIV diagnosis and prevention. In this sense, we chose passenger volume, highway length, overall passenger flow, and vehicle ownerships as indicators to measure the level of traffic development. Third, medical and health resources are an important means to deal with epidemic diseases, and differences in access to health care, treatment modalities, and outcomes among different ethnic, and ethnic groups have been confirmed [9]. Finally, education is a notable factor of the spread of epidemiology [16,36,37], especially in rural areas. Meanwhile, residents with lower education have a lower chance to receive diagnostic and medical checks without awareness and required knowledge [38]. The above-mentioned influencing factors are all from China Statistical Yearbook, as summarized in Table A1 in Appendix A.To understand the AIDS intensity level for individual provincial administrative units, the AIDS rate for each year were first calculated:(1)AIDS_Rateij=The number of AIDS patients in the administrative unit i for the year j Population in the administrative unit  i for the year j 
+        where i refers to the provincial administrative unit ID (ranging from 1 to 31 modified from the national standard geocodes at the provincial level), and j refers to the study year with a range from 2005 to 2017.To make visual comparisons in terms of AIDS intensity among individual provincial administrative units and between different study time points, we further calculated the distribution frequency of AIDS intensity for each individual provincial administrative unit:(2)AIDS_Intensityij=AIDS_Rateij∑i=131AIDS_Rateij/31
+        where ∑i=131AIDS_Rateij/31 is overall AIDS incidence intensity for year j for the entire Chinese mainland. AIDS_Intensityij refers to AIDS intensity for individual provincial administrative unit i for year j. An AIDS_Intensityij value of 1 represents an average intensity of the administrative unit i to the overall AIDS intensity for the Chinese mainland. Based on it, the AIDS intensity levels for the administrative unit i in year j were further categorized (Table 1).A Markov chain is defined as a process with a limited number of states with the Markovian property and some transition probabilities [36]. A Markov chain can either remain the same or transit to the other state for a time period in a Markov chain process. The length of a Markov chain is the segmentation times between two successive observations. For our AIDS intensity study, there are four states (i.e., intensity levels) for each administrative unit. In other words, one of the four values (1, 2, 3, and 4) can be assigned to the variable of AIDS intensity level in the series at each time t, corresponding to a different state. Given that the transition between two successive states only depends on the AIDS intensity level at the starting time, a first-order time dependence 4-state Markov chain was constructed (Table 2).There are 4 × 4 transition probabilities (mij) for our Markov chain illustrated in the Table 2. The diagonal of the matrix (mij, I = j) is a smooth transition, i.e., the state of AIDS intensity level is stable for one Markov chain process. The transition probabilities in the upper-right corner (mij, I < j) refers to an upward transition, and the transition probabilities in the lower-left corner (mij, I > j) refers to a downward transition. If the state of AIDS intensity level is stable for one Markov chain process, we assume that the AIDS intensity for the corresponding province was in control. An upward transition for the AIDS intensity level indicates that the AIDS epidemic situation was serious and a download transition for the AIDS intensity level indicates that the AIDS epidemic situation was reduced. The calculation formula of each element mij in the matrix is:(3)mij=nijNi
+        where mij represents the probability of a type i AIDS intensity level (i.e., code number) transforming to a type j AIDS intensity level; nij is the sum of the number of units converted from type i into type j for all the segmentation times during one Markov chain study period; and Ni is the sum of the number of units belonging to type i for the segmentation times during one Markov chain study period.In this study, we assume that the spatial patterns of AIDS were not randomly distributed, since AIDS infections are related to the interactions between cases in different places. In order to explore the spatial clustering and heterogeneity character of AIDS incidence rate for the entire Chinese mainland, the spatial autocorrelation analysis (Global Moran’s I) was conducted. Spatial autocorrelation can be used to describe the spatial difference between the spatial units and their adjacent units. Moran Index was proposed in 1950 [37] and have been intensively applied to determine spatial autocorrelation character. Here, the Global Moran Index measure for clustering (positive spatial autocorrelation) or dispersion (negative spatial autocorrelation) of the AIDS incidence. A Moran’s Index value near 1 indicates clustering, whereas a value near −1 indicates dispersion.As a measurement of spatial associations at local level, the Local Indicators of Spatial Autocorrelations (LISA) statistics allow us to detect AIDS clusters on the assumption that the local AIDS spatial pattern was not randomly distributed. Local Moran’s Index tests the agglomeration and difference chrematistics of AIDS incidence rates in different spatial locations. The Local Moran’s I is defined as:(4)Moran′s I=nS0*i∑in∑j=1nwij(xi−x¯)(xj−x¯)∑in(xi−x¯)2
+        where n is the number of the samples; xi is the attribute value of spatial unit i; x¯ is the mean value; wij is spatial weight; S0 is the sum of all elements of the spatial weight matrix. Local Moran’s I analyzes the correlation degree of spatial variables between the observed value and the adjacent space unit, and judges the hot spot area of the spatial object [39]. It ranges from −1 to 1, and the absolute value close to 1 means a high similarity between the tested unit and the adjacent unit (i.e., High-High or Low-Low aggregation) and an value close to 0 indicates a low correlation (i.e., High-Low or Low-High aggregation). In this sense, there are four kinds of combinations between the attribute value of spatial unit i and its neighborhood: High-High (H-H), High-Low (H-L), Low-High (L-H), and Low-Low (L-L) aggregation [40]. Local Moran’s I scatter plots for AIDS incidence in different study point were obtained using the GeoDa software environment [41].Here, AIDS incidence clusters were defined as geographic areas in which AIDS prevalence was disproportionately higher compared to neighboring areas. Hot-cold Spot Analysis (Getis-Ord Gi * statistics) [42] was adopted using the Spatial Analysis toolset in the ArcGIS environment.Unlike OLS, GWR is a spatial regression method to model spatial variation in the relationship between dependent and independent variables and estimates the influence degree [43]. By calculating the local parameters of the regression model, the spatial non-stationarity of each parameter in different spatial ranges can be revealed [44]. The model formula is as follows:(5)yi=β0(ui,vi)+∑k=1pβk(ui,vi)xik+εi
+        where: yi is the dependent variable for the location i; β0 is the intercept, (ui,vi) is the coordinate of the location i, β0(ui,vi) is the constant term; βk(ui,vi) is the coefficient of the kth independent variable of sampling point i. Instead of remaining the same everywhere, βk(ui,vi) varies in relation to location i, xik is the kth independent variable at location i, εi is the random error term at location i.Prior to modeling, correlation analyses were conducted among the potential explanatory variables (Table A1 in Appendix A) to assess the multicollinearity. The correlation analyses indicated that there was a strong multicollinearity issue for variables in the same group. Principal component analysis (PCA) was adopted to eliminate the multicollinearity issue. PCA transforms the group of the potential explanatory variables that may have a correlation into a group of linearly uncorrelated variables by orthogonal transformation [45]. Through data transformation and processing, the potential influencing factors of the AIDS incidence rate can be grouped into less integrated factors, which not only maintains the main information of original factors but also avoid the complexity of the correlation among them.The overall change trend AIDS incidence rate at the national level for the mainland of China from 2005 to 2017 can be divided into four stages (Figure 1). The first stage (2005–2008) witnessed a stable growth, with an average AIDS incidence rate of 0.06 per ten thousand persons (0.06/10000). In comparison, the AIDS incidence rate showed a slight increasingly tendency in the second stage (2008–2011), with an average AIDS incidence rate of 0.11 per ten thousand persons (0.11/10000). During the third stage (2011–2014), the AIDS incidence rate exhibited a remarkable fluctuating growth, with an annual growth of 105% from 2011 to 2012 as well as a steady stable growth, with an annual growth of 7.67% from 2012 to 2014. After 2014, the AIDS incidence rate indicated a stable growth again, with an average AIDS incidence rate of 0.38 per ten thousand persons (0.38/10000) and annual growth rate of 11.5%.The spatial distribution of AIDS incidence rates at the provincial scale was first visualized intuitively in ArcGIS10.2 software (ERSI, Redlands, USA, Figure 2 and Figure 3). The AIDS incidence rate had been increasing for most of the provincial units on the Chinese mainland (Figure 2). Apparently, the differences in AIDS incidence rates were very high during the entire study period, with high incidence rates in southwestern China, including Yunnan, Guangxi, Guizhou, Sichuan and Chongqing, and Xinjiang in northeastern China. Meanwhile, there were considerable differences among these provincial administrative units in the AIDS incidence rate from 2014 to 2017, and the AIDS incidence rate showed a decreasing tendency for Guangxi and Yunnan. In contrast, the AIDS incidence rates were relatively low, with a generally stable increase during the study period in central and eastern China during the study periods.The AIDS intensity levels for the individual provincial administrative unit for the above time points were further calculated (Table A2 in the Appendix A). The AIDS intensity level charts in 2005, 2008, 2011, 2014, and 2017 showed the difference in the spatial distribution and temporal variation (Figure 3). Different from other periods, the high AIDS intensity levels were relatively scattered in space in 2005. Yunnan, Guangxi, Beijing, Xinjiang, and three adjacent provinces in central China showed high AIDS intensity levels in 2005. From 2008 to 2017, the high incidence areas gradually monopolized in the southwestern region, and the spatial agglomeration was enhanced. The spatial pattern of AIDS intensity levels changed greatly from 2005 to 2008. The AIDS high-value zone expanded to Sichuan and Chongqing and the pattern of AIDS low-level intensity areas remained relatively stable from 2008 to 2011. There is a note that the AIDS intensity level changed from high (code 1) to the middle and low level (code 3) for Henan from 2011 to 2014. The spatial pattern of AIDS high-level intensity areas remained relatively stable from 2014 to 2017.Spatial Markov matrixes were calculated for the respective four stages to compare the AIDS intensity level evolution character classified above (Table 3). Notably, smooth transition probabilities of low-low (m11) and high-high (m44) AIDS intensity levels were above 0.86, and the highest was 1 for the four stages, which were far higher than the average level of other transition probability. It is also noted that the transition probabilities between adjacent levels were much greater than that between cross levels, indicating that the temporal evolution of AIDS intensity are gradual rather than jumping. The transition probabilities related to medium-high levels and medium-low levels transiting to other levels were relatively unstable, ranging from 0.17 to 1.In terms of the AIDS intensity evolution characteristics between different stages, the overall downward transition probabilities between 2005 and 2008 were higher than the upward transition probabilities, indicating that the AIDS intensity level gradually decreased; the upward transfer probabilities between 2008 and 2011 were greater than the downward transfer probabilities, indicating that AIDS risk had been gradually increasing. In the next two stages, the upward transition probabilities had been decreasing from 0.3 to 0.03 and the downward transition probabilities had been decreasing from 0.4 to 0.01, indicating that AIDS risk had been gradually decreasing.Table A3 in Appendix A provides the Global Moran’s I index statistical result for the AIDS incidence rates. In the year 2005, 2006, and 2007, the returned p-value is not statistically significant, and the global spatial autocorrelation characteristics of the AIDS incidence rates was not detected. Started from 2009, p values were all <0.05 (passed through the 95% confidence test) and z scores are positive (>1.65, the threshold set by the null hypothesis being rejected), indicating that the AIDS incidence rates had become spatially clustered. Especially, between 2011 and 2017, the global Moran’s I index raised rapidly from −0.114 to 0.312, indicating a growing spatial clustering tendency of the AIDS incidence rate for the enter study area.To further measure the spatial correlation and difference of AIDS incidence rates for each provincial unit and surrounding units, we used the GeoDa software to calculate Moran’s I scatter plots diagrams for five representative time points, as mentioned in Section 3.3. Figure 4 shows that there had been an increasing tendency for Moran’s I of AIDS incidence (0.15, 0.277, 0.266, 0.35, and 0.395) for the five-time points. It is also noted that the scatter points of the five nodes are basically concentrated in the first quadrant, namely H-H aggregation, indicating that there was an obvious spatial aggregation phenomenon in the AIDS high incidence areas.The Getis-Ord Gi * statistics results at the provincial level were interpreted in the ArcGIS environment. Similar to the interpretation of the Global Moran’s I statistics, the resultant Z score, and p-value for each unit determined the high or low values. Based on that, the spatial patterns of the Hot-Cold Spots were illustrated in Figure 5. The spatial agglomeration of hot-cold spots had been gradually enhanced, which was consistent with the global Moran’s index results above. In 2008 and 2011, the Hot Spots of AIDS incidence rates were mainly located at Yunnan, Guangxi, Guizhou, Chongqing, and Sichuan, which were also the Hot Spots in the following two notes with a higher statistical confidence level. Increasingly, starting from 2011, Shandong became the cold spots of AIDS incidence rates with a higher statistical confidence level compared to the sounding provinces, and this cold spot had been expanding continuously with a strong spatial aggregation characteristic. In 2017, the cold spot areas with a 99% confidence level were Shandong, Hebei, Beijing, and Tianjin, and the cold spots also extended to Inner Mongolia, Liaoning, Shaanxi, and Jiangsu with 95% confidence level.For the variables in four groups (i.e., economy, traffic and transportation, medical treatment and education), a comprehensive variable was extracted as the influencing factor of each group of potential explanatory variables. The eigenvalues were 3.79, 1.76, 4.52 and 2.73 respectively, and coefficients were 2.00, 1.46, 2.24 and 1.73 respectively (Table A4 and Table A5 in Appendix A). Taking the AIDS incidence rate in 2017 as the dependent variable, the GWR analysis was conducted on the variables of economy, transportation, medical care, and education level. The distribution of the variables and the homogeneity issue were checked and verified before the regression. The fixed spatial core was selected as Kernel type, and the corrected Akaike Information Criterion (AICc) was selected as the Bandwidth method. The AICc difference of GWR (i.e., 31.2) is much lower than that of the global model (i.e., 33.7), which indicates that GWR provides a better fit to the observed data. Figure 6 and Figure A1 in the appendix showed the overall spatial variation of the GWR coefficients and their standard errors, respectively. For the sake of comparability and interpretation, the influencing variables for each observation were normalized to [0, 1].For the four variables from different perspectives of socio-economy, the standard errors of coefficients for medical care and education were much higher than those of the other factors. The ratios of the corresponding coefficient value and the standard error were higher than 2.5 for most of the observations (Figure A1 in Appendix B and Figure 6), which indicated that these two variables were not significantly related to the AIDS incidence rate. Conversely, the economy had a generally negative effect on the AIDS incidence rate, based on the coefficient values and their standard error statistics. In comparison, the regression coefficients for traffic and transportation were generally positive and higher than the absolute value of the economy, that is, it had a stronger positive effect on the AIDS incidence rate.The economy had a negative effect on the AIDS incidence rate for most of the areas in the mainland of China. Relatively absolute higher estimated coefficients were observed in western China (<−0.1) and especially southwestern China (<−0.2). It is noted that the economy level is relatively low in these regions, however, the degree of impacts was relatively large, and attention should also be paid to these areas. It indicated that a lower level of economic development would lead to a decrease in the capacity to prevent and control the AIDS epidemic. An unexpected finding of a slightly positive effect of the economy on the AIDS incidence rate was exhibited in Northeast China. Since the coefficients were near 0, and they were much lower than the coefficients in other regions, the effect of the economy on the AIDS incidence rate is hard to predict. GWR also revealed a strong spatial heterogeneity in the relationships of traffic and transportation and AIDS incidence rate. The more developed, the more the AIDS spread prevalent. The effects of traffic and transportation on the AIDS incidence rate were positive and the coefficient decreased from the west to the east, indicating that the western region is greatly influenced by the traffic and transportation development.Our results provide a comprehensive quantification of subnational AIDS epidemic trends on the Chinese mainland with geospatial analytical methods. Our study first addressed the spatial variations and time evaluations in AIDS incidence rate and intensity levels. Around the year 2004, the Chinese government and local authorities carried out a series of HIV infection checks and blood tests, to intervene in further HIV transmission through blood [7]. Thus, the AIDS incidence rate had been relatively stable after 2008. The information on AIDS intensity levels “hot spots” of AIDS incidence rate throughout a relatively long period can be used to identify the high risks areas to target primary prevention strategies. For instance, Yunnan is located in the “Golden Triangle” region of Southeast Asia, where drug trafficking is rampant [46]. Our finding revealed that the AIDS prevalence in this region had been dominant and become prevalent in the past decade, which is consistent with the reports that the distribution of HIV/AIDS infection in China is evolving from border regions nearby the “Golden Triangle” to inland areas [47].Finding on the Markov chain matrix analysis indicated that the high AIDS intensity level area and low AIDS intensity level area tend to maintain the original state in the next period. In other words, these areas had a higher probability of maintaining a high-intensity level in the next coming years and vice versa. Besides, the transfer probability between adjacent two levels is greater than that of cross-level, indicating the space-time transition and evolution had been gradual rather than jumping. Our spatial agglomeration study also demonstrated that a large proportion of AIDS patients had been concentrated in a small number of administrative units with strong spatial correlation. Our analysis highlights this favorable situation to efficiently target resources and interventions for bringing AIDS and HIV infection under control on the Chinese mainland.Overall, the analysis of GWR revealed that the economy was most closely related to the AIDS incidence rate on the Chinese mainland. First, the economy was shown to be an essential and foremost factor that influences the AIDS incidence pattern. The regression coefficients of economic development were negative except for northeast China, which is consistent with the previous finding related to the spatial distribution of AIDS and main socio-economic driving factors in China [48]. In southwestern China, the economic and education level are relatively low, whereas the sexual transaction and drug abuse situation are severe, thus, more attention should be paid on AIDS and its prevention strategies. In some respects, the situation in the southwestern China confirmed with previous broadly worldwide finding that poverty exacerbates the AIDS/HIV risk [47,49]. The difference in influencing mechanism may also dependent on the cultural backgrounds. Early marriage and childbearing can be another concern for primary AIDS prevention for areas where ethnic minorities gather with a unique culture. Minorities in concentrated communities have a relatively open and tolerant attitude towards sexual behavior and a relatively low awareness of prevention, which increases the possibility of HIV infection. Together with the finding on the spatial clustering tendency for Sichuan, Guizhou, Yunnan, and Guangxi provinces, this area is potentially in need of diagnosis and treatment services.In terms of the traffic and transportation, the effect was more prevalent in western China than in eastern China, as evidenced by the higher estimated coefficients. The traffic and transportation in this study generally refers to population flow throughout the transportation networks. Even through the transportation by itself do not have a cause on AIDS, since the HIV transmission is different from respiratory epidemics. However, the GWR analysis indicated that the in transnational migrants accompanied by population flow due to trade cooperation and travel, and drug trafficking have a cause on AIDS. Previous study indicated that more than 60% of foreign AIDS cases travel to and from the border between Yunnan and Myanmar, which has also contributed to the spread of AIDS in the region [50]. The frequent cross-border exchanges and interactions is also the cause of the AIDS epidemic in Xinjiang, which is adjacent to the “Golden Crescent”, a drug producing area in Central Asia [51]. There are many drug users in this area and drug users share infected syringes. The population flow due to labor export from rural areas to cities in Sichuan, Chongqing, and Guizhou lead to dangerous sexual behaviors such as multi-sex couples and sex transactions is another susceptible factor for HIV transmission [52]. The evolution of the AIDS incidence levels (Figure 3) and the spatial patterns of AIDS incidence hot spots (Figure 5) might indicate the movement of migrants and an expanded transportation network across the country.Medical care and education level have been widely identified as important determinants of AIDS prevention at different scales [9,16,36,38]. However, our spatial regression analysis did not indicate a significant correlation at the provincial level. At the local scale, Liu et al. [17] showed that the increase in the number of medical institutions per capita can reduce the HIV infection rate in Henan Province. The improvement of medical care and monitoring systems can enhance the surveillance capability for spatial epidemiology [6]. A recent study of the HIV epidemic among men who have sex with men in China also showed that the number of medical institutions positively correlated with the number of reported HIV/AIDS cases depended on spatial regression at the county level [53].Previous studies also found that high-educated individuals should be targeted for AIDS prevention [16] and that the individuals with a college degree or above showed a high AIDS incidence rate [53]. Our study showed that the education effects on AIDS incidence rate varied with different regions and significant conclusions cannot be directly demonstrated. In this study, four groups of variables (i.e., economy, traffic and transportation, medical care, and education) were explored as macro comprehensive socio-economic indicators, however, the basic characteristics and information closely related to AIDS incidence may be ignored or missed. The examination of causal relationships of the related factors is needed, since the original selected variables can affect each other [15]. Meanwhile, the overall pattern at the provincial administrative level was estimated, whereas the variations at the local level tend to be masked and the scale issue is an indispensable limitation of this study. The effects of religious beliefs, sex education at high schools can be important factors to intervene HIV transmission. The fine-scale spatiotemporal investigation of AIDS prevalence based on GIS and spatial analysis can be a fundamental method for precise AIDS interventions. Further, we just examined the historical AIDS trends at the provincial administrate level, due to the privacy protection of the AIDS incidence data at local scale. Further investigation of the traffic and transportation for targeted areas may help to figure out the spread of the AIDS in the future.In this study, we used a spatially statistical method to explore the spatial distribution of the AIDS epidemic and its influencing factors on the mainland of China from 2005 to 2017. This study first provided a comprehensive quantification of AIDS epidemic trends at the provincial administrate level. According to the analyses of the AIDS intensity level mapping and the spatial clustering, we found that the prevalence of AIDS in southeastern China has remained high in the past decade. The period between 2008 and 2011 witnessed the apparent increasingly severe tendency of the AIDS incidence level, which gradually decreased afterward. Markov chain matrix analysis indicated that AIDS intensity level transfer probability between adjacent two levels was greater than that of cross-level, suggesting the AIDS epidemic has been generally in control on the Chinese mainland. In terms of influencing factors of the AIDS incidence, the economic development closely related to the AIDS incidence rate on the Chinese mainland. The GWR result further showed that the effects of medical care and education on the AIDS incidence rate differed from different regions. Our study contributes to an analytical framework of understanding AIDS epidemic trends and spatial variability of the underlying factors throughout a complex area to customize scientific prevention. In future research, the fine-scale spatiotemporal investigation of AIDS prevalence based on GIS and information related to the spatial variability of relationships can be a fundamental method to precise AIDS interventions.Conceptualization, Y.W. and C.Z.; methodology, Y.W. and C.Z.; software, Y.W. and C.Z.; validation, D.G., Z.L. and C.Z.; formal analysis, Y.W.; investigation, D.G., Z.L., and C.Z.; resources, Y.W., D.G., Z.L., and C.Z.; data curation, D.G. and C.Z.; writing—original draft preparation, Y.W.; writing—review and editing, D.G., Z.L., and C.Z.; visualization, Y.W.; supervision, D.G. and C.Z.; project administration, D.G.; funding acquisition, Z.L. and D.G. All authors have read and agreed to the published version of the manuscript.This research was funded by the National Natural Science Foundation of China (31901157), and the China Postdoctoral Science Foundation (2020T130088).Not applicable.Not applicable.We did not upload statistical data but they are available to any author who requests them.The authors declare no conflict of interest.Description of a synthesis of potential influencing factors of incidence rates for the mainland of China in 2017.AIDS intensity levels at the individual provincial administrative unit scale for the Chinese mainland for the year 2005, 2008, 2014, and 2017.The interpretation of Global Moran’s I index statistical result for the AIDS incidence rates in the entire Chinese mainland.Note: The Null hypothesis is that the AIDS incidence rates were randomly distributed in the entire Chinese mainland for the respective study time points.Results of principal component analysis.Note: PC-loadings refers to principal components (PC) loadings. Score coefficient refers to the degree that the case deviates from the mean of all samples.Eigenvalues, contribution rates and coefficients of the four principal components.PC_Eco refers to the selected principal component of the economy, and so on. Eigenvalue is the contribution of the corresponding eigenvector to the entire matrix, and it is generally greater than 1. The cumulative variance contribution rate is the sum of the variance contribution rates of all principal components, which is generally higher than 80%. Since only one principal component was selected in this study, the variance contribution rate was required to be higher than 80%. The PC_Coefficient refers to the eigenvector of the respective principal component, and the expression is:(A1)     Coefficienti=∑ PC_Loadingj/Eigenvaluei
+        where i represents different aspects of economy, traffic and transportation, medical care and education, and j represents the individual variables in each aspect (Table A4).The overall spatial variation of the standard errors for regression coefficients.Average AIDS incidence rate trend in the mainland of China during 2005–2017.The spatial distribution of AIDS incidence rate on the mainland of China in the year 2005, 2008, 2011, 2014, and 2017.The spatial differences and evolution of AIDS intensity levels in the mainland of China from 2005 to 2017.Moran’s I scatter plot of AIDS incidence rates on the Chinese mainland for different time points (2005, 2008, 2011, 2014, and 2017). The trend line passes through the first and third quadrants, so shows high-high aggregation and low-low aggregation.The spatial patterns of the Hot-Cold Spots of AIDS incidence rates in 2005, 2008, 2011, 2014, and 2017.Spatial distribution of regression coefficients: economy, traffic and transportation, medical care and education.The classification scheme for AIDS intensity level for the individual provincial administrative unit.Note: DFij is the distribution frequency of AIDS intensity for individual provincial administrative unit i for year j.An example of a Markov transition-probability matrix for the types of AIDS intensity level in a Markov chain process.Note: mij (1 ≤ I ≤ 4; 1 ≤ j ≤ 4) represents the probability of a type -i AIDS Intensity Level (code) transforming to a type-j AIDS Intensity Level (code) during one study period.Spatial Markov matrix (probability) of AIDS intensity level in the four stages: 2005–2008, 2008–2011, 2011–2014, and 2014–2018.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01044.txt

@@ -0,0 +1 @@
+A risk of cardiovascular disease (CVD) is increased by multiple factors including psychosocial stress and health behaviors. Sexual minority men who identify as Bears form a subculture distinguished by characteristics associated with increased CVD risk such as elevated stress and high body weight. However, none of the previous studies comprehensively investigated CVD risk in this population. Our study compared Bears (N = 31) with other gay men (N = 105) across a wide range of CVD risk factors. Logistic regression and analysis of covariance (ANCOVA) models were performed to compare both groups concerning behavioral (e.g., physical activity), medical (e.g., self-reported hypertension), and psychosocial (e.g., depressiveness) CVD risk factors. Bears were characterized by older age and higher body mass index (BMI) than the control group. We also observed higher resilience, self-esteem, as well as greater prevalence of self-reported hypertension, diabetes, and hypercholesterolemia in Bears. None of these differences remained statistically significant after adjusting for age and, in the case of self-reported diagnosis of diabetes, both age and BMI. Our study demonstrates that Bears are characterized by increased CVD risk associated predominantly with older age and higher BMI. Health promotion interventions addressed to this community should be tailored to Bears’ subcultural norms and should encourage a healthier lifestyle instead of weight loss.Cardiovascular disease (CVD) is the leading cause of death worldwide and annually accounts for most deaths associated with noncommunicable diseases [1]. Among Polish men, only in 2014, combined ischemic heart disease, stroke, and other CVDs accounted for as much as 2723 potential years of life lost (PYLL) per every 100,000 men aged 0 to 74 years [2]. It has been estimated that approximately 80% of premature heart disease and stroke is preventable [3].Risk factors for cardiovascular disease include both behavioral characteristics such as dietary habits, limited physical activity, tobacco use or alcohol consumption, and medical indicators such as high systolic blood pressure, high body mass index, high fasting plasma glucose, and high total cholesterol [2]. Another group of factors that has been studied extensively in the context of CVD is exposure to stress and related mental health adversities such as depressiveness [4]. Chronic stress has been linked in research to vascular pathology (e.g., hypertension), pathogenic changes of the metabolic function (e.g., dyslipidemia), and immune function (e.g., chronic low-grade inflammation), which are the key factors in the development of cardiovascular pathology [5,6]. Psychosocial factors combined with smoking, abnormal lipids, diabetes, hypertension, abdominal obesity, irregular consumption of fruits and vegetables, alcohol intake, and limited physical activity contribute to more than 90% of the CVD risk [4].Both the relative importance and the distribution of the CVD risk factors vary across different populations, and studies demonstrate that some subgroups may be disproportionately burdened with CVD risk and outcomes compared to others [2,4]. One of such populations characterized by elevated levels of various CVD risk factors is sexual minority men [5]. Compared to heterosexual men, sexual minority men are characterized by higher rates of tobacco use [7,8], excessive alcohol use [9,10], and increased exposure to chronic stress associated with prejudice and discrimination due to minority sexuality [11,12]. A recent systematic review of studies exploring determinants of cardiovascular health in various minority groups demonstrated significant relationships between cardiovascular health indicators and perceived discrimination [13]. Sexual minority men are also disproportionately burdened with health inequalities including a greater prevalence of mental health adversities such as depression and anxiety [14,15], and physical health problems such as cancer or hepatitis [16].However, the population of sexual minority men is far from homogenous and includes various subgroups and subcultures, which are distinguished by a unique constellation of practices and preferences important for general and cardiovascular health [17,18,19]. One such subculture includes sexual minority men who identify as Bears. Although the Bear subculture originated in the US approximately four decades ago, it has spread all over the world, and organizations of men who identify as Bears exist also in Poland [19]. According to the literature, one in every five sexual minority men may identify as a member of the Bear community [17].Men who identify as Bears are distinguished by their preferences concerning the look of the male body. This includes a larger, heavyset build which combined with more pronounced body and facial hair is regarded as sexually attractive and indicative of masculinity [20,21,22]. The Bear community is more accepting of individuals who do not conform to the mainstream, stereotypical images of gay men—on the contrary, it positively celebrates larger bodies and challenges fat-phobic social norms [23]. A few studies that have been focused on health outcomes and determinants in this population reveal a complex mosaic of factors that potentially shape the health of Bears. They point at both increased and decreased cardiovascular risk among Bears relative to other sexual minority men.Perhaps the most consistently reported characteristic of Bears, which has been associated in the literature with CVD risk, is increased body weight and related increased body mass index [19,24,25]. Significantly increased body mass index (BMI) values relative to other sexual minority men and/or mean BMI values in a general population indicative of obesity have been observed among both older and younger men from this community (the latter are called Cubs) and Bears living in various sociocultural contexts (i.e., Chinese, Australian, and Polish Bears) [19,24,25]. This disparity is also related to another source of increased CVD risk, which is greater exposure to psychosocial stress. In the case of Bears, greater psychosocial stress, as compared to other sexual minority men, results from exposure to not only sexual minority stigma but also weight stigma, which both may contribute to decreased well-being and increased prevalence of mental health adversities in this population [19,21,23,26]. For example, in a study by Mijas et al. [19], both exposure to weight stigma and sexual minority stigma negatively predicted self-esteem among men who self-identified as Bears.Interestingly, among Bear-identified men, higher BMI positively predicted self-esteem, which suggests that claiming Bear identity may inspire a change in self-perception and resistance toward dominant cultural body ideals [19]. Qualitative studies also indicate that Bears greatly benefit from becoming a part of an accepting and supporting community, which contributes to increased self-acceptance, enhanced sense of belonging, and strengthened individual resilience [20,26]. This interpretation is supported by studies that failed to observe significant differences between Bears and other sexual minority men concerning depressiveness or social phobia [27] as well as the frequency of being treated for depression or anxiety [25]. Despite being exposed to multiple stigmas, Bears do not seem to suffer from diminished mental health relative to other sexual minority men. Therefore, psychosocial CVD risk factors among Bears compared to other sexual minority men require more research.Several studies focused on comparing Bears and other sexual minority men concerning such health behaviors as smoking tobacco, alcohol use, and drug use [25,27,28]. Although men from this subculture may be more likely to use illicit substances before or during sexual contacts [27], they do not smoke tobacco more often than other sexual minority men [28] or even smoke less often compared to men from other subcultural groupings within the gay community [25]. Although there is some evidence that social gatherings for Bears are usually associated with increased consumption of alcohol [29], quantitative studies did not demonstrate significantly increased frequencies of excess or binge drinking among Bears [27,28].None of the previous studies explored quantitatively dietary habits or patterns of physical activity in this population. Only one qualitative study so far has investigated the ways through which Bears navigate their physical activity [23]. Participants in this study struggled with anticipated weight stigma, which caused them to avoid spaces associated with physical activity such as the gym or the beach. This suggests that Bears may be characterized by limited physical activity as compared to other sexual minority men. Similarly, the dietary habits of Bears have been explored by only one study, which was conducted among Brazilian Bears [29]. The study was focused on eating practices among Bears and revealed that certain food preferences (i.e., eating meat and drinking beer) were indicative of dominant subcultural definitions of masculinity and represented a strategy to build the ideal masculine body and avoid discrimination associated with minority sexuality [29]. This study also suggested that eating habits among Bears may increase CVD risk in this population and should be explored in further studies.Although observed differences between Bears and other sexual minority men suggest that medical CVD risk factors may also disproportionately burden this population, none of the previous studies investigated such indicators of CVD risk among Bears as the prevalence of hypertension, hypercholesterolemia, or diabetes. Our study aimed at filling this gap.In this study, we focused on comparing men identified as Bears with other gay men with regard to cardiovascular disease risk factors, including (i) psychosocial characteristics, such as exposure to stigma, self-esteem, resilience, and depressiveness; (ii) behavioral characteristics, such as dietary habits, physical activity, tobacco use, and patterns of alcohol consumption; and (iii) medical indicators, such as diagnoses of hypertension, diabetes, hypercholesterolemia, and body mass index.Based on the previous studies, we hypothesized that men who identified as Bears will be characterized by increased cardiovascular risk in case of most medical and behavioral factors, but not psychosocial ones.The study employed data from two research projects. The first one concerned health determinants among members of the Polish Bear community. The invitations for participation in the study were distributed among members and supporters of the Bears of Poland association through social media, mailing lists, and during various social gatherings addressed to members of the community. Participants’ questionnaire data and anthropological measurements were collected during research meetings that had been taking place from June to December 2017 in selected cities in Poland. The study was positively reviewed by the Jagiellonian University Bioethics Committee (122.6120.70.2017). The analyses were performed on the data obtained from 31 cisgender gay men who are members of the Bear community and who incorporated Bear subcultural identification into their identity.The second research project aimed at examining health determinants within the Polish LGBTQ community (lesbian, gay, bisexual, transgender, and queer community). The study was conducted through the Qualtrics internet platform. The invitations for participation were distributed using the snowball method, through emails, social media, and websites of non-governmental organizations supporting the LGBTQ community. Additional efforts were made to reach members of the LGBTQ community living in smaller towns and villages. The data had been collected between January and March 2018. The study gained a positive review from the Research Ethics Committee of the Institute of Psychology at Jagiellonian University (KE/02/052017). Data from 105 cisgender gay men, who did not identify as Bears, were used in the analyses.The analyses employed data obtained from a total of 136 cisgender gay men, including 31 men identifying as Bears and 105 gay men who do not identify as Bears (control group). The average age of the sample was 31.6 years (SD = 10 years). The youngest participant was 18 years old, and the oldest was 69 years old. Most of the participants (N = 103; 75.5%) reported a university education level and lived in cities with a population of over 500,000 inhabitants (N = 80; 58.8%). One in five participants was experiencing financial hardships (N = 26; 19.1%). A comparison of both groups of men in terms of demographic variables is shown in Table 1.In both studies, the demographic questionnaire included such information as year of birth, gender and sexual identity, the size of the city of residence, education level, as well as whether the participant’s income is sufficient to cover basic monthly expenses. The study also employed other questionnaire tools, including the Rosenberg Self-Esteem Scale (RSES [30]; in Polish adaptation by Łaguna et al., [31]), the Daily Heterosexist Experiences Questionnaire (DHEQ [32]; in Polish adaptation by Mijas and Koziara [33]), the Resilience Measurement Scale (Skala Pomiaru Prężności SPP-25) [34], the Center for Epidemiologic Studies Depression Scale–Revised (CESD-R [35]; in Polish adaptation by Koziara [36]), and the Dietary Habits and Nutrition Beliefs Questionnaire (KomPAN [37]).The Rosenberg Self-Esteem Scale (RSES) comprises 10 statements that are to be answered on a four-point scale in terms of how well they reflect the participants’ feelings. The general score of the questionnaire assumes values between 10 and 40 points; the higher scores indicate higher self-esteem. The Polish adaptation of the scale has good psychometric properties and is widely used in research [31].The Daily Heterosexist Experiences Questionnaire [32,33] was implemented to examine the perceived exposure to the stigma associated with gay identity in surveyed men. The questionnaire consists of 50 items describing various stigmatization experiences, which participants rate using a six-point scale in terms of the degree to which these experiences were stressful for the participants. The response format was as follows: 0 = “Did not happen/not applicable to me”, 1 = “It happened, and it bothered me not at all”, 2 = “It happened, and it bothered me a little bit”, 3 = “It happened, and it bothered me moderately”, 4 = “It happened, and it bothered me quite a bit”, 5 = “It happened, and it bothered me extremely”. The questionnaire covers nine factors, the following six of which were included in the study: Victimization—a factor describing experiences of physical abuse based on a non-heterosexual identity; Harassment and discrimination—a factor including inferior treatment, verbal abuse, and harassment based on sexual identity; Family of origin—a factor depicting experiences of rejection by the family of origin; Vigilance—a factor capturing the effort made to conceal one’s own sexual identity; Isolation—a factor marking feelings of alienation and loneliness experienced as a result of being a non-heterosexual person; and Vicarious trauma—strain resulting from learning about discrimination and abuse experienced by other LGBTQ people. The Polish adaptation of the questionnaire has good psychometric properties, and its general higher scores indicate a higher perceived exposure to stigma [33].Moreover, the study included the Resilience Measurement Scale SPP-25 [34] to assess the resilience, which is understood as the ability to cope with both daily and traumatic stressors. The scale consists of 25 items to which participants respond using a five-point scale in terms of how accurately the items describe the participants. Higher scores on the scale indicate higher levels of resilience.The Center for Epidemiologic Studies Depression Scale–Revised (CESD-R) questionnaire [35,36] comprises 20 statements describing a person’s mood and well-being over the last 2 weeks. The study participants rate each statement using a five-point scale in terms of the frequency of symptoms’ or behaviors’ occurrence (where 0 means that the symptom did not occur or lasted no longer than 1 day, and 4 means that it had been occurring almost daily for 2 weeks). Higher scores on the CESD-R scale mean higher intensity of depressive symptoms.To evaluate positive and negative eating habits, the Dietary Habits and Nutrition Beliefs Questionnaire (KomPAN) was used [37]. It takes into account the frequency of consumption of products from different food groups, which are rated by the participants on a six-point scale, in terms of the participants’ frequency in their consumption (where 1 means “never”, and 6 means “several times a day”). The study included statements necessary to calculate two indices: the healthy diet index (e.g., fruit, vegetables, and fish consumption), and the unhealthy diet index (e.g., fast food, red meat, and sweets consumption).To assess the prevalence of cardiovascular diseases risk factors, the study employed excerpts from the Behavioral Risk Factor Surveillance System questionnaire (BRFSS [38]) on alcohol consumption, smoking, and physical activity. Questions regarding the frequency of alcohol consumption have been amended so that they would (i) present the volume of different drinks in milliliters, and (ii) refer to the volume of drinks containing ≈10 g of pure ethyl alcohol. To measure the frequency of alcohol consumption, the participants answered the questions of how many days they had drank alcohol drinks during the 30 days preceding the study, and whether during the same period, there had been at least one occasion when they had consumed at least 60 g of pure alcohol (with the drink being defined as 250 mL of 5% beer, 100 mL of 12% wine, or 30 mL of 40% spirits). The participants also answered the question of whether they had ever smoked tobacco and if they smoke now.To assess physical activity, participants were asked to evaluate how much time (on average) they spend during the week on moderate-intensity physical activity (such as cycling, walking, dancing, or gardening), or vigorous-intensity physical activity (such as jogging, cardio exercises, roller skating, gym exercises, or cross-country skiing). The answers given by the participants have been dichotomized so that they would match the WHO-recommended physical activity for the prevention of chronic diseases, which is at least 150 min of moderate-intensity physical activity or at least 75 min of intense/vigorous physical activity (per week). The dichotomous variable indicated whether or not sufficient physical activity was taking place.The participants were also to answer the question of whether they had been diagnosed with diabetes, hypertension, or hypercholesterolemia by a medical doctor or other health specialists.The body height of men identifying as Bears was measured using a portable stadiometer, and their body mass was measured using the Tanita BC-454 Body Composition Monitor. The men in the control group who participated only in the questionnaire study provided information about their body height and current body weight by themselves. Based on the information obtained, the body mass index was calculated for men in both groups.The data analysis was performed using the Statistica 13 software (Statsoft, Inc., Tulsa, OK, USA) [39]. Comparisons between the group of Bears and men from the control group in terms of demographic characteristics (such as age, place of residence, education level, or financial hardships), depending on the type of variable, were conducted using an analysis of variance (ANOVA), or logistic regression, to be able to show regression coefficients together with the standard error, or odds ratio together with confidence intervals set at 95%. Comparisons between the groups in terms of the prevalence of health-related behaviors or cardiovascular risk characteristics (i.e., BMI values, patterns of alcohol use, tobacco use, physical activity, as well as mental health and resilience), were conducted while controlling for those demographic variables, with respect to which the groups of men differed significantly (p < 0.05). The above analyses were conducted using analysis of covariance (ANCOVA) for quantitative variables, and logistic regression for dichotomous variables. Comparisons of groups of men in terms of the prevalence of medical CVD risk factors (diabetes, hypertension, and elevated cholesterol levels) were performed using logistic regression with adjustment for age and BMI value, which significantly differed men of the two groups. The assumed significance level was 0.05.Table 1 presents the comparisons of both groups of men in terms of demographic variables. No statistically significant differences were observed between Bear men and men from the control group regarding the numbers of individuals with university education level, living in cities with a population greater than 500,000, or reporting financial hardships. However, men in the Bear group were on average older than men in the control group (Table 1). As a significant age difference was observed, further comparisons on CVD risk factors included adjustment for age. These comparisons are presented in Table 2.Men who identify as Bears were found to have significantly higher BMI values, whose mean value was characteristic for obesity (Table 2). This difference remained statistically significant after adjustment for age. No differences between Bear men and men in the control group were observed in terms of depression nor in terms of perceived stigmatization due to sexual identity. However, the men in the Bear group were characterized by both higher resilience and higher self-esteem. These differences were no longer statistically significant after adjustment for age.In the area of health behaviors (alcohol consumption understood as the number of days during the 30-day period preceding the study when men consumed alcohol and at least one heavy episodic drinking (HED) during the 30 days), no statistically significant differences between the groups were observed. The analogous results were found regarding the percentage of smoking men and the percentage of men undertaking weekly recommended physical activity. Moreover, no statistically significant differences were found in both negative and positive eating habits between the compared groups.Although men from the Bear group reported a significantly higher prevalence of medical CVD risk factors (diabetes, hypertension, or elevated cholesterol levels), none of these differences remained significant after adjusting models for age and BMI (Table 3).Our study aimed to compare Bears with other gay men across a range of CVD risk factors including behavioral, medical, and psychosocial ones. Based on previous research, we expected to observe increased levels of medical and behavioral CVD risk factors in Bears compared to other gay men. Consistently with our expectations, Bears were characterized by increased levels of all medical CVD risk factors, including higher BMI, increased frequency of self-reported diagnoses of hypertension, diabetes, and hypercholesterolemia. Of all investigated behavioral CVD risk factors, negative dietary habits index and alcohol consumption were marginally increased in Bears. Among psychosocial characteristics, individual resilience and self-esteem were significantly higher in Bears, indicating better mental health relative to other gay men. None of the observed differences remained statistically significant after adjusting for age and, in the case of self-reported diagnosis of diabetes, both age and BMI.Increased BMI and older age seem to be the distinguishing characteristics of Bears compared to the gay men from the control group. This result is consistent with previous studies demonstrating that indeed, Bear subculture is particularly welcoming toward older and heavier sexual minority men [25,28,40]. Both characteristics which distinguish members of the Bear subculture have been also consistently and strongly associated with cardiovascular disease [41,42].Although the body mass index is a convenient tool to monitor the prevalence of obesity at the population level and obesity has been consistently associated with increased CVD morbidity and mortality [41], it has been also demonstrated by numerous studies that obesity defined solely based on the BMI value is a heterogeneous health condition [43]. One of the major contributors to the cardiovascular and metabolic risk associated with increased BMI values and a key factor in obesity heterogeneity is body fat distribution and the tendency to accumulate visceral fat [43]. Another factor largely affecting the CVD prognosis associated with obesity is a physical activity and cardiorespiratory fitness levels [41]. There is some evidence that overweight and obese adults who meet physical activity guidelines (i.e., ≥150 min of moderate-to-vigorous physical activity per week) are characterized by similar or even lower risk for future CVD events as compared to physically inactive adults with normal body weight [44]. This phenomenon is referred to in the literature as a fat-but-fit paradox, and it underscores the importance of promoting physical activity among various populations, including those that are characterized by increased levels of overweight and obesity as well as normal body weight [41].Aging has also been described in the literature as one of the most important factors affecting cardiovascular homeostasis, since many of the processes that contribute to age-related changes in the vascular structure and function are involved in development of the cardiovascular disease [42]. The prevalence of metabolic syndrome and diabetes is also significantly increased in older populations [42,45]. Therefore, it is not surprising that the observed differences between studied groups concerning being diagnosed with hypertension, diabetes, or hypercholesterolemia were no longer significant after controlling for age, and in case of diabetes, both age and BMI.We observed similar effects in the case of resilience and self-esteem—significant differences indicating higher self-esteem and resilience in Bears were attributable to age differences between both groups. According to the literature, self-esteem increases during a lifespan and usually peaks between the ages of 60 and 70 years [46]. This result is also consistent with previous studies comparing Bears and other sexual minority men, which indicated significantly decreased self-esteem in Cubs—younger members of the Bear community [25]. It is possible that over the years, Cubs develop strategies to resist discrimination and stigma, which contribute to increased self-esteem in older age. Studies on sexual minority samples demonstrate that throughout their lives, LGBTQ persons indeed develop resilience in the face of experienced adversities [47,48]. Consistent with previous research comparing Bears with other sexual minority men, we also did not observe significant differences between both groups associated with exposure to sexual minority stigma and health outcomes such as depressiveness [25,27]. Given the associations between chronic stigma and cardiovascular health [13], observed exposure to sexual minority stigma in both groups underlines the importance of addressing this issue from a public health perspective. The situation of the LGBTQ community in Poland is particularly difficult including structural inequalities such as no legal recognition of same-sex relationships and recent rise of anti-LGBTQ rhetoric in Polish public discourse. In 2020, Poland has been rated by ILGA-Europe as a country with the worst human rights situation of LGBTQ people among EU Member States [49]. According to data collected by the European Union Agency for Fundamental Rights, of all EU LGBTQ respondents, Polish citizens most often reported being physically or sexually attacked due to their minority identity [50]. This calls for greater involvement of public health experts and policy makers in initiatives aimed at countering the negative impact of stigma on health of LGBTQ persons in Poland.Among behavioral CVD risk factors, only unhealthy dietary habits and number of drinking days during the 30 days preceding the study were increased in Bears and almost reached the level of statistical significance. However, these disparities may be associated with the way the data were collected and the fact that to reach more participants, we accompanied Bears during their social gatherings, some of which lasted for several days and were associated with increased alcohol consumption and less healthy diet. Consistent with previous studies, we also did not observe significant differences between Bears and other sexual minority men concerning smoking tobacco and excess (heavy episodic) drinking [27,28]. The prevalence of smoking in Bears was similar to corresponding data on the current prevalence of tobacco use in men in Poland which for current daily and occasional smokers was 25.9% [51]. The prevalence of smoking among gay men from the control group was disproportionately higher (Table 2), and as such, it requires more attention from public health experts.Our study was also the first to compare Bears and other sexual minority men concerning physical activity levels. We did not observe a significant difference between both studied groups in case of reaching physical activity levels recommended by the WHO [52]. Rates of insufficient physical activity in the case of Bears and other gay men were also similar to corresponding rates for Polish men (29% and 25% vs. 31.5% respectively) [53]. Given that an increase from being inactive to reaching recommended physical activity levels is associated with reduced risk of CVD incidence by 17% and diabetes incidence by 26% [54], both Bears and gay men from the control group may benefit from initiatives promoting physical activity.Our study has several limitations including a small sample size and convenience sampling, which limits the generalizability of the results. We also used self-reported data on CVD risk factors rather than conducting precise measurements. The cross-sectional design of the study limits conclusions on causality. Finally, we compared data from two studies with slightly different methods of data collection (online vs. paper–pencil survey). However, we conducted a study on a very rarely investigated population of Polish sexual minority men, including members of the Polish Bear community who would be extremely difficult to reach through probability sampling methods. Our preliminary observations indicate that future studies should further explore observed effects, preferably using longitudinal research designs and more precise indicators of cardiovascular health such as plasma glucose or low-density lipoprotein cholesterol levels. More studies preferably framed within the intersectionality perspective are also needed on associations between chronic stigma exposure and CVD risk in sexual minority populations. Another important line of research includes intervention studies aimed at testing the effectiveness of cardiovascular health promotion initiatives addressed to both sexual minority men and members of the Bear community.Our research was also the first to compare Bears and other sexual minority men across such a wide range of CVD risk factors and therefore provides a significant contribution to the literature on intersections of sexualities and health. We demonstrated that certain groups within the broader LGBTQ community may be particularly exposed to CVD health inequalities and that subcultural affiliations of sexual minority men should be included in studies investigating the health of this population. Our analysis has also important practical implications for public health experts and policymakers by indicating the need to further scrutinize observed health disparities.Our study demonstrated that men who identify as Bears constitute a distinct population among gay men, which is characterized by increased CVD risk associated predominantly with older age and higher BMI values. Given that more than a quarter of Bears in our sample did not meet the recommended physical activity levels, these differences may contribute to the disproportionate CVD morbidity and mortality in Bears. Health promotion interventions addressed to this community should be tailored to Bears’ subcultural norms, which associates larger body-build with physical attractiveness. Instead of focusing on weight loss, they should encourage a healthier lifestyle, including greater physical activity and healthier dietary habits. Vibrant and close Bear communities make members of this subculture easier to reach and offer an affirmative, non-stigmatizing space to talk about health and healthier life regardless of body size. Public health experts should design their interventions based on these premises.Conceptualization, M.M. and G.J. Methodology, M.M. Data analysis, M.M. and K.K. Investigation, M.M., K.K., A.G. Data curation, M.M. Writing—original draft preparation, M.M. Writing—review and editing, M.M., K.K., A.G., G.J. Supervision, G.J. Funding acquisition, M.M. All authors have read and agreed to the published version of the manuscript.This work was supported to Magdalena Mijas by the doctoral scholarship received from the Polish National Science Centre (Grant Number 2019/32/T/NZ7/00069).The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Jagiellonian University Bioethics Committee (122.6120.70.2017; 18 May 2017) and the Research Ethics Committee of the Institute of Psychology at Jagiellonian University (KE/02/052017; 29 May 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 restrictions.We are grateful to all participants who generously devoted their time to take part in this study. We also gratefully acknowledge The Bears of Poland Association for their support with recruitment and organization of data collection process.The authors declare no conflict of interest.A comparison of Bears with other gay men in terms of demographic characteristics.Comparison of Bears with other gay men in terms of selected behavioral and psychosocial risk factors for cardiovascular diseases.Note: * HED—heavy episodic drinking.Comparison of Bears with other gay men in terms of the prevalence of medical risk factors for cardiovascular diseases.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01045.txt

@@ -0,0 +1,11 @@
+The purpose of this study was to analyze the performance of pharmaceutical companies’ business diversification into medical devices in terms of their technical efficiency (TE) as compared to that of traditional pharmaceutical companies. For a total of 174 externally audited pharmaceutical companies engaged in the drug product business between 2008 and 2019, pharmaceutical companies were classified into two groups according to medical device business diversification. The TE of pharmaceutical companies that diversify the medical device business was lower than that of traditional pharmaceutical companies. However, in terms of the meta-technology ratio (MTR) calculated using meta-frontier analysis, pharmaceutical companies diversified into medical devices showed higher MTR than the traditional pharmaceutical company group. The results imply that the corporate performance growth potential of traditional pharmaceutical companies is lower than that of pharmaceutical companies that have diversified into the medical device business.Health issues caused by cancer, diabetes, population aging with low birth rate, infectious diseases, etc. are topics of interest to many people. The importance of the medical device and pharmaceutical industries has been emphasized due to infectious diseases such as influenza, Middle East respiratory syndrome, and coronavirus viral disease 2019 (COVID-19). In the case of COVID-19, the virus has serious adverse effects not only on human health but also the social economy. Thus, a diagnostic test device and a vaccine for the disease are being devised and demanded in countries around the world [1,2,3]. This is because medical devices and medicine for patient treatment are directly connected to each other. However, despite the importance of these industries, pharmaceutical companies are under significant pressure between the consumer associations that demand inexpensive products with good efficacy and investors who pursue high performance and profit [4]Interestingly, research and projects related to the pharmaceutical industry over the past two decades have been invested in discovering new drugs to generate profit [5]. However, as the cost for basic research and development (R&D) is gradually increased, much research and many projects at pharmaceutical companies were stopped [6]. Furthermore, poor R&D led to a decrease in the pharmaceutical industry’s productivity. Specifically, changes in the industrial structure such as the rise of the biotechnology field contributed to the increase in R&D costs of pharmaceutical companies [7]. As productivity declines continue, the pharmaceutical industry is facing unprecedented industrial challenges and surveillance with the extinction of monopoly products and the reduction of pipelines caused by patent expiration [8].To overcome the crisis of productivity decline, pharmaceutical companies first made an effort to improve the process for finding new drug candidates [9]. When productivity was significantly lowered, R&D through outsourcing was boldly chosen as an alternative to reduce costs [10]. Second, pharmaceutical companies aimed at increasing overall corporate productivity by improving the pharmaceutical manufacturing process [11] and optimizing the supply chain with inventory management [12]. These efforts in internal processes did not lead to improved R&D process for a new drug or higher profit that could raise corporate performance, because the causes of productivity decline in pharmaceutical companies are diverse and complex [13].Another effort of pharmaceutical companies to overcome the crisis was business diversification [14]. For example, the number of mergers and acquisitions for the pharmaceutical and biotechnology industry in 2018 recorded a total of 1438 and a total volume of $339.6 billion as the highest in the last decade. The medical device industry, which is indicated as a major factor in the decline of the pharmaceutical industry [15], has become a major target of business diversification from pharmaceutical companies. Interestingly, at this time, the number of acquisitions in the healthcare industry (131 cases), including medical devices, was the second highest after acquisitions of the homogeneous sector (449 cases) followed by distribution/logistics (57 cases) and the information and communication sector (30 cases) [16]. Pharmaceutical companies are highly interested in diversifying the medical device business because both medical device and drug are used by patients or doctors as end users for clinical purpose, and the distribution environment of products looks the same. The phenomenon by which multinational medical device companies are gradually developing and launching medical devices that contain medicines can also be attributed to the fact that the business environment is similar [17]. For these reasons, many pharmaceutical companies are expecting to raise overall corporate performance by improving their productivity through the business diversification of medical devices because of the productivity crisis.However, there are many discussions on how merge and acquisition (M&A) or business diversification will affect corporate performance due to complex factors such as market characteristics of industry and understanding of product, conflicting regulations, organization, etc. as well as additional cost in the process [18,19,20]. In terms of the pharmaceutical industry, the studies for pharmaceutical companies mainly focused on R&D synergy of M&A with the biotechnology industry [21,22,23]. The integration of biotechnology with a similar R&D environment also presents various uncertainties regarding the improvement of pharmaceutical companies’ performance [23]. When considering these studies, there is still an absence of research on whether the business diversification of pharmaceutical companies into medical devices positively affects their corporate performance in the crisis of productivity.The pharmaceutical business differs from medical device in the characteristics of the entire business cycle from product R&D through sales. In the traditional pharmaceutical business, when a product is released through a large R&D investment, it continuously generates high profits with improving the supply of raw materials and promoting sales within the protection of patent rights [24]. When compared to pharmaceuticals, the medical device business has a short product life cycle and risk about easy product duplication, so the market competition is overheated due to the low entry barrier and product profits are not very high. Due to these differences, when a pharmaceutical company diversifies its medical device business, the total sales of the company might increase, but the overall performance of the company might decrease. Furthermore, considering additional costs and time, such as new production line, labor, preparation for medical device regulations, and expenses from sales and management, manufacturers should be aware that pharmaceutical and medical device businesses differ not only in their product development and life cycles, but also in the nature of the business and legal factors [25].This study attempted to determine whether diversifying the medical device business can increase the performance of pharmaceutical companies by analyzing financial data from South Korea. According to the Ministry of Food and Drug Safety (MFDS), the Korean pharmaceutical market in 2019 increased by 5.2% from 2018 (21.24 billion USD) to 22.33 billion USD, ranking 12th in the world (1.6%). Although the Korean pharmaceutical market is growing at a high level as a mature market, overall pharmaceutical companies are facing a decline in corporate performance and a productivity crisis due to the cost of new drug development, patent expiration, regulation, and competition. The phenomenon is common among pharmaceutical companies around the world. For this reason, many pharmaceutical companies in South Korea are expecting to improve corporate performance through diversification of medical device business. South Korea became a country where the trend and phenomenon of pharmaceutical companies that are representatively diversifying medical device business is gradually expanded to overcome the decline in corporate performance and a productivity crisis. For this purpose, this study measured the performance of companies using technical efficiency (TE) indicators. Where sales or productivity is affected by the company size, technical efficiency has the advantage of not being affected by the firm size because it estimates the firm’s production function first and measures technical efficiency according to the distance from the production function. Recently Chung et al. [26], Jo et al. [27], Na et al. [28], and many others measured firm performance using technical efficiency in accordance. Because conventional TE has the disadvantage of not being available for comparison between companies using different production functions, we compared pharmaceutical companies that have expanded their business to medical devices with traditional pharmaceutical companies in terms of a firm’s TE.SFA expresses the relationship between input and output factors as a production function and estimates the TE using the frontier production function representing the maximum output from the input. At this time, the TE of a company indicates where the company’s technology level is relatively located, when compared to the efficiency technology level of the frontier production function. The farther the technology level of a company is from the frontier production function represents a lower level of efficiency. The production frontier can be estimated through both nonparametric and parametric methods. In this study, the production frontier for the parametric method was estimated using SFA. Also, FRONTIER Version 4.1 software provided by Coelli was used for estimation.According to Battesse and Coelli [29], the model of the stochastic production frontier is assumed as the following Equation (1) to reflect the change of time in efficiency.
+        (1)Yit=f(xit, β)eVit−Uit, i=1,2,…,N, t=1, 2,…, TAt this time, Yit is the output of company i at time t, xit is the input vector of company i at time t, f is the production function, β is a vector containing the parameters of the production function, Vit is independent of Uit with a random error with a distribution of N(0, σv2), and Uit is a nonnegative random variable representing the TE of company i at time t. If Vit is the general random error of the regression equation, then Uit represents the company’s inefficiency. To show that it is always inefficient, Uit itself is not negative and this study assumed that Uit follows a half-normal distribution. Because data from 2008 to 2019 were used, T is 12.From Equation (1), the efficiency, TEit, of company i at time t is given by Equation (2) below.
+        (2)TEit=e−Uit=Yitf(Xit,β)eVit, i=1,2,…,N, t=1,2,…, TIn general, the Cobb–Douglas function and the translog function are the most widely used SFA production functions. However, in the case of Cobb and Glass, there is a tendency to oversimplify it because the output variable is seen as a linear combination of input variables. Therefore, in this study, we used the translog function. In particular, a random-effects, time-varying production model was used. When assuming a translog type of production function, Equation (1) can be expressed as Equation (3) below.
+        (3)lnYit=β0+∑m=13βmlnxmit+∑m=13∑k≥m3βmklnxmitlnxkit+Vit−Uit
+        where x1it represents the amount of capital (K) of the i-th company at time t, x2it represents the cost (M) of the i-th company at time t, and x3it is the number of workers (L) who receive a salary from the i-th company at time t. The total assets for K, the number of employees for L, and the cost of revenue for M are used as input variables, and net sales (Y) are used as an output variable in this study.Since the TE of a specific company is difficult to compare with other companies using different technologies, comparisons of technological efficiency between each group cannot be performed through traditional SFA. Therefore, to compare the efficiency levels of different groups operating under different technical conditions, we used the meta-frontier production function that wraps the production functions of all groups [30]. From Battese, Rao, and O’Donnell [31], the meta-frontier production function model is defined as follows.
+        (4)Yit*=f(xit, β*)=exitβ*, i=1,2,…,N, N=∑i=1RNj, t=1, 2,…, T, s.t. xitβ*≥xitβ(j) for all jIn this case, β(j) is a vector composed of the parameters of the j-th group’s production function and j indicates each group. In this study, the two groups are traditional pharmaceutical companies that have only produced medicines (j = 1) and diversified pharmaceutical companies that also produce medical devices (j = 2). β* is a vector of unknown variables of the meta-frontier function that satisfies the following equation. From Equation (4) above, the graph of the meta-frontier production function is positioned above the graph of the production frontier function of each group for all periods. The meta-frontier production function becomes an envelope of the frontier functions of each group based on the same technology. For simplicity, if we assume that function f in Equation (1) is eXitβ(j), Equation (1) can be divided as shown in Equation (5).
+        (5)Yit=e−Uit(j)×exitβ(j)exitβ*×exitβ*+Vit(j)Dividing both sides of Equation (5) by exitβ*+Vit(j) yields the following.
+        (6)Yitexitβ*+Vit(j)=e−Uit(j)×exitβ(j)exitβ*In Equation (6) above, the right side, e−Uit(j) is the technical efficiency (TE) of Group j and the second is the j group frontier for the meta-frontier function. It is expressed as a ratio of a function, which is called the Technical Gap Ratio or Meta-Technology Ratio (MTR). TE*, representing TE in the meta-frontier function, is calculated by multiplying TE by MTR and can be expressed as Equation (7).
+        (7)TEit*=Yitexitβ*+Vit(j)=TEit×TGRitThere are two methods of measuring the parameters of a meta-frontier function: Linear Programming (LP) and Quadratic Programming (QP). LP is a method of minimizing the sum of the absolute deviation values, and QP is a method of minimizing the sum of the squares of deviations. According to Battese, Rao, and O’Donnell [31], LP and QP are defined as following Equations (8) and (9).
+        (8)LP: minβ*L*=∑t=1T∑i=1N|xitβ*−xitβ^(j)|, xitβ*≥xitβ^(j)
+        (9)QP: minβ*L*=∑t=1T∑i=1N(xitβ*−xitβ^(j))2, xitβ*≥xitβ^(j)Matlab 7.1 software was used to measure the parameters of the meta-frontier function using LP and QP.In this study, actual corporate financial data were secured from the KIS-VALUE database of the National Information and Credit Evaluation. Based on the information from the South Korean Ministry of Food and Drug Safety (MFDS), pharmaceutical companies are divided into traditional pharmaceutical companies and diversified pharmaceutical companies for medical devices according to their import certifications and licenses to manufacture medical device products. Specifically, the number of externally audited Korean pharmaceutical companies that acquire approval to manufacture and import medical devices for general treatment and surgery from the MFDS has increased rapidly since 2008. Because of a novel influenza outbreak in 2009, Korean pharmaceutical companies expanded to new businesses that manufacture, import, and distribute various types of medical devices, from advanced medical devices to diagnostic test kits and instrument. Thus, a total of 174 externally audited pharmaceutical companies in South Korea were identified for the period between 2008 and 2019.One hundred three traditional pharmaceutical companies that only conducted pharmaceutical business and 71 pharmaceutical companies that diversified into the medical device business were separated into two groups. The total number of samples with fiscal data used in the study was 1028 for traditional pharmaceutical companies and 728 for pharmaceutical companies diversified into medical devices. Table 1 contains details regarding the samples.FRONTIER 4.1 was used for SFA and meta-frontier analysis (MFA) with MATLAB 7.1 was carried out to verify corporate efficiency. Table 2 shows the estimated frontier production function for each group with meta-frontier production function parameters optimized through LP and QP methods.Table 3 shows the results of the TE for each group and the value of TE* in the meta-frontier production function with the MTR by using the estimates of the group frontier production function and meta-frontier production function.As a result, traditional pharmaceutical companies (76.28%) showed higher TE values when compared to diversified pharmaceutical companies for medical devices (59.11%). However, as mentioned earlier, comparisons between groups using different production functions are meaningless. Therefore, the TE of the two groups using different production functions should be compared through MTR. Conversely, a group of diversified pharmaceutical companies for medical devices showed a higher MTR value both with LP and QP (MTR_LP: 94.54%, MTR_QP: 88.82%) when compared to the group of traditional pharmaceutical companies (MTR_LP: 80.39%, MTR_QP: 76.67%).Traditional pharmaceutical companies have pursued high corporate efficiency through continuous research on pharmaceutical business models in diverse fields such as patents, regulations, distribution, R&D, and manufacturing innovation. With recent advances in marketing strategies and technologies, the pharmaceutical industry has been trying to overcome the decline in productivity by maximizing profits for many years [32,33,34]. Nevertheless, unlike increasing their market size, the problem of decreasing productivity has not yet been solved. For the pharmaceutical industry, this is expected to be a complex cause of the already mature business model, excessive market competition, regulation, and competition with similar businesses such as the biotechnology and medical device industries. Companies might seek to integrate with homogeneous firms to improve performance and expect corporate synergy through the combination and expansion of heterogeneous industries. The pharmaceutical industry is also making an effort to reinforce their business portfolio through mergers and acquisitions and business diversification due to the efficiency decline and low productivity. As mentioned in the introduction, in 2018, a total of 1438 transactions with a total volume of $339.6 billion in global mergers and acquisitions in the pharmaceutical and bio-industry industries were recorded. The number of acquisitions in the healthcare industry including medical devices (131 cases) was the highest after homogeneous industry acquisitions (449) [16].As mentioned earlier, pharmaceutical companies expect to improve overall productivity by reinforcing R&D pipelines between homogeneous industries [35], while expecting to improve corporate performance through business expansion into different industries. Globally, pharmaceutical companies’ business expansion of the healthcare business including medical devices has gradually increased, and Korean pharmaceutical companies, which are diversifying into the medical device business, also had a total of 31 (17.8%) in 2008. Currently, a total of 71 (40.8%) pharmaceutical companies have gradually increased their number to diversify their medical device business. Thus, the importance of research on how the medical device business affects the performance of pharmaceutical companies has been raised in South Korea.As a result of this study, the MTR of the pharmaceutical company group that conducted business diversification was higher than that of the traditional pharmaceutical company group. This higher MTR of the diversified group means that the group’s frontier production function is located closer to the meta-frontier production function. The frontier production function is determined by the technology used by the companies in the group and is the set of maximum outputs that the companies can produce. Therefore, the fact that the MTR of the pharmaceutical group that diversified into the medical device business is higher than the MTR of the traditional pharmaceutical group means that the maximum output that can be produced through the same input is higher, that is, it has a higher potential. Interestingly, in the MFA, which is the same production function, both TE* calculated with LP and QP showed higher results than traditional pharmaceutical companies diversifying their medical device business. As described above, TE* can be calculated as the multiplication of TE and MTR. In traditional pharmaceutical companies, although the MTR was lower than that of the pharmaceutical company group that diversified into the medical device business, the TE was much higher than that of the group. Thus, the TE* was higher. This means that traditional pharmaceutical companies are exhibiting higher efficiency under the current production function, but their potential is lower than that of pharmaceutical companies that have diversified into the medical device business.Based on research results, policy makers and corporate decision makers as well as future study should consider the following implications of this study for sustainable management in the pharmaceutical industry and the medical device industry.First, in terms of theoretical implications, while most of the existing studies for decades focused mainly on improving the R&D process of the pharmaceutical industry, the industry has been establishing diverse strategies to grow the productivity and performance potential of companies such as the business diversification of medical device. Future studies should be able to fill the blanks in these relevant research topics and provide new perspectives to improve the performance of pharmaceutical companies.Second, in terms of practical implications, the interaction between doctors and pharmaceutical companies is known to influence the prescription of medicines [36]. Although doctors are users, they have a large portion of the patents for medical devices [37]. Since they have already been involved from the earliest stages of development in clinical trials [38], doctors are interested in not only medicines but also medical devices. Furthermore, since the trend of the treatment process is rapidly changing due to the development of medical devices, the direction of drug development and productivity can be flexibly modified. This is the reason why medicines and medical technology are complementary to each other and technology advancement is simultaneously progressing. Therefore, business diversification into medical devices seems to increase the potential for corporate performance by improving the technological efficiency of pharmaceutical companies.Third, in terms of policy implications, policy makers considering and reviewing business promotion of the pharmaceutical industry should be encouraged to review diversified business models to improve corporate performance and potential. Furthermore, R&D for new drugs must be continuously conducted by promoting corporate policy with business portfolio expansion of medical device with complicated crisis of productivity.The medical device business clearly differs from the pharmaceutical business in terms of the product production cycle from R&D to product launch and follow-up management. However, when we examined the two businesses from the perspective of healthcare providers and users, not from manufacturers or regulatory agencies, medical devices and medicines in treatment are complementarily connected. In other words, both products are developed and used for the purpose of treatment based on the interaction between the doctor and patient. In conclusion, when compared to traditional pharmaceutical companies, pharmaceutical companies diversified into the medical device business are dominating the traditional pharmaceutical companies in the trends of treatment process and the contact point of customers with market changes. Therefore, these advantages appear to improve overall corporate performance growth potential.As the pharmaceutical industry may experience a decline affected by major developments in the medical device industry and the biotechnology industry, it is necessary to take insights on multiple businesses and have a view of political decision for pharmaceutical industry promotion. Policy makers should not encourage companies to carry out one business and a similar industry as their conservative general policy.This study has the following limitations. First, since this study used data from externally audited companies for the period between 2008 and 2019, smaller companies with low total assets were excluded from the group. Second, different variables that could affect the TE of pharmaceutical companies were not considered as well as business diversification. Third, the study is insufficient to provide global implications because it analyzed data only from Korean pharmaceutical companies. Therefore, future research could provide additional implications for the pharmaceutical industry if it continuously expands its importance by using worldwide data and considering other variables that can affect technological efficiency.Conceptualization, Y.E. and D.L.; Analysis, D.L.; Writing, Y.E. and D.L. All authors have read and agreed to the published version of the manuscript.This research was supported by the Ministry of Education of the Republic of Korea and the NRF (No. 2020S1A5A8045556, 2020R1F1A1048202).Not applicable.Not applicable.Restrictions apply to the availability of these data. Data was obtained from KIS-VALUE and are available from https://www.kisvalue.com/web/index.jsp with the permission of KIS-VALUE.The authors declare no conflict of interest.Sample statistics.Note: Numbers in the parenthesis are standard deviations. 1 USD is 1,105.0 KRW (Korean Won) as of 25 January 2021.Estimation results of group and meta-frontier production functions.Note: *, **, and *** mean p < 0.1, p < 0.05, p < 0.01 respectively.SFA estimates of technical efficiencies and meta-technology ratios.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01046.txt

@@ -0,0 +1 @@
+These authors contributed equally to this work.There are many epidemiological studies asserting that fine dust causes lung cancer, but the biological mechanism is not clear. This study was conducted to investigate the effect of PM10 (particulate matter less than 10 μm) on single nucleotide variants through whole genome sequencing in lung epithelial cancer cell lines (HCC-827, NCI-H358) that have been exposed to PM10. The two cell lines were exposed to PM10 for 15 days. We performed experimental and next generation sequencing analyses on experimental group that had been exposed to PM10 as well as an unexposed control group. After exposure to PM10, 3005 single nucleotide variants were newly identified in the NCI-H358 group, and 4402 mutations were identified in the HCC-827 group. We analyzed these single nucleotide variants with the Mutalisk program. We observed kataegis in chromosome 1 in NCI-H358 and chromosome 7 in HCC-827. In mutational signatures analysis, the COSMIC mutational signature 5 was highest in both HCC-827 and NCI-H358 groups, and each cosine similarity was 0.964 in HCC-827 and 0.979 in the NCI-H358 group. The etiology of COSMIC mutational signature 5 is unknown at present. Well-designed studies are needed to determine whether environmental factors, such as PM10, cause COSMIC mutational signature 5.Fine particles are particulate matter, a fine substance in a solid state. Fine particulate matter having a diameter smaller than 10 μm is called PM10. Most particulate matter (PM) comes from the combustion of fossil fuels. Recently, the concentration of fine particulate matter has been steadily increasing. The PM concentration in Korea has been measured as being greater than that of other countries, including the USA and Japan. In Korea, PM concentrations have been on the decline but are still high as compared to other countries [1]. PM enters humans through inhalation, and exposure to PM aggravates pulmonary inflammation because of its oxidative stress and direct toxic injury [2]. According to recent studies, exposure to PM has been found to affect the risk of cardiovascular disorders, cancer, neurologic disease, and many respiratory diseases, including chronic obstructive pulmonary disease, asthma and lung cancer [2,3,4]. Exposure to PM has been found to degrade lung function, and this affects many respiratory diseases. As the PM concentration increases, the number of patients with an acute exacerbation of asthma or chronic obstructive pulmonary disease (COPD) increases. As the PM exposure increases, the morbidity and mortality rate of COPD and asthma also escalate [5,6,7]. Other studies have reported an increased exacerbation frequency of idiopathic pulmonary fibrosis (IPF) because of decreased lung function [8]. Recent studies have shown that PM can also be a carcinogen [9,10], that PM concentration is associated with lung cancer development, and that there is positive correlation between the PM concentration and the incidence of lung cancer. Although several studies have investigated the epidemiological correlation between PM and lung cancer, the correlation between the molecular biological properties, including the mechanism of PM and the pathogenesis of lung cancer, is not clear. Mutations do not occur randomly. Depending upon their own preferred contexts, different mutational processes often result in different patterns of somatic mutations. These distinct patterns are called mutational signatures. Therefore, analyzing the pattern of the mutation can help us understand their causes. In addition to the mutational signature, kataegis, hypermutation confined to small genomic regions, is found in many types of cancer. Between transcribed and untranscribed gene strands, the efficiency of DNA damage and DNA maintenance processes can vary [11]. Thus, research about the mutational processes has a potential impact on the understanding and treatment of cancer.This study was conducted to investigate the effect of PM10 on single nucleotide variants in lung cancer through whole genome sequencing of the lung epithelial cancer cell line (HCC-827, NCI-H358) exposed to PM. We used the Mutalisk program (National cancer center, Goyang-si, Korea) for analysis [12]. Additional information about Mutalisk is available on the following website: http://mutalisk.org/.In this study, two lung cancer cell lines were used: NCI-H358 and HCC-827. The former, NCI-H358, carries the EGFR mutation while the latter cancer cell line, HCC-827, carries K-ras mutation together with strong epithelial properties [13]. We used these two cell lines in order to find common results in lung cancer cells, regardless of the characteristics of each cell. We purchased the two cell lines from the Korea Cell Line Bank. We selected ERM® CZ120 (Sigma-Aldrich, St. Louis, MO, USA) fine dust as our PM10 because it is easy to use and is used for Korean fine dust meter calibration, and the particle size is guaranteed. The ERM® CZ120 fine dust is mainly composed of arsenic, cadmium, lead, nickel, and other additional ingredients.Each of the two cell lines was dissolved in a 37 °C water for one minute in a frozen state and diluted in 10 mL of a medium (RPMI 1640 + 10% FBS + 1% antibiotics). After dilution, each medium was centrifuged for five minutes at 1500 rpm, and the supernatant was removed. After removing the supernatant, we diluted the medium again in 10 mL of culture medium and incubated it in 100Φ dishes. When the confluence of the cells in the 100Φ dishes reached about 80%, we used the following procedure to progress the cells to subculture. First, we removed 10 mL of medium. This was washed by adding 6 mL of PBS (phosphate-buffered saline) (Thermo Fisher Scientific, Waltham, MA, USA). The supernatant was removed, and then 0.25 trypsin-EDTA (Thermo Fisher Scientific, Waltham, MA, USA) 1 mL~2 mL was treated on the dish. After incubating this for five minutes in a CO2 incubator (Thermo Fisher Scientific, Waltham, MA, USA) at 37 °C, confirming the cells that had been separated through microscopy, the trypsin-EDTA was neutralized by mixing it with 5 mL of the culture medium. This was collected in a 10 mL tube, and the supernatant was removed after centrifuging. After adding 1 mL of the culture medium to the cell pellet, we calculated the number and viability of the cells by using the trypan blue method. Each cell was seeded in 3 × 105 cells in 100Φ dishes. We added 10 mL of the culture medium. After that, each cell line was exposed to PM10 so that the final concentration of PM10 was 50 μg/mL. We subcultured this five times every three days and added the PM10 in the same manner as before. Thus, the cell lines were exposed to PM10 for a total of 15 days. The untreated cells were grown simultaneously, built under the same parameters with which the treated cell lines were harvested at same time points. We performed NGS analysis in both groups. The workflow is as follows:We extracted the genomic DNA of the lung epithelial cells (Wizard Genomic DNA purification Kit, Promega, A1120).In order to glean a final library of 300~400 bp average insert size, each DNA fragment was sequenced in accordance with the Illumina Truseq DNA sample preparation guide. Using the covaris systems, 1 μg (TruSeq DNA PCR-free library) or 100 ng (TruSeq Nano DNA library) of genomic DNA was fragmented, which induced dsDNA fragments with 3′ or 5′ overhangs.An end repair mix converted the dsDNA fragments with 3′ or 5′ overhangs into blunt ends. The 3′ to 5′ exonuclease detached the 3′ overhangs, and the 5′ overhangs were repleted with the polymerase. The appropriate library size was chosen following the end repair, incorporating different ratios of the sample purification beads.To avoid the jointment of the dsDNA fragments during the adapter ligation reaction, a single “A” nucleotide was added to the 3′ ends of the blunted fragments.Ligation of numerous indexing adapters to the blunt ends of the DNA fragments was carried out to lay out the groundwork for hybridization onto a flow cell. The enriched DNA library was amplified utilizing the PCR for sequencing. The PCR primer solution was operated for PCR, which allowed for the annealment of the ends of each adapters (Truseq Nano DNA library only).The library needed to be loaded into a flow cell so as to capture dsDNA fragments situated on a lawn of surface-bound oligos, complementary to the library adapters, for the purpose of generating clusters of cells. By way of bridge amplification, each of the fragments underwent an amplification process, which resulted in distinct, clonal clusters. The completion of cluster generation was indicated by a signal that the templates were all set up for sequencing. Single bases were detected via Illumina SBS technology (Illumina, San Diego, CA, USA), as they were placed into DNA template strands, which operated based on the method utilizing proprietary reversible terminator.Raw images and base calling were produced by the Illumina Platform, which embodied an integrated primary analysis software referred to as RTA (real time analysis). Illumina package bcl2fastq2-v2.20.0 was utilized to convert the BCL/cBCL (base calls) binary into FASTQ and to mismatch barcodes by setting the demultiplexing option to perfect match (value: 0).The HiSeq sequencing system provided paired-end sequences, which are mapped to the human genome with Isaac aligner (iSAAC-04.18.11.09, Illumina, San Diego, CA, USA) where the reference sequence was the UCSC assembly hg19 (original GRCh37 from NCBI, Feb. 2009). Enabling the 32-mer seed-based search incorporated into the Isaac aligner, the foremost mapping candidates could be withdrawn and identified. Low quality 3′ end along with adapter sequences had been trimmed from the alignment in the process. A binary alignment output file (.bam) was also generated by the Isaac aligner, which was equipped with sorted and duplicate-marked data.Identification of single-nucleotide variants (SNVs) and short insertions and deletions (Indels) were performed through read processing, in which Strelka was used to effectively filter out both low quality reads and PCR duplicates. Such a read realignment process was carried out by Strelka so as to bolster the accuracy level. In addition, the germline probability model-based variant genotyping was analyzed. A block-compressed genomic variant call format (gVCF) file was the result file that contained information on these variants. Variant-only VCF was generated by the extract variant, one of the utilities included in gvcftools package, as all non-variant blocks from the gVCF file were eliminated along with the filterification of low-quality and high-depth variants. Another program, referred to as SnpEff (v4.3t), came into play to annotate variants from the Variant-only VCF file, to which the in-house program and SnpEff were applied for its annotation, with extra databases including ESP6500, ClinVar and dbNSFP3.5.Manta (1.5.0) (Illumina, San Diego, CA, USA) was first enabled to recognize the specific input data and options through the configuration step before going into the procedural execution on a single node. For identification of structural variants and large indels, the entire genome sequencing structural variant calling analysis, with default options, was performed by Manta. Control-FREEC (11.5) was used to determine and pinpoint copy number variant with 10,000 window size and no additional options. It also incorporated GC-content bias to normalize read counts and XY for sex. The CNV types were classified based on genome ploidy value 2. Thus, values above two indicated gain, whereas values below 2 signified loss.As mentioned earlier, in this study, two cell lines, HCC-827 and NCI-H358, were used. SNVs were analyzed with the Mutalisk program. For analysis, we used definitions and statistical tools as defined in Mutalisk for localized hypermutation, mutational signatures, transcriptional strand bias, correlation coefficients of somatic mutations, epigenomic and transcriptional features (including GC contents), DNA replication timing, DNase hypersensitivity, and histone modification. The program input was a standard VCF file, obtained from whole genome sequencing of the differences between the cell lines exposed to PM10 and the control group.A total of 12,388 differences in the SNVs, six copy number variations (CNVs), and 763 structure variations (SVs) in NCI-H358, and 13,348 SNVs, 27 CNVs, and 891 SVs in the HCC-827, were accounted for when the PM10 exposed cell group and the control group in two cell lines were compared. The underlying purpose of this study is to investigate SNVs. In total, 3005 SNVs were newly investigated in NCI-H358 and 4402 SNVs were examined in HCC-837 in the PM10 group.SNVs were closely examined in both treated and untreated cells. As shown in Figure 1, localized hypermutation or kataegis is visually laid out in rainfall plot format. To be more specific, in NCI-H358, chromosome 1 appeared to be the most localized with 516 mutations, followed by chromosome 5 with 413 mutations and chromosome 2 with 344 mutations. Meanwhile, chromosome 7 was shown to be the most localized with 519 mutations in HCC-827, which was followed by chromosome 14 with 445 mutations and chromosome 1 with 388 mutations.COSMIC mutational signatures, packaged with 30 different mutation types ranging from COSMIC signature 1 to COSMIC signature 30, trigger various mutational processes to create unique combinations of mutation types [14]. Figure 2 shows the mutational signature decomposition results generated by Mutalisk, using the differences between cell lines exposed to PM10 and the control group, in both the NCI-H358 and the HCC-827 groups. The number of total mutations was 4402 in the NCI-H358 group and 3005 in the HCC-827 group. In the NCI-H358 group, the seven best decomposition models ranked by Bayesian information criterion (BIC) were the Catalogue of Somatic Mutation in Cancer (COSMIC) mutational signature 5, 18, 1, 13, 30, 2, and 10. On the other hand, in the HCC-827 group, the six best decomposition models were the COSMIC mutational signature 5, 16, 1, 20, 30, and 13. The COSMIC mutational signature 5 was the dominant type of mutational signatures in both NCI-H358 and HCC-827, and the cosine similarity score between the observed and decomposed distribution of mutations was equivalent to 0.979 in NCI-H358, and 0.964 in HCC-827.Figure 3 shows the SNV transcriptional strand bias in the differences between those cell lines exposed to PM10 and the control group of the two cell lines (NCI-H358, HCC-827). By utilizing the RefSeq Gene dataset, Mutalisk facilitates the annotation of the transcription strand information of somatic mutant pyrimidine bases (reference alleles of C or T) and calculates the enrichment for each mutation class [12]. With regard to the NCI-H358 group, the C > T mutations were highly enriched in the untranscribed regions, whereas in the transcribed regions, the C > A mutations were enriched within the NCI-H358 group. The C > T mutations demonstrated high enrichment in the untranscribed regions within the HCC-827 group as well. The ** p-value was <0.05, and the p-value was obtained by using a goodness of fit test. Somatic mutations—each of which is mapped to genomic regions according to each feature, respectively—for each individual functional element (i.e., GC content, DNA replication timing, DNase I hypersensitivity, and histone modification) were fully examined utilizing Mutalisk as the program can recognize mutations in the vcf file that map to the coding region of the selected genome assembly. The GC-content, DNase I hypersensitivity and histone modification features are classified into low, intermediate, or high level, while the DNA replication timing is classified into three phases (early, intermediate, or late phase) [12]. Figure 4 shows the Pearson correlation coefficients between epigenetic modification density and the frequency of SNVs in NCI-H358 and HCC-827. The GC content has a positive correlation in the NCI-H358, and the H3K9me3 has a positive correlation in HCC-827. The other histone modifications have negative correlations. Figure 4 also shows the percentage of explained histone modification variance. The H3K27Ac, H3K9me3, and DNase HS were observed in the top third of the explained variance rate in both cell lines, and GC content, replication timing, and H2AZ were the same in the bottom third. According to a recent study, exposure to PM causes genetic and epigenetic change. Some studies analyzed peripheral blood of individuals exposed to air pollution which shows that PM can cause DNA damage [15,16]. The process of gene expression includes DNA methylation, post-translational histone modification, histone variation, chromatin remodeling, and noncoding RNA [17]. Epigenetic factors such as exposure to PM, affect DNA methylation, modifying gene expression [18]. PM causes increased histone H4 acetylation in human airway epithelial cell lines (A549, BEAS-2B), which acts on IL-8 and COX-2 promoters to increase gene expression [19,20]. Exposure to PM also increases miRNA-222 and miRNA-21 [21]. All in all, PM can serve as a casual factor that can lead to a wide range of variations associated with genetic and epigenetic changes. Since no research analyzing the mutational signature of PM could be found, we performed the analysis for the purpose of this study.In this study, we used the Mutalisk for COSMIC mutational signatures to identify somatic mutations. Some types of mutational signature etiologies are revealed. The COSMIC signature 5 has been found in all cancer types and in most cancer cells, but the etiology of the COSMIC signature 5 is unknown [22,23,24]. In the signature 5 mutations, FHIT loss was observed. This variation is related with factors such as air pollution (i.e., asbestos) and smoking [25]. Comparing the difference between the cell lines exposed to PM10 and the control group in NCI-H358 and HCC-827, the COSMIC mutational signature 5 was the highest in both NCI-H358 and HCC-827. Based on these results, we can conclude that COSMIC mutational signature 5 increases because of environmental factors, including PM. We can also safely conclude that PM10 acts as a carcinogen. Transcriptional strand bias, which has been identified in reporter gene assays and cancer genome sequences to reflect the activity of nucleotide excision repair (NER), along with somatic mutations for every functional element for genomic and epigenomic modification, were closely studied using Mutalisk. NER can be defined as a non-specific repair process that is activated when DNA distortions caused by mutagenic biochemical modifications are detected [26]. Our result from this study demonstrated that with regard to the NCI-H358 group, high enrichment of the C > T mutations in the untranscribed regions, and enrichment of the C > A mutations in the transcribed regions manifested. In the HCC-827 group, the high enrichment of the C > T mutations centered on the untranscribed regions. In addition, most of the functional elements associated with histone modification displayed negative correlations. From the changes spotted in functional elements, speculations can be made that some variations in the function or phenotype of the cells are likely to occur. Further research on the matter needs to be conducted with regard to the biological significance of these results. PM10 is not a single substance. It is composed of several substances. In this study, we used Sigma ERM-CZ 120 PM10. PM10 is composed of ionic components, metal components, and carbon components. Despite the fact that the two components of ERM® CZ120, arsenic and nickel, are commonly referred to as causal factors of cancer [27,28], it is yet to be seen whether these specific heavy metal substances significantly contributed to our results. Most of the studies on COSMIC mutations have looked at the association of a single substances and COSMIC mutations [29,30]. On the other hand, in this study, PM10, which is composite material, was used, and this may produce different results from those previously known. As COSMIC mutational signature 5 was confirmed in comparison between the group exposed to PM10 and the control group, PM10 can be considered as a candidate for causing COSMIC mutational signature 5. Unlike our results, previous studies show G:C → T:A transversions. The reason for this difference might be caused by the difference of the target gene and used material. They investigated the TP53 mutation and used a single substance such as 3-nitrobenzanthrone, benzo[a]pyrene [31,32]. However, we investigated the whole genome sequence and used a complex compound (PM10). Concerning kategis, chromosome 1 was found to be most localized in NCI-H358 and chromosome 7 in HCC-827. The genes expressed predominantly on chromosome 1 varied for each individual type of lung cancer [33]. As for chromosome 7, which is widely reputed for its significant relations with the initiation and growth of lung cancer [34], EGFR and MET, for this chromosome, were found to play a major role with regard to lung cancer [35]. Some genes of chromosome 7 can also be held accountable, to a certain degree, for the survival period of patients with lung cancer [36].There are several limitations in our study. First, we did not use normal cells in this study. Because there is an oncogene addiction to the driving mutation of a cancer cell, it could not be ruled out that COSMIC mutations were caused by its oncogene addiction. We used the cancer cell lines because normal cells were difficult to maintain in this study. Therefore, the number or type of mutations can be overestimated when using cancer cells as compared to normal cells. Second, this was not performed in one cell. This study did not proceed by the expansion of one cell. Some mutations may have occurred during the cell separation. Despite these limitations, many mutations were found in this study. Mutations occurring in both the HCC-827 and NCI-H358 cells were similarly observed and occupied a large portion of the total mutations. Considering these results, it can be concluded that PM affects SNVs. Third, the use of PM10 may be a limitation in this study because PM10 mainly settles in the upper airway. However, some studies describe the relationship between PM10 and lung disease [37,38], so this is unlikely to be a major limitation. Fourth, we used commercially available fine dust, which contains a lot of heavy metals. Therefore, the results may differ in some areas with different fine dust components.To sum up the results, kataegis was observed in chromosome 1 of NCI-H358 and chromosome 7 of HCC-827. Signature 5 was the dominant type in both cell lines. A considerable number of mutations in transcribed strands of genes, epigenetics, and genetic elements have been found. We suggest that these can affect life-control phenomena.In this study, PM10 exposure is illustrated as a causal factor that induces COSMIC mutational signature 5, which is found to have been present in many cancer types including lung cancer. This study needs further evaluation and confirmation through well-designed research.J.W.S. and S.J.K. designed the research (project conception, development of overall research plan, and study oversight). S.J.P., G.W.K. and S.Y.L. conducted the research (hands-on conduct of the experiments and data collection). D.K., W.J.H., I.B.J. and J.K. analyzed the data or performed bioinformatics analysis. S.J.P., G.W.K., S.Y.L., D.K., W.J.H., I.B.J., S.J.K., J.K. and J.W.S. wrote the paper. All authors read and approved the final manuscript.This study was supported by a National Research Foundation of Korea Grant funded by the Korean Government (NRF-2018R1D1A3B07048311) and the Konyang University Myunggok Research Fund of 2019.Not applicable.Not applicable.The data presented in this study are available on request from the corresponding author.Ethics declaration is not needed because we experimented with cell lines.The authors declare that they have no conflicts of interest.Localized hypermutation rainfall plot of the difference between the cell lines exposed to PM10 and control group in NCI-H358 (a), and HCC-827 (b). Each dot on the plot represents the physical genomic distance between each mutation. In NCI-H358, chromosome 1 appears to be most localized with 516 mutations, whereas chromosome 7 is most localized with 519 mutations in HCC-827. The figure was made using Mutalisk.Mutational signatures: The best mutational signature decomposition model came from NCI-H358 (a) and HCC-827 (b). Signature 5 was the dominant type of mutational signatures in both the NCI-H358 and HCC-827 cell lines. The figure was made using Mutalisk.Transcriptional strand bias: Transcriptional strand bias analysis results in the difference between the cell lines exposed to PM10 and the control group in NCI-H358 (a), HCC-827 (b) (** p-value < 0.05). In the NCI-H358 group, high enrichment of the C > T mutations were demonstrated in the untranscribed regions and the C > A mutations were also enriched in the transcribed regions. In the other group (HCC-827), high enrichment of the C > T mutations were identified in the untranscribed regions. The p-value is obtained by a goodness of fit test. The figure was made using Mutalisk.Pearson correlation coefficients and percentage of explained variance: Each Pearson correlation coefficient and the variance values show the computed differences between the cell lines exposed to PM10 and control group in NCI-H358 (a) and HCC-827 (b). Each Pearson correlation coefficient is computed between the read densities of 1-megabase bins and the mutation counts in the corresponding bins for each regulatory element. Bins without any mutation (=0) are excluded from the calculation, and mutations from both autosomal and sex chromosomes are included. The percentage of explained variance between the 14 elements and the mutations. The figure was made using Mutalisk.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01047.txt

@@ -0,0 +1 @@
+Patients with heart failure (HF) may not receive enough HF education from their clinicians throughout the course of the illness. Given that information is readily accessible on the Internet, patients with HF may seek HF information online. However, the relevance of online information for patients, the health literacy demand, and quality of the information is unknown. The purpose of this study was to compare the HF topics available online with topics HF patients perceived to be important and to evaluate the health literacy demand and quality of online HF information. The most popular search engines and a website that ranks the popularity of the websites were searched to identify websites with HF information. The health literacy demand and quality of the information were evaluated using the Patient Education Material Evaluation Tool for Print Materials and the DISCERN tool, respectively. First, the HF Patients’ Learning Needs Inventory (HFPLNI) was used to determine whether the websites included the 46 topics identified in this inventory. Patients with HF (n = 126) then completed the HFPLNI to rate the perceived importance on each topic. A chi-square test was used to compare the differences between the topics on the websites and those patients perceived to be important. Of the 46 topics, 39 were less likely to be included on the websites even though patients perceived that they were important topics. Information on the websites (n = 99) was not written could not be easily understood by patients and did not meet the overall health literacy demands of 58.0% and 19.8% of the patients, respectively. Only one-fifth of the websites were rated as fair to good quality. Online HF information had high health literacy demand and was poor quality with mostly generic HF information, which did not meet patients’ information needs. Websites need to be developed reflecting patients’ learning needs with low health literacy demand and good quality.With information readily on the Internet, it has become common practice for individuals to seek health information online [1,2]. People may seek online information to learn more about health conditions for themselves or others and/or to clarify information given by their healthcare providers [1,3]. Mobile health applications have also become widely available to support self-care with features that track activities and provide relevant information [4]. This information can influence patients’ decisions about their health behaviors [5,6].Given that heart failure is a chronic condition, patients with heart failure should adhere to recommended self-care regimens, which entail a number of life-long, complex activities [7,8]. To successfully engage in self-care, patients need a comprehensive understanding of heart failure, its symptoms, and self-care [9]. However, patients may not receive comprehensive heart failure information throughout the course of the illness because heart failure education is not a routine practice in many heath care systems, such as in South Korea. Therefore, it is not surprising that patients’ need for information about their condition and management strategies is not met by their healthcare providers.Another challenge for patients is the relevance of the information they find online. Although a vast amount of health information related to heart failure is available, much of it may be different from what the patients are searching for, even after they carefully choose relevant search terms. Even if patients find relevant websites addressing the topics they are interested in, the quality of the information might be inadequate or written at a health literacy level that exceeds the patients’ level. Cajita and colleagues (2017) found that the heart failure information found online written in English had only fair quality, and required readers to have a relatively high level of health literacy [10]. Orlowski and colleagues (2013) observed similar results about heart failure online information on 15 websites including limited credibility and readability [11].Clinicians are responsible for educating their patients by providing relevant information and suggesting additional helpful websites [12]. However, it is important to evaluate the relevance of the topics, quality, and literacy levels of online heart failure information so clinicians can recommend the best information. To the best of our knowledge, only one study, Cajita and colleagues (2017) has examined online heart failure information, but they did not evaluate the online information topics [10]. In addition, because an enormous number of new websites are created every day, their study results are unlikely to reflect current online heart failure information. Therefore, the purpose of this study is to systematically review online heart failure information. The specific aims are to (1) compare heart failure online information topics with topics patients with heart failure perceived to be important, (2) evaluate the required health literacy demand and the quality of online heart failure information, and (3) compare the health literacy demand and quality of online heart failure information among the websites.This was a descriptive survey study to systematically review online health failure information.To identify websites containing heart failure information, the following terms were searched: heart failure, congestive heart failure, acute heart failure, chronic heart failure, cardiomyopathy, and cardiac dysfunction. These search terms were entered into the three most popular search engines used in Korea (Google, Naver, and Daum). Potential websites for this study were restricted to the first ten pages of results for each search term. We also used a website (www.rankey.com) that ranks websites based on Internet users’ traffic over 12 weeks in diverse categories. To identify websites including heart failure information, we hand-searched websites identified from Rankey.com in the following six categories: (1) medical and health, (2) clinics, (3) academic or tertiary medical centers, (4) drugs, (5) western and oriental medicine, and (6) health management.Since the focus of this study was online heart failure information targeting patients with heart failure, the following inclusion and exclusion criteria were selected. Inclusion criteria were that the website (1) focuses on heart failure; (2) provides information about any aspect of heart failure; (3) is freely available without a log-in; and (4) is written in Korean. Exclusion criteria included: (1) unable to view due to technical problems after three attempts; (2) includes advertisements and other promotional materials; (3) targets health professionals or clinicians; (4) information about heart failure of animals; or (5) materials written by individuals without a clinical background (e.g., chatroom discussions).A pair of trained research assistants rated the content on each website using validated tools to assess the heart failure information. The heart failure topics were rated based on the Heart Failure Patients’ Learning Needs Inventory (HFPLNI). Health literacy demand was rated using the Patient Education Material Evaluation Tool for Print Materials (PEMAT-P). The quality of information was rated using the DISCERN tool. The supervising researcher along with the research assistants thoroughly read the user manuals of the PEMAT-P and DISCERN, and protocol for categorizing topics using the HFPLNI. Any unclear items were discussed until consensus was reached. To ensure rating consistency, the research assistants first rated the same five websites independently. Discrepant item scores were discussed and resolved with the supervising researcher. If their rating consistency did not reach a minimum of 95%, an additional five websites were rated until the target percentage for consistency was reached.Patients with heart failure were recruited from outpatient cardiology clinics affiliated with academic medical centers in Korea. Patients were eligible if they were 21 years or older, diagnosed with heart failure, and living independently. Patients were excluded if their condition was severe enough to be listed on the cardiac transplant list or had psychological or neurological conditions that could interfere with cognitive function (e.g., stroke).After obtaining Institutional Review Board approval from all sites, physicians referred eligible patients to the researchers. Trained research assistants explained the purpose and procedure to eligible patients, and obtained written, signed, informed consent from each patient who agreed to participate in the study. After consent, the participants were asked to complete the HFPLNI to examine patients’ information needs related to heart failure. In addition, patients provided demographic information.The HFPLNI was used to measure patients’ perceptions of the importance of 48 topics related to heart failure information [13]. Of the 48 topics, two topics were specific to hospitalized patients, so these two topics were removed. Each topic was rated on a five-point scale ranging from 1 (not important at all) to 5 (very important). The patients who were recruited from outpatient clinics were asked to complete the HFPLNI to measure their perceived importance of the 46 heart failure information topics. The Cronbach’s alpha coefficient of the Korean version was 0.97 in a previous study [14]. The trained research assistants also used this instrument to assess whether or not the websites included the topics listed in the HFPLNI (yes/no).PEMAT-P was used to evaluate the health literacy demand of the heart failure information on the websites [15]. PEMAT-P has two subdomains: (1) the level of understandability measuring how well the written material is understood by health consumers from diverse backgrounds with varying levels of health literacy, and (2) the level of actionability, which measures how well a health consumer is able to identify what they need to do based on the information presented. This instrument includes 24 items (17 items for understandability and seven items for actionability) with a binary scale (agree or disagree). Of the 24 items, 10 items were rated as not applicable (e.g., no numbers or visual aids were included in the material). The total scores were computed by averaging the items that were rated as “agree”, and then multiplying the result by 100. A score greater than 70% indicated that the material was understandable and actionable. The validity and reliability of the Korean version of the PEMAT-P were supported [16].The quality of the heart failure information was evaluated using DISCERN [17]. The DISCERN tool consists of 15 items to judge the reliability and quality of the information, as well as the overall quality. Each item was rated using a five-point scale: a score of 5 indicates that the item completely fulfills the quality criterion, scores of 4–2 indicate that the item partially fulfills it, and a score of 1 indicates that the item does not fulfill it at all. The overall quality rating ranges from 1–5, with 2 or below indicating poor quality with serious shortcomings, 3 indicating fair quality, and 4 or above indicating good quality. Based on the total DISCERN score from the 15 questions, the websites were grouped into the following categories related to the content: excellent (63–75), good (51–62), fair (39–50), poor (27–38), and very poor (15–26). The psychometric property of the Korean version was previously established [18].Data were analyzed using SPSS version 25 (Armonk, NY, USA: IBM Corp). Descriptive statistics including frequency distributions, means, and standard deviations were used to describe patient characteristics in addition to HFPLNI, PEMAT-P, and DISCERN scores. Patient ratings on each topic in HFPLNI were recoded to a binary variable to test whether important heart failure education topics identified by patients were significantly different from heart failure information topics found on the reviewed websites in this study. If patients’ ratings on the degree of importance were 4 (important) or 5 (very important) points, they were coded as “important”; if patients’ ratings were 1 (not important), 2 (somewhat important), or 3 (moderately important) points, they were coded as “unimportant.” Analysis of variance with Scheffe correction was used to compare the health literacy demand and quality of information among the types of website publishers.A total of 99 unique websites were included in this study (Figure 1). Among the 99 websites, seven were classified as government or professional organizations (e.g., Korean Heart Failure Society), 70 were affiliated with a hospital or clinic, 15 were commercial companies (e.g., pharmaceutical companies), and seven were physicians.The average age of patients (n = 121) was 59 years (SD 12.99) with a range from 25–85 years (Table 1). The majority of the patients were male, had a high school education and above, and were categorized as New York Heart Association functional class I/II, indicating no or minimal functional limitations due to symptoms of heart failure. All patients took at least one medication related to heart failure.The most frequently reported topics that the patients rated as important or very important were in the following order: the possibility of improvement of cardiac function (77.7%), general principles for taking medications (72.7%), actions to take when side effects of medications developed (72.7%), consequences of not following medical advice (72.7%), actions to take in case of worsening symptoms (72.7%) (Table 2). The least frequently reported topics that the patients rated as important or very important were in the following order: the presence of an available support group (38.0%), importance of sharing emotional distress (45.5%), when engaging in sexual activity is allowed (45.5%), strategies to fit in daily weight lifestyle strategies (52.1%), and contributors to the onset of cardiac disease (52.9%).Heart failure topics that were most frequently addressed in the reviewed websites included symptoms and signs of heart failure (97.0%), prognosis of heart failure (“what happens when someone has heart failure”) (97.0%), causes of heart failure (94.2%), risk factors of heart disease (84.5%), and function and anatomy of the heart (80.0%) (Table 2). However, none of the websites included information related to advice for family members in case of sudden death. A few websites addressed information about advanced directives (1.0%), where family members could learn about cardio-pulmonary resuscitation (1.0%), an available support group (1.0%), normal emotional responses to living with chronic illnesses (1.0%), and the importance of sharing emotional distress (1.9%).Of the 46 topics, 39 topics were less likely to be addressed on the websites although patients perceived that the information was important to learn (p-values < 0.05) (Table 2). Four topics that were more likely to be addressed on the websites compared to patients’ perceived importance to learn (p-values < 0.001) were as follows: symptoms of heart failure, anatomy and function of the heart, contributing factors to heart disease, and prognosis of heart failure.The overall mean PEMAT-understandability score was 58.0% (SD 15.1) ranging from 12.5% to 87.5% (Table 3). Over 90% of the websites included only information related to the main purpose (97.0%), displayed the information in logical order (91.9%), and grouped information into short sections (91.6%). However, only 6% of the websites included a summary of the content, and only 12% defined the medical terms they used. The overall mean PEMAT-actionability score was 19.8% (SD 19.9) ranging from 0.00% to 100%. In addition, 63.3% of the websites suggested at least one action that patients with heart failure could take. However, only 2.0% of the websites provided a concrete tool to help patients with heart failure take action or visual aids for patients to follow the instructions.The appraisal of the quality of the websites is summarized in Table 4 based on the DISCERN criteria. The total DISCERN scores indicated that 33.3% of the websites were rated as very poor quality, 46.5% were poor quality, 18.2% were fair quality, and only 2.0% were good quality. Of the 15 items, the mean scores of nine items were below 2, indicating that they did not fulfill the corresponding quality criterion. The three highest rated items were in the following order: achieved the aims if the websites clearly stated their aims (4.76, SD 0.73), provided additional support and information sources (3.46, SD 1.25), and provided shared decision-making support (3.20, SD 1.90). Items rated as the three lowest were in the following order: the effect of treatment choices on overall quality of life (1.30, SD 0.63), description of risks of each treatment (1.38, SD 0.82), and citations of the sources of information (1.49, SD 0.76).There were significant differences in website quality by website type (p-value = 0.009) (Table 5). Websites created by government or professional organizations had the highest DISCERN total scores, and were significantly different from hospital affiliated and commercial company websites (p-values < 0.05).Searching for health information on the Internet can help patients with heart failure learn about and manage their condition because routine comprehensive patient education is not always readily available in healthcare systems. For example, the average time clinicians spend with patients during a regular office visit is 20 min in the United States and 9 min in Korea [19,20]. A 60-min inpatient education session for heart failure patients is rare [21].Although online health information can be a good source for patients to supplement what they learn from their clinicians, we found evidence that patients’ expectations about the topics and quality of the information were not met. Our findings show that the reviewed websites mostly included generic information about heart failure, which was not the priority for most patients. In addition, the online heart failure information required a high health literacy level, and the information was rated as poor quality. Websites published by government or professional organizations provided heart failure information requiring relatively low health literacy demand and was rated as higher quality compared to websites by others (e.g., hospitals and commercial companies). The results of this study suggest that online information is not sufficient to fulfill patients’ learning needs, and much of the information is not written so patients can easily understand and take action to manage their condition based on high quality information. Our findings also echo the growing concern about the quality of much of the health information presented online.We found that online information did not meet the learning needs of patients with heart failure because there was a significant mismatch between what patients wanted to learn and the topics addressed in the websites. Similar results have been found in previous studies showing a disconnect between what topics and information clinicians perceived to be important compared to patients’ perceptions [13,22]. In our study, the topics patients ranked high tended to be related to self-care activities, and especially problems they might encounter such as strategies to use in case their symptoms became worse or they experienced medication side effects. However, these topics were rarely addressed on the websites. This finding implies that patients’ need to develop self-care skills to manage heart failure is not fulfilled online. It also highlights the problem of providing information without incorporating voices of the target population. Therefore, it is crucial for creators of online information to carefully assess what patients with heart failure want to learn.Both clinicians and patients have ranked psychological topics at the bottom in previous studies [13,22,23,24], even though psychological adjustment is an important part of living with heart failure. In our study, although more than 45% of the patients perceived that managing psychological issues was an important topic to learn, this information was rarely included on the websites (1–3%).One of the key results of this study is the scarcity of online heart failure information that is written in an understandable and actionable manner. The writing styles, organization, and use of visual aids and numbers were inadequate for patients to read and comprehend the information they needed. However, the most significant problem with the heart failure information from the websites we reviewed was that the information was not actionable, meaning that it was not written in a way that patients could apply what they read to manage their heart failure. This finding is consistent with previous studies showing that actionability of online health information including heart failure is a significant issue [10,25,26]. It is quite concerning that online information may not help patients engage in self-care, as patients with heart failure frequently experience challenges with self-care [27,28].We also found that the overall quality of the online heart failure information was poor in this study. Cajita and colleagues’ study also found that online heart failure information written in English was only fair quality [10]. Although the scores for most of the items in DISCERN were lower in our study compared to that of Cajita and colleagues, several items were scored low in both studies including items related to citations for the sources of information, the effect of treatment on quality of life, areas of uncertainty, and descriptions about the consequences of no treatment. The overall poor quality of online heart failure information is concerning because patients might not have the skills to evaluate the reliability and quality of the information. Therefore, it is necessary for clinicians to give clear guidelines about how to find credible online information.Individuals evaluate the credibility of online information based on a variety of factors, such as authority and credibility of the authors, recommendation by others, and references [29]. Although was it was beyond the scope of this study to identify the characteristics of websites with quality online information for readers with a lower health literacy level, we found that websites created by government or professional organizations provided better quality information that was more actionable compared to others. This finding is consistent with the findings of previous studies that have also found that government or professional organizations provided higher quality online information with an adequate level of health literacy demand [10,25]. Although the overall quality and actionability of the information created by government or professional organizations was not satisfactory in our study, guiding patients to search for these types of websites could increase the chance that patients will learn credible and actionable information.One of the limitations of this study is that the sample was from one country, Korea, and thus only websites written in Korean were included. This sample limits the generalizability of our findings to only online Korean information. In addition, our sample was not representative of the whole heart failure population because the majority of the patient in our sample were male in New York Heart Association functional class I/II.Another limitation of this study is the exclusion of non-Korean-language websites given our Korean sample. For example, health information on major international cardiology organizations’ websites (e.g., American Heart Association) have a dedicated section for patients and/or their caregivers and have been translated into a very limited number of languages [30]. However, these sites were not included in our study given the nature of our sample: Korean patients. Although we did not evaluate the quality and health literacy demands of the online health information on their websites, it can be assumed that the information is high quality with low health literacy demands since they are reputable and credible sources. Further investigations should evaluate if clinicians should inform patients about these international societies of cardiology as good resources of information.Some interactive websites also provide information based on patients’ situations (e.g., the duration of heart failure) and a mobile application to support patients’ self-care, but these sites and mobile applications were also not included in this study. Future studies should evaluate these applications given that the popularity of these mobile health applications has increased [3]. Several meta-analyses have reported that they are effective for improving self-care [31,32]. According to the study by Sohn and colleagues [33], more than 60% of patients with heart failure showed interest in using mobile health applications to support self-care. Thus, nurses and other health care providers could develop mobile health applications for patients with heart failure and use them to support patients’ self-care. The data from the applications could also be used when educating patients in follow-up visits. Health care providers should also include content based on patients’ learning needs and evaluate the quality and health literacy demands of the information when they develop these applications.In addition, although online videos (e.g., YouTube) and mobile applications are another popular online resource, we did not include these resources in this study. However, watching online videos might not be a common practice for older patients with heart failure who are searching for health information because a very small number of older people reported that they knew quite well about such video creation services [34]. We also did not evaluate whether the information on the websites was the most current, evidence-based information. Future research is needed to evaluate whether the online information reflects the most up-to-date evidence.Although the Internet is a popular source of health information for patients to learn about their conditions, we found, overall, that the available heart failure information online required high health literacy demand and had relatively poor quality. The sites also tended to include generic heart failure information, so the information did not adequately meet patients’ educational needs about their condition. However, government or professional organizations provided relatively better quality heart failure online information with higher actionability compared to other online information. Clinicians should inform patients with heart failure that the quality of online heart failure information can vary widely and direct them to websites published by government or professional organizations. Clinicians also need to suggest that online publishers (especially, government or professional organizations) include topics that reflect patients’ priorities when they are seeking information about heart failure. Future work is also needed to characterize what constitutes good quality online information with lower health literacy demand so clinicians can provide guidelines on how to search for credible online information.Conceptualization, K.S.L.; methodology, K.S.L.; software, K.S.L.; validation, J.Y.Y. and M.S.J.; formal analysis, K.S.L.; investigation, Y.M.C. and S.H.O.; writing—original draft preparation, K.S.L.; writing—review and editing, J.Y.Y. and M.S.J.; visualization, Y.M.C.; supervision, K.S.L.; project administration, K.S.L.; funding acquisition, K.S.L. All authors have read and agreed to the published version of the manuscript.This work was supported by the New Faculty Startup Fund from Seoul National University.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Samsung Medical Center (protocol code 2017-04-136 and 08.21.2017).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 authors declare no conflict of interest.HFPLNI: Heart Failure Patients’ Learning Needs Inventory.Search results using PRISMA flow diagram.Sample characteristics (n = 121).Comparison of topics perceived as important by patients and topics addressed on the websites.Health literacy demand ratings of online heart failure information (n = 99).Quality rating of online heart failure information (n = 99).Comparison of health literacy demand and quality of online heart failure information by website type.Note. Different superscripts indicate significant differences among the groups. Values are the mean (SD).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01048.txt

@@ -0,0 +1 @@
+The health crisis caused by COVID-19 has had a huge impact on the provision of physiotherapists’ services during the pandemic. Patellofemoral pain syndrome (PFPS) is a major health problem and one of the most common causes of pain in the front of the knee in outpatients. The objective was to evaluate the effectiveness of a therapeutic physical exercise (TPE) program supervised by a physiotherapist using telematic channels in reducing pain and disability in a sample of 54 patients with PFPS in the Physiotherapy Service of the San José de la Rinconada Health Center (Seville). Subjects were evaluated pre- and post-intervention (4 weeks—12 treatment sessions). An analysis was made of perceived pain—using the visual analog scale (VAS) and the DN4 neuropathic pain questionnaire—and functional balance—through the Kujala Score test and the Lower Extremity Functional Scale. The supervised TPE program in patients with PFPS produced a reduction in pain: VAS F1, 52 = 8.68 (p = 0.005) η2 = 0.14 and DN4: F1, 52 = 69.94 (p = 0.000) η2 = 0.57; and in Lower Extremity Functional Scale (LEFS) disability: F1, 52 = 19.1 (p = 0.000) η2 = 0.27 and KUJALA: F1, 52 = 60.28 (p = 0.000) η2 = 0.54, which was statistically significant (p = 0.000 for p < 0.05). Hence, the TPE program presented was effective in reducing pain and disability in patients with PFPS.The period of time covering the present study is unprecedented. On 14 March 2020, the Government of Spain decreed a state of emergency, thereby beginning a period of confinement in order to stop the spread of the SARS-CoV-2 virus. This situation forced face-to-face care of patients attending the Physiotherapy Rooms in Primary Care Centers on an outpatient basis to be cancelled [1]. From that precise moment, physiotherapists in the primary care basic teams began to think of alternatives in order to maintain physical activity in their patients. Therapeutic exercise routines would make it possible to maintain the therapeutic goals of the different programs being carried out depending on their pathology.It was a question of developing a labor of health education in the population being cared for by said physiotherapy rooms. Explaining the basic measures of prevention of the spread of the virus (hand hygiene, face masks, avoiding social contact, etc.) was necessary, as well as eliminating the techniques that would require contact, whilst at the same time providing self-management and self-care tools that matched the pathology of each patient, all of this while prioritizing telecare, recurring to telematic resources (telephone calls, emails, audio-visual resources, links to online material, etc.) [2]. Some studies have suggested a high social acceptance and confidence of patients toward telehealth in trauma care, especially for real-time diagnosis and remote treatment [3]. In addition, the specific e-Health and telemedicine programs implemented in the evaluation and treatment of musculoskeletal problems can reduce health costs. These programs may generate significant impact on patients living in rural or remote areas and increase adherence to treatment [3]. At the initial stage of the global pandemic declaration by the WHO, this was the only possibility of continuing to provide services and avoiding risks to health professionals. In this case, primary care physiotherapists, given the lack of personal protective equipment at the beginning of the pandemic, needed to suspend in-person services.Patellofemoral pain syndrome (PFPS) is a common musculoskeletal problem, characterized by pain in the front of the knee and tending to become chronic [4]. Although it affects the whole population, its incidence is greater in adolescents and young adults [5,6].PFPS has been associated with osteoarthritis of the knee and a high body mass index, but in a recent study, it was observed that this is not true in adolescents [7]. Patients usually describe an increase in symptoms on going up and down stairs, squatting, running, or sitting for a long time; since these activities increase compressive load forces in the patellafemoral articulation [8].PFPS continues to be one of the most common and challenging musculoskeletal issues facing physiotherapists and sports medicine professionals [9].Even though the pain associated with PFPS is characteristic, the cause of this pain remains unknown. Traditionally, PFPS has been related to damage of the articular cartilage; however, we know that articular cartilage is aneural [10]. There exist a variety of pathologies that can present signs and symptoms similar to PFPS, wherefore it is used to refer to all pain in the front of the knee [11].The lack of understanding of the etiology and pathology associated with patellofemoral pain and dysfunction is reflected in the great number of therapeutic options to deal with PFPS. Conservative treatments are common, above all initially, and physiotherapy is one of the commonest interventions used [9].Physiotherapeutic treatments often include the strengthening of the vastus medialis muscle of the quadriceps to promote active stabilization of the patella inside the trochlea of the femur, as well as manual therapy procedures, patellar realignment through taping, stretching, and therapeutic exercise [12,13,14]. Although these treatments seem to be based on theoretical reasoning, the evidence as to the effectiveness of these interventions is not well established [11].Therapeutic physical exercise (TPE) constitutes one of the most valuable tools in our therapeutic arsenal. Adequate prescription tailored to each patient depending on their pathology, prior state of health, and prior levels of physical activity is a demand that the health system and society places on our backs. This entails the configuration of the physiotherapist as an expert in TPE. They must supervise and optimize the TPE program, adapting it to each patient, monitoring the prior and aspirational levels of physical activity in terms of both quantity and quality. In addition, the physiotherapist must identify any psychosocial barriers that might hinder the adherence of patients to the prescribed TPE programs. In short, the physiotherapist becomes the guide of the patient who must take the reins of their health and take on a proactive role in the face of their pathology [4,15].TPE and the integration of individual and group health education, in the activity of the primary care physiotherapist, afford greater control of repetitive demands on physiotherapy in the medium-term. Hence, supervised TPE has been hailed as one of the most effective tools in resolving problems deriving from chronic conditions where the reduction of pain and disability are the main objectives to be achieved [16]. The purpose of this study is to evaluate the effectiveness of a TPE program in reducing pain and disability, to quantify the decrease in pain in the front of the knee, and to appraise the improvement in functional disability in patients with PFPS after treatment with TPE supervised by a physiotherapist via telematic channels.The present study is a longitudinal and prospective clinical trial conducted in the Physiotherapy Service of the San José de la Rinconada Health Center (Seville, Spain), dependent on the Andalusian Health Service. This study was supervised by the institutional ethics committee CEI University Hospital Virgen Macarena and Virgen del Rocio, with ethics approval number 41255cfcgab78019fga8501bb354b0ba4b15a804, registered in the Australian New Zealand Clinical Trials Registry: ACTRN 12619001457134, available at http://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=377666&isReview=true and found to be in agreement with the Declaration of Helsinki.The target population that this protocol was aimed at was made up of all patients diagnosed with unilateral or bilateral PFPS belonging to the health area of the Basic Health Area of la Rinconada (Seville), with over 36,000 potential users. Inclusion criteria were patients diagnosed with PFPS of over a 6 month evolution, aged between 18 and 70 years. The exclusion criteria were presenting cognitive alterations and having undergone conservative or surgical treatment of the affected knee less than 6 months previously.The convenience sampling method was used based on previous studies [17], an appropriate formula for the comparison of two proportions. The total number of patients who began the study was 60 (30 + 30), and only 6 patients (3 + 3) did not finish the study (due to dropping out or not answering the appointment or supervision calls). The final estimated sample size with a 10% dropout was equal to 27 individuals in each group (α = 0.05, β = 0.2). Eligible patients were randomly divided into two groups after providing written informed consent with the epidemiological data analysis program (EPIDAT) [18] program. All of them were Caucasian subjects. The first control group (C group) received indications via telematic channels from their physiotherapist following the Spanish Society of Rehabilitation and Physical Medicine (SERMEF) indications [17], whereas the second TPE group, as well as receiving the indications via telematic channels from their physiotherapist following SERMEF indications [19], also received telephone advice at least three times a week over the four-week period that was the object of the study.Measurements of the different variables of the study were made at two times, at the beginning thereof and after a four-week period, which coincided with the end of the intervention or treatment plan (12 home-based TPE sessions supervised by the physiotherapist via telematic channels).The dependent variables considered in the heart of our research were (i) the perception of pain by the visual analog scale (VAS) which consists of using a validated scale, 10 cm long (0–10), where subjects place themselves from (0) lack of pain to (10) the worst pain bearable; (ii) perception of neuropathic pain, established through the DN4 questionnaire [20]; (iii) functional disability, quantified using the Kujala Score [21] and the Lower Extremity Functional Scale (LEFS) [22]; and (iv) an anthropometric study of the knee was also performed [5].Interventions. Patients in the TPE group learned strengthening, endurance, flexibility, and an active range of motion exercises. A training program was created (Table 1) based on SERMEF indications [19]: above all, to activate the vastus medialis of the quadriceps muscle; to produce an authentic effective break to the lateralizing tendency of the patella during knee extension movements; as well as to work on the hamstrings, which are essential so that there is a balance of forces acting on the knee [12]; and to work on the tibia internal rotator muscles and strengthening of the triceps surae, thereby decreasing patellar pressure, owing to the posterior displacement of the lower extremity of the femur (synergy of quadriceps and triceps surae).Then, they received a pamphlet containing descriptions and pictures detailing the above exercises (Table 1). Patients were asked to continue these exercises three times a week for four weeks (total of 12 sessions). They were told to place a cold pack (or similar) on their knees for 20 min before each session. A physiotherapist was responsible for making contact with patients via phone call and email three times a week where patients explained their exercise program and consulted their doubts about the execution. In this manner, the specialist remotely monitored the progress of the exercises and maintained principles of daily activities and symptom improvements. The physiotherapist asked patients to carry out their exercises as instructed by the pamphlet they received.In the C group, patients performed the same exercises as the TPE group, after being instructed by the physiotherapist on the exercises to be done at home but with no phone call control.Pre- and post-treatment evaluations. In both groups, after acquiring demographic data, the intensity of knee pain was assessed with the VAS before the first session of treatments. Then, the KUJALA score, LEFS, and DN4 test were filled out by the physiotherapist to measure neuropathic knee pain, symptoms, and physical function throughout the day. Patients were also provided with X Index for evaluation of physical function. The KUJALA and DN4 scores were normalized ((acquired score/total possible score) ×100). In accordance with previous studies [17], participants were re-evaluated using the same scales at the end of four weeks to assess the effects of each treatment (Figure 1).The sample size calculation was based on the detection of (1) an improvement of 15% in self-perceived pain intensity [23] and (2) a difference of >9 points in LEFS score at inter-group comparison after the treatment [24] and >10 points in the Kujala score. Taking into account a one-tailed hypothesis, an alpha value of 0.05, a desired power of 80%, and a large effect size (f = 0.2), 26 participants were required per group of treatment (G * Power, version 3.1.9.2).The statistical analysis of the data was carried out using PSW Advanced Statistics (SPSS Inc, Chicago, IL) version 24.0. Data were reported as mean (standard deviation) and confidence intervals (IC 95%). Firstly, the normal distribution of variables was verified by the Shapiro–Wilk test after descriptive analysis. The Levene test was used to assess the homogeneity of variances. Linearity was assessed by bivariate dispersion graphics of residual values observed from the expected values. Comparisons between groups were made for demographic and clinical data of reference using Student’s t-test for continuous variables and Pearson’s chi-square test for categorical variables. All analyses followed the intention-to-treat principle, and the groups were analyzed as randomized.Differences in the measurements were detected by analysis of covariance (ANCOVA) to evaluate each dependent variable by age, height, weight, and body mass index, and as a variable factor the group (group TPE versus C). Eta square (η2) was used to calculate the effect size (small when 0.01 ≤ η2 ≤ 0.06; medium when 0.06 ≤ η2 > 0.14; large when η2 > 0.14). Statistical significance was determined at p < 0.05.A total of 60 subjects, aged between 25 and 67 years, were selected for the trial. After the inscription phase, the final sample included 54 individuals (n = 54), 26 men and 28 women; mean age (SD) was 51 years (10.1). Of the affected knees, 29 were right (54%) and 25 left (46%). Of the subjects, 24 had an educational level of primary studies (44%); 18, secondary (34%); and 12, university (22%). Of the sample subjects, 38 were employed in the service sector (82%) and 10 in construction (18%).In relation to the patient diagnoses: 23 with patellar tendinitis (44%), 21 with gonarthrosis (40%), 7 with chondromalacia patella (12%), and 3 with bursitis (4%) took NSAIDS and analgesics, while 28 subjects (52%) took nothing.Table 2 shows the mean values and standard deviation of the main variables, both as a whole and for each study group (experimental group and control group) and their statistical significance (p > 0.05), confirming that the groups had equivalent values at the beginning of the study.Although the knee extension variable was measured by goniometric measurement, as all the knees in the study presented a value = 0°, or full extension, it is neither presented nor analyzed in the present study.Table 3 includes the initial and final scores of the variables in the study, and the differences between measurements before and after treatment.Statistical significance was found in favor of the experimental group in the between-groups comparison and in the interaction in the covariables age, height, weight and body mass index, in both the perception of pain: VAS F1, 53 = 4.77 (p = 0.001) η2 = 0.33 and DN4 F1, 53 = 6.99 (p < 0.001) η2 = 0.42; and in the knee disability index: LEFS: F1, 53 = 4.38 (p = 0.002) η2 = 0.31 and KUJALA: F1, 53 = 10.58 (p < 0.001) η2 = 0.52; as well as, finally, in range of movement: Flexion: F1, 53 = 6.35 (p < 0.001) η2 = 0.50. No statistically differences were found between before- and after-treatment measurements regarding the use of basic analgesic drugs. Finally, it must be considered that no side effects were observed.In this study, we evaluated the effectiveness of a telematic TPE program, supervised by a physiotherapist, in reducing pain and disability in patients with PFPS. In all participants, a significant improvement was observed in pain intensity, range of movement, and disability from baseline to four weeks of treatment, according to VAS, LEFS, DN4, and KUJALA indices.According to the most recent systematic reviews on the approach to PFPS, TPE is by far the most used therapeutic alternative [24]. Both due to its effectiveness as a treatment, and due to the ease of carrying it out in the home environment—needing few resources and given the situation of confinement that our target population was subjected to—practically no other alternative was left.In the abovementioned reviews, the primary option is pointed out as eccentric muscle work, demonstrating a more important improvement in functional balance, as well as patient satisfaction, than other alternatives [25]. This effectiveness would be related to the increase in production of collagen at the quadriceps tendon level and correct alignment of the new fibers synthesized, thereby improving their structure. Likewise, the formation of new vessels is produced, which would lead to an attenuation of the levels of pain perceived by the nociceptive nerve fibers present in the structure. Despite the numerous studies conducted, exercise type, frequency, and load are highly variable in the different studies included in the systematic reviews [26].This makes it absolutely essential to carry out more comprehensive studies in the future [6,27]. Meanwhile, isometric exercise has also been pointed out as one of the most interesting therapeutic alternatives to tackle PFPS. It is above all due to the analgesic effect it induces nearly immediately, which lasts for 45 min after being carried out, practically from the first few sessions. This analgesia is related to muscular inhibition at the cortical level. This makes isometric exercise a tool of great clinical usefulness. For this reason, we guide our patients to do this before and after carrying out the therapeutic exercise program, recommending that they leave the exercises that involve concentric or eccentric work or subject the knee joint to loads to the middle of the program. Thus, we make the most of this induced analgesia by using it to our advantage to boost the therapeutic adherence of our patients [4,27].One of the limitations of our study was the relatively short-term follow-up of four weeks due to the lack of adherence to the treatment of our subjects, which in our case was limited to three subjects (who did not finish the study) per group.Another limitation was the methods used for telecommunications, i.e., phone and email (and audiovisual resources and links to online material only if it was necessary) but not videoconference. It is obvious that by increasing the attractiveness of home-based programs, we can facilitate patient adherence to tele-rehabilitation [17].One of the strengths of our study was the fast adaptation of the physiotherapy service to the pandemic, with low resources available, so as not to interrupt the treatment of patients. Another of the strengths is the similarity between our participants’ characteristics and those of participants in other studies. This may help other researchers to achieve a common conclusion by reviewing these study results.The perceived pain in patients with PFPS after the application of a TPE protocol supervised by a physiotherapist via telematic channels was reduced on average by 1/10, measured using the VAS and DN4 scales.Application of TPE supervised by a physiotherapist via telematic channels was effective and produced changes in the degree of disability of patients afflicted with PFPS measured through the LEFS with an increase of 15 points, rising from 45 to 60.Measurement of this disability using the KUJALA questionnaire presented a mean reduction of 20 points, going from 49 to 69.Finally, it was determined that range of movement was increased by an average of 9°.It can be concluded, therefore, that therapeutic exercise supervised by a physiotherapist via telematic channels demonstrates greater effectiveness in pain reduction and disability in patients afflicted with PFPS than merely providing information sheets to the patient.Conceptualization, M.A.-C. and I.E.-P.; methodology, M.A.-C. and L.E.-A.; software, A.M.B.-Q.; validation, C.J.B.-Q., A.M.B.-Q. and M.d.l.A.C.-D.; formal analysis, M.A.-C.; investigation, C.J.B.-Q.; resources, C.J.B.-Q.; data curation, C.J.B.-Q., A.M.B.-Q. and M.d.l.A.C.-D.; writing—original draft preparation, I.E.-P.; writing—review and editing, I.E.-P.; visualization, L.E.-A.; supervision, I.E.-P.; project administration, M.A.-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 committee CEI UNIVERSITY HOSPITAL VIRGEN MACARENA and VIRGEN DEL ROCÍO, (with ethics approval number 41255cfc9ab78019f9a8501bb354b0ba4b15a804, dated 06/04/2019), and registered in the Australian New Zealand Clinical Trials Registry (registration number: ACTRN 12619001457134, dated 22/10/2019).Informed consent was obtained from all subjects involved in the study.The authors would like to thank all of the patients in the Physiotherapy Service of the San José de la Rinconada Health Center (Seville), dependent on the Andalusian Health Service, for their collaboration.The authors declare no conflict of interest. The authors alone are responsible for the content and writing of the article.Design and flow participants through the trial. PHCC, primary healthcare centre; VAS, visual analog scale; TPE, therapeutic physical exercise group.Training program. Patellofemoral pain syndrome basic plus stretches.Baseline characteristics of study groups.Data are reported as mean (SD); * between-groups statistical significance (one-factor ANOVA).Baseline, post-intervention, and mean score changes of knee pain and lower extremity function.LEFS: Low extremity functionality scale; DN4: Neuropathic pain in four questions; VAS: visual analogue scale. Data are reported as mean (SD) or (95% confidence level); * indicates statistically significant within-group differences (p < 0.05); ** indicates statistically significant within-group differences (p < 0.001); † indicates statistically significant within-group differences (p < 0.05); †† indicates statistically significant between-groups differences (p < 0.001).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01049.txt

@@ -0,0 +1 @@
+These authors contributed equally to this work.This study aimed to investigate the annual incidence of parapharyngeal and retropharyngeal abscess (PRPA) based on 10-year population-based data. Patients with PRPA were identified from the Taiwan Health Insurance Research Database, a database of all medical claims of a randomly selected, population-representative sample of over two million enrollees of the National Health Insurance system that covers over 99% of Taiwan’s citizens. During 2007–2016, 5779 patients received a diagnosis of PRPA. We calculated the population-wide incidence rates of PRPA by sex and age group (20–44, 45–64, and >64) as well as in-hospital mortality. The annual incidence rate of PRPA was 2.64 per 100,000 people. The gender-specific incidence rates per 100,000 people were 3.34 for males and 1.94 for females with a male:female gender ratio of 1.72. A slight increase in incidence rates among both genders over the study period was noted. Age-specific rates were lowest in the 20–44 age group with a mean annual incidence of 2.00 per 100,000 people, and the highest rates were noted in the age groups of 45–64 and >64 years with mean annual incidences of 3.21 and 3.20, respectively. We found that PRPA is common in Taiwan, males and older individuals are more susceptible to it, and incidence has increased in recent years.Parapharyngeal and retropharyngeal abscesses (PRPAs) arise at one of two adjacent anatomical spaces in the neck, often collectively termed as deep neck space abscesses [1]. These infections often develop as sequelae of upper respiratory infections and are potentially life-threatening because of the possibility of the bacterial invasion of the carotid sheath and critical structures within it (e.g., common carotid artery, internal jugular vein, and vagus nerve), potential for airway obstruction, and systemic sepsis [2]. The infection may easily spread to other contiguous spaces, especially to the “danger” zones, possibly leading to mediastinitis and death [3]. Patients may suffer from sore throat, trismus, dysphagia, odynophagia, stridor, dyspnea, hoarseness, and unilateral paresis of the tongue. Hospitalization for appropriate intravenous antibiotics and surgical drainage are imperative for a positive outcome [4].Despite the aforementioned impacts of PRPA and the high cost burden [5], there is a paucity of long-term population-based epidemiological data on the incidence and mortality of PRPA, especially among the Chinese population. Previous studies have documented incidence rates of PRPA, as well as age and gender distributions, longitudinal trends, and in-hospital mortality [2,5,6]. However, some of these studies have not been representative of the population due to small sample sizes and hospital-based study samples [2,6,7,8,9]. Therefore, this study aimed to estimate the incidence of PRPA among the Taiwan population including age and gender distributions, as well as longitudinal trends, using insurance claims data in Taiwan’s National Health Insurance (NHI) databases.We obtained data from Taiwan’s National Health Insurance Dataset (NHIRD)—published by the National Health Research Institute in Taiwan—covering a period of ten years (from January 2007 to December 2016). The NHIRD consists of curated administrative claims data from the Taiwan NHI program. The NHIRD includes claims data and registration files of over 99% of all Taiwanese residents (n = 23.72 million in 2015–2016). The NHIRD provides an excellent opportunity for Taiwan researchers to perform large-scale clinical epidemiology investigations on diseases.This study was approved by the Institutional Review Board of Taipei Medical University (TMU-JIRB N202011013). This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines of research reporting standards [10].Because the hospitalization of all patients with PRPA is standard medical care policy in Taiwan, we used the NHIRD inpatient database to identify all patients aged ≥20 years with a principal diagnosis of PRPA (International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 478.21, 478.22, or 478.24 or ICD-10-CM codes J39.0 or J39.1) admitted between 1 January 2007 and 31 December 2016 to investigate yearly incidence rates and trends of PRPA. We identified a total of 5779 PRPA patients and categorized them into three 3 age groups: 20–44, 45–64, and >64.We obtained population census data from the Population Affairs Administration at Ministry of the Interior in Taiwan, which are publicly released annually, including age and gender distributions. We used the population data to calculate the national incidence rates per 100,000 people and by age and gender over the ten-year study period from January 2007 to December 2016.The annual PRPA incidence rate was calculated as the sum of new PRPA cases divided by the size of the total Taiwanese population in a year. Furthermore, yearly PRPA incidence rates per 100,000 people were calculated for 10 years, by gender, and for the age groups of 20–44, 45–64, and >64 years. We used a t-test to examine the difference in annual PRPA incidence rate between genders. In addition, we used the annual percent change (APC) and 95% confidence intervals (CIs) to quantify trends in annual PRPA incidence using a linear model to determine whether the APC was statistically significantly different from zero.A total of 5779 hospital admissions with PRPA occurred between 2007 and 2016, with a mean patient age of 52.16 years (standard deviation = 16.61) and 64.2% male. There were 33.7%, 42.2%, and 24.1% of PRPA patients in the age groups of 20–44, 45–64, and >64 years, respectively. The majority of PRPA patients (48.8%) resided in Northern Taiwan, and only 2.1% were in Southern Taiwan. In-hospital mortality among the study patients over the 10-year period was 2%. Table 1 presents the demographic characteristics of the study patients.The mean annual incidence rate of PRPA over the 10-year period was 2.64 per 100,000, varying between 2.43 and 2.81 in the first five years and between 2.47 and 3.71 in the later five-year periods (data not shown in tables). However, there was no statistically significant increase in annual incidence rate of PRPA over the study period with an APC of 1.26% (95% CI = −13.28–18.25%).Gender-wise annual incidence rates are shown in Figure 1. The t-test revealed that incidence rates among males were consistently and statistically significantly higher compared to females (p < 0.001). Gender-specific rates over the study period were 3.34 per 100,000 for males and 1.94 per 100,000 for females, with a male to female gender ratio of 1.72. There was a slight increase in PRPA rates among both genders over the study period.The age-specific incidence over the study period is presented in Table 1. As expected, the age-specific rate was lowest in the 20–44 age group with an average annual incidence of 2.00 per 100,000 over the study period. Incidence rates among the 45–64 and >64 years age groups were similar (3.21 and 3.20, respectively). Figure 2 presents year-wise incidence rates among each age group, showing consistently higher rates among the older age groups compared to distinctly lower rates among younger adults.Reliable population-based incidence data on PRPA including long-term trends and age and gender distributions have not been published because most available data are from hospital-based case series and therefore not representative. To our knowledge, this was the first population-based survey to investigate the incidence of PRPA using Taiwan’s NHIRD, which includes data on the entire population of Taiwan.Our study showed a higher mean age of occurrence of PRPA (52.16 years old) than the published literature, which has shown mean ages of 44–48.5 years [5,6,7,11,12]. Our finding of a consistent male preponderance in every study year was in line with previous studies [2,7,11]. We also found incidence differences between age groups that were observed consistently each year over the 10-year study period (Figure 2). Our estimate of an in-hospital mortality rate (around 2.0%) was in line with previous studies that examined mortality rates (reporting rates between 0.3–3.4% of all deep neck infection) [5,6,9,13,14,15,16,17,18]. However, one study from Germany reported no mortality of PRPA in their cohort [19].Our study showed an annual incidence rate of PRPA of 2.43–3.17 per 100,000 adult people during 2007–2016, which was higher than similar population-based studies. A study from Germany showed a much lower PRPA incidence, 1.32–1.94 per 100,000 people with an incidence of 1–2 per 100,000 above 15 years old during a comparable study period [19]. A Scottish study using the United Kingdom’s National Health Service (NHS) data of 2012 and 2016 that excluded patients aged under 16 years reported annual rates ranging between 0.55 and 2.45 per 100,000 in 2012–2016 for deep neck infections after including infections in the submandibular space (10.8%) and multiple spaces (12.6%) [5]. Another NHS study based on data from England for 1991–2011 estimated PRPA rates of 3–5 per 100,000 with the inclusion of pediatric patients [20]. The discrepancies in incidence rates between our study and other studies may be due to the included age groups, geographic regions, ethnicities, and/or life styles. Notably, both the British NHS and Germany’s health system have a universal health care coverage system with full access to high-quality medical care to all citizens without financial barriers, similar to Taiwan. The higher incidence rate in this study compared with other countries addresses the importance of the disease entity, especially for health care workers’ early alertness. Furthermore, it might warrant further studies of the impact of health system in Taiwan.Interestingly, during the study period, we found an increasing incidence of PRPA, from 2.43 to 3.17 per 100,000 during our study period, with the increasing trend observed in all age and gender groups. Many recent studies have shown a similar trend [5,11,13,20]. Similarly, the study from Germany reported a 47% increase in the incidence rate of PRPA from 1.32 cases per 100,000 in 2005 to 1.94 per 100,000 in 2017 [12]. The Scottish study also found a similar rising trend for deep neck space infection. A US study demonstrated an increasing incidence rate from 2003 to 2010 with a peritonsillar abscess complicated by a retropharyngeal abscess [11]. Some authors have proposed that older age may be a risk factor for PRPA [6,11]; an aging population may result in increasing trends of PRPA among adults, which may account for our finding. Another potential factor is the trend towards increased tonsillectomy or uvulopalatopharyngoplasty rates in Taiwan [21,22]. Further studies are needed to explore the reasons for the observed increasing trend in PRPA incidence rates.Infections in retropharyngeal space are easily spread down to the mediastinum and then to the lung and heart, compressing the airway and invading both the carotid artery and jugular vein due to proximity. Widespread infections may easily cause septic shock and disseminated intravascular coagulopathy [23]. Therefore, retropharyngeal space infections are likely to develop life-threatening complications, making PPRP more lethal than deep neck infections [12]. Using a population-based design, our study was ultimately able to calculate an in-hospital mortality for PRPA of 2.0%. The discrepant mortality rates from studies of deep neck infections may be results of unavoidable selection bias brought by hospital-based designs.Several strengths in our study should be noted. The NHIRD provides an excellent opportunity for Taiwan researchers to perform large-scale clinical epidemiology investigations on diseases. Taiwan’s NHI offers a fully accessible and affordable health care system for every citizen, which limits selection bias in PRPA ascertainment due to socioeconomic status. Furthermore, the NHIRD consists of all claims of the entire Taiwanese population, about 23 million, including outpatient visits, emergency department visits, and inpatient admissions during 2007–2016. Our study identified the PRPA patients by ICD codes from inpatient admission claims. PRPA is a serious and emergent condition that requires patients to seek immediate medical help and hospitalization. The NHI, providing affordable and accessible health care across the nation, enabled us to gather almost all cases of PRPA. As a result, we believe that we included all PRPA patients, regardless of their socioeconomic status. The use of claims data also avoided recall bias and misclassification bias that may result from studies based on self-reported data. However, there have only been three prior population-based studies which used ICD code as a diagnosis for PRPA [11,13,20]. Other such studies have all been hospital-based with findings that may not be generalized to the whole population.The study had some limitations. Claims data do not have medical data on disease severity and the results of physical examinations and diagnostic procedures, characteristics of the abscess, family history, genetic parameters, lifestyle, diet, and tobacco and alcohol usage, all of which may have a bearing on its incidence. In addition, the diagnoses of PRPA could have been made by physicians from various subspecialties with potential for mistaken diagnoses. Further, some patients with PRPA may have been diagnosed as cellulitis or acute lymphangitis of the face and neck (L03.2), cellulitis and acute lymphangitis of the neck (L03.22), cellulitis of the neck (L03.221), or acute lymphangitis of the neck (L03.222), which may result in underestimates of incidence.Our study suggests a higher incidence rate of PRPA in Taiwan compared to that in other countries, as well as an increasing incidence of PRPA in recent years, thus warranting further research into the possible causes such as an increased amount of performed tonsillectomy/uvulopalatopharyngoplasty procedures or lifestyle changes.Conceptualization, T.-H.Y. and H.-C.L.; methodology, H.-C.L.; validation, H.-C.L. and Y.-W.K.; formal analysis, Y.-W.K.; data curation, H.-C.L.; writing—original draft preparation, T.-H.Y., C.-S.W., Y.-F.C., H.-C.L.; writing—review and editing, S.X.; project administration, Y.-F.C.; funding acquisition, Y.-F.C. All authors have read and agreed to the published version of the manuscript.This study was supported by grants from Taipei Veterans General Hospital (V108C-145).This study was approved by the Institutional Review Board of Taipei Medical University (TMU-JIRB N202011013). This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines of research reporting standard.Not appliable.The National Health Insurance Research Database, which has been transferred to the Health and Welfare Data Science Center (HWDC). Interested researchers can obtain the data through formal application to the HWDC, Department of Statistics, Ministry of Health and Welfare, Taiwan (http://dep.mohw.gov.tw/DOS/np-2497-113.html).The authors declare no conflict of interest.Annual incidence rates of parapharyngeal and retropharyngeal abscess in Taiwan by gender during 2007–2016.Annual incidence rates of parapharyngeal and retropharyngeal abscess in Taiwan by age group during 2007–2016.Demographic characteristics of 5779 patients diagnosed with parapharyngeal and retropharyngeal abscess (PRPA) in Taiwan during 2007–2016.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01050.txt

@@ -0,0 +1,21 @@
+This study examines the impact of a newly developed structured training on Singapore paramedics’ psychological comfort before the implementation of a prehospital termination of resuscitation (TOR) protocol. Following a before and after study design, the paramedics underwent a self-administered questionnaire to assess their psychological comfort level applying the TOR protocol, 22 months before and one month after a 3-h structured training session. The questionnaire addressed five domains: sociocultural attitudes on resuscitation and TOR, multi-tasking, feelings towards resuscitation and TOR, interactions with colleagues and bystanders and informing survivors. Overall psychological comfort total (PCT) scores and domain-specific scores were compared using the paired t-test with higher scores representing greater comfort. Ninety-six of the 345 eligible paramedics responded. There was no statistically significant change in the mean PCT scores at baseline and post-training; however, the “feelings towards resuscitation and TOR” domain improved by 4.77% (95% CI 1.42 to 8.13 and p = 0.006) and the multi-tasking domain worsened by 4.11% (95% CI −7.82 to −0.41 and p = 0.030). While the structured training did not impact on the overall psychological comfort levels, it led to improvements in the feelings of paramedics towards resuscitation and TOR. Challenges remain in improving paramedics’ psychological comfort levels towards TOR.In suitable patients, field termination of resuscitation (TOR) by paramedics has been shown to be effective in reducing conveyance of futile resuscitations to the hospital [1], conserving limited resources [2] and averting dangerous high-speed transports [3]. TOR protocols have been practiced in Canada and the United States of America [2] for more than 20 years, and there is a trend towards worldwide adoption of such TOR protocols. Despite this, prehospital TOR protocols have not been adopted in many Asian countries for various reasons. These include differences in the sociocultural context surrounding death [4] and legal barriers [5]. The “rescue culture” mentality and the perception of death as an unsuccessful outcome of many prehospital emergency medical services (EMS) [6] also limit the uptake for any TOR protocol.There are inherent challenges to the application of the TOR protocol itself. Applying TOR protocols involves the fulfilment of one pre-requisite and two components [5,7] (Figure 1). The pre-requisite is the performance of cardiac resuscitation on the patient, while the two components include assessing for suitability of TOR application based on pre-determined criteria and the breaking of bad news to the patient’s loved ones. The complexity of the pre-requisite and the two components may be a source of stress to the paramedic. The negative impacts on practitioners such as stress and burnout that arise from death pronouncement and breaking bad news have been well documented [8]. Prior research showed that terminating resuscitation poses difficulties to paramedics from external factors such as perceived societal roles of paramedics [9] and challenges in executing the protocol [10] and internal factors such as those governed by personal emotions and beliefs. Furthermore, the lack of formal training in breaking bad news has been a barrier to the uptake of the TOR protocol [11].The five key domains that make up the construct of paramedics’ psychological comfort with field pronouncement of death have been previously described and validated in Morrison’s questionnaire [12]. These domains include sociocultural attitudes on resuscitation and TOR, multi-tasking, interactions with colleagues and bystanders, feelings towards resuscitation and termination of resuscitation, and informing survivors [12].The pre-requisite of training for cardiac arrest scenarios is common and fundamental to most EMS training systems in the world [13]. However, there has been a lack of focus in Singapore EMS training on the latter two components addressing the communication and the education in applying the TOR protocol (Figure 1). Training to improve communication skills and confidence in breaking bad news has been shown to be essential [14] and previously described in a single pilot study to have the potential to improve paramedics’ psychological comfort with TOR [11]. The second component of TOR training addresses the understanding of the basis of the TOR protocol: the science behind TOR criteria and the assessment of patient suitability. This gap in TOR training has not been previously described in the literature.A 3-h structured training session was designed to focus on the latter two components of TOR training as the pre-requisite had been covered in the core syllabus of Singapore paramedic education. This study is the first in Singapore to quantitatively assess the impact of structured training involving coaching in breaking bad news and the education of the basis of TOR in improving paramedic’s psychological comfort. The study was conducted in Singapore, a Southeast Asian island city-state nation of 5.7 million (2019) people [15], with approximately 345 certified paramedics from the Singapore Civil Defence Force (SCDF) and private operators responding to EMS calls from a single national emergency number [16]. There were 191,468 EMS calls made in 2019 [17] and 5094 deaths that occurred in residence in 2018 [18]. Prehospital care is well established with a tiered emergency medical service (EMS) framework, and paramedics are from various ethnic, social, cultural and religious backgrounds. The paramedics are fluent in English and undergo an 18-month training in paramedicine before licensure and practice. No patients or members of the public were involved in the design, conduct, reporting or dissemination plans of this research.This is a before and after study of paramedics’ psychological comfort (Figure 2 and Figure 3). The pre-training test was conducted 22 months before a 3-h structured training session on TOR (Figure 2). The post-training test was carried out one month after the structured training session.The 3-h training session had two pedagogical objectives:(a)To impart skills in breaking bad news;(b)To provide scientific evidence for the TOR protocol.To impart skills in breaking bad news;To provide scientific evidence for the TOR protocol.Results from a baseline questionnaire [16] were used to guide the structured training. Emphasized domains included “feelings towards resuscitation and termination of resuscitation”, “sociocultural attitudes on resuscitation and TOR” and “informing survivors”. Various formats and delivery methods including online videos, lectures, role-play simulations and group learning sessions were utilized (Table 1). (* Summary of training content provided in Appendix A).The original Morrison questionnaire consists of 22 questions measuring the construct of “psychological comfort with field pronouncement of death” [12]. Responses to the close-ended questions in the Morrison questionnaire are designed as a 5-point Likert scale response questionnaire (5 = never to 1 = always). Morrison developed her 22-item questionnaire based on data from focus group discussions with paramedics and feedback reviews from 41 stakeholders. Six questions were added to the original Morrison questionnaire to address four possible additional predictors of psychological comfort: knowledge of survival probability, religious affiliation, the location of the patient and perceived trust of the family. These four factors are pertinent to the local sociocultural context in Singapore.The questionnaire (Box A1) had shown good construct validity and had been revalidated in the local setting with a good Cronbach alpha score (α = 0.896) and moderate test–retest reliability (r = 0.627) [16]. The psychological comfort total (PCT) score was the sum of the total responses to 28 questions with higher scores representing greater comfort. Both PCT score and individual domain scores were re-scaled into a percentage score (0 to 100), with a higher score representing greater comfort for easy interpretation.Sample size: We postulated that overall PCT scores would improve by a moderate effect size following the training. A sample size of 90 participants was required to detect a difference of 0.3 SD (moderate effect size) in the overall PCT score to achieve 80% power with a 5% (two-sided) level of significance [19]. The effect sizes of the five domains were calculated.Both the pre-test and post-test were conducted as self-administered questionnaires distributed to the paramedics during their compulsory monthly educational sessions. Informed consent was obtained as participation in the questionnaire was voluntary. Missing data were presumed to be at random (<5%) and were imputed using the median value of other questions in the same domain.The pre-test and post-test data were compared using the paired t-test and the 95% confidence interval (CI). A p value of less than 0.05 was considered statistically significant. No multiplicity correction was conducted considering the exploratory nature of the study.The analysis comprised the following assumed predictors of psychological comfort: personal characteristics (age, gender, religious affiliation, previous personal experience with death and dying); and professional experience (years of professional experience as a paramedic, number of out-of-hospital cardiac arrests responded to in the past year and employment type (public vs. private sector)). Further regression analysis was performed on the change in overall PCT percentage to the respondents’ characteristics.Ninety-six out of 345 eligible paramedics (27.8%) responded. Approximately half of the respondents were male, and a majority (85.4%) professed a religion. The mean experience of the respondents as practicing paramedics was eight years. Most participants (96.9%) had at least one death pronouncement in an out-of-hospital setting cardiac arrest in the previous 12 months (Table 2). Regression analysis of the change in overall PCT percentage to respondents’ characteristics demonstrated no statistically significant results apart from “age”, which had a p value of 0.049 (analysis provided in Supplementary Materials).Moderate effect size (0.356) was observed for the “feelings towards resuscitation and TOR” domain and the remaining domains yielded small effect sizes with their absolute numbers smaller than 0.3 (Figure 4). The overall PCT score at baseline and post-training did not show any statistically significant change: mean change = −0.58%, 95% CI (−2.73 to 1.53), p = 0.576 (Table 3). The PCT score in the feelings towards resuscitation and TOR domain increased by 4.77%, 95% CI (1.42 to 8.13), p = 0.006, while the PCT score in the multi-tasking domain decreased by 4.11%, 95% (−7.82 to −0.41), p = 0.030 (Table 3).This study showed an insignificant change in paramedics’ psychological comfort scores following a structured 3-h training as measured by the Morrison questionnaire. The statistically significant change in the psychological comfort based on the respondents’ age did not yield any meaningful impact as this was a paired longitudinal study where the post-training responses were compared with the initial responses. While the training improved feelings towards resuscitation and the need for a TOR protocol, there was a decrease in the multi-tasking domain PCT scores. When dissected, three out of five domains did not show a significant change. The varying levels of emphasis on different targeted domains may have contributed to this finding (Table 3). There was a lower level of emphasis on the “interaction with colleagues and bystanders” domain; therefore, it was unsurprising that this domain’s post-training score change was insignificant.The training had attempted to address the other two domains of “sociocultural attitudes on resuscitation and TOR” and “informing survivors”. Both domains relate primarily to the confidence in communicating with the patient’s family on field TOR and death notification. Studies have shown that breaking bad news training increased healthcare workers’ comfort in accepting death [11,20]. However, this analysis did not replicate such an outcome. Compared to other healthcare settings, Singapore paramedics face a different set of challenges such as a time-constrained communication due to operational demands, legal constraints in death pronouncement such as non-binding “do-not-resuscitate” orders and differing public expectations regarding transport to the hospital for further resuscitation [21]. In light of these circumstances, a single structured training session may have been insufficient to increase the paramedics’ confidence when communicating with families. The repeated questionnaire was performed before the actual implementation of TOR, limiting the ability to see the impact of structured training on actual practice.As expected, education on the evidence for TOR led to an improved score in the “feelings towards resuscitation and TOR” domain. Part of the difficulty in adopting such protocols has been the “rescue mentality” of EMS. It is demonstrated in this study that perceptions towards death and TOR could change when the scientific evidence behind the protocol is thoroughly understood. The training design via didactic lectures was able to achieve the transmission and acceptance of this information.The outcome in the multi-tasking domain was unexpected as the domain score decreased after the structured training. However, the p value in the non-parametric Wilcoxon signed-rank test was insignificant compared to the paired t-test. A statistically significant finding in this domain could be a false positive as a significant decrease in the score was not expected. Furthermore, the non-clinically meaningful effect size for this domain could explain the above finding.Questions in this domain centered around core paramedic skills such as application of automated external defibrillator pads. The study participants were reasonably experienced, and a majority (61.5%) had attended to six or more cardiac arrests in the past 12 months. It is unlikely that this finding represents a reduction in the actual core resuscitation skills. A probable explanation for this finding could be the cognitive overload experienced by paramedics in executing a complicated multi-step cardiac arrest protocol that now incorporates an additional component of TOR. Adding to the cardiac arrest protocol’s complexity, the need to remember and execute additional steps could have led to increased stress as participants feared that routine procedures were now comparatively neglected. Where appropriate to the setting, efforts should be made by policymakers to simplify TOR protocols. This finding deserves to be explored and replicated in further studies.Educational programs that focus on breaking bad news differ in length and format but share similar components such as didactic teaching, role-playing or simulation [22], group discussion [23] and viewing of instructional videos [24]. The literature suggests that such training programs have been effective in improving the delivery of news and practitioners’ confidence [25]. Beyond the format, this study highlights the unique needs of training for TOR. While the ability to break bad news is a key component, other factors such as multi-tasking and addressing feelings towards the TOR protocol need to be similarly addressed. There is no “one size fits all” approach to TOR training, and it is imperative to understand and clarify the learning needs so that training programs can address these needs accordingly.Prehospital educators ought to be clear-eyed regarding the utility of short single-session training. Some studies have suggested that continuing medical education may be more important than a single session for practitioners to gain the ability to internalize and execute the protocol smoothly [26]. Further training sessions should take into account feedback from paramedics and could take the form of specific case discussions and role-plays.Lastly, it is possible that scores could be affected by an active “run-in” period following implementation. With practice, the informing survivors and multi-tasking domains could see some improvement as paramedics gain experience and confidence in the execution of the protocol. Future studies should address the effect of a run-in period and evaluate the need for continual medical education.In summary, this study provides new information about the usefulness and limitations of structured TOR training and can inform settings that are considering introducing similar protocols. In the implementation of a new TOR protocol, factors other than paramedics’ psychological comfort should be considered, and these include the complexity of the protocol itself [7], public expectations of not dying at home [4] and legal barriers towards field death pronouncement by paramedics [5,27].The strength of the study lies in its longitudinal nature that allows a sequential observation of the changes after an educational intervention in Singapore paramedic participants. As mentioned in the discussion, this study was conducted before the actual implementation of the TOR protocol and the findings may change significantly once the paramedics have real-life experience with the application of the TOR protocol.The response rate (27.8%) was broadly in line with other studies of paramedics’ attitudes and perceptions on clinical trials [28,29]. Historically with all researchers who employ questionnaires, many studies are confronted regularly with the issue of non-responders and its impact on conclusion inferences. The low response rate in this study was a significant limitation and may predispose the questionnaire’s outcome to non-response bias. The sample size determination strategy adopted in this study partially mitigated the effect of non-response bias.Furthermore, the length of this study spanned almost two years with a one-year interval between the pre-test questionnaire and the training, and confounding factors such as the time sequence effect, the turnover of paramedic staff and the voluntary nature of the self-administered questionnaire may have influenced the response, thus leading to an unanticipated outcome.The structured training program is the first of its kind in Singapore and, to our knowledge, in Asia that addresses the gaps in the training of TOR application. This study has uncovered many contextual challenges and the limitations of such approach in improving paramedics’ psychological comfort levels towards TOR. While structured training did not significantly impact the overall psychological comfort levels, it led to improvements in feelings towards resuscitation and TOR of paramedics. TOR protocols may lead to cognitive overload, thereby leading to a reduction in confidence in multi-tasking.The inherent design challenges of a questionnaire, compounded with the training program’s heterogeneous nature and the study’s voluntary nature, limit a robust conclusion to be drawn. Nevertheless, this approach has shown a positive impact on paramedics’ psychological comfort in applying the TOR protocol. Further research should focus on this area to ascertain the effect after a run-in period. Other healthcare systems that wish to develop a similar training program should also tailor it to their local and cultural practice.The following are available online at https://www.mdpi.com/1660-4601/18/3/1050/s1.Authors C.B., D.R.H.M., S.A. and J.H.P. were involved in the concept and the design of the study, acquisition of data and writing the first and subsequent versions of the manuscripts. R.C.Y.C. and S.Q. were involved in data acquisition and revising the manuscript. M.E.H.O., S.Q. and M.G. provided guidance in the methodology, statistical support and revised the manuscript. All authors contributed to, and agreed on, the final submission. M.E.H.O. and S.Q. gave the final approval of the submitted manuscript. All authors have read and agreed to the published version of the manuscript.This study was supported by Alexandra Health Enabling Grant (Grant number: AHEG1902).The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the National Healthcare Group Institutional Review Board (NHG DSRB Ref: 2018/00728). 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.The authors declare no conflict of interest.Summary of Training Content.Serial Number: ______                                                                                                  Paramedic Survey on Termination of Resuscitation (TOR)
+            Introduction
+           The termination of resuscitation (TOR) protocol will be
+implemented nationwide soon for Emergency Medical Services. We would be grateful if you accept our invitation to participate
+in this survey. The survey questions aim to collect your views, feelings, and experiences
+implementing the TOR protocol as prehospital service providers. This survey will
+take about 20 min to complete. Participation is voluntary and completing the survey
+implies consent for the study team to use the data for research purposes. This study
+is conducted solely for research purposes and is not part of your routine training
+or assessment. Your identity and your answers to this survey will be
+kept strictly confidential. We seek your honest feedback. Your collaboration is
+very important. It is the intent of this study that the results may help all prehospital
+care providers to be comfortable and to provide optimal care and compassion when
+performing termination of resuscitation in the field. If there are any questions or comments regarding this
+survey, you may contact the principal investigator (Dr Desmond Mao) at 6602553. Please answer each question based on your personal experience
+as a paramedic and on what you feel would be the effect of the TOR protocol upon
+implementation. For each question, please circle or tick [√] your chosen
+answer when options are provided, or write your answer in the space provided. 
+            DEFINITIONS
+          For your convenience, these are some definitions of terms
+used:-A termination of resuscitation may include any of the following paramedic actions at the scene of death: failed attempt at resuscitation, contacting online medical control physician, possible pronouncement of death, informing the patient’s survivors of the death and completing the necessary paperwork.-Field pronouncement (obviously dead) allows withholding treatment in the presence of rigor mortis, lividity, decomposition or decapitation-Survivor(s) is/are family member(s) or friend(s) of the patient, present at the scene of the cardiac arrest, who is/are informed by the paramedics of the death of their loved one.-Emotional difficulty means something that is “psychologically” difficult, that is, it makes you feel stressed, anxious, depressed or emotionally upset (for example ‘sad or tearful or angry for your own loss’).A termination of resuscitation may include any of the following paramedic actions at the scene of death: failed attempt at resuscitation, contacting online medical control physician, possible pronouncement of death, informing the patient’s survivors of the death and completing the necessary paperwork.Field pronouncement (obviously dead) allows withholding treatment in the presence of rigor mortis, lividity, decomposition or decapitationSurvivor(s) is/are family member(s) or friend(s) of the patient, present at the scene of the cardiac arrest, who is/are informed by the paramedics of the death of their loved one.Emotional difficulty means something that is “psychologically” difficult, that is, it makes you feel stressed, anxious, depressed or emotionally upset (for example ‘sad or tearful or angry for your own loss’).Paramedic Survey on Termination of Resuscitation—2018.Pre-requisite and components of termination of resuscitation (TOR) training.Chronology of events. Figure is not drawn to scale.TOR study flow diagram.Effect size of the domains.Training with targeted domains.Respondents’ characteristics.Baseline and post-training psychological comfort total percentage in the various domains and the paired difference (n = 96).PCT: psychological comfort total item percentage ranges from 0 to 100%, with a higher percentage indicating greater comfort. The paired difference is obtained by deducting the pre-training PCT % from the post-training PCT %. A negative value in paired difference indicates a drop in psychological comfort level post-training. * p value for multi-tasking domain in Wilcoxon signed-rank test >0.05.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01051.txt

@@ -0,0 +1 @@
+The movements of the affected upper limb in infantile hemiplegia are slower and clumsy. This leads to a decrease in the use of the affected hand. The visual effect obtained using the mirror box and the observation of actions in another individual can activate the same structural neuronal cells responsible for the execution of these actions. This research will study the affected upper limb functionality in hemiplegia infantile from 6 to 12 years old after the application of two intervention protocols: observation action therapy and mirror therapy combined with observation action therapy. Children with a diagnose of congenital infantile hemiplegia will be recruited to participate in a randomized controlled trial with two intervention protocols during four weeks (1 h per/day; 5 sessions per/week): Mirror Therapy Action Observation (MTAO) or Action Observation Therapy (AOT). The study variables will be: spontaneous use, measured with the Assisting Hand Assessment (AHA); manual ability measured with the Jebsen Taylor Hand Function Test (JTHFT); surface electromyography of the flexors and extensors muscles of the wrist and grasp strength through a grip dynamometer. Four assessments will be performed: At baseline situation, at the end of treatment, 3 and 6 months after treatment (follow-up assessments). This study will study the effects of these therapies on the use of the affected upper limb in children with hemiplegia.Infantile hemiplegia is a subtype of infantile cerebral palsy (CP), characterized by the involvement of one of the halves of the body due to a brain injury. Its prevalence is one case per 1300 live births [1]. The upper limbs are affected more than the lower limbs. The affected hand has a deficit in proprioception and tactile perception, which hinders fine motor skills, generally those of the fingers and the strength exerted by them [2]. Movements in the affected upper limb are slower and clumsy and are accompanied by mirror movements (MM) which are involuntary symmetrical movements observed in the contralateral upper limb to the voluntary upper limb movements. These movements are temporally observed in the development of the healthy child and then disappear. When these movements are subsequently present, they are considered abnormal [3]. Mild MM are common in preschool children but often appear in children with hemiplegia [4]. Although the mechanisms of mirror movements are not fully understood, it is likely that they are due to abnormal organization of the development of the motor system. Thus, alterations in the maturation of the transcallosal pathways have been suggested [4,5]. Moreover, children with hemiplegia have deficits in the selective control of the affected fingers and a diminished use of them. [6]. Consequently, the use of the affected hand is reduced, which is commonly known as “developmental neglect”, affecting activities of daily living [7,8,9].Spasticity is defined as “a sensory-motor control disorder resulting from an injury to the upper motor neuron, which presents as an involuntary intermittent or sustained activation of the muscles” [10]. The spasticity includes the velocity-dependent increase in tonic stretching reflexes (muscle tone) and phase stretching reflexes (exaggerated tendon jerk) [11], as well as flexor and extensor spasms, flexor reflexes and altered motor control. Clinically, it is often difficult to differentiate the spasticity from the symptoms and signs caused by structural changes in the muscles [10,11].Children with spastic hemiplegia show significant restriction in their daily activities due to the limited active range of movement of the affected upper limb, specifically in the elbow, as well as the alterations in muscle activity observed by electromyography. Thus, the affected side shows more co-contractions than the non-affected side. Surface electromyography is a good evaluation system to assess the activation of the motor unit in children with hemiplegia [12,13,14]. The restriction on the child’s participation in the environment due to the non-use of the affected upper limb is caused by the lack of exteroceptive and proprioceptive information at cortical level, as a consequence of the brain injury [15,16]. Therefore, the deficits in the structure, as well as the decrease in the movement amplitude or the spasticity would not be the exclusively causes of the activity limitation [16]. This developmental restriction decreases the learning to use the affected upper limb, producing exclusive dominance of the unaffected upper limb. The child uses compensatory strategies to avoid using the affected side [16]. This non-use increases muscle tone, reduces active movement, affects the growth of the affected upper limb and reduces the strength of this arm [16,17].Increased muscle tone due to spasticity is often reduced using antispastic treatment (botulinum toxins, phenol…) [18]. However, these treatments do not necessarily improve upper limb functionality. An improvement in the quality of movement can appear, but it would not increase the spontaneous use of the affected upper limb. Therefore, it may be necessary to use treatments that promote the function and activity-participation to improve the functionality of the affected upper limb and its performance in daily life [19].Mirror therapy is based on visual stimulation [20]. During mirror therapy, a mirror is placed on the midsagittal plane of the person, reflecting the unaffected side as if it were the affected side [20]. Therefore, the movements of the unaffected side create the illusion of normal movements of the affected side [21]. The illusion of the mirror can prevent or reverse a learned non-use of the affected limb [22]. One of the advantages of this therapy is its use as a self-administered home therapy [21]. Bruchez et al. [23] have shown the efficacy of mirror therapy in children with hemiplegia aged 6 to 18 years, improving the quality of movement and perception of the affected upper limb through feedback visualization of the healthy limb [23]. The replacement of the visual information of the affected upper limb with a mirror reflection of the non-affected upper limb improves motor control of children with unilateral cerebral palsy or hemiplegia [24] and obstetrical brachial palsy [24]. In hemiparesis, visual feedback increases the excitability of the ipsilateral motor cortex, optimizes the congruence between afferent (visual) feedback and promotes an internal representation of motor movements [25].The positive effects could result from the coupling between the two upper limbs. This is supported by evidence that such simultaneous movements with both affected and unaffected limbs modulate interhemispheric inhibition, allowing the unaffected hemisphere to facilitate activation of the affected hemisphere. [26]. On the other hand, it has been shown that the observation of actions in another individual can activate the same neuronal structures, responsible for the execution of these actions, known as the mirror neuron network. From here appears the AOT, which stimulates the motor behaviour of the affected upper limb through the observation of sequences of systematic activities and their subsequent execution [27,28]. Transcranial magnetic stimulation (TMS), used while the subject was observing the actions, showed an increase in the excitability of the corticospinal tract, resulting in muscle contraction patterns very similar to those observed in action, favouring motor activity. Hence, as there is a greater corticospinal activation, there could be a greater cortical representation and therefore an increase in the spontaneous use of the affected upper limb in infantile hemiplegia [28].No literature describing research combining mirror therapy and action observation could be found. Their combined use could increase the perception of the affected upper limb and facilitate motor recruitment and quality of movement, through visual illusion. In addition, with action observation, the spontaneous use and participation of the affected upper limb in daily activities through repetition of observed tasks can be increased. The combination of mirror therapy and the action observation therapy (MTAO) could facilitate the quality of movement of the upper limb, improving its structure and function and providing greater cortical activation and spontaneous use, through the repetition of specific movements. Furthermore, it can be done at home using a single intervention protocol. Conducting the therapy in the child’s home can involve the child in repeated practice of tasks and involve the family in the treatment process [29]. It has been reported that parents of children with disabilities have higher levels of stress than parents of typically developing children [30], which may suggest that the home is a viable setting for implementing therapy with parents. Their collaboration in producing target settings for treatment programs that involve family values and priorities may decrease the parents’ stress level [31]. The application of MTAO has a great clinical implication, since it is a low intensity program, which can be carried out at home with the follow up of a therapist, reducing economic costs and improving the interaction family—child—therapist.The aim of the research will be to study the functionality of the affected upper limb in hemiplegia in children aged 6 to 12 years after applying two intervention protocols: MTAO and AOT”.The hypotheses are: (1) “Both MTAO and AOT improve the functionality of the affected upper limb, with the effect of MTAO being greater” (2) “MTAO increases the spontaneous use and the manual ability in the affected upper limb more than AOT”. (3) “MTAO improves the muscle activity of the affected hand and increases grip strength more than AOT”.A simple-blind (rater) randomized controlled pilot study will be carried out using two groups: AOT (control group) and MTAO (experimental group).The sample size calculation has been based on the projected effect of the treatment on the main outcome measure, the AHA. To detect an effect size of 1.40 [2], with a significant level of 0.05% and 80% power, a minimum of 10 participants per group will be recruited. To compensate for possible non-responses, 20% more participants will be recruited.The participants will be recruited in Spain. The recruitment will be from October 2019 to March 2021. To obtain the sample, a dissemination campaign has been carried out through social networks, the Spanish infantile hemiplegia association Hemiweb and different health professionals involved in the management of these patients, following these inclusion and exclusion criteria.Inclusion criteria:Congenital infantile hemiplegia.Aged between 6 and 12 years.Lack of use of the affected upper limb.Level I–III of the Manual Ability Classification System (MACS). Level I: handles objects easily and successfully; level II: handles most objects but with reduced quality and/or speed of achievement and level III: handles objects with difficulty; needs help to prepare and/or modify activities [32].Level I–III in the Gross Motor Function Classification System (GMFCS). Level I: can walk indoors and outdoors and climb stairs without using hands for support, can perform usual activities such as running and jumping and has decreased speed, balance and coordination; level II: can climb stairs with a railing, has difficulty with uneven surfaces, inclined or in crowds of people and has only minimal ability to run or jump and level III: walks with assistive mobility devices indoors and outdoors on level surfaces, may be able to climb stairs using a railing and may propel a manual wheelchair and need assistance for long distances or uneven surfaces [33].Congenital infantile hemiplegia.Aged between 6 and 12 years.Lack of use of the affected upper limb.Level I–III of the Manual Ability Classification System (MACS). Level I: handles objects easily and successfully; level II: handles most objects but with reduced quality and/or speed of achievement and level III: handles objects with difficulty; needs help to prepare and/or modify activities [32].Level I–III in the Gross Motor Function Classification System (GMFCS). Level I: can walk indoors and outdoors and climb stairs without using hands for support, can perform usual activities such as running and jumping and has decreased speed, balance and coordination; level II: can climb stairs with a railing, has difficulty with uneven surfaces, inclined or in crowds of people and has only minimal ability to run or jump and level III: walks with assistive mobility devices indoors and outdoors on level surfaces, may be able to climb stairs using a railing and may propel a manual wheelchair and need assistance for long distances or uneven surfaces [33].Exclusion criteria:Disease not associated with congenital hemiplegia.Low cognitive level compatible with attending a special education school.Presence of contractures in the affected upper limb affecting the functional movement.Surgery in the six months previously to the treatment.Botulinum toxin in the two months previously to or during the intervention.Pharmacologically uncontrolled epilepsy.Disease not associated with congenital hemiplegia.Low cognitive level compatible with attending a special education school.Presence of contractures in the affected upper limb affecting the functional movement.Surgery in the six months previously to the treatment.Botulinum toxin in the two months previously to or during the intervention.Pharmacologically uncontrolled epilepsy.Patients who meet the inclusion criteria will be referred by email from the recruitment centres to the researcher responsible for the project. An information meeting about the intervention protocols of both therapies will be held at the research centre with the involved families. They will then have to sign an informed consent form. Families who will accept to be included into the research will be trained in the execution of each therapy (AOT or MTAO) by an independent therapist for each therapy, teaching them how they should carry out the activities at home. In addition, they will receive a video with the activities included in the protocol. The treatment will be only initiate when the families and children will be confident about it. A weekly follow-up will be used to avoid any complications and increase the treatment adherence. The families will be requested to fill in a table with the execution time of each proposal activity in the demonstrative videos and how it is performed by the child and the behavior towards it. The follow-up with the families will be conducted online, reviewing all the problems in the activities, modifying them if it is necessary to improve the performance. The family is a key element in improving critical components (intensity, repetition, feedback) of the therapies established with the child [34]. Hadders-Algra et al. [34] state that a family-centred approach creates a richer and more varied range of opportunities by training the family to encourage the children to use the affected upper limb in their usual environment.Two protocols are designed to be carried out at home, each lasting 20 h, applied during a four-week period (1 h per day from Monday to Friday) at home. The protocols are created based on the previous studies [25,35,36,37]. For the AOT group uni/bimanual activities are designed including 15 sets of daily life upper limb exercises: eight sets for bimanual activities and seven sets for unimanual activities (Table 1).In the execution, the child will have to watch the video and execute the observed action in the best possible way. The observed action will be repeated for 4 min. All the activities included in the protocol will be observed and imitated during a total time of 60 min without resting periods. During the sessions, each child should sit on a chair with both arms on a table in front of a monitor screen placed 1 m away. The parent will sit on the child’s affected side and should encourage the child by giving verbal suggestions without making any demonstration.For the MTAO group, the first 15 min will be assigned to the mirror therapy and the remaining 45 min to the action observation therapy. In the first part of the mirror therapy, six activities to be performed with both upper limbs/hands symmetrically have been designed. Even if the affected upper limb cannot successfully complete the proposed action, the child will be asked to do it to the best of his or her ability. (Table 2).The action observation therapy will be carried out for 45 min (after mirror therapy), using the same activities as in the AOT group (Table 2) and the same positioning of the child, but in this group 3 min are allocated for the visualization and execution of each set without resting during the execution.Both protocols will include 60-min sessions. The MTAO protocol, begins with 15 min of mirror therapy, followed by 45 min of AOT with a duration of 3 min for each activity, in order to complete 60 min.The randomization method will be carried out using the Epidat v. 4.2 software (Consellería de Sanidade, Xunta de Galicia, 15703, Spain), which, after a simple randomization process and consecutive sampling, will divide the participants into two groups (TOT and MTAO) with a number of 10 participants per group. A numerical sequence will be obtained which will be kept in opaque sealed envelopes. The envelopes will be opened by a blind researcher, who will decide at random, according to the randomization sequence. Although subject recruitment has begun, due to the Coronavirus pandemic, the trial has been delayed until the follow-up of participants can be ensured.Four assessments will be performed. The first assessment will be focused on obtaining the data before the treatment and sociodemographic variables (including age and sex), i.e., in week 0 (baseline situation, immediately before starting treatment), whereas the second assessment will be conducted at the end of the treatment, i.e., in week 4 (a total dose of 20 h). Moreover, two follow-up assessments will be done: 3 month and 6 months after treatment in both protocols (Figure 1).Spontaneous use will be measured using the AHA v. 5.0 scale [38], which is validated for children with infantile hemiplegia and obstetric brachial palsy from 18 months to 12 years. Considering the age range of the participants in this study, the School Kids AHA v. 5.0 Scale will be used, a valid and reliable tool that includes 20 items, with a score from 1 (non-performance) to 4 (effective use) [39,40,41]. The play session will be videotaped and subsequently scored by a trained evaluator and blinded to the group assignment. The Rasch model provides equal interval measures in logits (logarithmic probability units) by converting ordinal rating scale observations into interval levels. In order to facilitate the interpretation of the results, the logit scale is converted into a user-friendly scale from 0 to 100 that continues to be based on the Rasch model and presents interval level data (i.e., AHA) [39].Manual ability will be assessed with the JTHFT. This standardized test of simulated functional tasks that quantifies the time to complete a battery of one-handed activities [34]. Activities performed with the paretic hand include turning cards, placing objects, simulating eating, stacking tokens, and manipulating empty and full cans. The JTHFT has shown excellent reliability (0.95–0.99) in children with hand disabilities. [42,43]. The session will be also videotaped and subsequently scored by a trained evaluator and blinded to the group assignment.Surface EMG (mDurance Solutions, Granada, Spain) will be used to analyze the muscle recruitment and coordination during maximum isometric contraction. The children will be seat comfortably in a chair. Both shoulders will be at a position of slight flexion and abduction. The elbow will be at flexion, the forearm at neutral position and the hand will be placed naturally on the table. To produce the maximum isometric contraction, children will be asked to grip a cylindrical wood piece. The surface electrodes (36 × 45 mm) will be attached to the skin of the flexor and extensor muscles of the wrist and being parallel to the muscle fibers direction. Previously, the skin will be cleaned using 70% alcohol and fine paper. Three trials of 10 s will be recorded, with a rest of 10 s between trials. Before collected the EMG signals, no contraction signals will be on the screen. The children must avoid intense exercise prior to the test to exclude the effect of residual fatigue. The affected hand will be evaluated first. The root mean square (RMS), integrated EMG (iEMG) and cocotraction ratio (=iEMG of wrist flexors/[iEMG of wrist extensor + iEMG of wrist flexors] × 100) will be calculated and analyzed [14,44,45].In addition, a grip dynamometer (Kern, Balingen, Germany) will be used to assess the grasp strength. The children will be seated in the same position as for the EMG evaluation. The involved hand will be assessed first and then, the non-involved hand. The children will be asked to grasp the dynamometer for 10 s. Three trials will be recorded with a resting interval of 10 s. The grasp dynamometry is a valid and reliable test in cerebral palsy children [46,47].Statistical analyses will be performed with SPSS v.24.0 (SPSS Inc., IBM, Chicago, IL, USA). Statistical significance will be set at p < 0.05. The test Shapiro-Wilk test will be used to study normality of the sample.A descriptive analysis of demographic variables will be performed using means and standard deviation for normal distribution variables, and medians and interquartile ranges for non normal distribution variables. The categorial variables will be described using frequencies and %.To evaluate statistically significant differences intergroup in the variables at baseline, the independent Samples T-test (normal distribution) or the Mann-Whitney U test (non normal distribution) will be used. To study the effects of the interventions on the outcomes measures at four time points, for normal distribution variables, a two-way (intervention × time) repeated measures ANOVA test will be conducted. The Bonferroni correction will be employed for pairwise post-hoc comparisons to further analyze significant interactions. The Greenhouse–Geisser adjustment will be applied to correct for the lack of sphericity (Mauchly’s sphericity test, p < 0.05) whenever is necessary. For non-normal distribution variables, the Friedman test will be used. Bivariate correlation coefficients will be estimated to examine the relationship between the subjects’ age, sex, or affected upper limb.The study was approved (Reference N: 420-20-24) by the CEU-San Pablo university Ethics Committee and will be conducted in accordance to the World Medical Association Declaration of Helsinki. The current Spanish data protection law will be complied with. Before the study began, the consent of the families and children’s caregivers will be provided. The data collected will be used exclusively for this research and anonymisation will be carried out. To ensure the confidentiality and privacy of the data, a numerical code will be assigned to each participant. The outcome measures will be associated with the numerical codes, and not with the personal data of the participants. The identifying list of the numerical codes will be kept by the principal researcher, using a password-protected file. The data of the outcome measures will therefore be anonymised. The principal researcher will be the only person to have access to the entire set of data, which will be kept in password-protected files. In both cases, they will be stored on computers protected by a security password. Under no circumstances will personal information be divulged.The aim of this research is to compare the effect of two intervention protocols (MTAO and AOT) on the functionality of the affected upper limb in infantile hemiplegia, from 6 to 12 years old.The motor acts are usually combined to form target-related actions. The same motor acts can be included in actions with different final objectives. It has recently been shown that the visual responses of a subset of the lower parietal and premotor neuron system, studied both during observation and in the execution of grasping acts in different actions (e.g., grasping to place something or grasping to eat), are modulated by the target of the final action (placing or eating) [48,49]. Therefore, during the observation of a person performing an action, the watcher codes the final objective of the observed action (corresponding to the person’s intention) and the objective of the motor act included in that specific action. Thus, the child can internalize the intentionality of the activity observed and the motor act performed. These findings suggest that AOT will increase the spontaneous use of the affected upper limb through observation and imitation of actions, and thus encourage the child’s participation within their natural context and in daily activities. When the proposed activities are targeted at a functional goal or represent everyday activities, this automatic use is encouraged within the everyday environment. Considering the reduction of selective movements and the difficulty of executing adequate movements, mirror therapy could help to improve movement quality in dynamics and manual dexterity measured through grasping [37]. In our study, improvements in the coordination of wrist flexor and extensor contraction and more fluid manual dexterity can be obtained. Therefore, the MTAO protocol could induce changes in the spontaneous use and in the quality of movement of the affected upper limb.This research has some limitations. Firstly, the lack of cross-cultural validation in Spanish of some of the measurement tools used, such as the AHA. In addition, a group with conventional intervention has not been included. As strengths, it highlights that it is the first research to evaluate the effect of mirror therapy combined with action observation therapy on upper limb functionality. In addition, the intervention protocols will be carried out at home, which means reducing economic costs and optimizing family resources. Finally, the results will improve clinical practice.The results of our randomized controlled trial will be disseminated to the scientific community through scientific (peer-reviewed) journals and congress. In addition, research results will be presented to stakeholders, including patients, families and physicians.The study of functional changes in the affected upper extremity in children with hemiplegia after the use of AOT or MTAO will help clinicians in the design and selection of appropriate therapeutic strategies to improve the functionality of the affected upper extremity. In addition, this will allow the home-based implementation of the interventions with the involvement of the family.Conceptualization, J.C.Z.-E. and R.P.-C.; study design, J.C.Z.-E., R.P.-C. and C.B.M.-C.; review of the literature, J.C.Z.-E., R.P.-C., M.C.-G., and P.B.-M.; writing—original draft preparation, J.C.Z.-E. and R.P.-C.; writing—review and editing, C.B.M.-C., M.C.-G. and P.B.-M.; project administration, J.C.Z.-E. and R.P.-C.; funding acquisition, J.C.Z.-E. All authors have read and agreed to the published version of the manuscript.This work was supported by Banco Santander, grant number FUSPBS-PPC22/2017 (IV Convocatoria de proyectos precompetitivos CEU-Banco Santander).The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the CEU-San Pablo university Ethics Committee (Reference N: 420-20-24).Informed consent was obtained from all subjects involved in the study.The authors would like to extend thanks to all enrolled children, their families and people who contributed to this study.The authors declare no conflict of interest.Recruitment, assessment and follow-up.Activities for Action Observation therapy.Activities for mirror therapy into MTAO group.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01052.txt

@@ -0,0 +1,2 @@
+A patient’s survival may depend on several known and unknown factors and it may also vary spatially across a region. Socioeconomic status, accessibility to healthcare and other environmental factors are likely to contribute to survival rates. The aim of the study was to model the spatial variation in survival for colorectal cancer patients in Malaysia, accounting for individual and socioeconomic risk factors. We conducted a retrospective study of 4412 colorectal cancer (ICD-10, C18-C20) patients diagnosed from 2008 to 2013 to model survival in CRC patients. We used the data recorded in the database of the Malaysian National Cancer Patient Registry-Colorectal Cancer (NCPR-CRC). Spatial location was assigned based on the patients’ central district location, which involves 144 administrative districts of Malaysia. We fitted a parametric proportional hazards model in which the spatially correlated frailties were modelled by a log-Gaussian stochastic process to analyse the spatially referenced survival data, which is also known as a spatial survival model. After controlling for individual and area level characteristics, our findings indicate wide spatial variation in colorectal cancer survival across Malaysia. Better healthcare provision and higher socioeconomic index in the districts where patients live decreased the risk of death from colorectal cancer, but these associations were not statistically significant. Reliable measurement of environmental factors is needed to provide good insight into the effects of potential risk factors for the disease. For example, a better metric is needed to measure socioeconomic status and accessibility to healthcare in the country. The findings provide new information that might be of use to the Ministry of Health in identifying populations with an increased risk of poor survival, and for planning and providing cancer control services. Cancer is a major health burden across the world, with over 14.1 million new cancer cases worldwide in 2012. Of these, around 1.35 million cases (9.6%) are new cases of colorectal cancer. The number of colorectal cancer cases is expected to increase by 80% by 2035, climbing to approximately 2.4 million new colorectal cancer cases and contributing to 1.3 million deaths worldwide [1]. A patient’s survival may depend on several known and unknown factors and it may also vary spatially across a region [2]. Socioeconomic status, accessibility to healthcare and other environmental factors are likely to contribute to survival rates [3,4].A number of studies have been carried out to investigate and model the spatial variation in survival for various types of cancer. For example, the spatial model of spatial variation in leukaemia survival in north west England [5], the spatial variation in prostate cancer survival in England [6] and an evaluation of the factors affecting the spatial variation in breast cancer survival in Queensland, Australia [7]. Each of these studies suggest that spatial variation in cancer survival exists.Spatial variation in the survival of colorectal cancer patients has been observed in several studies conducted in developed countries. A study observed disparities in survival across New Jersey among more than 25,000 people diagnosed with colorectal cancer between 1996 and 2003. The lowest survival rates were found mostly in economically deprived areas while those in affluent areas had longer survival times; lack of healthcare accessibility is assumed to be one of the key predictive factors here [8].Existing research has quantified the geographical variation in survival using a discrete-time multilevel logistic survival analysis for colorectal cancer patients [9]. This research, which included over 400 Statistical Local Area (SLA) regions in Queensland, Australia, found that patients had substantially lower survival rates in rural and deprived areas than patients in urban and affluent areas, after controlling for individual characteristics and cancer stage.In Spain [10], space-time trends of colorectal cancer mortality risk during the period from 1975 to 2008 were mapped by sex and by two age groups: the middle-aged (50 to 69 years) and the elderly (≥70 years). Their findings demonstrated spatial variation in mortality risk across the region by both sex and age group.Material deprivation and geographical accessibility to healthcare were found to influence survival in colorectal cancer in a study involving cases from three cancer registries in France and one cancer registry in England [11]. This study showed that both of the above factors were relevant to patient survival, but that the effect differed between the two countries. Material deprivation was significantly associated with cancer survival in England while lack of accessibility to healthcare lead to poorer survival in France. In Korea, a study found that low survival was associated with the late stage at diagnosis found in poorer socioeconomic group patients [12]. Despite the national screening program, those with the lowest socioeconomic status were significantly more often diagnosed with late stage cancer compared with the highest socioeconomic group (the ORs were 1.29; 95% CI; 1.03, 1.61) [12]. On the other hand, the inequalities of survival may also be explained by the fact that the lower socioeconomic status groups received less treatment [13]. The group with the lowest socioeconomic status had a 24% higher risk of death than that of the highest socioeconomic group(hazard ratio 1.24; 95% CI 1.16, 1.32) [13].The findings from these studies suggest that it is important to investigate spatial variation in cancer survival in Malaysia.There have been several previous studies that have examined the spatial distribution of colorectal cancer incidence in areas of Malaysia, and which have described variation in colorectal cancer incidence, but none have investigated the spatial variation in survival from this disease [14,15,16]. To our knowledge, no studies have examined the epidemiology of this cancer using spatial modelling, and in particular, none have extended research to include the whole of Malaysia or the whole Malaysian population.Our aim is to model the spatial variation in survival for colorectal cancer patients in Malaysia while accounting for individual and socioeconomic risk factors. We also aim to investigate how individual and socioeconomic factors might affect survival from colorectal cancer, adjusting for spatial location.Identifying the factors that influence the difference in survival rates across the region may help public health authorities to better plan healthcare delivery, and thus eventually reduce disparities in colorectal cancer survival in Malaysia.There are two national cancer registries in Malaysia: the National Cancer Registry (NCR) and the National Cancer Patient Registry (NCPR) [17]. Both are managed by the Ministry of Health with the NCR being administered by the Disease Control Division and the NCPR by the National Institute of Health [18].The NCR captures the data on diagnoses from all regions in Malaysia. Diagnoses are reported to a state registry and from there to the National Registry. However, reporting cases to the state registries from hospitals is voluntary and therefore it is not always completed. However, the NCR is not passive; it conducts active case finding and routine checks. Assessment of the completeness of registration in the NCR is difficult because it is not clear how many of the 165 main hospitals in Malaysia are sending records to the registry, or how accurately diagnoses are recorded even when they are sent [19].The NCPR collects data on registrations of cancer from participating sites. These participating sites include 34 hospitals that diagnose and treat cancer patients in Malaysia. The objectives of the NCPR are to describe the natural history of cancers and to determine the effectiveness of treatments, to monitor safety, and to evaluate access to treatments. The NCPR collects data on four cancers: colorectal cancer, blood cancers, breast cancer and nasopharyngeal cancers. The NCPR records diagnoses and collects clinical data on risk factors, treatments and patient outcomes. This makes the NCPR data useful for research into the effects of treatments and survival from cancers [20]. This study used data from the National Cancer Registry-Colorectal Cancer (NCPR-CRC) [17].Our data consisted of 4501 patients with histologically verified primary colorectal cancer diagnosed between 2008 and 2013 (ICD-10, C18-C20). After excluding patients without Malaysian citizenship, patients with negative age and negative survival time, there were 4412 subjects’ data available for analysis.There are many instances in this dataset where information was missing or incomplete, but these are recorded inconsistently with a variety of indicators such as “missing”, “not applicable”, “not available”, “unknown” or “NA”. These do not always match up with the categories defined for each variable in the data. We therefore decided to combine the various missing data as ”unknown” or ”not applicable” if these categories were stated as such in the patients’ form, or otherwise just as ”missing”. We also obtained advice from our data provider about the uncertain category recorded in the database to justify our decision regarding data categorization. For example, even though the variable ”tumour differentiation” has categories of ”not applicable”, ”not available” and ”missing”, most of the data in these categories did not tally with the data definition. Our data provider clarified this, and assured us that all data in those categories of this variable were actually just ”missing”.We collected exposure data on ethnicity as Malay and non-Malay. All others were coded as other because of the small numbers. Smoking status was categorised as: non-smoker, former smoker, active smoker and missing status. Education level was classified as “Nil”, “Primary”, “Secondary”, “Tertiary” or “missing”.Clinical data was recorded as three categories: yes, no and missing. However, cancer stage was recategorised to “Stage I”, “Stage II’, “Stage III”, “Not staged”, and missing. Since the “site of distant metastases” contains many categories, each of which only contained a small number of data points, we decided to combine all the metastases regardless of their specific site. Irrespective of where and when it has been detected, the “presence of distant metastases” was therefore reclassified into yes, no and missing. ”Tumour differentiation” was recorded as “well”, “moderate”, “poor” and “missing”.The treatment modalities were categorised into four types of treatment received by the patients. They are patients who underwent surgery alone, patients who underwent surgery followed by chemotherapy and/or radiotherapy, patients who underwent chemotherapy and/or radiotherapy and patients who got other alternative treatments or palliative care. Patients without information of the treatment received were recorded as an unknown group.The specific cause of death provided in the data was not verified and therefore could not be deemed reliable, so, we decided to perform the analysis on all-cause mortality. The data records whether each patient is dead or alive at the end of the study period; we relabelled each patient’s status as either dead or censored. The censored group are patients who were alive until the end of the study period as recorded in the database. We obtained ethical clearance from the Ministry of Health Medical Research Ethical Committee (MREC), Malaysia and the Faculty of Health and Medicine Research Ethics Committee (FHMREC), Lancaster University (Ref no: NMRR-15-311-24656(IIR).The spatial data involved at the district level analysis includes all districts in Peninsular and East Malaysia and was based on polygonal data. The analysis for Peninsular Malaysia and East Malaysia was done separately as they are physically separated. We therefore separated the shapefile of Malaysia into Peninsular Malaysia and East Malaysia.In order to conduct analysis at the district level, each participant was assigned to their correct district based on the town variable recorded in the data. Each of the 520 unique names of towns seen in the data belong to one of the 144 administrative districts in Malaysia, of which 87 districts are in Peninsular Malaysia and 57 districts are in East Malaysia.We included a measure of socioeconomic status [21] and the density of hospitals as covariates in our model. The latter was assessed as a proxy for access to healthcare facilities, the lack of which may delay patients seeking or obtaining medical attention early on and which may decrease survival. The socioeconomic status categorisation was based on information from a census in 2000 as a measure of socioeconomic status in Peninsular Malaysia [21]. The index has a positive or negative value. A more positive index for a particular area indicates that the facilities in that area go beyond basic needs, and vice versa for more negative indices [21]. This index was available for 82 of the 87 districts in Peninsular Malaysia, and we used the “autoKrige” function from the ”gstat’ R package to impute the value of the index for the remaining five districts [22]. The autoKrige function implements the technique of ordinary kriging, a method for smoothing spatial data and predicting values for new locations (in this case, the centroids of the districts with missing socioeconomic index).On the other hand, we created a proxy-measure for accessibility of healthcare using the estimated number of hospitals per unit area (the density), calculated by the “density.ppp” function from the “spatstat” package in R software. Then, we included this density as one of the parameters in the spatial survival model. To model the correlation in space, we used an exponential correlation function, using the distance between the centroids of regions to determine the correlation.We used the spatial survival model to analyse our spatially referenced survival data. We created our spatial survival model using the “spatsurv” package [23]. The spatsurv package uses parametric models for the baseline hazard function and correlated log-Gaussian frailties to model spatial dependence. The hazard function takes the following form:h(ti; ψ, Yi) = ho(ti; ω) exp{Xiβ + Yi},
+        where ho is the baseline hazard function, t is the observed time for the i th individual, Xi is a vector of covariate values for the i th individual, ψ = h(β,ω,η) are covariate effects, the parameter of the baseline hazard and the parameter of the covariance function of a spatially latent Gaussian field Y, respectively. Yi is the value of the field at the location of individual i.In the Bayesian model, a prior density for the parameter of interest was inputted and the data modified the prior by using the likelihood to arrive at the posterior.The spatsurv package uses a Markov chain Monte Carlo (MCMC) algorithm to perform Bayesian inference for the parametric proportional hazards model. The idea is to use MCMC to draw samples from the posterior and estimate the parameters in our model. The spatsurv package implements the Markov chain Monte Carlo (MCMC) inferential algorithms because although they are typically slower than approximate methods, such as those based on the Laplace approximation for example, they deliver an unbiased, joint inference for all model parameters and are relatively easily extensible to wider model classes with additional hierarchies [23]. We looked for evidence of satisfactory convergence and mixing in the MCMC chain by considering the mcmcplot of β, ω, η and Y. We compared plots of prior and posterior to check that our data were sufficient to allow identifiability of the parameters in our model. The model has been fitted using three different distributions for the baseline hazards: Weibull, Exponential and B-spline. The models were compared using the Watanabe-Akaike information criterion (waic) value.We plotted the posterior baseline and cumulative hazard for each model, as well as the spatial covariance function and correlation against distance, for Peninsular Malaysia and East Malaysia. The plots of the posterior spatial correlation function show how similar the hazard is across space, and how fast that similarity decays. A correlation plot with a fast drop (small Ø) shows that there is little dependence between the hazard and distance. On the other hand, if the correlation plot has a slow drop, this shows that there is strong spatial dependence in the hazard or the risk of death from colorectal cancer is highly associated with place. It means that even though there may be a large distance between places, the correlation in their hazard is high. The interpretation of Ø is that for distances over Ø apart, there is little dependence on space.We also mapped the probability that the covariate-adjusted relative risk exceeds certain thresholds. These plots represent the risk over space that is not accounted for by the covariates in our model. All analysis were carried out using R software.We produced two spatial survival models, one for Peninsular (West) Malaysia and one for East Malaysia. Table 1 shows the parameter estimates for the hazard of death from the covariates used in these models. The covariates with a credible interval of hazard ratio(HR) that are marked with an asterisk (*) are significantly associated with the hazard of death from colorectal cancer in our spatial survival model. After controlling for spatial location and socioeconomic factors, cancer staging still plays an important role in determining the risk of death from colorectal cancer in Malaysia. Patients diagnosed at Stage IV had six times (median 6.37, 95% CRI (4.34, 9.75)) and seven times (median 7.33, 95% CRI (2.99, 24.00)) higher risk of death from colorectal cancer in West and East Malaysia, respectively, than patients diagnosed at Stage I. Each year of increase in age led to a slight increase (median 1.01, 95% CRI (1.01, 1.02)) in the relative risk of death. Other factors that significantly affect survival in colorectal cancer patients in Malaysia were ethnicity, tumour differentiation and the presence of distant metastases. We included two parameters to represent the socioeconomic distinctions in the population; the density of the hospitals in each district and the middle-class household item index (socioeconomic index). However, the socioeconomic index was only available for Peninsular Malaysia. Both of these variables showed an association with a decrease in the risk of death as the value of the parameters increases, however neither result was significant.The table shows the hazard of death from colorectal cancer for the variables chosen in the spatial survival model. It is represented by the median HR and 95% credible interval.The spatial correlation function plot (Figure 1 and Figure 2) shows how similar the hazard is across space, and how fast that similarity decays.Figure 3 shows the risk map (leaftlet plot) for the probability of exceedance risk of hazard of 1.1 and 1.25 in Peninsular Malaysia. These plots show P[exp(Y) > 1.1] and P[exp(Y) > 1.25]. Three regions in Peninsular Malaysia were identified as having a higher probability that the hazard of death would exceed 1.1, that is, those with a probability greater than 0.75 of exceeding the stated hazard. The probability starts to lessen when we increase the exceedance threshold to 1.25.Figure 4 shows the risk map for the probability of exceedance risk of hazard of 1.25 and 1.5 in East Malaysia. East Malaysia has two states, Sabah and Sarawak. The areas with the highest probability of exceeding the stated hazard of death in East Malaysia are all located in Sarawak. There was one district in Sarawak, Limbang, that was highly likely to have a hazard of death exceeding 1.25. Upon increasing the threshold to 1.5, only one part of the Sarawak region still had a probability of exceedance of between 0.50 to 0.75, and none had a probability greater than 0.75.In this study, we investigated the survival model of colorectal cancer in Malaysia, and incorporated geographical location. Our findings show that there is spatial variation in survival prognoses, or the hazard of death, for colorectal cancer in Malaysia even after adjusting for individual-level and area-level covariates. Cancer staging, tumour differentiation and the presence of distant metastases have a significant effect on survival for colorectal cancer patients. However, we also found that Malays had a significant 26% higher risk of dying from colorectal cancer than non-Malays. An increase in age slightly increased the risk of death from colorectal cancer.We found that a high socioeconomic index in an area did not significantly affect the risk of death from colorectal cancer. Our findings were similar to those of [24] who found that the socioeconomic status of the population did not significantly influence outcomes in patients with colorectal cancer. In comparison, a systematic review by [25] found that socioeconomic status had a significant impact on survival of colorectal cancer, where the risk of death was greater among patients with low socioeconomic status. Regarding socioeconomic status, recall that our measure, based on the 2000 census, was aggregated to district level. Hence a potential explanation for not observing a significant effect may be due to the presence of ecological bias, that is, our measure of socioeconomic status did not pertain to individuals. In fact, considering that education is often directly related to socioeconomic status, our results do show evidence that higher socioeconomic status is protective. However, further research in this area is required, and a more finely-resolved spatial map of deprivation could help us to better identify this effect.It is possible that there is a positive association between our variable “education”, an individual-level variable, and our socioeconomic index, which is an area-level factor measure with education as one of its domains. We think that these two measures are not likely to be well correlated for the following reasons. The socioeconomic index includes a measure, at an area level, of the proportion of the population with tertiary education in a district (area level). Education is one of five domains by which the socioeconomic measure is comprised. We realize that education might drive socioeconomic status, but it is not the sole driver of socioeconomic status. The widely used UK Index of Multiple Deprivation (IMD) is comprised of seven domains, one of which is based on education, and the index measures small areas across England, called Lower-Layer Super Output Areas (LSOAs) [26].The education variable in the model represents the individual education level of the patients in this study, which is classified into five categories, nil, primary, secondary, tertiary and missing status. Thus, we think that the education and the SE index variable in the model represent different things. To see if this is likely, we checked if there was any correlation between the two variables but it was not significant with a correlation coefficient of ρ = 0.01.We noticed that there were patients that had changed their address after diagnosis, but we decided not to take the distance of the patients’ addresses to the hospital as our measure of accessibility to healthcare. Address at diagnosis is important as an “exposure” or proxy for unmeasured exposures, but we had no record of length of residence at the address at diagnosis, which leaves open the likelihood of misclassifying cases by exposure to place. For example, people move house or job for many reasons, but sometimes for health reasons. They may, for example, move nearer a hospital when ill, or away from an exposure when concerned and there is evidence that population movement can lead to misclassification in epidemiological studies, such as when a birth address is used in studies of birth defects [27]. Instead, we decided to look at how the density of the hospitals in the area affected the patients’ survival in our study.The relationship between survival and distance to the treatment hospital is not clear cut: it is not necessarily the closest hospital to patients that they will choose to go to seek treatment and for this reason, we used the smoothed density(number of hospitals per unit area) as a covariate in our model. We expected the coefficient of this covariate to be positive, since, in places where there is a greater concentration of hospitals, patients tend to get diagnosed quicker as early symptoms are recognised and acted upon. We found that greater accessibility to healthcare decreased the risk of death from colorectal cancer but the effect was not significant in our study, which contradicts a previous study reporting that lack of access to healthcare was significantly associated with being diagnosed at a more advanced stage of colorectal cancer [28], which we know adversely affects patient survival. Different ways of assessing accessibility to healthcare may influence the direction and significance of the effect. For example, the previously mentioned study measured the shortest time taken to travel to the nearest appropriate health facility, while our study measured the number of hospitals per unit area. Furthermore, other factors that could be mediators to the accessibility to healthcare, such as the transportation system, affordability of care and cultural barriers were not considered in our study.Three areas were found to have a high risk of death (where the risk of death > 1.1 relative to other areas in Peninsular Malaysia). They were Jeli district in the North-East, Kinta district in the North-West and Melaka Tengah located in the West of Peninsular Malaysia. In East Malaysia, Sabah had better survival compared to Sarawak as shown by the probability of exceedance risk maps. The reasons for these differences could be such things as delay in chemotherapy treatment [29] and comorbidities [30], which were not assessed in our study, and this would be an interesting direction for further work in this area.We have a limitation as Malaysia does not have a formal socioeconomic status instrument that is consistent across the population (for example, the UK have the Index of Multiple Deprivation, IMD) [26]. Currently, the Malaysian Statistics Department only produces data on income at the district level; data on income at finer levels does not exist. We hope that in future research, it will be possible to consider household income and employment status as additional indications of socioeconomic status.The spatial analysis method used here has several strengths. First it allows us to combine data from the individual level with data from aggregated levels and map our findings. It also allows us to estimate and map the effect of unobserved environmental confounders through the use of a spatially correlated random effect terms. We applied MCMC for our analysis as it delivers full joint inference for all model parameters. Some limitations to be considered is that this analysis assumes a particular model form and correlation structure for the spatial variation and that it can be difficult in practice to identify the parameters of the spatial process. Attaining good convergence and mixing of MCMC can be difficult without access to bespoke software but again this is not an issue for us. Despite these limitations, this is the first study to examine the variation in survival for colorectal cancer in Malaysia. Having controlled for the potential individual and area-level factors, we found there is variation in the survival or risk of death from colorectal cancer in the population.Our findings indicate there is wide spatial variation in colorectal cancer survival across Malaysia, after controlling for individual and area-level characteristics. Better healthcare provision and higher socioeconomic index in the districts where patients live decrease the risk of death from colorectal cancer, although the associations were not statistically significant in this study. To obtain reliable SES data from individuals would require a range of questions to be answered as part of the initial data entry form. This might seem an unnatural thing to try to elicit routinely as part of a colorectal cancer diagnosis by medical practitioners. Similar comments apply to our chosen measure of healthcare accessibility; again, this information would be better obtained from each patient.In the future, the following suggestions may improve the quality of data and its resulting analysis. Reliable measurements of environmental factors are needed to provide good insight into the effects of potential risk factors for the disease. For example, a better metric is needed to measure the socioeconomic status and accessibility to healthcare in the country. Ensuring complete, accurate and consistently recorded data in the National Cancer Registry database is vital for reliable analysis and meaningful results. Perhaps ascertainment of active cases across the country and also better training of staff that interact with the database is required in order to minimise the amount of missing data.This study will provide new input for the Ministry of Health to target the population with a high risk of poor survival in providing cancer control services and to enhance health activities that are cancer-related in order to improve survival in the population.Conceptualization, A.K.G., B.M.T. and T.K.; methodology, A.K.G., B.M.T. and T.K.; validation, B.M.T. and T.K.; formal analysis, A.K.G.; Data curation, A.K.G.; Writing—original draft preparation, A.K.G.; Writing—review and editing, A.K.G., B.M.T. and T.K.; supervision; B.M.T. and T.K. 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 Institutional Review Board(or Ethics Committee) of Medical Research Ethical Committee (MREC), Ministry of Health, Malaysia (Ref no: NMRR-15-311-24656(IIR).Not applicable. This research was used the data available from the National Cancer Patient Registry Colorectal Cancer(NCPR-CRC) database with no interaction with the human subject.Restrictions apply to the avaialability of these data. Data was obtained from the National Cancer Patient Registry Colorectal Cancer(NCPR-CRC) database with permission and ethical approval from Medical Research Ethical Committee (MREC), Malaysia (Ref no: NMRR-15-311-24656(IIR).The authors would like to thank the National Cancer Patient Registry-Colorectal Cancer (NCPR-CRC) Malaysia team for the data given for this research. The authors declare no conflict of interest.The figure above shows the plot of the posterior spatial correlation function for Peninsular Malaysia. It shows that cases within a distance of less than 17 km (Ø) had a high correlation of hazard in space. The correlation of hazard starts to decrease when the distance for cases is more than 17 km apart; this is supported by the Ø value shown in Table 1.The figure above shows the plot of the posterior spatial correlation function for East Malaysia. It shows that the cases within a distance of less than 45 km (Ø) had a high correlation of hazard in space. The correlation of hazard starts to decrease when the distance of cases was more than 45 km apart; this is supported by the Ø value shown in Table 1.The risk map for the probability of exceedance risk of hazard of 1.1 and 1.25 in Peninsular Malaysia: ℙ[exp(Y) > 1.1] (left) and ℙ[exp(Y) > 1.25] (right).The risk map for the probability of exceedance risk of hazard of 1.25 and 1.5 in East Malaysia: P[exp(Y) > 1.1] and P[exp(Y) > 1.25].Table of parameter estimates from the fitted spatial survival model.* significant estimates; ** CRI: credible interval. ** Median Hazard Ratio.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01053.txt

@@ -0,0 +1 @@
+(1) Objective: To describe men’s experiences as acute myocardial infarction sufferers from a social phenomenological perspective, a year after the event (2) Methods: The phenomenological interview was used to capture the participants’ discourse. The data were analyzed according to the theoretical methodological approach of social phenomenology. (3) Results: The discourse analysis of the content produced the following categories, set out according reasons “why”: personal biography, knowledge set, warning signs prior to the illness, experience at the intensive care unit, and rehabilitation process; and reasons “for”: expectations as regards the illness, health professionals, and future social life and work prospects. (4) Conclusions: Participants had not established a healthy condition one year after myocardial infarction, perceiving a very thin line between life and death. Personal biography influences the coping of the disease. They feel like the illness helped them to create new meanings and value of life. They envisage a future full of great restrictions and uncertainty. The results of this study have underlined the need to involve care at all stages of the illness: the physical and emotional dependence upon admittance at the intensive care unit, the need to be cured, the constant demand for information about the illness, the difficulties encountered upon returning home, uncertainty about the future, etc. All these moments indicate that proper nursing care adapted to the specific needs of each individual and their family members must be provided in order to help them to overcome all the stages involved in this process. It is necessary to individualize care because the sense of reality is common and universal, but the ways of expressing are subjective, and it depended on the totality of experiences accumulated throughout life.Heart disease has remained the leading cause of death at the global level for the last 20 years. However, it is now killing more people than ever before. The number of deaths from heart disease has increased by more than 2 million since 2000. The European region has seen a relative decline in heart disease, with deaths falling by 15% [1,2]. In Spain, cardiovascular disease is the main cause of hospitalization and death [3], although evidence of a drop in mortality over the last two decades [4] and a rise in morbidity [5] clearly indicates that the lethal nature of the disease has diminished.Individuals suffering from a heart disease, such as myocardial infarction (MI), may undergo major changes in their lives. Several research works have involved longitudinal studies about the perception of the illness over time [6,7,8]. Many men described a series of losses after the cardiac diagnosis; these include loss of physical strength, emotional health, depression, fatigue, and problems with physical functioning [9,10]. In the immediate aftermath of the MI, patients restructured and re-evaluated their attitudes towards self, life, religious beliefs, and others [7,11,12].This study forms part of the field of phenomenology as a structure of reference and, in keeping with our main aim, is based on a social phenomenological approach that will enable us to discover and understand the situation that people experience after suffering a heart attack.One study attempted to incorporate the phenomenological method into the social sciences based on the architecture established by Husserl, proposing an organized knowledge of social reality as a primary objective [13].For social phenomenology, people live in the world in natural attitude. They have the ability to intervene in this world naturally, influencing and being influenced, transforming themselves continuously and changing social structures [13].The aim was to understand the world in relation to others in the realm of intersubjectivity, analyzing social relationships as mutual relationships that involve people. The fundamental purpose is to understand certain social actions that have a contextualized effect on shaping the social, and not just individual aspect.Action is interpreted by the subject from their existential motive, derived from the experiences recorded in subjectivity, constituting leading wires of the action in the social world. Those relating to the achievement of objectives, expectations, and projects are called “reasons for”, and those based on history, body of knowledge, and experience within the biopsychological context of the person are called “reasons why”. The set of reasons for and why refers to typical situations with means and purposes [14].The biography of people who suffered a myocardial infarction occupies the space of social action, characterized by the baggage of available knowledge that is constructed through previously accumulated experiences and which characterizes the typifications of the world, generated through social life.Therefore, we understand that part of existential motivation has both a meaning that is subjective (experienced by subjects) and objective (referring to a situation which shows itself significant for those who experience the phenomenon investigated).Various research related to the field of health has applied the social phenomenological approach in order to gain an insight into the phenomenon [15,16].The aim of this study was to describe the men’s experiences as acute myocardial infarction sufferers from a social phenomenological perspective, a year after the event.Qualitative and phenomenological research was chosen for this study, conceived in the light of social phenomenology [13]. This approach permitted the investigation of this social group of men who experience a given typical situation.The participants were chosen using purposeful sampling in accordance with the maximum variation sampling strategy [17]. The total number of people who participated in the study was 14.The following criteria were applied when selecting participants for the study: (1) men who had received a medical diagnosis of acute myocardial infarction, (2) those who had been admitted to the intensive care unit within the previous year, (3) those who were able to communicate properly, excluding those who suffered sensory impairments, (4) those currently living at home and diagnosed as stable as regards their illness, (5) patients who had suffered a relapse of their illness were included in the study, provided that this was not the first time they had suffered this pathology.A total of 20 participants were contacted, 4 of them declined to participate in the study and 1 had difficulty hearing. The other 15 participants met the inclusion criteria, one of whom died days before the interview.This study was carried out in the Region of Valencia (SE Spain), in Health Department number 20. The reference of this department is the General University Hospital of Elche, which has twelve beds for intensive care.After receiving the relevant permission from the Departmental Management Office and approval for the study from the Clinical Research Ethics Committee, we accessed the database of patients admitted to the intensive care unit, thus preselecting those subjects who met the inclusion criteria. We then contacted each of the preselected subjects by telephone to ask for their voluntary collaboration in the study, explaining our aim and informing them that the data would be collected from them at their home during an interview that would deal with issues related to their life and illness.There, participants gave their authorization and signed the informed consent, guaranteeing therefore the ethical research principles in accordance to the Law 02/2016 of Biomedical Researches.Data collection was carried out from June to December 2017.Confidentiality was guaranteed in the processing and use of any data supplied was exclusively for the purposes established in this research. In order to obtain an overall view of each participant, the semi-structured interview was chosen as the most appropriate tool for capturing the true essence of the process experienced since the onset of the disease (a myocardial infarction) until reaching a certain degree of stability, i.e., until the patient was at home without any of the symptoms associated with myocardial infarction. The interview began with general questions such as: “Could you tell something about the experience of your illness?”, “How would you describe your experience at the intensive care unit?”, “What has suffering this illness meant to you?”, and “What changes have happened in your life as a result of this illness?” Other questions complemented the interview (Table 1).The researchers were able to ask for clarification of unclear answers at any time throughout the interview. Interviews were digitally recorded and later transcribed for attentive reading of the statements.As the interview progressed, themes that had emerged throughout the discourse were discussed in further detail.To ensure the anonymity of participants in this study, they were identified as follows: D1, D2, D3, etc., corresponding to Discourse 1, Discourse 2, Discourse 3, and so on.The data analysis was carried out using the steps proposed by Parga Nina and other researchers of social phenomenology. These steps were the following [15,16,17]:(a) Reading of discourse in order to discover the experiences lived by the participants, first with view to identifying the global sense of these. (b) Re-reading of the transcripts, identifying specific categories that express significant aspects related to the participants’ experiences as regards in-order-to and because motives. (c) Grouping of meaning units extracted from the discourse that represent a point of convergence within the content, creating identified categories. (d) Establishing the meanings of social action based on the participants’ discourse in order to obtain the typical experience.After gathering the convergences that emerged from interviews, the appearing categories were warning signs prior to the illness, origin of the illness, fear of dying in the intensive care unit, association of the disease with family and friends, expectations as regards their illness, expectations as regards health professionals, and social expectations. That information was discussed according to the framework developed for social phenomenology, in addition to literature in experiences of life following of myocardial infarction field to enrich it.The total number of people who participated in the study was 14 men. Age ranged between 41 and 66 years (mean = 54.7). Only one of the participants was single and did not have children, seven were married and five separated and all of these participants have children (mean = 2.2). As regards education, nine had completed primary education, four had completed secondary education, and only one had further education qualifications. All except one of the participants, who was a pensioner, were employed at the time of their illness, which had since changed to pensioner for four of them, five were still on sick leave, and only four had gone back to work. Most of the participants (10) had a middle socioeconomic status and the rest low (Table 2).The convergence of the reasons why and the reasons for, emerging from the analyzed statements, permitted capturing the meanings that patients gave to having a heart attack after one year. Figure 1 shows the main findings that correspond to the explanatory model.Warning signs prior to the illness: most of the participants claimed to have had some complaint prior to the illness but had not paid too much attention to it or had blamed it on some other cause.“A month before, I walked out of work, I couldn’t stand it there any more, I just couldn’t, I sat down, I knelt, as if I was out of it, I lay on my back. Then I blamed it on stress and it never occurred to me that it might have been a heart attack ….”Origin of the illness: the participants attributed the origin of their illness to work-related stress and an excessive workload throughout their lives.“Everything is coming out now, I’ve worked a lot, I’ve had to work in the fields and the suchlike since I was 11, I’ve worked so much my whole life, so very, very much”They claimed to have had an extremely busy social life, accompanied by an excessive consumption of alcohol and tobacco; factors which they believe favored the appearance of their illness.“My life in general was really crazy, I mean, I’ve always liked partying although I don’t think I partied to excess and I’ve always liked know going out and disappearing for two or three days, and making the party last all weekend”All participants except one claimed to have been heavy smokers and were aware of the consequences.“I smoked since I was 12 or 13 until last year”. “I ate a lot of salt, a lot of fat. I never cared”Fear of dying in the intensive care unit: the people interviewed expressed fear and anxiety when admitted to the intensive care unit and lived that period as if close to death. Those who had previously been admitted to the unit felt calm and drowsy under the effects of sedation.“When I was in the ICU I cried a lot because I was afraid that I was already dead”Association of the disease with family and friends: they knew of the disease as family members or friends had been afflicted and consequently died.“I knew someone who had a heart attack and it’s like I say sometimes, (…) it’s precisely those who have had a heart attack that die from one”Expectations as regards their illness and reordering goals: all participants were in a state of anticipation in relation to their illness; they stated that they were afraid of becoming disabled and that they may have another attack that will lead to their death.“When I got home, I was afraid and we’ve been back to the hospital twice”“Before myocardial infarction, I had decided to change my 100 m2 home to 150 m2, or change my car, after, I thought about it. I said to myself that I am all running around for a better life, if I am to be bed-ridden, does it really make any difference if my home is 100 or 150 m2 or if my car is a Peugeot or Elegance Mercedes”Expectations as regards health professionals: most of the participants commented on the lack of information provided by health professionals. Although the participants were greatly reassured by the way the health professionals had acted during an emergency situation, they would expect more humane care and treatment so that unnecessary invasive techniques may be avoided.“I was given that information in the corridor and I… I was really scared, it’s not that I don’t want more information but I thought that they should be the ones to tell me how I should live my life from now on”Social expectations: all participants had reduced their social activities mainly due to limited physical activity, lack of confidence when driving, changes in eating habits, and having given up smoking and alcohol consumption. They expressed the need to spend more time with their loved ones; giving great importance to being surrounded by those who care.“Now I drive less, before we didn’t used to go out a lot but we did go away at Christmas”Work prospects: Only four participants had gone back to work, at the same post they had held prior to their illness; four were classed as unable to work through disability and the rest were on temporary sick leave.“I get very tired, the stairs kill me and if not them, the hills, I can’t manage the hills, I’m a pensioner now”Reflections on life, weighing up the lifecycle: The illness has marked a turning point in the participants’ lives. At the time of the interview (one year after the event), all participants claimed to have taken stock of their lives.“Life has changed me a lot. Life has no meaning when you are sick, let alone having a painful heart. You will value life when you get sick, even if you have the world, it will be valueless”The results from our study show two clearly distinct ways of perceiving the same phenomenon: acute myocardial infarction. These two main categories are “positive and negative coping experiences” [7]. On the one hand, there are those who consider it a threatening disease that interrupts their daily life, generating negative feelings and disapproval towards it. This illness is marked by a loss of spontaneity in actions and lack of energy, which means adapting oneself to an unhealthy heart [11,12,13,14,15,16,17,18,19,20,21,22]. In contrast, the other group is made up of people who treat this phenomenon as a warning sign, a chance to change their lifestyle, and who view it in a positive light [22], patients restructured and re-evaluated their attitudes towards self, life, religious beliefs, and others [7].Admittance to the intensive care unit causes participants to become scared about death, making them feel that they are very close to the end. Fear and anxiety are two key emotional reactions that accompany the patient during their stay in the ICU [12,23]. The patients took into consideration the MI as a warning that made them aware of the very thin line between life and death [21]. During their stay, the sufferers positively rate the professionalism with which the health workers act, the reassurance provided by both the staff and the environment, as well as the quick and efficient manner with which emergency situations are dealt [24]. Despite good interpersonal relationships, the participants expect health workers to be able to transmit appropriate information about the process to the patients. Poor communication not only caused anxiety while the patient was in ICU but also contributed to less than optimal recoveries after discharge. The participants in the study expressed the need for a more humane treatment that would prevent unnecessary invasive techniques or waiting times when faced with an emergency situation.The importance of family support as a source of affection and the unique link with the outside world has been made obvious. The need to feel supported, the demand of constant company, the security and sense of being provided with calm by their loved ones as well as the importance of support firstly, from their partner, and secondly, from their children [23,24] have been some of the points set forth by those who have participated in our study.Most participants in the study had not had direct prior contact with the illness, despite stating that they had experienced certain symptoms related to ischemic cardiomyopathy. They consider themselves to be very healthy and do not perceive any risk factors associated with this disease. Patients tend to ignore or reinterpret symptoms through rationalization and denial of their vulnerability to this illness [25]. On the whole, participants attribute the origin and cause of their illness to work stress and an excessive workload throughout their lives [12]. They considered smoking to be a secondary risk factor. Although they do not consider eating habits to be associated to risk factors, they do realize that their diet is not adequate. These results coincide with some authors’ statements, which say that patients had their own view of the onset of disease. The patients spent much time and energy considering the causes and seeking explanations for their disease [12].Just like those who had previous heart problems, attributed the appearance of the illness to the fact that they had not looked after themselves properly. This generated a sense of guilt [6,25].The weakness and fatigue experienced during the rehabilitation process hinders everyday activities. The global fatigue was associated with concurrent symptoms, such as breathlessness and stress, and coping strategies such as changes in values, intrusion, and isolation. The patients felt emotionally and physically powerless and described their life as “transfixed” [6,7].Poor physical health and low levels of physical activity after myocardial infarction affect negatively when returning to work [26]. It is very difficult for men, who considered themselves the family breadwinners [6]. They felt that they had let down their wives and experienced guilt. Our results show that inability to work often forces people to take early retirement or, in some cases, to change to a job that requires less physical effort.Due to the sufferers’ lack of confidence about driving their cars, their leisure activities are limited. Consequently, this stage is perceived as a turning point—a time for changing lifestyles, applying restrictions, and steering life into a new phase [21,26]. This means assuming the responsibility necessary to bring about such changes, the search for professional help, and how to take care of one’s self in the future [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21].Health education of patients by nursing professionals has been proposed as the main intervention tool in order to encourage them to adopt lifestyle changes. Healthcare professionals and patients do not always have the same priorities regarding what information is most important to consider. The patients declared that they would have needed clearer guidelines concerning the rehabilitation process [7].The participants in the study are currently living in a state of expectation as regards their illness [7]. Being constantly on guard means that sufferers will tend to visit their health center to check any sign they may have detected, which in most cases bears no relation to their coronary disease. Being afraid of having another heart attack is the biggest concern sufferers have during this stage of the illness [21]. They envisage a future full of great restrictions and uncertainty.They are worried about resuming sexual activity as, although they are aware that there is no need to restrict themselves in this sense, they are concerned about how to act in this situation and they express the need to be informed accordingly. When people express concern, decreasing frequency, and less desire, this is usually due to a lack of knowledge or ability to look for alternatives regarding the new conditions imposed by the illness [27].Different authors [28] affirm the need to provide telephone help lines to facilitate monitoring and provision of the necessary patient information. An integrative review of the research suggests that people with cardiac disease showed some benefits from nurse-led/delivered telephone interventions [29].The illness stabilization phase involves a process of introspection and looking back over the whole life cycle. New values and incentives appear to stimulate change: they appreciate time spent at home, feeling the need to enjoy their family as much as possible, and state that they are living a more organized and quiet life [6,11]. Some authors concluded that patients realize that they have the chance to live a new and better life by reconsidering their daily existence. They considered the illness a positive event which encourages a change in physical and eating habits which in turn reaps benefits [6,9].We can therefore state that those who have had a heart attack have passed a resilience survival test, characterized by the ability of an individual or social system to live well and to develop positively despite difficult living conditions and, furthermore, to come out stronger and transformed as a result [24]. In combination with their experience of going through a key life transition, these younger people also search for a new personal meaning in life [14]. To sum up, they have been able to recognize the illness itself and have strengthened their own ability to face up to the serious problems caused by the illness and, from here on, make the most out of life [8].Most of the patients interviewed for this study mentioned their wife as the main source of support throughout all the stages of the illness. On the other hand, spouses are often aware that everything is not in order and may become overprotective inadvertently increasing the patients’ sense of inferiority or feeling that they are not relied on anymore [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]. They acted as his guardian by phoning home when at work, being available, being on the alert, being attentive to the patient, and being prepared [21,22,23,24,25,26,27,28,29,30].An important detail discovered in our study was the support perceived by those people who lived in an environment where there were small children, as they stated that the latter were the main source of help in overcoming the process as they believed they must continue living for the children.The application of phenomenological sociology theory [13] has allowed us to understand a way of thinking, reasoning, and acting of people who suffer from AMI. This theory’s objective is to focus in the social relations in the world of life. People are immersed in a social, historical, and cultural context. Health care professionals must consider the amount of knowledge and experience acquired over a lifetime, as well as biographical situation, in order to help themselves understand the lived phenomenon.The persons, during their life, see the world from the perspective of their own interests, motives, desires, ideology, and religious commitments. The reality of common sense is culturally considered universal, however, the way these forms are expressed in individual life depends on the totality of experience that the subject constructs in the course of his concrete existence.This paper has important implications for the team of professionals that make up the cardiology unit. The results of this study have underlined the need to pay attention on all stages of the illness: the physical and emotional dependence upon admittance at the intensive care unit, the need to be cured, the constant demand for information about the illness, the difficulties encountered upon returning home, uncertainty about the future, etc. All these moments indicate that proper nursing care adapted to the specific needs of each individual and their family members must be provided in order to help them to overcome all the stages involved in this process.It is necessary to provide a new structure that would allow the development of patient-centered and family interventions after AMI that embrace patients’ preferences, needs, values, and goals, that are tailored to each stage of each person’s recovery.This study was carried out in the national and local context of the Spanish society and its results should be interpreted cautiously due to its qualitative nature of study. The generalization of our data to other countries such as the United States can be complicated by cultural differences between the two countries.The main limitation of this study is given by the composition of the sample. Future research should investigate the perception of the disease in people with a high socioeconomic profile or with university studies. It would be interesting to expand our findings with women and establish gender differences with respect to the disease.The results of this study show the different situations experienced by people who have suffered a heart attack. A period of one year after MI is considered by scientists as a stability moment, but our findings show us how people feel that their lives drastically changed. The participants regarded the MI as a warning that made them aware of the very thin line between life and death. They feel like the illness helped them to create new meanings and value of life. They envisage a future full of great restrictions and uncertainty.It can be concluded that in spite of problems and stresses experienced by patients, myocardial infarction may also have positive effects for them. That is, patients may draw some positive experiences from their illness.Conceptualization, M.S.-R. and J.S.-G.; methodology, J.S.-G.; validation, M.S.-R., G.F.d.F. and M.S.-R.; formal analysis, M.I.U.-G. and M.S.-R.; investigation, M.I.U.-G.; resources, S.G.-C.; data curation, G.F.d.F.; writing—original draft preparation, M.S.-R. and M.I.U.-G.; writing—review and editing, J.S.-G. and G.F.d.F.; visualization, G.F.d.F.; supervision, J.S.-G.; project administration, J.S.-G. 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 Ethics Committee Clinical Research of AGENCIA VALENCIANA DE SALUD (27/02/2016).Informed consent was obtained from all subjects involved in the study.The authors would like to thank the patients who participated in the study.The authors declare no conflict of interest.Analysis of the content according reasons “why” and reasons “for”.Interview guide.Sociodemographic description of participants.* AMI: acute myocardial infarction. ** IDDM: insulin dependent diabetes mellitus.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01054.txt

@@ -0,0 +1 @@
+During the lockdown declared in Spain to fight the spread of COVID-19 from 14 March to 3 May 2020, a context in which health information has gained relevance, the agenda-setting theory was used to study the proportion of health advertisements broadcasted during this period on Spanish television. Previous and posterior phases were compared, and the period was compared with the same period in 2019. A total of 191,738 advertisements were downloaded using the Instar Analytics application and analyzed using inferential statistics to observe the presence of health advertisements during the four study periods. It was observed that during the lockdown, there were more health advertisements than after, as well as during the same period in 2019, although health advertisements had the strongest presence during the pre-lockdown phase. The presence of most types of health advertisements also changed during the four phases of the study. We conclude that, although many differences can be explained by the time of the year—due to the presence of allergies or colds, for instance—the lockdown and the pandemic affected health advertising. However, the effects were mostly visible after the lockdown, when advertisers and broadcasters had had time to adapt to the unexpected circumstances.Almost every aspect of human life has been affected by the SARS-COV-2 pandemic since its declaration in March 2020, and academia has rapidly focused on this unprecedented topic and its effects, causes, and derivatives. This has also happened in the field of communication, with numerous studies published regarding the manner in which information about the virus has been consumed during the pandemic [1] or about the potential spread of “fake news” [2].In general terms, one of the most relevant branches within the field of communication studies is that which focuses on the effects of television media. Some of the most important and influential theories of the field, such as the cultivation theory [3] or the uses and gratifications theory [4], apply to this area. Although television has lost some of its predominance in recent years due to the rise of online and social media, it is still the most consumed medium in most countries. The Estudio General de Medios [5], the most relevant media audience study in Spain, shows that television is still the most used medium, with a penetration of 85%, superior to the 81.1% penetration of the Internet (these data were collected during the first months of 2020, with fieldwork completed just before the lockdown in Spain in mid-March).Despite the good Internet connection quality in Spain—the penetration rate [6] and average speed of the Internet [7] in Spain are above the European average—it is not as widespread as television, and its use is not as equally distributed among the population groups, with a large underrepresentation of older age groups. In this context, following the lockdown declared in Spain on 14 March 2020, which confined most of the population to their homes, television consumption rose to record levels in March and April, the most restrictive months of the lockdown [8,9]. Due to this important role played by television, it is relevant to focus on this medium.Given the significant relevance of TV consumption during the lockdown, and the key role of advertising for the business and communication model of this medium, it is of interest to discover how the lockdown has affected the types of advertisements present in television. In addition, in a context in which health has become a constant presence in the media, it is of interest to study whether the same impact has occurred in advertising, particularly during a period in which the consumption of most non-essential goods and services was not possible. Thus, the main goal of this study was to quantify the presence of health advertising in its different forms within the advertisements broadcasted on commercial television in Spain before, during, and after the lockdown. More specifically, we attempted to analyze the kinds of health products or campaigns that were prevalent during this time.In this way, we hope to fill the existing knowledge gap regarding advertising during the health crisis because, despite the increase in articles related to COVID-19 and its connection with a broad spectrum of objects of study, no relevant works have focused on advertising in the Spanish setting. Thus, this article focuses on a key element during a period in which the awareness of citizens about health-related issues has become dominant, and in which consumption has plummeted and drastically changed, together with the habits, state of mind, and media consumption of a vast majority of citizens. Additionally, the focus on television advertising can be justified not only by the still superior penetration rate of television in the Spanish media market but also by the decreasing investment of advertisers in television, money that is increasingly going to digital campaigns, which have overtaken television as the main advertising platform [10].Health communication, in its broadest sense, was present in the daily life of citizens during the lockdown and in the days before and the weeks and months afterward. However, this field of study is one of the most relevant and traditional in the area of communication studies [11]. The presence of health content in the media has been broadly studied [12,13]; however, some of the most productive objects of study of this area have been the effects of the media coverage of health issues [14], and those campaigns or strategies to promote particular behaviors or awareness using mass media [15,16]. The use of narrative persuasion in health campaigns is one of the most frequently used approaches among these studies [17].It is clear that communication during a health crisis is more relevant and intense [18]. As Vaughan and Tinker [19] show, health communication plays an important role during a pandemic, especially if a cooperative public is required. A relevant aspect that has been observed during the SARS-COV-2 pandemic is how an emerging infectious disease is communicated, something that authors such as Holmes [20] have studied for a number of years. Relevant theories or models have been developed to study health communication applied to HIV/AIDS [21] and severe acute respiratory syndrome (SARS) [22,23].Scholars have also shown significant interest in recent years in the role of media during previous health crises, such as the Ebola epidemic of 2014 [24] or the H1N1 flu in 2009 [25]. In recent years, the interest in social media has grown [26], and many studies related to health and risk communication during the current coronavirus pandemic have focused precisely on these media [27]. Severe acute respiratory syndrome (SARS), a disease belonging to the same family of viruses as COVID-19, showed how electronic media make it possible to rapidly disseminate infectious disease prevention messages [28]. Although television is less present in the current academic scenario, it is nonetheless a highly relevant medium, and its role during the COVID-19 pandemic also has been an object of study [29]. The present article contributes to this line of work, focusing on the medium that, to date, still has the highest penetration rate in Spain.The other key aspect of the present work is advertising, a relevant field of work within communication. The study of advertising in the media is a highly interesting topic, and several investigations have tackled how advertising affects viewers [30]. In general, the analysis of the content depicted in advertising [31] and its effects [32,33] has a high degree of relevance, especially regarding television.Regarding the connection between health and advertising, the focus has traditionally been on public service announcements (PSAs) to promote healthy habits [34]. However, several studies have also tried to address the presence of advertisements for medicines or medical products and services [35,36] on television. Following this line of work, the present study attempted to discover the frequency of health-related advertisements during the lockdown in Spain, a period during which health was the central aspect in most citizens’ lives, and in which television consumption grew, but advertising investment decreased.Given the context outlined above, this study took place under the theoretical framework of the first level of the agenda-setting theory, that is, the capacity of media to set the relevance of topics in the public discourse, which is higher if their media presence is also high [37]. Many previous works have used this theory to study the agenda-setting effects of advertising [38,39], although they have focused mostly on political advertising and the second level of the agenda, and few analyses have been undertaken using the first level of agenda-setting to study other types of advertising. Similarly, agenda-setting has been previously used to analyze the coverage of health crises [40], and previous studies have examined the role of the media during the COVID pandemic from the perspective of the agenda-setting theory [41]. However, these approaches have focused mostly on the media coverage or information rather than on the advertising. This study is focused on the application of agenda-setting to advertising outside of politics and, more specifically, during a health crisis. Therefore, based on the agenda-setting theory and the potential effects that health advertising can have during a health crisis in the public discourse, this study attempted to answer the following research questions:RQ1: What was the proportion of health advertisements broadcasted daily before, during, and after the lockdown by the main Spanish commercial television broadcasters?RQ2: Did the presence of different types of health advertisements on the main Spanish commercial television broadcasters increase during the lockdown compared to the same period in the previous year?RQ3: What types of health advertisements were prevalent before, during, and after the lockdown on the main Spanish commercial television broadcasters?RQ1: What was the proportion of health advertisements broadcasted daily before, during, and after the lockdown by the main Spanish commercial television broadcasters?RQ2: Did the presence of different types of health advertisements on the main Spanish commercial television broadcasters increase during the lockdown compared to the same period in the previous year?RQ3: What types of health advertisements were prevalent before, during, and after the lockdown on the main Spanish commercial television broadcasters?Although the types of advertising that television offers are broad, our study focuses on advertisements, that is, brief messages shown in the breaks of television programs and the most common and traditional form of television advertising. Following the definition of the Ley General de Comunicación Audiovisual [42], we considered advertisements those messages whose goal is to “promote the supply of goods or services, including real estate, rights and obligations”. In this research, we analyzed the advertisements aired in blocks during, before, or after television programs, excluding sponsorship, telepromotion, advertorials or infomercials, and product placement.A content analysis was conducted of the advertisements broadcasted by the two main television broadcasters in Spain, representing more than 25% of the market share: Antena 3, owned by Atresmedia group, with 11.2% of accumulated share from January to June 2020, and Telecinco, owned by Mediaset, with 14.4% of accumulated share from January to June 2020, according to data from Kantar Media. La 1, the main public television channel and the third most viewed with 9.6% accumulated share until June 2020 was not considered because it is forbidden by law to broadcast commercial advertisements.Data were collected in four blocks of 50 days each, previously predefined to answer the research questions of the study. Those blocks were:Lockdown: Since the lockdown was decreed in Spain on 14 March 2020, following the declaration of the state of alarm by the government, until Phase 0 of the de-escalation started on 3 May 2020, and the restrictions started to be progressively eased. This was the most dramatic period, during which the harshest restrictions were applied, and it is central to our analysis because the other periods of the study were designed around this one.Pre-lockdown: Fifty days before the beginning of the lockdown, that is, from 24 January 2020 to 13 March 2020. This includes the weeks before the pandemic was declared by the World Health Organization when the virus was gaining media relevance, but, as expected, without having yet influenced the advertising behavior.Post-lockdown: Fifty days after the de-escalation started, that is, 4 May 2020 to 22 June 2020. This period includes the ending of the state of alarm on 21 June 2020. It includes the weeks in which the restrictions were being eased, and people could start resuming some of their normal activities, reaching the so-called “new normality”.Equivalent dates of 2019: From 14 March 2019 to 3 May 2019, that is, the fifty days included in the central period of analysis. During this period in 2019, however, the SARS-COV-2 had not been detected, thus allowing comparison with a period during which there was no influence of the virus.Lockdown: Since the lockdown was decreed in Spain on 14 March 2020, following the declaration of the state of alarm by the government, until Phase 0 of the de-escalation started on 3 May 2020, and the restrictions started to be progressively eased. This was the most dramatic period, during which the harshest restrictions were applied, and it is central to our analysis because the other periods of the study were designed around this one.Pre-lockdown: Fifty days before the beginning of the lockdown, that is, from 24 January 2020 to 13 March 2020. This includes the weeks before the pandemic was declared by the World Health Organization when the virus was gaining media relevance, but, as expected, without having yet influenced the advertising behavior.Post-lockdown: Fifty days after the de-escalation started, that is, 4 May 2020 to 22 June 2020. This period includes the ending of the state of alarm on 21 June 2020. It includes the weeks in which the restrictions were being eased, and people could start resuming some of their normal activities, reaching the so-called “new normality”.Equivalent dates of 2019: From 14 March 2019 to 3 May 2019, that is, the fifty days included in the central period of analysis. During this period in 2019, however, the SARS-COV-2 had not been detected, thus allowing comparison with a period during which there was no influence of the virus.The collection of the data took place between 1 June 2020 and 6 August 2020 using the Instar Analytics application, developed by Kantar Media, which is the company that conducts audience measurements in Spain. All of the advertisements broadcasted each day in each of the two studied channels were collected and later classified as explained in the following section, obtaining a total sample of 191,738 advertisements (2107 unique cases).In addition to the date and time of transmission, the number of contacts, and the channel on which they were broadcasted, the sample of advertisements was content analyzed following a codebook that adapted the consumption categories defined by the European classification of individual consumption by purpose (ECOICOP), which is used to compile the consumption price index (IPC) in Spain and other European countries. This is an expenses classification including all of the different types of goods and services that can be acquired and donations or other possible money expenditures; thus, every commercial good or service, and the advertisements promoting them, can be allocated within one of the categories. Because advertising mostly aims to promote goods and services that are purchased and paid for by consumers, and due to the wide acceptance of this classification, it was considered an adequate basis for our study. However, two modifications were made to adapt the codebook to our study, as follows:First, there are public service announcements or institutional advertising that does not have an associated expense—antiracism or antitobacco campaigns, for instance—that are also broadcasted as television advertisements. With the aim of including this type of advertising in the study, it was decided to include them within the last category. Similarly, advertisements that improve the corporate image of a brand, although not directly advertising a good or service, can be allocated within the line of the business of the company; in the case of very diverse companies with several branches of activity, these advertisements were allocated within the “miscellaneous goods and services” category. Thus, the general classification used is as follows, with each advertisement classified exclusively under one of the following 12 categories provided by the ECOICOP:Food and non-alcoholic beverages: includes all food products and non-alcoholic beverages that can be purchased for consumption at home;Alcoholic beverages: includes all alcoholic beverages that can be purchased for consumption at home, including low or non-alcoholic beverages that are generally alcoholic, such as non-alcoholic beer. The original ECOICOP category also includes tobacco, but given the prohibition on tobacco advertising, it was not included in the study;Clothing and footwear: includes clothing materials—including fabrics and accessories such as buttons or sewing threads—garments, accessories, footwear and cleaning, repair and hire of clothing, and footwear;Housing, water, electricity, gas and other fuels: includes real estate acquisition or renting, and repair services or materials (painting, minor plumbing, electricians, etc.), other services relating to the dwelling (gardening, security, etc.), and supplies for the dwelling (heating, electricity, water, gas, fuels, heat energy, etc.). In general, this category includes all products and services for the acquisition or maintenance of a dwelling;Furnishings, household equipment and routine household maintenance: includes all household or garden furniture and furnishings, carpets and other floor coverings, lighting equipment, household textiles (bed linen, curtains, table linen, etc.), household appliances, glassware, tableware and household utensils, tools and equipment for house and garden (electric drills, lawnmowers, alarms, etc.), non-durable household goods (cleaning products, candles, nails, fire extinguishers, etc.). It also includes all rental, repair or associated services and cleaning or maintenance services;Health: This category is further discussed below, but in general terms, it includes health products and medicines, health services and health communication, and PSAs;Transport: includes bicycles and motor vehicles (excluding recreational vehicles such as camper vans or boats) and all services and products for their use, reparation, cleaning, or maintenance (including parking expenses, fuel, driving lessons, GPS). It also includes transportation services, such as buses, trains, taxis, flights, or any other private or public transportation method;Communication: includes postal and parcel delivery services (it does not include Amazon or food delivery apps, for example), and telephone and Internet providers, services, or equipment (including smartphones or bundled telecommunication services, but excluding computers or video-on-demand services);Recreation and culture: includes audiovisual, photographic, and information processing equipment (in general, all technological devices, such as computers or televisions, except for smartphones), and all software, applications and Internet-based services and business that do not specifically apply to any other category (for instance, an Internet-based clothes shop would be included in “clothing and footwear”, but Amazon, with a broader scope, would belong in this category). It also includes major durables for recreation (camper vans, canoes, golf carts, etc.), sports equipment, musical instruments, major durables for indoor recreation (gaming machines, billiard tables, etc.), any other recreational items and equipment (toys, games, celebration articles), and flowers and other garden products. It also includes any repair, maintenance, or complementary service or good applied to the previous goods. It also includes pets and any food, product, or service associated with them. Additionally, it includes all recreational or cultural services and attendance, such as sports events, cinema, theater, concerts, museums, television subscriptions, or any other game of chance, both online and offline. Finally, it includes newspapers, books, and stationery, and package holidays;Education: includes all educational services, including language courses, schools, or universities.Restaurants and hotels: includes catering services, including restaurants, bars, and take-away local food services. Furthermore, includes accommodation services;Miscellaneous goods and services: includes personal care (hairdressing, grooming, hygiene, wellness, etc.), baby or child care (nurseries, babysitters, etc.), personal effects (jewelry, clocks, watches, travel goods, articles for babies, etc.), and counseling, insurance, financial, legal, or funeral services. Similar to the case of tobacco, the original ECOICOP category also includes here services such as prostitution, which, given their illegality, are not advertised and, therefore, were not included in the final codebook.Food and non-alcoholic beverages: includes all food products and non-alcoholic beverages that can be purchased for consumption at home;Alcoholic beverages: includes all alcoholic beverages that can be purchased for consumption at home, including low or non-alcoholic beverages that are generally alcoholic, such as non-alcoholic beer. The original ECOICOP category also includes tobacco, but given the prohibition on tobacco advertising, it was not included in the study;Clothing and footwear: includes clothing materials—including fabrics and accessories such as buttons or sewing threads—garments, accessories, footwear and cleaning, repair and hire of clothing, and footwear;Housing, water, electricity, gas and other fuels: includes real estate acquisition or renting, and repair services or materials (painting, minor plumbing, electricians, etc.), other services relating to the dwelling (gardening, security, etc.), and supplies for the dwelling (heating, electricity, water, gas, fuels, heat energy, etc.). In general, this category includes all products and services for the acquisition or maintenance of a dwelling;Furnishings, household equipment and routine household maintenance: includes all household or garden furniture and furnishings, carpets and other floor coverings, lighting equipment, household textiles (bed linen, curtains, table linen, etc.), household appliances, glassware, tableware and household utensils, tools and equipment for house and garden (electric drills, lawnmowers, alarms, etc.), non-durable household goods (cleaning products, candles, nails, fire extinguishers, etc.). It also includes all rental, repair or associated services and cleaning or maintenance services;Health: This category is further discussed below, but in general terms, it includes health products and medicines, health services and health communication, and PSAs;Transport: includes bicycles and motor vehicles (excluding recreational vehicles such as camper vans or boats) and all services and products for their use, reparation, cleaning, or maintenance (including parking expenses, fuel, driving lessons, GPS). It also includes transportation services, such as buses, trains, taxis, flights, or any other private or public transportation method;Communication: includes postal and parcel delivery services (it does not include Amazon or food delivery apps, for example), and telephone and Internet providers, services, or equipment (including smartphones or bundled telecommunication services, but excluding computers or video-on-demand services);Recreation and culture: includes audiovisual, photographic, and information processing equipment (in general, all technological devices, such as computers or televisions, except for smartphones), and all software, applications and Internet-based services and business that do not specifically apply to any other category (for instance, an Internet-based clothes shop would be included in “clothing and footwear”, but Amazon, with a broader scope, would belong in this category). It also includes major durables for recreation (camper vans, canoes, golf carts, etc.), sports equipment, musical instruments, major durables for indoor recreation (gaming machines, billiard tables, etc.), any other recreational items and equipment (toys, games, celebration articles), and flowers and other garden products. It also includes any repair, maintenance, or complementary service or good applied to the previous goods. It also includes pets and any food, product, or service associated with them. Additionally, it includes all recreational or cultural services and attendance, such as sports events, cinema, theater, concerts, museums, television subscriptions, or any other game of chance, both online and offline. Finally, it includes newspapers, books, and stationery, and package holidays;Education: includes all educational services, including language courses, schools, or universities.Restaurants and hotels: includes catering services, including restaurants, bars, and take-away local food services. Furthermore, includes accommodation services;Miscellaneous goods and services: includes personal care (hairdressing, grooming, hygiene, wellness, etc.), baby or child care (nurseries, babysitters, etc.), personal effects (jewelry, clocks, watches, travel goods, articles for babies, etc.), and counseling, insurance, financial, legal, or funeral services. Similar to the case of tobacco, the original ECOICOP category also includes here services such as prostitution, which, given their illegality, are not advertised and, therefore, were not included in the final codebook.Second, the health category, the most relevant category for our study, was subdivided into more specific subcategories for a more detailed analysis. It must be noted that the original ECOICOP classification has multiple subcategories within each of its 12 broad categories; this subclassification was also used in the case of the “health” category. Although this already existing subclassification was followed, some adaptations were required to obtain a more detailed comparison of the different types of advertisements:Health insurance: although originally included in “other goods and services”, health insurance was moved into this category so that more detailed observations of this type of service could be conducted. When insurance companies were advertised, focusing on other types of insurance or in a general sense, without a focus on the health aspect, they were left outside the health category;Care services, such as caretakers or residences for ill or old people, were also moved from the “other goods and services” to this category;Institutional communication or PSAs related to health issues, both focused on the coronavirus or on other health aspects, were included in two ad hoc subcategories created for the study: one for the advertising of public institutions and one for the advertising of private institutions;Different types of medicines, not differentiated in the original subclassification, were taken into account in the new codebook to enable the study of how medicines that could be used to counter COVID-19 or other specific illnesses were more or less present during the period of analysis. These subcategories were developed following the classification proposed by the World Health Organization in its List of Essential Medicines [43], together with an exploratory observation and the ECOICOP classification.With the goal of focusing on products particularly associated with COVID-19 prevention, a specific group was also added, in which facemasks, hydrogel or gloves were included.Health insurance: although originally included in “other goods and services”, health insurance was moved into this category so that more detailed observations of this type of service could be conducted. When insurance companies were advertised, focusing on other types of insurance or in a general sense, without a focus on the health aspect, they were left outside the health category;Care services, such as caretakers or residences for ill or old people, were also moved from the “other goods and services” to this category;Institutional communication or PSAs related to health issues, both focused on the coronavirus or on other health aspects, were included in two ad hoc subcategories created for the study: one for the advertising of public institutions and one for the advertising of private institutions;Different types of medicines, not differentiated in the original subclassification, were taken into account in the new codebook to enable the study of how medicines that could be used to counter COVID-19 or other specific illnesses were more or less present during the period of analysis. These subcategories were developed following the classification proposed by the World Health Organization in its List of Essential Medicines [43], together with an exploratory observation and the ECOICOP classification.With the goal of focusing on products particularly associated with COVID-19 prevention, a specific group was also added, in which facemasks, hydrogel or gloves were included.Based on this process, the advertisements classified in the previous variable as “health” were subclassified into one of the following 27 subgroups. All of the other advertisements were coded as “0” in this classification.Slimming products, such as drugs and other treatments, excluding food products with a health component, such as anti-cholesterol yogurts;Analgesic and anti-inflammatory products, such as ibuprofen or paracetamol;Antacids, including all types of stomach protectors;Contraceptives, including pills and non-oral forms (such as condoms);Antihistamine products;Antibiotics, anti-fungal, and other anti-infective products;Antipyretic drugs specifically designed to fight fever, excluding products that could have such an effect, but not as the main goal, such as analgesics;Antitussive, mucolytic, or anti-flu products: includes drugs and syrups, and nose-sprays or products such as Vicks VapoRub;Vitamin supplements, including drugs and chemical products, but not food products with a health plus, such as calcium-rich milk;Medical creams and spray, including anti-varicose vein, vaginal, and anti-inflammatory creams and similar;Laxatives and antidiarrheal products;Homeopathic products;Throat lozenges;Sleeping pills and other sleeping treatments;Other drugs and medical or pharmaceutical products, including only non-durable products that could not be allocated in any of the previous categories;Facemasks, gloves, and hydrogels, and similar products specifically recommended to prevent COVID-19;Therapeutic equipment, including glasses, hearing aids, wheelchairs, crutches, stairlifts, etc.;Other medical products, including bandage strips, adhesive dressing, syringes, merbromin, and pregnancy tests;Medical or hospital services, including private doctors, plastic surgery, clinics, etc.;Dental services and clinics;Paramedical services, including blood tests, thermal treatments, rehab, physiotherapy, opticians, oculists, or otolaryngologists when the services and clinics were advertised rather than glasses or hearing aids;Private health insurance;Care homes, including retirement homes for elderly persons, residences for disabled persons, or rehabilitation centers;Residential care and assistance, including home help or daycare for elderly or disabled persons at home;Other medical or aid services, including other health-related services not included in the previous categories, for example, medical credits;PSAs by public institutions;PSAs campaigns by private institutions.This classification is shown in Table 1. Additionally, all items were recoded into dichotomous variables (dummy variables), with 1 indicating the presence of a category or subcategory, and 0 its absence, thus allowing measurement and comparison of the proportion of health advertisements and of each group within this category during each phase or in each channel.To verify the reliability of the instrument, a randomly selected subsample of 108 different spots was coded by two independent coders. This subsample is around 5% of the sample of different advertisements (N = 2107); because the majority of the 191,738 advertisements were repetitions that would be equally classified, only the sample of unique cases was used to measure the reliability of the instrument. We used Cohen’s kappa and Krippendorff’s alpha (measured from 0 to 1, being 1 total agreement) and obtained an average of 0.797. This was considered to be adequate because it is above the 0.70 threshold [44,45]. This can be seen more in detail in Table 2.Once the reliability of the instrument was confirmed, the content analysis was conducted. Subsequently, inferential statistics—mostly one-way ANOVA tests and bivariate correlations—were used to answer the research questions. All of the statistical analyses were conducted in IBM’s SPSS (v. 26, IBM, Armonk, NY, USA). A 0.001 (99.9%) type one error was used in this study for stronger rigor of the inferential tests.Before answering the research questions, an exploratory, descriptive analysis was conducted to study the general distribution of the variables. Thus, it was observed that Antena 3 broadcasted more advertisements (101,490, 52.9% of the total) than Telecinco (90,248 advertisements, representing 47.1% of the total). Regarding health advertisements, Telecinco paid significantly more attention to this type of advertisement (M = 0.16, SD = 0.36) than Antena 3 (M = 0.15, SD = 0.36; t (188,302.71) = −3.992, p < 0.001, d = 0.02), although Antena 3 still led overall, with 15,201 advertisements about health, compared to the 14,111 of Telecinco, due to the larger number of adverts broadcasted in general on this channel.Of the four periods of study, the most advertisements were broadcasted during the 50 days before the lockdown (63,707 advertisements, 32.7% of the total) and in the 50 days of 2019 (61,854 advertisements, 32.3%), considerably higher than the lockdown (29,466 advertisements, 15.4% of the total) and the post-lockdown (37,711, 19.7% of the total) periods. The daily distribution does not show strong deviations, but it should be noted that, on 16 March 2020, two days after the lockdown was declared, 2356 advertisements were counted, more than twice the average of 958.69 advertisements per day. It is also relevant to note that after the lockdown period and, in particular, during the lockdown, the total number of broadcasted advertisements was significantly smaller; in the case of the lockdown weeks, the number of advertisements was less than half of those broadcasted in the previous phases.The general distribution of the sample of advertisements is shown in Table 3 in more detail. It should be highlighted that 43,027 advertisements (22.4%) were about food and non-alcoholic beverages, 29,853 (15.6%) were about leisure and culture, 29,312 ads (15.3%) were health-related, and 49,880 (26.0%) fell in the category of other goods and services. Although food and health are rather specific categories, it is important to keep in mind that leisure and culture includes all mobile apps and websites, and most technological devices (with the exception of smartphones, which fall into communication), whereas other goods and services includes all PSAs not related to health, general advertising for supermarkets or shopping chains, and all products of personal hygiene, jewelry, and perfumery. These specific types of advertising are relatively frequent, which partly explains the significant presence of these two broad categories.Focusing on health advertising, Table 4 shows that the most common type of advertisement during the 200 days of analysis was the PSAs of public institutions with 8574 advertisements (4.5% of the whole sample), followed by medicinal creams or sprays (3170 advertisements, 1.7% of the sample) and dental services, such as clinics or treatments (3064 advertisements and 1.6% of the sample. In all of the periods of study, no advertisements were shown for antibiotics, anti-fungal and other anti-infective products, antipyretics, homeopathic products, other medical products, or care homes.To answer RQ1 and RQ2, we recoded the general classification so that we could measure the presence of health advertisements (coded 1) compared to all other categories (coded 0). This allowed us to determine the proportion of health advertising during the lockdown and comparison it with the other analyzed periods. Given that the equality of variances was not assumed, Welch’s F test showed significant differences in the proportion of health advertisements between the four studied periods (F(3, 89,438.41) = 136.26, p < 0.001). Dunnett’s T3 test proved that the proportion of advertisements related to health broadcasted during the lockdown (M = 0.16, SD = 0.37) was significantly bigger than the proportion of health advertising in the same period of 2019 (M = 0.14, SD = 0.35; d = 0.06) and in the de-escalation phase (M = 0.13, SD = 0.34; d = 0.08), although it is significantly smaller than in the pre-lockdown phase (M = 0.17, SD = 0.38; d = 0.03). Figure 1 shows the proportion of health advertisements in each phase.Without taking into account the days of 2019, during which the pandemic played no role, it should be noted that the presence of health advertising decreased significantly with time, with the highest proportion found before the lockdown and the smallest proportion after. This is reinforced by the significant and negative correlation between the day of emission and the presence of health advertisements (R = −0.06; p < 0.001), meaning that the proportion of health advertising decreased during these 150 days.Next, we focused on the health subclassification to compare the proportion of the different types of health advertisements during the different phases, thus answering RQ3. Most types of health advertisements showed significant differences between the four phases. For each subgroup of health products and services, parametric or nonparametric tests were conducted.The Kruskal–Wallis test was conducted for testing the distribution of slimming products in the different phases, because in one of the periods no advertisements of this kind were found, and non-parametrical tests were required. Significant differences were found (K(3) = 602.82, p < 0.001). According to Mann–Whitney U, this type of advertisement was found significantly more frequently after (M = 0.01 (The exact mean values of this category are: 2019 (M = 0.0065), pre-lockdown (M = 0.0000), lockdown (M = 0.0005), and post-lockdown (M = 0.0076)), SD = 0.09) than before (M = 0.00, SD = 0.00) and during (M = 0.00, SD = 0.02; d = 0.12) the lockdown. Furthermore, the sample of the previous year (M = 0.01, SD = 0.08) had a significantly greater presence of these advertisements than the pre-lockdown and the lockdown (d = 0.10) phases. These results are shown in Figure 2.Regarding analgesics and other anti-inflammatory products, the Welch’s test (F(3, 86,101.54) = 30.69, p < 0.001) showed that during the post-lockdown period there were significantly more advertisements (M = 0.01 (The exact mean values of this category are: 2019 (M = 0.0046), pre-lockdown (M = 0.0065), lockdown (M = 0.0061) and post-lockdown (M = 0.0101)), SD = 0.10) than during the pre-lockdown (M = 0.01, SD = 0.08; d = 0.04), during the lockdown (M = 0.01, SD = 0.08; d = 0.04), and during the sampled period of 2019 (M = 0.00, SD = 0.07; d = 0.06). During the pre-lockdown there were also significantly more advertisements in this category than in the 2019 period (d = 0.03) (Figure 3).Welch’s test showed antacids were differently present in the four studied periods (F(3, 87,755.69) = 8.14, p < 0.001). As can be seen in Figure 4, and according to Dunnett’s T3 test, before the lockdown there were significantly fewer of these advertisements (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0048), pre-lockdown (M = 0.0032), lockdown (M = 0.0042), and post-lockdown (M = 0.0044)), SD = 0.06) than after the confinement (M = 0.00, SD = 0.07; d = 0.02) and in 2019 (M = 0.00, SD = 0.07; d = 0.03).Advertisements of antihistamine products also experienced, according to the Kruskal–Wallis nonparametric test, (K(3) = 79.79, p < 0.001), differences in the studied phases. A Mann–Whitney U test showed that during the lockdown (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0011), pre-lockdown (M = 0.0000), lockdown (M = 0.0009), and post-lockdown (M = 0.0003)), SD = 0.03) significantly more adverts of this kind were broadcasted than afterward (M = 0.00, SD = 0.02; d = 0.03) and before (M = 0.00, SD = 0.00). Similarly, the sample of 2019 (M = 0.00, SD = 0.03) also had a significantly greater presence of these advertisements than the pre- (d = 0.03) and post-lockdown phases. These differences can be seen more clearly in Figure 5.Anti-flu products also showed significant differences according to Welch’s test (F(3, 85,081.00) = 458.04, p < 0.001). As can be seen in Figure 6, the proportion during the period of 2019 (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0068), pre-lockdown (M = 0.0027), lockdown (M = 0.0154), and post-lockdown (M = 0.0044)), SD = 0.04) was significantly smaller than during the lockdown (M = 0.01, SD = 0.09; d = 0.10) and during the pre-lockdown (M = 0.02, SD = 0.15; d = 0.19) phases. In addition, the proportion during the post-lockdown weeks (M = 0.00, SD = 0.03) was significantly smaller than during the lockdown (d = 0.11) and before it (d = 0.20). Finally, the proportion of these advertisements was significantly smaller during the lockdown than before (d = 0.12).Regarding vitamin supplements, Welch’s test showed also significant differences (F(3, 81,565.62) = 18.03, p < 0.001). Dunnett’s T3 test showed differences between all groups, with the lockdown period showing the greatest presence of this type of advertisement (M = 0.02 (The exact mean values of this category are: 2019 (M = 0.0068), pre-lockdown (M = 0.0027), lockdown (M = 0.0154), and post-lockdown (M = 0.0044))., SD = 0.12), and significantly more than the same period in 2019 (M = 0.01, SD = 0.08; d = 0.08), the post-lockdown period (M = 0.00, SD = 0.07; d = 0.11), and the pre-lockdown period (M = 0.00, SD = 0.05; d = 0.13). Similarly, the sample of 2019 had a significantly greater proportion of these advertisements than the post lockdown (d = 0.03) and the pre-lockdown (d = 0.06) phases, whereas the post-lockdown phase had significantly more presence of advertisements about vitamin supplements than the pre-lockdown phase (d = 0.03). This is also shown in Figure 7.The presence of medical creams and sprays also experienced differences, according to Welch’s test (F(3, 86,697.18) = 26.28, p < 0.001), although Dunnett’s T3 test showed that it was only during the pre-lockdown period (M = 0.01 (The exact mean values of this category are: 2019 (M = 0.0174), pre-lockdown (M = 0.0131), lockdown (M = 0.0197), and post-lockdown (M = 0.0182)), SD = 0.11) when there were significantly fewer of these advertisements, compared to the lockdown (M = 0.02, SD = 0.14; d = 0.05), the post-lockdown (M = 0.02, SD = 0.13; d = 0.04), and the 2019 (M = 0.02, SD = 0.13; d = 0.04) periods. Figure 8 shows these differences.Advertising of laxatives was also significantly different according to Welch’s test (F(3, 81,687.35) = 257.15, p < 0.001), and Dunnett’s T3 test showed that during the post-lockdown (M = 0.00, SD = 0.01) significantly fewer laxatives and antidiarrhea advertisements were broadcasted than in the lockdown (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0083), pre-lockdown (M = 0.0047), lockdown (M = 0.0045), and post-lockdown (M = 0.0002)), SD = 0.07; d = 0.09), the pre-lockdown (M = 0.00, SD = 0.07; d = 0.09), and the sample of 2019 (M = 0.01, SD = 0.09; d = 0.12). Furthermore, the sample of 2019 had a significantly greater presence of these advertisements than the lockdown (d = 0.05) and the pre-lockdown (d = 0.04) periods. This is shown in Figure 9.It was observed that the presence of throat lozenges also experienced significant changes according to the nonparametric Kruskal–Wallis test (K(3) = 1034.04, p < 0.001). As can also be seen in Figure 10, the Mann–Whitney U test showed there were significantly more advertisements of this kind before (M = 0.01, SD = 0.10) than during the lockdown (M = 0.00, SD = 0.05; d = 0.10), afterward (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0004), pre-lockdown (M = 0.0104), lockdown (M = 0.0026), and post-lockdown (M = 0.0000)), SD = 0.00), and the sample of 2019 (M = 0.00, SD = 0.02; d = 0.14). In addition, the lockdown period had a significantly greater presence of these advertisements than the following phase and the same period of 2019 (d = 0.07).Sleeping products also showed significant differences according to Welch’s test (F(3, 77,618.59) = 140.523, p < 0.001), and these differences were significant between all phases according to Dunnett’s T3 test: the greatest presence of these sleeping products could be found during the lockdown (M = 0.01 (The exact mean values of this category are: 2019 (M = 0.0020), pre-lockdown (M = 0.0060), lockdown (M = 0.0114), and post-lockdown (M = 0.0090)), SD = 0.11), which was significantly bigger that afterward (M = 0.01, SD = 0.09; d = 0.02), the pre-lockdown period (M = 0.01, SD = 0.08; d = 0.06) and in 2019 (M = 0.00, SD = 0.04; d = 0.12). Additionally, the post-lockdown phase had significantly more presence of this advertisements than the pre-lockdown (d = 0.03) and the 2019 (d = 0.10) phases, whereas the presence before the lockdown was bigger than in the sample of 2019 (d = 0.06). This is visible in Figure 11.The miscellaneous remainder of pharmaceutical products also showed significant differences according to Welch’s test (F(3, 88,480.72) = 26.28, p < 0.001). Dunnett’s T3 test showed that the pre-lockdown period (M = 0.01 SD = 0.12) was when most advertisements of this kind were found, significantly more than in the lockdown period (M = 0.01 (The exact mean values of this category are: 2019 (M = 0.0072), pre-lockdown (M = 0.0141), lockdown (M = 0.0103), and post-lockdown (M = 0.0089)), SD = 0.10; d = 0.03), the post-lockdown period (M = 0.01, SD = 0.09; d = 0.05), and the 2019 studied phase (M = 0.01, SD = 0.08; d = 0.07). Similarly, during the sample of 2019 there were significantly fewer of these advertisements than during the lockdown (d = 0.03) and the de-escalation phase (d = 0.02). These differences are shown in Figure 12.Due to the lack of relevance of the phenomenon before the lockdown, no advertisements for masks, gloves, and other products to prevent the spread of COVID were broadcasted, so nonparametric tests were needed. The Kruskal–Wallis test showed significant differences (K(3) = 766.57, p < 0.001) among this category, and the Mann–Whitney U test showed that after the lockdown there were significantly more advertisements of this kind than in 2019 and before the lockdown (both with no cases; M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0000), pre-lockdown (M = 0.0000), lockdown (M = 0.0002), and post-lockdown (M = 0.0130)), SD = 0.00), in addition to during the lockdown (M = 0.00, SD = 0.01; d = 0.10). This is shown in Figure 13.Therapeutical products also had a significantly different presence in the different studied phases, as Welch’s test showed (F(3, 99,450.70) = 189.94, p < 0.001). Dunnett’s T3 test showed the lockdown (M = 0.00, SD = 0.05) had a significantly lower presence of these advertisements than the pre- (M = 0.01 (The exact mean values of this category are: 2019 (M = 0.0146), pre-lockdown (M = 0.0063), lockdown (M = 0.0020) and post-lockdown (M = 0.0064)), SD = 0.08; d = 0.07), and post-lockdown periods (M = 0.01, SD = 0.08; d = 0.07), and the 2019 period (M = 0.01, SD = 0.12; d = 0.14). Furthermore, the sample of 2019 had also a significantly higher presence than before (d = 0.08) and after (d = 0.08) the lockdown. These results are shown in Figure 14.Welch’s test also showed significant differences (F(3, 86,697.18) = 26.28, p < 0.001) regarding the presence of dental services advertisements. Dunnett’s T3 showed that all phases were significantly different: the lockdown period (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0238), pre-lockdown (M = 0.0195), lockdown (M = 0.0031), and post-lockdown (M = 0.0075)), SD = 0.06) had a smaller presence of these advertisements than the de-escalation phase (M = 0.01, SD = 0.09; d = 0.06), the pre-lockdown weeks (M = 0.02, SD = 0.14; d = 0.16), and the sample of 2019 (M = 0.02, SD = 0.15; d = 0.18). The sample of 2019 had more presence of these advertisements than before (d = 0.03) and after (d = 0.13) the lockdown. Finally, the proportion of advertisements of dental services was bigger in the pre-lockdown phase than after the lockdown (d = 0.10). This can be observed in Figure 15.Although less present in general, paramedical services also showed significant differences according to Welch’s test (F(3, 102,707.98) = 47.85, p < 0.001). As can be seen in Figure 16, all periods were significantly different from each other: the studied sample of 2019 (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0025), pre-lockdown (M = 0.0018), lockdown (M = 0.0002), and post-lockdown (M = 0.0009)), SD = 0.05) had a greater presence of these advertisements than the pre-lockdown phase (M = 0.00, SD = 0.04; d = 0.02), the post-lockdown phase (M = 0.00, SD = 0.03; d = 0.04), and the lockdown weeks (M = 0.00, SD = 0.02; d = 0.06). In particular, during the lockdown, the presence of these advertisements was significantly smaller than before (d = 0.05) and afterward (d = 0.03), whereas during the pre-lockdown phase there was a greater proportion of these advertisements than in the post-lockdown weeks (d = 0.02).Health insurance advertisements showed significant differences according to the Kruskal–Wallis nonparametric test (K(3) = 182.35, p < 0.001). A Mann–Whitney U test showed that the sample of 2019 (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0034), pre-lockdown (M = 0.0028), lockdown (M = 0.0005) and post-lockdown (M = 0.0000)), SD = 0.06) had a significantly greater presence than the lockdown (M = 0.00, SD = 0.02; d = 0.07) and the post-lockdown (M = 0.00, SD = 0.00) phases. The same was observed with the pre-lockdown phase (M = 0.00, SD = 0.53), with a significantly greater presence of these advertisements than during (d = 0.06) and after the lockdown. Figure 17 shows this more clearly.Residential care and assistance advertisements were only present in 2019 and in the post-lockdown phase, so non-parametrical tests were conducted; the Kruskal–Wallis test showed that the differences were significant (K(3) = 173.03, p < 0.001); the Mann–Whitney U test showed that the post-lockdown period had a significantly greater presence (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0001), pre-lockdown (M = 0.0000), lockdown (M = 0.0000), and post-lockdown (M = 0.0013)), SD = 0.04) than the sample in 2019 (M = 0.00, SD = 0.01; d = 0.05), in addition to the lockdown and the weeks before, during both of which no advertisements of this kind were broadcast (M = 0.00, SD = 0.00). This is shown in Figure 18.As shown in Figure 19, similar to the previous case, other medical services were only advertised during the post-lockdown phase (M = 0.00 (The exact mean values of this category are: 2019 (M = 0.0000), pre-lockdown (M = 0.0000), lockdown (M = 0.0000), and post-lockdown (M = 0.0006)), SD = 0.03), which resulted in significant differences according to the Kruskal–Wallis test (K(3) = 87.13, p < 0.001), and the Mann–Whitney U test confirmed that the only significant differences existed between the post-lockdown phase and the other periods, which did not broadcast any advertisement of this kind (M = 0.00, SD = 0.00).Public PSAs also showed significant differences according to Welch’s test (F(3, 88,304.88) = 152.35, p < 0.001). Dunnett’s T3 test showed that during the lockdown phase (M = 0.05 (The exact mean values of this category are: 2019 (M = 0.0343), pre-lockdown (M = 0.0563), lockdown (M = 0.0528), and post-lockdown (M = 0.0362)), SD = 0.22) there were significantly more advertisements of this kind than in the post-lockdown (M = 0.04, SD = 0.19; d = 0.08) and the 2019 (M = 0.03, SD = 0.18; d = 0.09) phases. This was also observed between the pre-lockdown (M = 0.06, SD = 0.23) weeks and the post-lockdown (d = 0.10) and the 2019 (d = 0.11) periods. This is shown in Figure 20.The presence of private PSAs also experienced significant changes according to Welch’s test (F(3, 71,001.16) = 265.75, p < 0.001). All phases were different from each other according to Dunnett’s T3 test: the greatest presence of these advertisements was observed during the lockdown (M = 0.02 (The exact mean values of this category are: 2019 (M = 0.0005), pre-lockdown (M = 0.0025), lockdown (M = 0.0169), and post-lockdown (M = 0.0086)), SD = 0.13), which was significantly more than during the post-lockdown (M = 0.01, SD = 0.09; d = 0.07), the pre-lockdown (M = 0.00, SD = 0.05; d = 0.15), and the 2019 (M = 0.00, SD = 0.02; d = 0.18) phases. During the sample of 2019, significantly fewer advertisements of these kind were broadcasted than before (d = 0.05) and after (d = 0.12) the lockdown, whereas after the lockdown there was also a significantly greater presence these advertisements than before (d = 0.08). This is shown in Figure 21.Finally, no significant differences were found regarding the presence of advertisements for contraceptives and medical services. In addition, as already shown, no advertisements for antibiotics and other anti-infective products, antipyretics, homeopathic products, other medical products, or care homes were found in the studied sample.These results showed the differences in the presence of health advertisements during the 50-day long lockdown in Spain due to COVID-19 in Spring 2020, the previous and subsequent phases, and the same period of 2019. The most relevant observations are the important differences between all phases, which can be explained by the pandemic, but also by the different times of the year, with health issues associated with one or the other. An example of this is the greater presence of health advertisements during the pre-lockdown phase, which could be explained by the fact that this occurred during Spain’s flu season, when colds, for example, are more prevalent. However, the smallest presence of health advertisements during the post-lockdown period seems harder to explain by the seasonal changes; for example, recent studies have observed an increase in joint pain or headaches with higher temperatures [46]. This phenomenon could be better explained by the effects of the pandemic and the lockdown, for instance, by the tiredness of viewers of finding health issues in the media, which resulted in less health advertising, or by the sharp decrease in the presence of health products or services that were not central during the crisis, such as dental or paramedical services and therapeutical products. Similarly, it can be expected that significantly fewer advertisements about health were shown during the 50 days of 2019 compared to the lockdown period when health became a central aspect of peoples’ lives. Nonetheless, the potential effect of the different volume of advertisements in each phase should be noted because a similar number of health advertisements during the lockdown or post-lockdown phases would result in a significantly greater proportion than in the two previous phases, given that the total amount of advertisements is bigger.Regardless of this general analysis, the fluctuations of this broad category can be explained by how each of the subcategories changed. Thus, the same divergence of motifs can be observed within some of the health subcategories. For example, the presence of antihistamine products, mostly used against spring allergies, was not broadcasted before the lockdown because these weeks took place during winter. For the same reason, it appears reasonable that anti-flu products and throat lozenges were prevalent before the lockdown, during winter months when the “traditional” flu and cold season takes place.However, some differences appear to be more clearly explained by the effects of the lockdown. Among these, the smaller number of advertisements broadcasted during and after the lockdown can be explained by the decrease in advertising investment due to the complex economic situation. More specifically, focused on the different types of health products and services, we can highlight, for example, how vitamin supplements were prevalent during the lockdown, when citizens were not able to leave their houses, resulting in a lack of sun and deterioration in other healthy habits. Similarly, sleeping products were prevalent during and after the lockdown, perhaps because new schedules and a lack of exercise or work had a negative effect on the sleeping routines of many citizens [47]. The clearest case may be that regarding masks and other COVID-related products, which were mostly present after the lockdown and, to a lesser extent, during the lockdown. Similarly, the strong increase in the PSAs of private companies during the lockdown appears to be clearly related to the fact that companies joined the effort of raising awareness about the importance of staying home, as has been observed in previous studies [48]. In addition, the differences in the presence of advertisements of dental services appear to be consistent with the lockdown because the values during and after the lockdown were far below those of 2019 and before the lockdown, indicating the decrease in interest in this non-COVID19-related health issue. Indeed, dental health was strongly affected by the pandemic because dental procedures were considered to increase the risk of spreading the virus, thus forcing many practitioners to close during the lockdown [49]. These differences also show how the changes in advertising campaigns took place after some time, and the effects were mostly observed after the lockdown, primarily because it took time for companies and broadcasters to adapt to the new situation.An interesting case is that of slimming products, which were present to a greater extent after the lockdown. This could be due to the gains in weight during this period [50], but also because of the summer season and the desire to have a “beach body”. However, the significant difference between the presence of these adverts between the lockdown, which was close to zero, and the same period of 2019, indicates the potential effect of the lockdown on the reduced desire to have an “attractive” body.These effects of the pandemic and the lockdown on health advertising showed more attention was paid to some issues, such as COVID prevention, sleeping, or nutrition, than others (for example, dental health). According to the agenda-setting theory, these impacts may have affected the attention paid to the less immediate health issues that were not part of the COVID agenda, with the associated potential risk. Although this effect is not expected to be high, especially in comparison to the agenda effect of news or other media content, more research is needed on how the health advertisements studied in this article affected the visibility of the sickness and health issues they contain among the public.In general, it can be concluded that the smallest proportion of health advertisements was found precisely during the lockdown, while the biggest effects produced by the pandemic and the lockdown were visible in the post-lockdown phase, once the de-escalation started. Nonetheless, many of the limitations and the effects of the pandemic continued for many subsequent months. The effect on advertising could be caused by the need for advertisers and broadcasters to adapt to an unexpected situation, for which new advertisements had to be filmed, and new advertising contracts had to be signed. Thus, it would be interesting to continue studying the posterior phases to identify the potential long-term effects. As a hypothesis for future studies, it can be posited that facemasks or other prevention products, strongly connected with the effects of the lockdown and the pandemic, continued to gain presence after 22 June and after the period of study.Future studies could also investigate another of the relevant conclusions of this work: that the time of the year and the season strongly influence the types of health products and services that are advertised. Thus, for instance, spring-related issues, such as allergies, have a strong influence on the presence of antihistamine products. Similar effects could be expected with slimming products and their increasing presence nearer to the summer season.To summarize, we conclude that the lockdown had stronger effects on some health-related advertisements, whereas the time of the year had a stronger influence on others. Overall, however, numerous significant differences, both in the presence of health advertisements and in the presence of the different subtypes, indicated a considerable disparity between phases.Finally, some limitations of the article should be noted, mostly related to the tools used for the study. First, the ECOICOP, as previously mentioned, is a classification developed for consumption rather than advertising. Although the classification was adapted and the health elements that constitute the central aspects of the study were specifically modified, this factor should be taken into account, particularly regarding the exploratory results. It should also be highlighted that some advertisements relate to a whole brand, making it harder to assign them to a specific category; for example, advertisements of insurance companies in which health insurance was not the main subject or supermarkets that advertised the chain rather than a specific product. In these cases, the advertisement was allocated to the “miscellaneous goods and services” (12) category. Finally, it is also relevant to note that drug and medicine advertisements are required to include a disclaimer about the correct use of medicines. This disclaimer is considered in the Instar Analytics application as an independent advertisement and so was classified as a public health PSA, suggesting that caution should be used when analyzing this category. This also helps to explain why this category had the strongest presence among health advertisements. However, no significant effect is expected among other categories because advertisements from within each category follow the same rule, so the proportions in the different phases were not affected by this factor. Similarly, there was no distinction between PSAs relating to COVID-19 and those dealing with other issues. Although a clear predominance of COVID-related PSAs should be expected during the lockdown, future studies are required to further investigate this specific category.D.B.-H. analyzed the data and wrote the reports; J.G.-C. collected the data and contributed with the rationale and discussions; C.A.-C. designed the study, the RQs and the theoretical background. All authors have read and agreed to the published version of the manuscript.This research was funded by Camilo José Cela University through the 6th Call for Research Aid, as part of the project “Interaction 3.0”.Not applicable.Not applicable.The data presented in this study are available on request from the corresponding author. The data are not publicly available due to their commercial condition, as they are provided by a private company.The authors declare no conflict of interest. Proportion of health advertisements.Proportion of advertisements of slimming products in each phase.Proportion of advertisements of analgesics and other anti-inflammatory products in each phase.Proportion of advertisements of antacids in each phase.Proportion of advertisements of antihistamine products in each phase.Proportion of advertisements of anti-flu products in each phase.Proportion of advertisements of vitamin supplements in each phase.Proportion of advertisements of medical creams and sprays in each phase.Proportion of advertisements of laxatives and antidiarrheal products in each phase.Proportion of advertisements of throat lozenges in each phase.Proportion of advertisements of sleeping products in each phase.Proportion of the rest of pharmaceutical products in each phase.Proportion of advertisements of masks and other anti-COVID products in each phase.Proportion of advertisements of therapeutical products in each phase.Proportion of advertisements of dental services in each phase.Proportion of advertisements of paramedical services in each phase.Proportion of advertisements of health insurance in each phase.Proportion of advertisements of residential care and assistance in each phase.Proportion of advertisements of other medical services in each phase.Proportion of advertisements of public PSAs in each phase.Proportion of advertisements of commercial public service announcements (PSAs) in each phase.Classification of the advertisements.Reliability of the measures.General distribution of the sample.Distribution of health advertisements.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01055.txt

@@ -0,0 +1 @@
+As longevity is increasing, the 65-year-old and older population is projected to increase in the next decades, as are the consequences of age-related muscle deterioration on the quality of life. The purpose of this study was to examine the associations of the ACTN3 R577X polymorphism with quality of life and muscular strength in an older Spanish population. In total, 281 older adults participated in this study. Anthropometric measurements, chronic diseases, prescribed medications, quality of life, hand grip strength, and physical activity and nutritional status data were collected. ACTN3 R577X genotyping was determined using Taqman probes. Multivariate regression analysis revealed in adjusted model that, in men, the ACTN3 R577X genotype was significantly associated with hand grip strength (HGS), regression coefficient (β) = 1.23, p = 0.008, dimension 1 of the five-dimension questionnaire EuroQoL (EQ-5D, mobility), (β) = −1.44, p = 0.006, and clinical group risk (CGR) category (β) = −1.38, p = 0.006. In women, a marginal association between the ACTN3 R577X genotype and the CGR category was observed, with a regression coefficient of (β) = −0.97, (p = 0.024). Our findings suggest that the ACTN3 R577X genotype may influence the decline in muscle strength and quality of life in older Spanish adult males.Ageing increases the levels of functional dependency of older adults, which has individual and social implications. Therefore, a new way of thinking about and measuring aging in a socioeconomic and sanitary context is necessary [1]. Societies have to improve the robustness of assistance systems in Europe and encourage older adults to be active and healthy [2].In recent studies, physical activity (PA) has appeared to be associated with better physical health and has thus become a priority of public health systems for a better quality of ageing [3,4]. Successful ageing is a multidimensional concept defined as good physical, psychological, and social functioning in old age in the absence of major diseases. PA influences a person’s physical and psychological health and functional status as well as the self-perception of “aging well”, maintaining a good quality of life [5].The age-related progressive deterioration in skeletal muscle mass, strength, and physical function is known as sarcopenia [6,7]. Sarcopenia is associated with multiple adverse health events, including cardiovascular problems, functional disability, and increased fall incidence, hospital admissions, and mortality [8,9,10,11,12,13]. Subjects with sarcopenia have demonstrated a significantly high proportion of problems related to several dimensions of quality of life. More than 10% of individuals aged 60–69 years and approximately 40% of adults over 80 years of age are affected by sarcopenia [14,15]. The loss of autonomy and the increasing risk of additional diseases caused by sarcopenia represents a significant problem also for public health systems, which, as longevity increases, is projected to increase in the next decades [16]. While skeletal muscle properties are known to be highly heritable, evidence regarding the specific genes related to muscle strength and aging is currently inconclusive. In the past 20 years, attention has been paid to the identification of specific genes and single-nucleotide polymorphisms (SNP) in elite athletes, attributing heritable characteristics to muscle strength and physical state [17,18,19,20]. One of the most studied genetic polymorphism is ACTN3 NM_001104.4 (ACTN3_v001):c.1729C>T at exon 15 or ACTN3 [rs1815739] where arginine (R) becomes a stop (X) codon at position 577 (R577X) [21]. The protein α-actinin-3 encoded by the ACTN3 gene is one of the main structural components of the muscle fiber Z disc, which can anchor actin filaments in the sarcomere [22] and bind to a variety of structural, metabolic, and signaling proteins [23]. The main function of α-actinin-3 seems to be structural. This protein is only expressed in type II muscle fibers. Therefore, individuals with the ACTN3 577XX genotype are deficient in α-actinin-3 protein, which is associated with a lower fast-twitch fiber percentage [24], and cannot produce α-actinin-3 protein in muscle. It is estimated that the incidence of this genetic variation is 16–18% in the total population [25,26]. At first, α-actinin-3 deficiency in the general population seemed to be related to the decline of physical strength with age [27]. The relationship between strength and muscle mass in elderly people was studied to reduce mortality [28], and the influence of the ACTN3 gene R577X polymorphism on muscle phenotype and bone mineral density in this population is not well established. In addition, the role of this polymorphism in health-related quality of life (HRQoL) or morbidity in this population is of great interest.Knowledge of individual ACTN3 genotypes could provide valuable information for the management of risk factors in the elderly and promote preventive measures aimed at improving quality of life during ageing through the personalization of preventive interventions [29].Two hundred and eighty-one older adults (over 65 years old) were recruited for the study. The population was selected by simple random sampling at different primary-care centers. All recruited participants were Caucasian descendents from three or more generations. Written, signed informed consent was obtained from all subjects. The inclusion criteria were adults over 65 years old, not institutionalized, and not affected by dementia or mobility impairments. The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Area de Salud de Burgos y Soria Ethics Committee (Ref. CEIC 1446).Data were collected by a research nurse, and the following socio-demographic characteristics were included: body mass index (BMI), age, gender (male, female), prescribed drugs, falls and hospital admissions during the last year, and clinical group risk (CGR) category. The CRGs category is defined using a claims-based classification system for risk adjustment that assigns each individual to a single risk group (among mutually exclusive ones) based on historical clinical (morbidity and chronicity) and demographic characteristics, to predict the future use of healthcare resources [30].To evaluate quality of life, the EuroQoL five-dimension questionnaire (EQ-5D) was used. This questionnaire, which has been validated in Spanish [31,32,33], allows a standardized measure of HR-QoL and the use of a EQ-5D visual analogue scale (EQ-VAS), that can be applied for a wide range of health conditions and treatments. This descriptive system evaluates the patient state of health in five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension has three levels: no problems, some problems, and severe problems, and the patient has to evaluate each dimension. The results are combined in a unique parameter corresponding to the participant’s health state and then a final EQ-5D index is calculated. The EQ-VAS records an individual’s self-rated health on a vertical visual analogue scale. This is used as a quantitative measure of health outcome that reflects the subjects’ own judgement. The EQ-VAS measures the patient’s self-rated health on a 20 cm vertical, visual analogue scale with endpoints labelled as “the best health you can imagine” and “the worst health you can imagine” [34]. The advantages of this questionnaire are that it is short, easy to complete, and simple to understand [33].For the analysis of physical activity levels, the participants completed the questionnaire called Physical Activity Scale for the Elderly (PASE), which is a validated 12-item questionnaire that is designed to measure the level of physical activity in individuals over the age of 65.The PASE questionnaire assesses basic activities of older adults (walking, recreational activities, exercise, housework, yard work, and caring for others). It recovers frequency, duration, and intensity level of the activities over the previous week to assign a score ranging from 0 to 793, with higher scores indicating greater levels of physical activity [35].To assess the nutritional status, the Mini Nutritional Assessment (MNA) questionnaire was used. Older adults were classified as well nourished, at risk for malnutrition, or malnourished. The MNA has 18 questions for the evaluation anthropometric, general, dietary, and self-care parameters. We performed the full MNA for all subjects [36].Muscular strength was assessed using the hand grip strength (HGS) test. HGS was measured in the dominant hand (the average score of three measures was used in the analyses) by a maximal isometric test using a hand dynamometer [37]. Analyses of HGS were undertaken by age and gender. The European Working Group on Sarcopenia in Older Persons defined weakness on the basis of a HGS less than 30 kg in men and less than 20 kg in women [38]. For the identification of participants with clinically meaningful weakness, HGS was classified in two categories as follows: weak-intermediate and normal, according to cut-off values published by Alley et al., 2014 [39].Genomic DNA was purified from total blood using a specific extraction kit (G-spin™ Total DNA Extraction Mini Kit, Intronbio, Seongnam, Korea), and genotyping analyses were performed in the Genetics Laboratory of the Universidad de Valladolid (Soria, Spain). Our study followed recent recommendations for replicating genotype–phenotype association studies [40]: genotyping was performed only for research purposes, and the researchers responsible of genotyping were totally blinded to personal identities.For ACTN3 R577X genotyping, we used real-time PCR and Taqman probes with a Step One Real-Time PCR System (Applied Biosystems, Foster City, CA, USA).The main characteristics are presented as the mean ± standard deviation (SD) or as a percentage. Student’s t-test or analysis of variance (ANOVA) was used for continuous variables, and the Chi-square test was used for categorical variables. Deviation from Hardy–Weinberg equilibrium for the ACTN3 R577X polymorphism was tested by the chi-squared test. Values were considered statistically significant when p < 0.05, and all the p-values were two-sided.Interaction analysis between the R577X genotypes and age and gender in relation to physical performance, QoL, or chronicity phenotypes was conducted using a general linear model and further by stratification analysis. To estimate the associations of the genetic variants with physical phenotypes, a regression coefficient was derived from linear regression models in SPSS, version 19.0 (SPSS Inc., Chicago, IL, USA).We analyzed a total of 281 subjects without missing values; 46.6% of them were men, and 53.4% were women, and their age ranged from 64 to 94 years, with an average of 76.1 (± 7.1).The characteristics of the participants divided by gender are shown in Table 1. Significant differences were observed between men and women in number of falls and HGS values (p < 0.001 and p < 0.001, respectively). Males presented greater HGS values than females, and females reported more falls during the last year. The mean values of HGS were 45.6 kg for men and 30.7 kg for women. For men, a grip strength less than 32 kg was classified as “intermediate-weak”; 33.6% of men were intermediate-weak. For women, a grip strength less than 20 kg was classified as “intermediate-weak”; 24% of women were intermediate-weak (data not shown).The average PASE score suggested no differences between males and females (Table 1). The MNA score suggested that 89.4% of the participants had a normal nutritional status and 10.6% were at risk for malnutrition, with no significant sex differences.As expected, for both men and women, after Spearman correlation analysis, we observed a negative correlation between age and HGS (p < 0.001 and p < 0.001, respectively) and a positive correlation with the number of prescribed drugs (p < 0.001 and p = 0.004, respectively). In addition, for both groups, the PASE value was negatively correlated to EQ-VAS (p = 0.018 and p < 0.001, respectively) and EQ-5D (p < 0.001 and p < 0.001, respectively). For men, we also observed a marginal negative correlation between number of prescribed drugs and PASE value (p = 0.016), as shown in Table 2 and Table 3.For the selected population, the genotype distribution of the ACTN3 R557X polymorphisms was 31.5% RR genotype, 48.5% RX genotype, and 20% XX genotype. The genotype frequencies were in Hardy–Weinberg equilibrium (p = 0.865).As shown in Table 4, multivariate regression analysis revealed in adjusted model that in men, the ACTN3 R577X genotype was significantly associated with HGS, regression coefficient (β) = 1.23, p = 0.008; dimension 1 of EQ-5D (mobility), (β) = −1.44, p = 0.006, and CGR category (β) = −1.38, p = 0.006. In women, a marginal association between the ACTN3 R577X genotype and CGR category was observed, with a regression coefficient of (β) = −0.97, (p = 0.024) (Table 4). However, in women, no significant association was observed between the ACTN3 R577X genotype and HGS or mobility (see Table 4).In addition, in both men and women no significant association was observed between the ACTN3 R577X genotype and the remaining dimensions of EQ-5D and EQ-VAS (Table 4).There is not much information available regarding the health conditions (EQ-5D, EQ-VAS, HGS, chronicity, and use of medications, among others) and the physical activity levels in older adults (65 or older).The process of aging increases the risk of a number of diseases. Some studies suggest that increasing the levels of PA in the elderly population could not only postpone the development of chronic diseases but also optimize healthcare systems [41,42]. Drugs consumption, in general, can be an indicator of the overall health status, and some studies show that individuals who practice low levels of PA tend to consume more medications or more healthcare resources, as compared to individuals with higher levels of PA [43]. We also observed a negative correlation between number of prescribed drugs and PASE, EQ-VAS, and EQ-5D.In addition, PA can benefit the QoL and also shows a positive impact on depression. Our results agree with previous published studies, showing that participants with higher levels of PA presented a higher EQ-5D index and self-rated health outcome (EQ-VAS) [44,45,46].This study, to our knowledge, is the first to examine the relationship between QoL (EQ-5D), muscular strength (HGS), and chronicity/morbidity (CGR category) and the ACTN3 R577X polymorphism in an older Spanish population.In our study, men older than 65 years of age showed differences in HGS, mobility (dimension 1 of EQ-5D), and CGR category according to the ACTN3 R577X genotype. On the other hand, for women we only found a statistically significant association between CGR category and ACTN3 R577X genotype.In man, we observed that the ACTN3 577XX genotype was associated with higher HGS values, not having any problem in mobility, and being in group 0 or 1 of CGR category (p = 0.006, p = 0.008 and p = 0.006, respectively). For women, the ACTN3 577XX genotype was marginally associated with being in CRG group 0 or 1 (p = 0.024). Recently, Ma et al. found evidence of gender- and age-specific associations of ACTN3 R577X genotypes with physical performance phenotypes (including HGS) in older populations [47]. In addition, Dato and colleagues reported that the genetic component of frailty was higher among males than among females and higher in older subjects [48].Physical performance measures including HGS are associated with healthy aging, and lower scores increase the risk of mortality [49,50]. Muscle strength and mass are protective against all causes of mortality in elderly [28].Previous studies have shown that the ACTN3 genotype is a modulator of muscle mass and function and of sarcopenia risk in elderly adults, initially being the allele R of ACTN3 R577X associated with greater maintenance of strength and function or with sarcopenia protection [24,51,52]. Better strength associated with allele R has been frequently studied in athletes [53]. From a physiological point of view, this relationship could be due to the association of the R allele with an increase in type II muscle fibers and the ability to maintain fast-twitch fiber size and mass with age in these subjects [24]. Surprisingly, in elderly people, some authors have reported that better results in strength tests were associated with the ACTN3 XX genotype, while for other researchers this genotype appeared as the least favorable [54]. Lifestyle habits of the elderly, such as regularly practicing physical exercise, seem to be determinant. Recently, Romero-Blanco published that women with the ACTN3 XX genotype improved their muscle strength after 24 months of training (in the study they tried to homogenize the characteristics of the participants, such as training, gender, age and lifestyle) [55]. Seto et al. hypothesized that the absence of α-actinin-3 produces an increase of calcineurin activity, which reprograms the metabolic phenotype of fast muscle fibers and results in better adaptation of skeletal muscles to training [56]. Moreover, Garton et al. suggested that α-actinin-3 deficiency also protects against muscle wasting [57].While the deficiency of α-actinin-3 has no apparent association with muscle diseases, there is an established relationship with morbidity in people who are frail, suggesting that in centenarians, it could provide a survival advantage [58].Deschamps et al. reported that centenarians with the ACTN3 XX genotype may be less predisposed to chronic diseases [59]. We have also observed an association between the ACTN3 XX genotype and being in the G0 or G1 CGR (healthy or with an acute process (G0) and single minor chronic disease (G1)). While several methods are available to identify frail patients, there are no guidelines for the identification of complex elderly patients, who often present high levels of multi-morbidity. Yet, it is well known that multi-morbidity in the elderly is associated with poor outcomes, and the commonly used CGR classification system is a good tool to evaluate this situation.On the other hand, as far as we know, there are no published studies evaluating the relationship between QoL and the ACTN3 genotype. We found a statistically significant association between dimension 1 of EQ-5D and the ACTN3 R577X polymorphism. Participants with the XX genotype had a higher probability of not having any problem in mobility than those without this genotype. These results are also in accordance with the association of the ACTN3 XX genotype with HGS values that we observed. Sarcopenia is associated with healthy outcomes and an obvious decline in QoL.Seto et al. [60] reported that genotype differences in fast muscle force production result in fast-twitch fibers developing slower activities, suggesting that the lack of alpha-actinin-3 may cause a faster decrease in muscle function with increasing age. The loss of type II muscle fibers may be particularly important concerning the influence of the ACTN3 R577X in the elderly, as ACTN3 is mainly expressed in this fiber type. Moreover, people without α-actinin-3 show better adaptation to resistance training [61].This study has several limitations. For example, muscle strength is a complex phenotype, which is likely influenced by numerous genes and genetic variants, as well as other environmental factors that may be interacting with these genes in several pathways. The sample group was divided by gender, which may have reduced the statistical power. On the other hand, despite the small sample size of the current study, our population was homogeneous and well defined in terms of phenotype assessment, and it is known that sarcopenia may be gender-dependent. To our knowledge, this is the first study to evaluate the relationship between QoL and the ACTN3 genotype. Moreover, the study follows the STREGA guidelines, all participants were randomly recruited by a research nurse, genetic polymorphism was selected considering its prevalence and functional impact, no departure from Hardy-Weinberg equilibrium was detected, nor multiple testing and reporting of quantitative (continuous) outcomes were used. A better reporting in studies facilitates the synthesis of research results and the further development of study methods in genetic epidemiology improving the understanding of the role of genetic factors.In this study, we found evidence of gender-specific associations of the ACTN3 R577X polymorphism with muscular strength, QoL, and morbidity in the older population. Our results support the hypothesis that the lack of alpha-actinin-3 may cause a faster decrease in muscle function with increasing age. Nevertheless, the specific underlying mechanisms will require further investigation. Establishing the influence of the ACTN3 R577X variant on functional health status or on quality of life in older adults is necessary to determine if this genotype could be useful for identifying individuals who may be more susceptible to sarcopenia and who may need specific global health interventions.We found an association between the ACTN3 R577X genotype and muscular strength in older men.We found an association between the ACTN3 R577X genotype and the dimension of mobility of EQ-5D in older men.Our results support the hypothesis that the lack of alpha-actinin-3 may cause a faster decrease in muscle function with increasing age.We found an association between the ACTN3 R577X genotype and chronicity and multimorbidity in Spanish older adults.Formal analysis, Z.V.; Funding acquisition, Z.V.; Investigation, A.F.-A., A.G.-A. and Z.V.; Methodology, A.F.-A., A.G.-A., J.A.R.-D., S.C.-M. and J.L.-L.; Resources, A.F.-A.; Supervision, Z.V.; Writing – review & editing, Z.V. All authors have read and agreed to the published version of the manuscript.This research was funded by Fundación Científica Caja Rural de Soria: 2018.The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by Area de Salud de Burgos y Soria Ethics Committee (Ref. CEIC 1446).Informed consent was obtained from all subjects involved in the study.Data sharing not applicable.The authors declare no conflict of interest.Characteristics of the Study Participants (N =281).Note: Values are percentages for categorical data and mean and standard deviation for continuous data. SD, standard deviation; BMI, body mass index; CRG, Clinical Risk Groups; EQ-5D, EuroQol 5-Dimension questionnaire; EQ-VAS, EuroQol Visual Analogue Scale; VAS, Visual Analogue Scale; PASE, Physical Activity Scale for the Elderly; HGS, Hand Grip Strength; MNA, Mini Nutritional Assessment. Statistically significant variables are in bold.Pearson correlation coefficients of the analyzed variables in men.Note: Each cell contains two values: (a) Pearson correlation coefficient; (b) p value, indicating if the correlation is significant. Statistically significant variables are in bold.Pearson correlation coefficients of the analyzed variables in women.Note: Each cell contains two values: (a) Pearson correlation coefficient; (b) p value, indicating if the correlation is significant. Statistically significant variables are in bold. Logistic regression of the association between ACTN3 R577X polymorphism (recessive model XX/(RR+RX) (rf)) and each dimension of EQ-5D, EQ-VAS, HGS, and CRG.* Problems in each dimension of EQ-5D. Statistically significant variables are in bold.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01056.txt

@@ -0,0 +1,6 @@
+Urban resilience in the context of COVID-19 epidemic refers to the ability of an urban system to resist, absorb, adapt and recover from danger in time to hedge its impact when confronted with external shocks such as epidemic, which is also a capability that must be strengthened for urban development in the context of normal epidemic. Based on the multi-dimensional perspective, entropy method and exploratory spatial data analysis (ESDA) are used to analyze the spatiotemporal evolution characteristics of urban resilience of 281 cities of China from 2011 to 2018, and MGWR model is used to discuss the driving factors affecting the development of urban resilience. It is found that: (1) The urban resilience and sub-resilience show a continuous decline in time, with no obvious sign of convergence, while the spatial agglomeration effect shows an increasing trend year by year. (2) The spatial heterogeneity of urban resilience is significant, with obvious distribution characteristics of “high in east and low in west”. Urban resilience in the east, the central and the west are quite different in terms of development structure and spatial correlation. The eastern region is dominated by the “three-core driving mode”, and the urban resilience shows a significant positive spatial correlation; the central area is a “rectangular structure”, which is also spatially positively correlated; The western region is a “pyramid structure” with significant negative spatial correlation. (3) The spatial heterogeneity of the driving factors is significant, and they have different impact scales on the urban resilience development. The market capacity is the largest impact intensity, while the infrastructure investment is the least impact intensity. On this basis, this paper explores the ways to improve urban resilience in China from different aspects, such as market, technology, finance and government.Urban resilience is the ability of a city to recover in the face of various disasters. Covid-19 is one of the most serious “black swan” outbreaks in global public health security in recent years. The outbreak, which has brought many cities of China and around the world to a standstill, has severely tested the ability of cities to repair themselves in the face of acute shocks and long-term stress. Faced with the rebound of the epidemic in many countries in Europe and the US, China has taken a series of commendable measures to combat the epidemic in the early stage of the outbreak. However, it is undeniable that the shortage of public health resources and the lagging early warning mechanism of the medical and health system have also been exposed one by one. In fact, this is already the sixth “public health emergency of international concern” (PHEIC) declared by the World Health Organization in 15 years, and public health events have become one of the great threats to contemporary urban development. However, with the increasingly complex urban operation system, for China in the critical period of structural adjustment and social transformation, the threats may also come from a series of risks and challenges such as periodic economic crisis, global climate change, and urban terrorist attacks. Faced with the constraints of many uncertain factors, how to ensure the stable survival and sustainable development of cities has become the focus of attention of governments and scholars all over the world [1,2]. In essence, the difference in the ability of cities to cope with risks is caused by the urban resilience of different cities [3]. As the ability of urban system to adapt to uncertainty, urban resilience can not only break the sustainability paradox from the perspective of imbalance [4], but also help urban system to digest and absorb external interference to the greatest extent, so as to maintain its original features and key functions [5]. For China, which is undergoing continuous industrialization and rapid urbanization, it undoubtedly provides us with a new research perspective, replacing “short-term pain relief” with “long-term pain treatment” [6], and maintaining urban operation in a more systematic and long-term way under the normal situation of the epidemic. Therefore, scientific measurement of the level and spatial evolution of China’s urban resilience and in-depth analysis of the spatial differentiation pattern of its driving factors will help speed up the clarification of the current development status of China’s urban resilience and point out the direction for the healthy development of cities in the new era.This paper takes 281 cities in China from 2011 to 2018 as research objects. On the one hand, the evaluation system is constructed from the five aspects of economy, engineering, society, ecology and institution to investigate the spatiotemporal differentiation pattern of China’s urban resilience development. On the other hand, the driving factors are selected from the perspectives of government, technology, market, openness and finance. The multi-scale geographical weighted regression model (MGWR) is used to deeply analyze the effects of various variables on urban resilience at different impact scales, so as to provide reference basis for improving urban resilience in China.The term “resilience” first appeared in the field of mechanics to describe the ability of metals to recover after being deformed under external forces [3]. With the deepening of the research, the concept has undergone two thorough modifications and improvements from the initial engineering resilience to the ecological resilience and then to the evolutionary resilience. In a sense, engineering resilience is the closest to the concept of resilience commonly understood by people. It refers to the ability of the whole system to recover to the equilibrium or stable state before disturbance after being disturbed [7]. With the deepening understanding of system and environmental characteristics and their mechanism of action, ecologist Holling first applied the idea of resilience to ecology in 1973 and proposed ecological resilience. Ecological resilience breaks the limitation that engineering resilience holds that the system has a single equilibrium state, and emphasizes the sustainable development ability of the system [7]. On this basis, based on the adaptive cycle theory of Gunderson and Holling [8], the concept of evolutionary resilience was proposed, which pays more attention to the adaptability, learning and innovation ability of the system. Compared with engineering resilience and ecological resilience, evolutionary resilience is more convincing in theory, and it is also regarded as the reference benchmark for urban resilience research. Therefore, urban resilience can be regarded as the inheritance and redevelopment of the traditional resilience theory. It takes the urban ontology as the research object and emphasizes the ability of urban system to maintain its original main features, structure and key functions after encountering external harassment [5]. The research focuses on deconstructing the interactive logic relationship between modern city and its crisis risk, as well as discussing the systematic response measures taken by cities to deal with complex disturbances. At the same time, urban resilience is also regarded as an innovative approach to achieve sustainable development goals. Compared with the fail-safe concept emphasized in the early stage of sustainable development, urban resilience focuses on the integrity of the overall urban pattern and the sustainability of functional operation and is a safe-to-fail approach [9].The evaluation method and index system of urban resilience are the in-depth study of the theoretical framework of resilience, which is helpful to translate the theoretical analysis into urban construction. As a new research subject, the academic circle has not put forward a unified evaluation method for urban resilience. However, by definition, resilience can be measured by the amount of disturbance that the urban system can absorb if its original functional structure remains unchanged [10,11]. As the core of resilience measurement, disturbance threshold is difficult to be obtained directly and is mostly represented by multi-functional indexes. As interdisciplinary studies, many scholars try to build evaluation systems from different starting points, which are mainly divided into three categories: different dimensions, different characteristics, and different scales. Among them, urban resilience evaluation can construct index system from multiple dimensions and single dimension. From the perspective of multiple dimensions, scholars usually select comprehensive indicators from the aspects of economy, society, nature and infrastructure, etc. [12,13], while the single dimension usually takes a specific resilience [14] or natural disaster [15] as the object and constructs highly targeted evaluation indicators. The evaluation system based on different characteristics is mainly based on the seven characteristics of resilient cities proposed by Rockefeller Foundation. On this basis, some scholars have made a comprehensive classification description from individuals, organizations and localities [16]. However, it should be noted that there is still lack of quantitative models and unified scales to fit the relationship between different system characteristics and quantify the resilience of urban system. The evaluation scale of resilient cities can be divided into macro metropolitan area, medium single city and micro community. They are respectively represented by the Resilience Capacity Index (RCI) for metropolitan areas proposed by Berkeley Research Institute, the resilience city index system launched by Rockefeller Foundation of the United States, and the community resilience index which includes six elements of community infrastructure, security, environment, economy, system, society and population [17].Targeted promotion strategies are the ultimate goal of urban resilience evaluation research. The increasing research on urban resilience strategies reflects the rapid development trend of research from theoretical exploration to practical application. To be specific, urban resilience practice at the present stage has two main footholds. The first is based on the characteristics of resilience. Jack Ahern, a professor at the University of Massachusetts, United States, believes that resilient cities should have five characteristics: multi-function, redundancy and modularity, (biological and social) diversity, multi-scale networks and connectivity, and adaptive planning and design [18]. And on this basis, five strategies of urban resilience planning and design are proposed [9]. This emphasis on building urban resilience by some means, including providing sustainable ecosystem services in view of the limited space, the distributed system to provide the same functionality to strengthen risk dispersion ability, based on the diversity of the biological system reaction to form low impact development pattern, repeated set loop to maintain functional connectivity and the use of adaptive design decisions to the problem. The second is to refine to the practical level. According to the theory of urban resilience and relevant studies by scholars [19], the main ideas of urban resilience planning are as follows: Vulnerability analysis and evaluation, uncertainty planning for cities, strengthening the leading role of the government in urban governance and formulating specific implementation strategies, etc. In this way, we can formulate policies and measures based on future scenarios on the basis of clearly understanding the mechanism of vulnerability, so as to effectively coordinate the conflict of interests in the implementation process of resilient cities and avoid the occurrence of basic dysfunction and fault of service supply when the urban system is exposed to external threats.In conclusion, the research on urban resilience has shifted from Bohr paradigm to Pasteur paradigm, and there are more practical scholars, making the research on resilience more timely and targeted, and overcoming the gap between “knowledge generation” and “knowledge transformation”. In terms of concept, although unified connotation has not been formed within the discipline, it does not affect the establishment of relevant theoretical framework and the development of research practice, but makes the research gradually extend from theoretical analysis to internal mechanism. In terms of evaluation, although no consensus has been reached on the quantitative method of urban resilience, it has made some initial achievements in the construction of resilience index system. The evaluation system is relatively comprehensive, covering different dimensions, characteristics and scales. Nevertheless, the index system that accords with China’s national conditions still needs to be improved. In terms of strategy, most of the existing studies provide guiding principles for decision-making and policy making of urban management. However, there is no in-depth study on the formation mechanism and influencing factors of urban resilience, and there is also a lack of discussion on spatial scale. Compared with the existing literature, the innovations and contributions of this paper are: (1) Evaluation index: Aiming at China’s current national conditions, this paper integrates traditional data and emerging data, and adds institutional dimension on the basis of four dimensions of economy, society, engineering and ecology to highlight the main role of government and non-governmental organizations in urban resilience development. (2) Research method: MGWR model is used to analyze the driving factors, which can effectively analyze the influencing factors from different spatial scales. Firstly, the city is placed in a macro background to consider the role of influencing factors on the overall dimension. Secondly, it is dynamic across multiple scales from regions to communities, even to families and individuals, presenting its spatial role. (3) Optimization path: In order to reflect the continuous relationship between promotion strategy and evaluation system, the mechanism of action of urban resilience in China is deeply explored, and the internal relationship between optimization path and evaluation system is further clarified according to its influencing factors.The entropy method provides a basis for multi-dimensional comprehensive evaluation and is more objective, avoiding the influence of human factors in subjective assignment. Therefore, the entropy method is adopted in this paper to give weight to the urban resilience index system of China’s cities, and the urban resilience index, economic resilience index, engineering resilience index, social resilience index, ecological resilience index and institutional resilience index are calculated.First, the range method is adopted to standardize the original data. According to the positive and negative of each index of urban resilience, the standardization is as follows:(1)Positive indicators: xij′=xij−min(xij)max(xij)−min(xij)
+          (2)Negative indicators: xij′=max(xij)−xijmax(xij)−min(xij)
+          where: xij is the original value of index i in region j, xij′ is the standardized value, max(xij) and min(xij) are the maximum and minimum values of this index respectively.Secondly, calculate the entropy value ej of the index xij:(3)ej=−k∑i=1npij×ln(pij)In Equation (3), Pij=xij′∑i=1nxij′, k = 1/ln(n), n is the sample size, which satisfies ej > 0.Thirdly, calculate the urban resilience index and sub-resilience index Si of each city:(4)Si=∑j=1mwj×xijIn Equation (4), wj=dj∑j=1mdj represents the weight of index j, where dj = 1 − ej.ESDA is an ideal data-driven analysis method. It takes spatial association degree as the core and discovers spatial clustering and spatial anomaly by describing and visualizing spatial phenomena of objects, thus revealing the spatial interaction mechanism between research objects. It usually includes global spatial autocorrelation analysis and local spatial autocorrelation analysis.Global Moran’s I index measures the overall trend of spatial correlation of unit attribute values in adjacent or adjacent regions in the whole study area.
+            (5)I=∑i=1n∑j=1nwij(xi− x ¯)(xj− x ¯)S2∑i=1n∑j=1nwijIn Equation (5), I is the global Moran’s I, n is the number of cities, xi and xj respectively represent the urban resilience of region i and j, and Wij is the spatial weight matrix.Local spatial autocorrelation mainly analyzes the different spatial association patterns that may exist at different spatial locations in order to find the spatial heterogeneity among data. In this paper, Local Moran’s I, also known as Local Indicator Spatial Association (LISA), is used to indicate the significance of spatial difference.
+            (6)LISA=(xi− x ¯)Sx2 ∑i=1n[wij(xi− x ¯)]In Equation (6): Sx2=∑ (xi− x ¯)2/n. A positive LISA means a high value surrounded by a high value or a low value surrounded by a low value, namely, H-H or L-L. A negative LISA means that a low value is surrounded by a high value or a high value is surrounded by a low value, namely, H-L or L-H.Compared with the traditional classical GWR model, MGWR makes its bandwidth specific by allowing variables to have different spatial smoothing levels, which solves the problem of limiting the optimal bandwidth of all variables to be the same in GWR. At the same time, the specific bandwidth of each variable can also be used as the index of the spatial scale of each spatial process, which makes the spatial process model generated by this multi-bandwidth method more real and useful [20]. Therefore, this paper uses the MGWR model to discuss the spatial differentiation and spatial scale differences of the driving factors of urban resilience in 281 cities of China from 2011 to 2018. The formula is as follows:(7)yi=∑j=1kβbij(ui ,vi)xij+εiIn Equation (7), βbij represents the regression coefficient of the local variable, bij represents the bandwidth used by the regression coefficient of the variable j, and (ui ,vi) represents the spatial coordinates of the sample point i, xij is the observed value of the variable j at the sample point i, and εi is the random disturbance term.The “sustainability” of the world in the 21st century largely depends on the sustainability of cities, and urban construction planning is bound to transition from “quantitative accumulation” to “qualitative transformation”. In this historical tide, urban resilience has provided a new paradigm for urban development, enabling urban development strategies to respond to economic, social or environmental changes. In the future, urban resilience and urban development will be characterized by interactive coupling, and urban resilience can be enhanced through the regulation of urban construction. Before this, it is necessary to accelerate the development of an evaluation system to measure urban resilience. Urban resilience is a multi-dimensional interdisciplinary subject involving a wide range of areas [21]. Based on data availability, this paper classifies urban resilience into four dimensions: economic resilience, engineering resilience, social resilience and ecological resilience. In addition, different from previous urban security strategies, urban construction in the post-epidemic era should focus more on unknown and unpredictable risk shocks, and it is necessary to incorporate epidemic prevention and control planning into the territorial space planning system. In view of the important role played by the government and some non-governmental organizations in the epidemic prevention and control work, this paper adds the dimension of institutional resilience, so as to comprehensively analyze the development status of China’s urban resilience (Table 1).Economic resilience emphasizes that the level of urban economic development should match the resilience to ensure the coordination and symbiosis among urban subsystems, so as to ensure the rapid resumption of production and construction in the event of external risks. Therefore, the indicators are selected from three aspects of economic strength, economic diversity and economic extroversion. Among them, the economic strength reflects the ability of economic security that a city can provide when dealing with external turbulence, and it is a direct manifestation of the city’s resilience. Economic diversity advocates the change of single economic structure and the adoption of diversified economy as a new development model, so as to maintain the diversity of urban economy and stratum. Economic extroversion requires cities to actively participate in the global economic division of labor. In the context of the global impact of the COVID-19 epidemic, China is at the forefront of the world economic recovery. All regions should make full use of their distinctive urban personalities and continue to consolidate their positions in the global industrial chain.Engineering resilience is to evaluate whether the infrastructure and community construction are capable of rescue and emergency response in the event of occasional extreme events and crises in the city. It mainly selects indicators from the three criteria: urban evacuation ability, basic supply and drainage ability and external communication ability. The urban evacuation ability is mainly reflected in the traffic accessibility, which reflects the city’s ability to shelter and evacuate people when the crisis occurs. The basic supply and drainage ability reflects the service level of urban hardware facilities. Edwin Chadwick, in his Reporton the Sanitary Condition of the Labouring Population of Great Britain, pointed out that infectious, endemic and other diseases were caused, aggravated or spread by crowded, damp and dirty living conditions, and that the incidence of these diseases could be effectively reduced by means of drainage, electricity supply and cleaning [22]. External communication ability emphasizes that the response time should be compressed to the greatest extent when the epidemic or risk occurs, and the loss should be reduced by timely transmission of disaster information so as to effectively organize the urban social system.Social resilience is the integration of factors such as demographic characteristics, organizational structure and human capital, which reflects the ability of cities to deal with the problems caused by social, political and environmental changes [23]. It mainly includes population adaptation ability, social security ability and risk response ability. Population adaptation ability reflects residents’ sense of belonging to the city and their ability to withstand crises. Social security ability and risk response ability emphasize that social resilience should attach importance to both the hard environment and the soft environment. The report of the 19th CPC National Congress clearly stated that to implement the healthy China strategy, cities must build a strong public health system under the normal situation of the epidemic, which is firstly reflected in the basic conditions of disease prevention and control, and secondly in the construction of the talent team for disease control.Ecological resilience refers to the extent to which natural resources and ecosystems can resolve crisis changes before they are reorganized and formed into new structures. It can be measured in terms of environmental pressure, governance ability and service ability [24]. Environmental pressure is generally represented by the environmental emission intensity index of the industrial sector in the process of ecological restoration. Governance ability is a measure to buffer and repair these environmental pressures. Service ability emphasizes the enhancement of urban ecological service capacity through planning transformation so as to alleviate the overload operation of the ecosystem.Institutional resilience refers to the guiding capacity of governments and non-governmental organizations to govern communities, the most important of which is the need to change political awareness and motivation so that cities can overcome obstacles under their guidance to achieve resilience and sustainability. Compared with other countries, China’s excellent performance in epidemic prevention and control is related to three factors. First, basic guarantee. In the face of the epidemic, the government eliminated the worries of patients by exempting treatment fees, thus avoiding the large-scale outbreak of anxiety among residents. Second, manpower input. After the outbreak of the epidemic, people from the national level to the local level have responded one after another. Strong grass-roots mobilization capacity provides effective guarantee for the realization of the full coverage of mass prevention and control or joint prevention and control. Third, financial input. In the early stage of the epidemic, the guiding principle of “proactive fiscal policy should be more proactive” is repeatedly emphasized, which provided a solid financial foundation and guarantee for epidemic prevention and control and stable economic operation.Therefore, this paper selects corresponding indexes based on the above criteria layer, as Table 1 showing.The time node of the data used in this paper is 2011–2018, and the original data of the basic indicators are from China City Statistical Yearbook and statistical yearbooks of provinces and cities, and part of the environmental index data is obtained from the bulletin of environmental quality of cities. Considering the availability of data, the research sample of this paper is 281 cities in China.Identifying the spatial and temporal differentiation of urban resilience can effectively reveal its evolution process and potential formation mechanism [25], which is an important basis and premise for optimizing and improving urban resilience. Therefore, it is necessary to explore the temporal and spatial attributes of urban resilience. In order to effectively analyze spatiotemporal evolution of urban resilience, on the one hand, the urban resilience index of China from 2011 to 2018 is obtained by using the entropy method, and its evolution state in the time dimension is described in more detail by combining the α value and the global Moran’s I index. On the other hand, in order to observe the difference of urban resilience in spatial distribution more intuitively, ArcGIS10.3 software was used to draw the spatial pattern evolution map, and the hierarchical structure analysis and local Moran’s I were used to calculate and clarify the different development modes and spatial correlation of urban resilience in various regions.From the perspective of overall evaluation, the overall urban resilience index showed a continuous decline, from 0.0885 in 2011 to 0.0773 in 2018 (Table 2). The analysis shows that since 2011, China has gradually entered into the period of concentrated outbreak of “urban disease”. In the context of “emerging + transformation + rapid urbanization”, “acute, chronic, complications” in cities co-occur and resonate, triggering systemic risks of urban operation and resulting in continuous decline of urban resilience.In terms of the sub-resilience, except the social resilience remained basically unchanged, the resilience of the remaining subsystems all decreased, among which the ecological resilience showed the most significant decline, with a decrease of 24.58% compared with 2011. The main reason is that China is in the critical stage of industrial development and transformation, and industries characterized by extensive, low-level and high energy consumption will be eliminated and transferred, which will have a significant impact on urban spatial layout and urban operation and management.From the perspective of sequence transformation, economic resilience began to dominate, while engineering resilience decreased significantly. On the one hand, economic development plays an important role in urban resilience. The rapid economic development helps cities gather social wealth to improve their disaster response capacity. On the other hand, the engineering resilience dropped from the first place in 2011 to the third place. This may be due to the rapid expansion of population, the lack of predictability in urban planning, as well as the lag of infrastructure construction, which leads to the phased decline in the engineering resilience.α convergence test is generally measured by the coefficient of variation, which can be used to directly reflect the changes in the distribution pattern of urban resilience in a region. For any t, there is αt + 1 < αt, indicating that the gap of urban resilience is gradually narrowing and there is α convergence [26].On the whole (Table 2), the α value of China’s urban resilience is on the rise, which means that the α convergence does not occur, while the regional differences become more and more significant as time goes on. The α value of the sub-resilience is generally consistent with the overall resilience and shows an upward trend, while only the engineering resilience shows an obvious sign of convergence. This is because the outbreak of “urban disease” has accelerated the pace of infrastructure construction in various regions, thus gradually narrowing the gap of engineering resilience between regions.In terms of regions (Figure 1), urban resilience in eastern, central and western China shows no obvious signs of convergence. The reason lies in that the policy support at the national level has not yet met the overall urban development needs of each region, while the advantageous resources, driven by the tendency of profit and preference, are mostly concentrated in the relatively developed cities in each region. Therefore, the widening trend of differences between regional cities is presented.The urban resilience of China is characterized by positive spatial autocorrelation, and Moran’s I value rises from 0.1717 in 2011 to 0.1812 in 2018 (Table 2). It indicates that the agglomeration trend in spatial distribution is gradually strengthening. However, in general, the overall Moran’s I index shows a trend of “decline first and then rise”, which may be due to the lack of strategic guidance in the overall planning of regional urban resilience development in the early stage, and the similarity degree of urban resilience in various regions decreases.From the perspective of sub-resilience, economic resilience also has a trend of first decreasing and then rising, showing a significant positive correlation with the overall urban resilience. This once again proves that economic resilience has a great impact on urban resilience. The Moran’s I value of engineering resilience is low and presents an inverted “V” type, while the Moran’s I value of social resilience and ecological resilience are both higher than that of engineering resilience and present a positive “V” type. It indicates that the development of social resilience and ecological resilience will compress the improvement space of engineering resilience to some extent. However, the institutional resilience has been continuously declining, indicating that the agglomeration trend is weakening. This is mainly because cities rely too much on their own development needs and lack unified policy guidance in improving the resilience level, leading to an increasingly obvious regional trend of institutional resilience.In order to reflect the spatial pattern changes of urban resilience, this paper divides the resilience levels into low level, medium low level, medium high level and high level according to the regional economic classification standard of the World Bank and according to the mean values of resilience index of 50%, 100% and 150% (Figure 2) [27].From the perspective of the quantity of each grade, there is no obvious change in the proportion from 2011 to 2018. In addition to the increase from 13% to 16% in the proportion of cities with high resilience, the proportion of cities with low and medium high resilience declined slightly, from 13% to 12%, and from 14% to 11%, respectively. Cities with medium low resilience levels have been dominant, with 169 in both 2011 and 2018. The main reason lies in the gradual exposure of low-end industrial structure in China, and the inability of ecological carrying capacity to adapt to the rapid development of urban economy, which leads to the low level of resilience.From the perspective of spatial distribution, the spatial heterogeneity of urban resilience is obvious. The resilience level of eastern cities is generally higher than that of the central and western regions, and the high-value areas are steadily clustered in agglomerates in the Beijing-Tianjin-Hebei urban agglomeration, Shandong Peninsula and Yangtze River Delta urban agglomeration. This is because the eastern region relies on a reasonable industrial structure and perfect infrastructure to gather a large number of social resources, which ensures that the city can quickly respond to the disaster and bear the high cost of recovery. In contrast, urban resilience in the central and western regions is distributed in a “Mosaic” pattern. This is mainly reflected in the cluster agglomeration of cities with medium low-level resilience and point-like dispersion of cities with high level resilience. Moreover, there was no significant change in the point-like high-value areas in 2011 and 2018, which are dominated by provincial capitals and central cities in urban agglomerations, forming a certain spatial shielding effect [27]. Most of these cities produce strong siphon effect by virtue of policy advantages and constitute a self-consistent closed structural system. It does not have strong expansibility, and only through strong policy intervention can it enhance the radiation power.By analyzing the resilience levels of eastern, central and western China at the provincial level (Figure 3), the development model of urban resilience in various regions of China can be better understood. By observing the urban resilience index of various provinces in China, it is found that the resilience levels of the eastern and western provinces are greatly different, and there are significant polarization phenomena in both provinces. These high-value polarization areas are highly coincident with the distribution of Chinese municipalities directly under the central government, among which Shanghai is the most prominent. On the one hand, Shanghai has been making continuous efforts in urban planning, operation and management in recent years. On the other hand, it has increasingly strengthened its leading role as a central city through structural adjustment, innovative development and other measures, which has made a qualitative improvement of urban safety construction. In contrast, the level of development among the central provinces is very balanced, with no more prominent or backward provinces.Combined with the ranking of urban resilience of each province, the three regions have different development patterns (Figure 4): (1) The eastern region has formed a three-core driving model supported by Beijing, Tianjin and Shanghai. Shanghai, Beijing and Tianjin, with their status as centers of politics, culture and international exchanges, have amassed resources and wealth, forming a relatively complete urban system. And with the gradual growth of the middle resilience class, it has some gradation, which enables the urban resilience of other provinces to ascend the ladder and alleviates the antagonistic emotions caused by the gap in resilience. (2) The central region is a rectangular structure. All the provinces in the region have medium low-level resilience, with the largest inter-provincial difference less than 0.01, lacking the “leader” with high resilience. This indicates that although the central region has a good development momentum in recent years, it cannot quickly catch up with the national average due to the original insufficient economic strength and insufficient supply of public goods and services, resulting in the low level of urban resilience. (3) The western region formed the pyramid structure. In this region, there are distinct layers and obvious polarization. Chongqing and Xinjiang have “fault exits” with high resilience levels of 0.1974 and 0.1466 respectively, while the rest of the provinces have gradually decreased. At the same time, the western region is also the only region with a low level of resilience, which undoubtedly exposes the imbalance of its urban resilience development. Due to the strong Matthew effect of Chongqing and Xinjiang themselves, the polarization of urban development in western China is extremely serious, which needs to be solved by the government to formulate more complete resource allocation policies in the future.Local spatial autocorrelation analysis can effectively reveal the spatial characteristics of the correlation between local and surrounding urban resilience. The study found that the spatial characteristics of the city on the time cross section did not change substantially. And according to the adaptive cycle theory, this type of association was divided into four stages of urban development (Figure 5) [28]. (1) H-H type represents the city in the mature stage. This type of city is mainly concentrated in the developed regions dominated by the urban agglomeration of the Yangtze River Delta and Pearl River Delta. The economic leading position of these cities determines the high maturity of infrastructure and social development of these cities. In addition, in recent years, the pattern of coordinated development and the effect of environmental governance gradually appear, making the overall urban resilience higher in the region. (2) L-L type represents the city in the growth stage. A large part of them are located in city clusters in the Yellow River Basin. Different from cities in the Yangtze River Basin, cities along the Yellow River have not formed closely related economies, and many regions are resource-based cities, seriously restricting their economic, engineering and ecological resilience development. In addition, some cities are located in the southwest of China. The modern industry, which has been blank for a long time, makes the local economic structure, infrastructure construction and human quality backward. It is the “congenital deficiencies” that leads to the overall low urban resilience. (3) L-H type represents cities in the updating stage. There are only a few cities of this type. In 2011, Xuancheng and Langfang were included, while in 2018, Xuancheng was the only city left, and Langfang successfully changed into H-H type. It may be that the rapid development of Beijing-Tianjin-Hebei region has an increasingly strong radiation-driven effect on Langfang, making it more resilient. (4) H-L type represents the city in the gestation stage. It can be found that they all belong to the capital cities of western provinces, such as Kunming, Lanzhou and Chongqing. Compared with other cities, these cities can use the siphon effect to obtain more resources. However, there is not enough radiation effect to drive the development of small and medium-sized cities around them.As an important realization path for sustainable development, urban resilience is not simply composed of a single subsystem but formed through the coordination and optimization of multiple elements, and the correlation strength between each element and the resilience index is different. Therefore, the enhancement of urban resilience must be based on a clear understanding of its impact mechanism. Referring to the research of existing scholars on influencing factors of urban resilience, this paper selects five indicators from government, technology, market, openness and financial factors to explore the mechanism of action of urban resilience (Table 3). Among them, the government factor played an important role in the battle against COVID-19, mainly through financial support. Changes in fiscal and tax policies often have the effect of “the slightest nudge causes the widest chain reaction” [29]. Moreover, the “real public needs”, with health care and education as the core, highlighted in the epidemic are difficult to be solved by market drive alone, and government public investment must be increased, so the infrastructure investment and financial level are used to represent its impact. The variables involved in technical factors include innovation input and human capital [30]. In Roosevelt’s New Deal period, the United States advocated “introducing the scientific spirit into the political and industrial fields”, which led to the peak of material environment construction in western developed countries. China began to lay out new infrastructure long before the epidemic, and this process must be supported by massive investment of innovation capital and human capital in the field of digital technology. Market factors focus on highlighting the domestic market capacity and market potential. The COVID-19 superimposed geopolitical disputes have challenged the external circulation model. However, with the domestic economy stepping to a new level, the market size and market potential are constantly expanding. All these have laid a foundation for the arrival of the great circulation era in China in the post-epidemic era [29]. Openness shows that the future city should actively respond to the call of domestic and international double circulation through foreign investment and foreign trade. The emphasis on “major domestic circulation” is not to close the door on the country, but to make the Chinese market more attractive to the world by expanding and strengthening the domestic market. Finally, “market exchange” is used to push forward globalization in the headwind, which is a process that forces cities to open up and reform. Financial factors include financial institutions scale and financial institutions efficiency. By expanding the financial institutions scale, social idle funds can be effectively gathered to support urban engineering construction, however, the acceleration of financial institutions efficiency will adversely affect urbanization and restrain the improvement of urban resilience [30].Table 4 shows that the goodness of fit of MGWR is 0.948, which is significantly better than the GWR model. However, AICc value is far lower than the result of classical GWR model, indicating that it has a better measurement of urban resilience. On the one hand, the residual sum of squares of MGWR is smaller, indicating that it can obtain regression results closer to the true value with fewer parameters. On the other hand, MGWR reduces as much noise and bias in the regression coefficient as possible by allowing the existence of multiple action scales, thus enhancing its robustness. Therefore, the MGWR model is superior to the classical GWR model.Different influencing factors have different heterogeneity and scale, that is, within a certain range, the effect size is similar, but beyond this range, the effect size is significantly different [20,31]. According to the action scales of different variables in MGWR and classical GWR (Table 5), GWR calculated the average value of each variable scale, so that all variables had the same action scale, with a bandwidth of 110, accounting for 39.1% of the total sample. In MGWR, differences are allowed in the action scale of all variables, and regression results show that all other variables are significant to different degrees except financial level. Specific manifestations are: (1) The bandwidth of innovation input, foreign trade and financial institutions scale are 87, 44 and 47 respectively, which are all micro-scale variables. The scale is close to the provincial level of China’s Xinjiang Autonomous Region. On the one hand, this indicates that they have a large spatial heterogeneity. Once the scale is exceeded, the coefficient will change dramatically. On the other hand, it also proves that urban resilience is very sensitive to these variables. (2) Market capacity, human capital, foreign investment and financial institutions efficiency are macro-scale variables, accounting for 44.8%, 39.9%, 42.7% and 52.7% of the total sample size respectively. This scale is close to the regional scale of northwest China, indicating that they are relatively stable in space. (3) The bandwidth of market potential and infrastructure investment reached 200 and 199 respectively, both of which are global scale variables. That is to say, they have little spatial heterogeneity.The coefficient of market capacity is between 0.3379–0.6216, which has a significant positive effect on urban resilience, and the intensity is the highest among all variables, with an average value of 0.437. In the spatial layout (Figure 6a), there is an obvious polarization trend, with some cities in the southwest as the center of high-value polarization, while the low-value areas are mainly distributed in the central and eastern coastal areas. The high value areas include Guangxi, Hunan, Hainan, Jiangxi, Guizhou, Yunnan and Guangdong, which are all poor provinces except Guangdong. The poor in these areas have the strongest marginal propensity to consume and the lowest real consumption capacity, while the implementation of poverty alleviation and rural revitalization strategies directly drives investment in local agriculture and rural areas. On the one hand, by reducing the burden of medical care, health care, elderly care and education, we have accelerated the process of transforming rural residents’ economic income into production and living consumption, stimulated the regional consumer market, and injected strong momentum into urbanization. On the other hand, the expansion of market capacity directly stimulates the internal driving force of economic growth and strengthens the degree of marketization in a specific region. It will not only help improve the resilience of local cities, but also promote the resilience of surrounding areas through the radiation and diffusion of markets. The reason why low-value areas cluster in the economically developed central and eastern regions is probably because some cities try to secure long-term economic growth by acquiring market resources of neighboring cities. It not only weakens the market vitality of neighboring cities and makes them difficult to resist risks, but also causes the continuous decrease of spatial carrying capacity of their own systems and the prevention of external impact.Market potential has a significant positive impact on urban resilience. The coefficient is between 0.0438 and 0.1167, with the mean value of 0.087 (Table 6). It indicates that 1% increase in population density will increase urban resilience by 0.087%, and the coefficient value indicates that its influence intensity is in the middle. There is an obvious step distribution trend in the space (Figure 6b), and some areas in south and east China are less affected by the market potential. The middle and lower reaches of the Yangtze River are the most prominent. Highly resilient cities in the region often contribute to urban development by gathering populations to replenish internal social capital and expand local markets. However, the increase of population density in specific cities will not only gradually “overdraw” the local environmental capacity, but also cause downward pressure on the engineering resilience of surrounding cities due to the outflow of resources. The areas most affected by the market potential are concentrated in the central and eastern areas dominated by the Yellow River Delta. Different from the above regions, the Yellow River Delta takes the ecological economic zone as an independent economic unit, devotes itself to the development of efficient ecological economy and insists on the synchronous improvement of productivity driven by scientific and technological innovation. By connecting a large number of major innovation platforms, projects, industries and talents, we can truly leverage high population concentration to increase productivity, thus stimulating the potential of cities and improving the social development momentum of neighboring cities.The coefficient of infrastructure investment ranges from 0.0319 to 0.0389, and the impact on urban resilience is mainly concentrated in parts of southwest China and South China (Figure 6c). Although there is a significant positive impact, the intensity is the weakest. The mean coefficient and standard deviation are 0.016 and 0.018, indicating that the differences between regions are not significant. The spatial distribution shows a gradually decreasing trend from Yunnan and Guangxi to the northwest. The western regions have made full use of the development opportunities of urbanization and rural revitalization to get rid of the constraints of insufficient infrastructure. Based on the consideration of the public and external nature of infrastructure, the government has increased investment in the “real public needs” such as education, employment, medical treatment, social security and cultural life. It not only reduces the vulnerability and risk of local development, but also strengthens the urban hardware conditions and software environment, making it possible to carry out post-disaster restoration in a short time. In contrast, the Pearl River Delta region, as the vanguard of a new round of scientific and technological revolution and industrial transformation, the original “old infrastructure” can no longer meet the digital needs of traditional industries, only through the “new infrastructure” can we completely reshape the production relationship and release the digital productivity. The COVID-19 outbreak has also shown the world how digital technology can play a key role in shaping urban resilience. Different from “old infrastructure”, the investment in “new infrastructure” usually does not require the government to disclose the ins and outs. Instead, the market subject should guide the investment independently according to the policy and be responsible for the profits and losses. Therefore, the influence of government infrastructure investment on urban resilience is gradually weakened.The coefficient of innovation input is between −0.1467 and 0.2103, and the local effect of innovation input on urban resilience is generally positive. This conclusion can be confirmed by the eastern coastal urban agglomerations dominated by Beijing-Tianjin-Hebei and Yangtze River Delta, but it has a negative spillover to the Pearl River Delta urban agglomeration (Figure 6d). As one of the three world-class urban agglomerations being cultivated in China, the negative effect of innovation input in the Pearl River Delta region is mainly reflected in the following points: (1) Lack of enough scientific and educational strength and innovation ability. The Pearl River Delta’s scientific and educational strength is inferior to that of the Beijing-Tianjin-Hebei region and the Yangtze River Delta in terms of the number of academicians of the Chinese Academy of Sciences and the number of the world’s top 500 universities. In addition, the technological development of the Pearl River Delta mainly relies on the introduction of advanced technologies, which leads to the lack of sufficient independent innovation ability and perfect innovation system, thus missing the opportunity to enhance urban resilience by promoting industrial structure optimization through innovation. (2) Competition for capital flow. On the one hand, the innovation of the Pearl River Delta is heavily dependent on the path of “self-sufficiency”, often through land finance to achieve “self-sufficiency”, which ultimately makes the new innovation city and new district become stale. On the other hand, the profit-driven capital makes enterprises more inclined to invest capital in real estate development after making profits, rather than innovation and research with a long return cycle. (3) Lack of collaborative innovation mechanism. Compared with the Beijing-Tianjin-Hebei collaborative innovation and the integrated development of the Yangtze River Delta, the essence of the negative spillover of innovation input in the Pearl River Delta lies in the liquidity of innovation resources. As the only urban agglomeration under the jurisdiction of just one province among the three, there are regional and institutional barriers among the innovation elements of the Pearl River Delta. For example, Although Shenzhen accounts for more than 50% of the patents in Guangdong, it is extremely short of basic innovation resources, while Guangzhou has nearly 70% of the universities and scientific research institutions in Guangdong, but it does not have enough innovation and transformation ability. Therefore, inefficient regional linkage further leads to the failure of technological progress to play a role in improving urban resilience.The coefficient of human capital ranges from 0.0526 to 0.3272, which has a significant positive impact on urban resilience, with an average value of 0.165, indicating a strong impact intensity (Figure 6e). It is noteworthy that the contribution of labor factor input to China’s economic growth is more than 5 times that of capital factor input. It indicates that China’s dependence on capital factor in the process of economic growth is reduced, while the demand for labor factor input, especially the high level of human capital, is increased. In addition, the intensity of human capital’s influence on urban resilience is manifested strong in eastern region while weak in central China, which is consistent with the regional distribution of scientific and technological resources in China. The eastern cities in the Yangtze River Delta are trying to break down the barriers to build a “sharing model” for scientific research talents. By breaking through rigid constraints such as household registration and identity, the Yangtze River Delta is committed to establishing a flexible flow mechanism of talents between cities to accelerate the integration process of scientific research talents, so as to stimulate the potential high-quality labor force to contribute to the improvement of urban resilience. In contrast, provinces and cities in central China lack the policy advantages and economic basis of places like Beijing, Shanghai, Guangzhou and Shenzhen, and many talent development agreements are only stay on file, leading to the loss of a large number of scientific research talents and achievements. The lack of relevant supporting facilities for innovation activities, to a large extent, hinders the effective aggregation of self-owned knowledge and knowledge flow and inhibits the improvement of local innovation ability. At the same time, the strength to engage in basic research is obviously insufficient, the output of original scientific research achievements with world leading level is not high, and there are many shortcomings in science and technology that hinder the industrial transformation. This disconnect will make the industrial structure lag behind the economic development, which is not conducive to improving the resilience.The coefficient of foreign investment ranges from 0.0407 to 0.1589, and the impact on urban resilience is only significant in the western region (Figure 6f). From the perspective of regression coefficient, the mean value is 0.04. Increasing foreign investment is conducive to improving the city’s ability to respond to the crisis. Based on the west development and the “One Belt and One Road” policy, the inland areas such as the northwest and southwest will no longer be restricted by the coastal radiation, but gradually become the frontier of the new opening up, which will greatly promote the agglomeration effect of foreign investment in the west. On the one hand, foreign investment to a large extent shows a positive spatial spillover effect. The increase of FDI in neighboring cities will promote that in the city, which will have a positive effect on the transfer of surplus agricultural labor force to urban non-agricultural industries. Meanwhile, foreign capital attracted by China is mainly concentrated in labor-intensive manufacturing industry, and the organic combination of labor-intensive FDI and rural surplus labor force will become a huge driving force for rapid economic development. On the other hand, foreign investment has positive influence on urban economic growth through technology spillover effect and regional innovation network effect. “One Belt and One Road” encourages foreign investment to gather in the western region in a large scale, thus accelerating the formation of the western innovation network. In addition, foreign-funded enterprises indirectly promote the technological progress and industrial upgrading of domestic enterprises through the spillover effect of technology, which makes the economy more and more diversity and thus enhances the vitality and resilience of urban economy. From the perspective of spatial distribution, Chengdu-Chongqing urban agglomeration has become the polarization center of western investment attraction. With “south channel” traffic location advantage, urban agglomeration of FDI with a strong “Matthew effect”, enhance its depth and breadth of opening to the outside world, FDI will no longer limited to the equipment field, but the whole industry chain layout, which can maximize the spillover effects of FDI, and promote urban resilience. Foreign trade level has a significant positive impact on the urban resilience. The coefficient is between 0.0583 and 0.9024, with an average of 0.347 (Figure 6g). Improving the level of foreign trade is conducive to the accumulation of resources and upgrading of industrial structure, thus forming the agglomeration effect. Urban agglomeration can not only improve the productivity of enterprises and cities, but also generate knowledge spillover effect to accelerate the accumulation of human capital so as to improve the innovation ability and R&D expenditure level of cities. Cities with high productivity and talent concentration tend to show stronger resilience in the face of external shocks. In addition, reducing the level of foreign trade between cities by expanding the openness can effectively stimulate the effect of foreign trade competition between neighboring cities, and also improve the urban resilience to a certain extent. The influence of foreign trade level has a certain circle structure, which is mainly concentrated in the Central Plains Urban Agglomeration and surrounding cities. The reason for this phenomenon may be that cities in central China are usually dominated by large manufacturing enterprises. This type of enterprise usually focuses on exporting one or several major products, but when faced with a sudden crisis such as serious epidemic, they tend to have more mobility and resource scheduling capability than coastal trading enterprises. At the same time, due to the lack of direct foreign ports, most central cities have special supervision areas with policy advantages, which have a stronger agglomeration effect on foreign trade activities and thus a stronger positive effect on urban resilience.Financial institutions scale has a significant positive effect on urban resilience, and the coefficient is between 0.0736 and 0.5175, with an average value of 0.115 (Figure 6h). Expanding the scale of urban finance can effectively pool the social idle funds for industrial development and social construction, and its positive influence mechanism on urban resilience mainly includes three ways: (1) The appropriate expansion of financial institutions scale is conducive to enterprises for financing and improvement of technology to achieve rapid expansion of non-agricultural industries. Through the agglomeration of secondary and tertiary industries, the process of economic development and urbanization is continuously accelerated to improve urban resilience. (2) A larger financial institutions scale can effectively make up the gap between the supply and demand of urban public service facilities construction funds, by fully mobilizing all kinds of investors and operators to participate in the infrastructure construction to constantly strengthen urban engineering resilience. (3) The expansion of financial institutions scale helps to lower the minimum investment threshold of poor people in human capital, so that they can make human capital investment through financing to jump into the high-income groups. The high level of educational talents accumulated in this process can help the city to adjust the structure quickly and realize economic growth, and continuously improve the urban social resilience. The coefficient standard deviation of financial institutions scale is 0.114, which reflects that its influence on urban resilience varies greatly among different regions. Pearl River Delta is the most affected region. The reason may be that the Pearl River Delta, as the leading region of China’s economic reform and opening up, has a relatively good financial development foundation, and its superior geographical location makes economic extroversion obvious. At the same time, in recent years, the division of regional financial functions has been gradually clarified, and the dual-core features have become more and more prominent. These not only effectively promoted the Pearl River Delta financial integration process, but also strengthened its “siphon effect” on the financial scale, making the financial institutions scale present a significant “central-peripheral” structural feature in space.The coefficient of financial institutions efficiency is between −0.1730 and −0.0569, which has a significant negative impact on urban resilience. However, the mean value of the coefficient is −0.057, indicating that the enhancement of financial institutions efficiency will reduce regional urban resilience to some extent (Figure 6j). The reasons for this phenomenon may lie in: From the perspective of short-term impact, financial institutions efficiency will restrain the development of employment urbanization in the short term. Because financial resources are too much in favor of urban economy, state-owned enterprises or large and medium-sized enterprises, they cannot effectively absorb surplus rural labor force and improve the urbanization rate, resulting in its adverse impact on urban society and economic resilience. In the long run, first of all, accelerating financial institutions efficiency may induce financial risks. The duality of urban economic development in China will lead to the inflow of capital into highly resilient urban enterprises under the guidance of policies, which will lead to unfair competition in financial resources between highly resilient and low-resilient cities, thus exacerbating financial risks and reducing urban resilience. Secondly, the development of financial institutions efficiency cannot provide long-term support for urban construction. Considering the risk-return of their own investment, most commercial financial institutions will not invest in urban construction, so the financial institutions efficiency will hardly have a significant impact on the resilience of urban engineering. However, the biggest negative impact of financial institutions efficiency is mainly concentrated in the Yangtze River Delta, which is probably due to the serious lag of financial integration in this region, leading to the divergence of financial institutions efficiency. The unbalanced distribution of financial resources makes the conversion of savings and investment also has great differences. For example, in recent years, the shortage of assets in Shanghai and the shortage of funds in Ningbo have further aggravated the regional financial risks and challenges and restricted the improvement of urban resilience.Based on the panel data of 281 cities in China from 2011 to 2018, the spatiotemporal differentiation characteristics of urban resilience in China are analyzed, and the driving mechanism of urban resilience is analyzed by using MGWR model. The results show that:(1)From 2011 to 2018, the overall urban resilience and the sub-resilience have decreased to different degrees, and the sequence has changed significantly: economic resilience has gradually replaced the dominant position of engineering resilience. No significant α convergence occurred in China and some different regions, but it is noteworthy that the only convergence phenomenon occurred in the engineering resilience. In addition, the global Moran’s I show that the spatial agglomeration trend of China’s urban resilience has been continuously strengthened. In the subsystem, the spatial agglomeration degree of engineering resilience and the institutional resilience have decreased, while other subsystems have increased.(2)In terms of spatial layout, cities with medium low-level resilience have always been in the core position. The distribution pattern of the eastern region being superior to the central and western regions has also been stable. Among them, some provinces in the eastern and western regions also show significant polarization. In addition, the development modes among the eastern, central and western regions are “three-core driving model”, “rectangular structure” and “pyramid structure” respectively. The Moran scatter plot explains the different development stages in different cities: The dominance of L-L (growth stage) cities has not changed. L-H (updating stage) and H-L (gestation stage) cities are mostly small and concentrated in the western region. H-H (mature stage) cities are mainly located in the Yangtze River Delta and Pearl River Delta.(3)There are significant differences among scales of driving factors, reflecting that different variables have different levels of spatial heterogeneity. Among them, innovation input, foreign trade level and financial institutions scale are all micro-scale variables. Market capacity, human capital, foreign investment and financial institutions efficiency are macro-scale variables. Only market potential and infrastructure investment are global scale variables. In addition, most variables have a significant positive impact on urban resilience, while effect of innovation input and financial efficiency is different. And the order of influence intensity is: Market capacity > Foreign trade level > Human capital > Financial institutions scale > Market potential > Innovation input > Financial institutions efficiency > Foreign investment > Infrastructure investment. It can be further summarized as: Market factors > Opening factor > Technical factors > Financial factors > Government factor.
+			  From 2011 to 2018, the overall urban resilience and the sub-resilience have decreased to different degrees, and the sequence has changed significantly: economic resilience has gradually replaced the dominant position of engineering resilience. No significant α convergence occurred in China and some different regions, but it is noteworthy that the only convergence phenomenon occurred in the engineering resilience. In addition, the global Moran’s I show that the spatial agglomeration trend of China’s urban resilience has been continuously strengthened. In the subsystem, the spatial agglomeration degree of engineering resilience and the institutional resilience have decreased, while other subsystems have increased.In terms of spatial layout, cities with medium low-level resilience have always been in the core position. The distribution pattern of the eastern region being superior to the central and western regions has also been stable. Among them, some provinces in the eastern and western regions also show significant polarization. In addition, the development modes among the eastern, central and western regions are “three-core driving model”, “rectangular structure” and “pyramid structure” respectively. The Moran scatter plot explains the different development stages in different cities: The dominance of L-L (growth stage) cities has not changed. L-H (updating stage) and H-L (gestation stage) cities are mostly small and concentrated in the western region. H-H (mature stage) cities are mainly located in the Yangtze River Delta and Pearl River Delta.There are significant differences among scales of driving factors, reflecting that different variables have different levels of spatial heterogeneity. Among them, innovation input, foreign trade level and financial institutions scale are all micro-scale variables. Market capacity, human capital, foreign investment and financial institutions efficiency are macro-scale variables. Only market potential and infrastructure investment are global scale variables. In addition, most variables have a significant positive impact on urban resilience, while effect of innovation input and financial efficiency is different. And the order of influence intensity is: Market capacity > Foreign trade level > Human capital > Financial institutions scale > Market potential > Innovation input > Financial institutions efficiency > Foreign investment > Infrastructure investment. It can be further summarized as: Market factors > Opening factor > Technical factors > Financial factors > Government factor.At present, the CoviD-19 epidemic is still out of control in the world. Although the prevention and control effect of the epidemic has been further shown, the spread trend is still very serious. From the perspective of the harm of the COVD-19 epidemic, the global spread and further deterioration of the epidemic will cause serious economic impact and social impact, which also indicates that China will be in the state of epidemic prevention and control for a long time. In order to meet the safe operation of Chinese cities under the background of normal epidemic, we put forward the following suggestions:First, establish domestic circulation as the main to improve the market activity. From the analysis of influencing factors, it can be seen that market factors have the greatest impact on the development of urban resilience, so it is necessary to give full play to their role in improving urban resilience in the future. In addition, due to the multiple impacts of COVID-19 and trade protectionism, China’s consumer market has come to a new juncture of restructuring. Given that the world economy is in a state of global decoupling, China should stick to the domestic cycle as the main body and continue to expand domestic demand. On the one hand, we should release the purchasing power on the basis of improving social security system. (1) Perfect the market and the government combined with the flexible implementation mechanism, use of government administrative expenditure to strengthen the urban engineering resilience in poor areas. At the same time, with the help of the market force to improve the economy and social resilience, reduce the living burden of residents, stimulate consumption. (2) Actively develop online consumption, the spread of the epidemic ushered in the growth of various types of “residential economy”, so we must force enterprises to speed up the digital transformation, and strive to adapt to and promote the innovation of consumer mode. (3) For poor and backward areas, the rural cultural tourism industry can be built to strengthen the resilience of urban-rural relations, so as to realize the villagers’ nearby employment and increase farmers’ income, and fundamentally strengthening the power of domestic demand. On the other hand, we should rationally distribute urban population density and vigorously develop regional ecological economy. The central and eastern coastal areas should avoid the downward pressure of high population density on urban engineering and ecological resilience, and give full play to the positive role of public resource allocation in urban construction through unified market construction and rational population distribution, and strengthen green orientation in urban spatial planning. It also makes use of industrial upgrading, education promotion and innovation drive to enhance the endogenous development capacity of urban resilience.Second, stick to open up and promote the domestic and international circulations. It can be seen from the above studies that openness plays an irreplaceable role in strengthening urban resilience development, so its influence on urban resilience development should be continuously exerted. By focusing on domestic economy, we will not shrink passively. Instead, we will counter the “anti-globalization” with high-level opening-up, counter the “divestment theory” by improving the business environment, and counter the “decoupling theory” by attraction of super-large markets, so as to speed up the formation of a more benign international economic cycle. First of all, we should actively seek new ideas for the development of foreign trade. Relying on the construction of free trade zones and free trade ports in various regions, we should speed up the cultivation of industrial clusters, bases or trading platforms with international influence, as well as high-quality business environment and sound market mechanism, and continuously expand the height, depth and breadth of opening-up by forming replicable and popularizable results. For exports, we need to make good use of the three major export outlets: domestic sales, cross-border e-commerce and import substitution. As for imports, we should take the initiative to reduce the tariff level so as to reduce the import cost of consumers, promote the balance of international payments, and enhance China’s voice in the world economy. Secondly, we should formulate differentiated opening-up policies. For the eastern cities, the primary task is to further expand the opening of logistics, research and development, digital economy and other service industries by relying on the construction of urban agglomerations, so as to accelerate the introduction of capital to help them quickly complete the “chain repair, chain expansion, chain strengthening” and form an efficient industrial chain and supply chain. We will build a world-class business environment as soon as possible, bring it into line with high-level international economic and trade rules, and gradually meet the requirements of legalization and facilitation of the business environment. For inland cities in the central and western regions, we should take “One Belt and One Road” as an important starting point and give full play to the advantages of the super-large market to speed up the construction of cooperative innovation networks. At the same time of relaxing the foreign exchange control of all kinds of capital in outbound acquisition, establish a cooperative innovation demonstration platform oriented to “One Belt and One Road”.Third, break down the barrier of factor flow and improve urban “technical resilience”. The 21st century is the era of knowledge-based economy, and the fourth industrial revolution with intelligence as the core has just begun to appear. The role of technological factors in the resilient development of cities is self-evident, and the need to use technological innovation to shape a “new city” is increasingly urgent. From the perspective of economic resilience, the first step is to accelerate the formation of regional collaborative innovation networks conducive to the effective flow of innovative resources such as knowledge, technology, capital and equipment. The Pearl River Delta and other regions can start with strengthening top-level design, improve the strategic framework, support universities, laboratories, high-tech zones and other innovation carriers to form innovation alliances, and jointly push forward the joint tackling of key and core technologies. We should also give full play to the leading role of core cities. While complementing each other’s advantages, we should respect the regional agglomeration law of scientific and technological innovation, and jointly create regional innovation poles to accelerate the formation of an innovation space with global attraction and radiation power. At the same time, joint innovation fund will be set up to fund common regional science and technology and industrial projects, providing special funds for inter-provincial cooperation among enterprises, universities and research institutes. The second is to continuously promote the construction of talent integration, especially for the central region, the key lies in the formation of a balanced, complementary and dependent talent structure within the region. This not only requires all regions to coordinate their talent planning, train, allocate and guide the flow of talents in accordance with the principles of complementing each other’s advantages and strengthening their characteristics, but also actively explore a flexible mechanism for the flow of scientific and technological talents with permanent residence, no transfer of relationship, recognition of identities, and ability to leave and enter. In terms of engineering resilience, COVID-19 highlights the ability of “digital technology” to respond quickly to public health emergencies. In the future, the public health—geographic information data system can be built based on big data to warn potential risks of various urban public activities. In addition, regional intelligent governance can be carried out to continuously improve the scientific nature and high efficiency of regional resource allocation through data sharing and common information.Fourth, we should improve the local financial development system and stimulate the endogenous driving forces for urban resilience. There is a close parallel relationship between urban economy and financial development, and in order to strengthen urban resilience, the healthy development of financial system must be realized. As an important factor in promoting urban resilience development, the role of financial factors still needs to be further strengthened. From the perspective of the financial system itself, the key lies in the realization of regional financial integration. Specific paths include: (1) Improving the financial resources allocation system. Firstly, it is necessary to establish a multi-level and wide-ranging financial market system to create favorable market conditions for regional financial cooperation and innovation. Secondly, regional financial infrastructure system should be improved, including trans-regional enterprise credit investigation service platform, electronic payment and settlement platform and financial information sharing system. (2) Jointly building a financial innovation service system. Its focus is to build a technology and financial service system to support the transformation of scientific and technological achievements and the industrialization of scientific and technological projects. It is also necessary to cultivate leading cities of fintech and standard innovation as soon as possible and promote the agglomeration of fintech talents, technologies and capital to the region. (3) Create a characteristic financial industry system. Set up pilot zones for financial reform and demonstration zones for characteristic financial industries around modern industries, and gradually form a characteristic financial service system of “finance + modern industry”. In order to promote urbanization, the role of finance in social, economic and engineering resilience should be emphasized. (1) To guide more financial resources to the field of education, to meet the floating population children’s education and vocational training needs, so as to transform the dividend of population resources into human capital dividend. (2) Control financial resources do not excessively concentrate on state-owned enterprises, large and medium-sized enterprises, and for small and medium-sized enterprises or industrial clusters “share soup”, so as to effectively promote the employment of rural surplus labor force. (3) Through providing financial support for high-quality agricultural products and agricultural production technology to promote agricultural industrialization and rural industrialization development. (4) Establish a diversified and market-oriented urban construction financing system to raise funds for urban infrastructure construction.Sixth, strengthen government functions and release the institutional resilience. As a new field of urban public management, resilient urban construction has not fully exerted its institutional resilience. Therefore, it is necessary to establish a decision-making and coordination mechanism dominated by the government and involving multiple departments and subjects. Firstly, improve the assessment classification system. It includes two parts: classification assessment index and emergency treatment index. On the one hand, we should divide the urban resilience into different risk levels and development stages to construct scientific assessment indexes, so that all cities can “fit into each other”. On the other hand, the emergency treatment indicators should be improved to provide differentiated policy support for different types of cities. Among them, developed cities should focus on experience exploration and innovation, while high-risk cities should avoid increasing social and economic vulnerability or disaster amplification effect in the rapid urbanization process. Secondly, construct public security system. The system should not only include the four subsystems of early warning system, prevention system, emergency response system and reconstruction system, but also include four levels of prevention and control: national—provincial—city—county—township (community). Thus, it forms a network mode with national macro control, provincial regional coordination, perfect city and county layout facilities and township (community) support. Thirdly, combine political responsibility and legal responsibility. In the whole process of undertaking resilient urban construction, the government should run through the legal warning line, and formulate legal basis for disaster response, especially for disaster prevention and mitigation at the community level, so as to effectively define the responsibilities of governments at all levels. Fourthly, improve the infrastructure resilience as a guarantee. The shortage of medical facilities is the biggest pain point of the COVID-19 outbreak. This reminds us that while improving the infrastructure, we should not only consider the normal functional requirements, but also consider the various requirements in case of unexpected risks. Pay attention to the construction and use of mobile facilities and equipment, and the construction of lifeline projects, so as to increase the redundancy of urban shock resistance.Conceptualization, Y.C. and Q.Z.; data curation, M.Z. and Y.Q.; formal analysis, Y.C., M.Z., and Q.Z.; funding acquisition, Y.C. and Q.Z.; methodology, M.Z., Y.C. and Q.Z.; software, M.Z.; supervision, Y.C. and Q.Z.; writing—original draft, M.Z. and Y.C.; writing—review and editing, M.Z., Q.Z. and Y.C. All authors have read and agreed to the published version of the manuscript.This work was supported by grants from National Social Science Foundation of China (16CJY045); Philosophy and Social Science Innovation Team Building Program of Henan Universities (2021-CXTD-12); Good Scholar in Philosophy and Social Sciences in Henan Institutions of Higher Learning (2019-YXXZ-20); Support Plan for Scientific and Technological Innovation Talents in Henan Institutions of Higher Learning (humanities and social sciences) (2018-cx-012).Not applicable.Not applicable.The data presented in this study are openly available in National Bureau of Statistics, reference number 978-7-5037-9120-8, 978-7-5037-8770-6, 978-7-5037-8432-3, 978-7-5037-8082-0, 978-7-5037-7706-6, 978-7-5037-7350-1, 978-7-5037-7019-7, 978-7-5037-6754-8.This work was supported by grants from National Social Science Foundation of China (16CJY045); Philosophy and Social Science Innovation Team Building Program of Henan Universities (2021-CXTD-12); Good Scholar in Philosophy and Social Sciences in Henan Institutions of Higher Learning (2019-YXXZ-20); Support Plan for Scientific and Technological Innovation Talents in Henan Institutions of Higher Learning (humanities and social sciences) (2018-cx-012).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.Trend chart of China’s urban resilience level from 2011 to 2018.The spatial pattern evolution map of China’s urban resilience in 2011 and 2018.Urban resilience index of eastern, central and western provinces.Urban resilience hierarchy in eastern, central and western China.Local Moran’s I of urban resilience in China.Spatial coefficient distribution of driving factors based on MGWR model.Urban resilience evaluation index system.Changes of urban resilience index, α value and Moran’s I index in China.Indicators of driving forces of urban resilience.Comparison of GWR and MGWR model indicators.Bandwidth of GWR and MGWR models.Statistical description of MGWR regression coefficient.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01057.txt

@@ -0,0 +1,2 @@
+This study aimed (I) to compare the number of repetitions that can be completed to failure (XRM) and before reaching a 15%, 30%, or 45% velocity loss threshold (XVLT) in the bench press exercise performed using different grip widths, and (II) to examine the inter-individual variability in the percentage of completed repetitions with respect to the XRM when the set volume is prescribed based on a fixed number of repetitions (FNR) and several velocity loss thresholds (VLT). Nineteen men performed four separate sessions in a random order where there was a single set of repetitions completed to failure against 75% of the one-repetition maximum during the Smith machine bench press exercise using a narrow, medium, wide, or self-selected grip widths. The XRM (p = 0.545) and XVLTs (p ≥ 0.682) were not significantly affected by grip width. A high and comparable inter-individual variability in the percentage of completed repetitions with respect to the XRM was observed when using both an FNR (median CV = 24.3%) and VLTs (median CV = 23.5%). These results indicate that Smith machine bench press training volume is not influenced by the grip width and that VLTs do not allow a more homogeneous prescription of the set volume with respect to the XRM than the traditional FNR.Training volume is one of the most critical variables when designing resistance training programs because it affects the resulting neural and morphological training adaptations (i.e., muscular endurance, hypertrophy, maximal strength, or power) [1,2]. The volume (number) of repetitions in a resistance training set has been traditionally prescribed based upon the one-repetition maximum (1RM) or the maximum number of repetitions that can be completed before reaching muscular failure (XRM) [1,2,3]. For example, athletes are instructed to perform a fixed number of repetitions (FNR) against a given relative load (e.g., 5 repetitions at 70% of 1RM) or considering the XRM (e.g., 3 repetitions at the 6RM load). However, performing a FNR against the same %1RM may induce different internal responses among individuals while the use of XRM loads may cause excessive fatigue [4]. An alternative approach to prescribe the training volume consists of recording the velocity at which the repetitions are performed [5,6]. While there are a number of ways to prescribe training volume using repetition velocity, the most commonly used approach involves termination of the set as soon as a pre-determined velocity loss threshold (VLT) has been reached [7,8,9]. For example, athletes are instructed to perform repetitions until a 20% VLT is reached. Recently, it has been postulated that VLTs allow for a greater control of the inter-individual variation in perceptual, metabolic, and neuromuscular responses to resistance training than more traditional methods of prescribing resistance training volume [10,11].The bench press (BP) exercise is commonly used within resistance training programs for the development of upper-body strength and power both in athletes and in various clinical and elderly populations [12,13,14,15]. One of the frequently overlooked aspects when performing the BP exercise is the impact of grip width on the performance of the exercise [16]. Previous studies have explored the effects of different grip widths on specific muscle activation patterns [17,18], 1RM performance [19,20,21], or kinetic and kinematic outputs [22,23,24]. However, little evidence exists regarding the effects of the grip width on training volume. For instance, Wilk et al. [25] compared the XRM during the free-weight BP exercise using narrow and wide grip widths against 75% of the 1RM. While the authors found no differences in XRM between the grip widths, it is important to note that the loads were prescribed based upon the 1RM achieved with each grip width. Furthermore, although VLTs are now increasingly used to prescribe resistance training volume [7,8,9], it is currently unknown whether the number of repetitions that can be completed before reaching a predetermined VLT (XVLT) is affected by the grip width.It is well documented that a large inter-individual variability (CV ≥ 15%) exists for the XRM in the Smith machine BP exercise performed against a range of relative loads (50-85% 1RM) [11,26]. Several studies have tried to identify the sources of this inter-individual variability [11,27,28,29]. In that regard, relative BP strength (i.e., the 1RM relative to the individual’s body mass) has been shown to be a poor predictor of the variability between the %1RM and the XRM using either a free-weight [27] or a Smith machine [11], while research on the influence of anthropometric characteristics (e.g., body mass, total arm length, biacromial width, or chest girth) on the XRM is scarce and presents inconclusive evidence [11,28,29].When looking at the Smith machine BP, the percentage of completed repetitions with respect to the XRM before exceeding certain VLTs presents a lower inter-individual variability than the XRMs (CV = 8.9% vs. 20.1%) [11]. However, more recently, García-Ramos et al. [30] reported a considerable amount of inter-individual variability for the percentage of completed repetitions with respect to the XRM before exceeding a predetermined VLT in the Smith machine BP exercise (CV = 18.8%). These conflicting findings highlight the need for further research on this topic. An argument for using VLTs could be made if a lower inter-individual variability in the percentage of completed repetitions with respect to the XRM exists for the XVLT in comparison to prescribing an FNR. However, no known study has compared the inter-individual variability in the percentage of completed repetitions with respect to the XRM between both approaches of prescribing resistance training volume (FNR vs. VLT).To address these gaps in the literature, subjects in the present study performed, on separate occasions, single sets of repetitions to failure against the 75% 1RM load in the Smith machine BP exercise using four different grip widths (narrow, medium, wide, and self-selected). The current study aimed to examine: (I) the effects of the grip width on XRM and XVTL (15%, 30%, and 45%) as well as slowest (MVslowest) and fastest (MVfastest) repetition in the set; (II) the association between different grip widths for XRM and XVLT, as well as between XRM and XVLT separately for each grip width; (III) the effects of relative strength and anthropometric characteristics on XRM and XVLT; and (IV) the inter-individual variability in the percentage of completed repetitions with respect to the XRM when the number of repetitions is prescribed based on FNR and VLT. Based on the findings of previous studies [11,25,27,30], we hypothesized that: (I) the XRM, XVTL, as well as MVslowest and MVfastest in the set would not be affected by the grip width; (II) significant correlations would be detected between the different grip widths for XRM and XVLTs, as well as between XRM and XVLTs for each grip width; (III) neither the relative strength nor the anthropometric characteristics would be significantly correlated with the XRM or XVLT; and (IV) a lower inter-individual variability in the percentage of completed repetitions with respect to the XRM would be observed using VLTs compared to a FNR.Nineteen male resistance-trained sports science students volunteered to participate in this study (Table 1). Inclusion criteria for the subjects were: (i) having at least two years of resistance training experience (2–5 sessions per week) with the goal of developing muscular force; (ii) having a relative 1RM strength higher than 0.70 in the BP exercise; and (iii) being free from any physical limitations that could compromise the study procedures. The sample size was similar to that considered in previous studies that also examined the acute effects of different VTLs (n = 16 to 20) [10,31] and an a priori power analysis was not performed due to the multiple statistical analyses performed. All subjects were informed of the study procedures and signed a written informed consent form before the commencement of the study. The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the institutional review board (IRB approval: 491/CEIH/2018).A crossover design was used to examine the effects of different grip widths on the XRM and XVLT during the BP exercise performed in a Smith machine (GervaSport, Madrid, Spain). Subjects came to the laboratory on five occasions separated by 48–72 h. The first session was used to determine subjects’ anthropometric characteristics and the BP 1RM. Sessions 2–5 consisted of performing a single set of repetitions to failure against the 75% 1RM load. A single BP grip width (narrow, medium, wide, or self-selected) was used in each session in a random order. Subjects were instructed to avoid any strenuous physical activity over the course of the study. All testing sessions were performed at the same time of the day for each subject (±1 h) and under similar environmental conditions (∼22 °C and ∼60% humidity).Body height, body mass, biacromial width (measured as the distance between the left and right acromioclavicular joints), mesosternal perimeter (measured as the contour of the thorax at the level of the mesosternal point), anteroposterior chest diameter (measured as the distance between the mesosternal point and the spinous process located at that level), transverse chest diameter (measured as the distance between the most lateral points of the thorax at the level of the mesosternal point), and total arm length (measured as the average distance of both arms from the acromioclavicular joint to the ulna’s styloid process) were measured at the beginning of the session following the protocol of the International Society for the Advancement of Kinanthropometry [32].Following the anthropometric assessment, a standard incremental loading test was used to determine the Smith Machine BP 1RM using the narrow grip width [33]. The warm-up consisted of jogging, dynamic stretching, upper-body joint-mobilization exercises, and 1 set of 5 repetitions with an external load of 15 kg. Thereafter, the external load was progressively increased in 10 kg increments until the mean velocity was lower than 0.50 m·s−1. From that moment, the load was increased from 5 to 1 kg until the 1RM load was reached. The rest between sets was set to 4 min, and 1–2 repetitions were performed with each load.Each session consisted of a single set of repetitions to failure against the 75% 1RM load. Only one grip width was tested in each session and the same absolute load was used in all sessions. All sessions began with the same general warm-up described for session 1. The specific warm-up included the specific grip width of the session with subjects performing 1 set of 10, 5, and 3 repetitions at 30%, 50%, 70% of 1RM, respectively, followed by 1 repetition at 90% of 1RM. The set of repetitions to failure ended when the subjects: (i) were unable to complete a repetition with the full range of motion; or (ii) paused for more than one second with the arms in the extended position [34]. Subjects were instructed to perform as many repetitions as possible and velocity performance feedback was verbally provided after each repetition to encourage them to perform all repetitions at the maximal intended velocity.The BP was performed according to the standard five-point body contact position technique (head, upper back, and buttocks placed firmly on the bench with both feet flat on the floor). Subjects started the task lying supine on a flat bench, with their feet resting on the floor, their elbows fully extended, and their hands placed on the bar using either a narrow, medium, wide, or self-selected grip width. From this position, they lowered the bar in a controlled manner until it made contact with the chest, held this position for approximately two seconds, and then lifted the bar as fast as possible until their elbows reached full extension [23,24]. The position of the bench was adjusted so that the vertical projection of the bar corresponded to each subject’s intermammary line. The distance between the index fingers was recorded and marked on the bar with a tape and kept constant for each subject throughout all lifts [21]. The narrow grip width represented a 100% of the biacromial width (38.6 ± 2.6 cm (35–45 cm)), the medium grip width a 150% of the biacromial width (57.9 ± 3.8 cm (52.5–67.5 cm)), the wide grip width a 200% of the biacromial width (77.3 ± 5.1 cm (70–90 cm)), and the self-selected grip width a 173 ± 22% of the biacromial width (66.7 ± 8.7 cm (44–78 cm)) (mean ± standard deviation (range)).Anthropometric measurements were performed by means of a steel flexible tape (Rosscraft Anthrotape; Rosscraft Innovations Inc., Vancouver, Canada) and a large sliding calliper (Campbell 20; Rosscraft Innovation Inc., Vancouver, Canada). Body height was measured using a wall-mounted stadiometer (Seca 202; Seca Ltd., Hamburg, Germany), while the body mass was assessed using a contact electrode foot-to-foot body fat analyzer system (TBF-300A; Tanita Corporation of America Inc., Arlington Heights, IL, USA). The BP exercise was performed in a Smith machine (GervaSport, Madrid, Spain). A linear velocity transducer (T-Force System; Ergotech, Murcia, Spain) was used to collect the mean velocity of all repetitions. The T-Force System interfaced to a personal computer by means of a 14-bit resolution analog-to-digital data acquisition board and custom software. Instantaneous velocity was sampled at a frequency of 1000 Hz and subsequently smoothed with a 4th order low-pass Butterworth digital filter with no phase shift and 10 Hz cut-off frequency. Validity and reliability of the T-Force system for the recording of mean velocity during the BP exercise has been reported elsewhere [35].Descriptive data are presented as mean ± SDs, while the coefficient of variation (CV) and Spearman’s rho correlation coefficients (rs) are indicated as the median value and range. The Shapiro–Wilk test revealed a violation of the normal distribution assumption for some variables (p < 0.05). Consequently, the Friedman test was used to compare the XRM and XVLTs as well as MVslowest and MVfastest in the set between the four BP grip widths. The rs was used to quantify the associations between: (I) the different grips widths for XRM and XVLTs; (II) the XRM and XVLTs for each grip width; and (III) the relative strength and anthropometric characteristics with the XRM and XVLT for each grip width. Finally, the inter-individual variability in the percentage of completed repetitions with respect to the XRM was calculated when performing a fixed number of repetitions (FNR of 5, 8, and 10) or when reaching certain velocity loss thresholds (VLT of 15%, 30%, and 45%) (Equation (1)). The FNR of 5, 8, and 10 were selected based on the average number of repetitions completed by the subjects when reaching the 15%, 30%, and 45% VLT, respectively. Note that the velocity loss limit was determined from MVfastest within the set until the threshold was exceeded for the first time (e.g., the set would stop below 0.59 m·s−1 for a MVfastest of 0.69 m·s−1 and a VTL of 15%). Statistical analyses were performed using the software package SPSS (IBM SPSS version 25.0, Chicago, IL, USA). Statistical significance was set at p < 0.05.
+        (1)CV (%) = Between−subjects SDSubjects’ mean score × 100Friedman tests did not reveal a significant effect for XRM (χ2(3, N = 21) = 2.13, p = 0.545), XVLT-15% (χ2(3, N = 21) = 1.14, p = 0.768), XVLT-30% (χ2(3, N = 21) = 1.50, p = 0.682), or XVLT-45% (χ2(3, N = 21) = 0.79, p = 0.853), MVfastest (χ2(3, N = 21) = 0.54, p = 0.909), or MVslowest (χ2(3, N = 21) = 0.68, p = 0.879) (Table 2). MVfastest was always observed among the first four repetitions (1st repetition = 63%, 2nd repetition = 28%, 3rd repetition = 8%, 4th repetition = 1%), while MVslowest was always attained during the last repetition. Subjects completed repetitions faster than the VLTs once the thresholds were exceeded for the first time in 13 out of 76 occasions for the 15% VLT (+1 repetitions = 6 occasions, +2 repetitions = 4 occasions, +3 repetitions = 1 occasion, +4 repetitions = 2 occasions), in 4 out of 76 occasions for the 30% VLT (+1 repetitions = 4 occasions), and 4 out of 76 occasions for the 45% VLT (+1 repetitions = 3 occasions, +2 repetitions = 1 occasion).There were positive correlations between the grip widths (rs = 0.823 (0.795–0.898)) for both XRM and XVLTs (Table 3). In addition, regardless of the grip width, the XRM and XVLTs were also positively correlated (rs = 0.623 (0.380–0.902)) (Table 4).The relative strength was negatively correlated with the XRM (rs = −0.697), XVLT-30% (rs = −0.533), and XVLT-45% (rs = −0.600) during the BP performed with a medium grip width. Body height was positively correlated with the XRM (rs = 0.587), XVLT-30% (rs = 0.515), and XVLT-45% (rs = 0.546) during the BP performed with a medium grip width. Furthermore, the total arm length was positively correlated with the XRM (rs = 0.522) during the BP performed with a medium grip width and with the XVLT-15% (rs = 0.487) during the BP performed with a wide grip width. No significant correlations were found for the remaining 120 comparisons (Table 5).A high and comparable inter-individual variability in the percentage of completed repetitions with respect to the XRM were observed when using both a FNR (CV = 24.3% (19.2%–27.7%)) and VLTs (CV = 23.5% (15.8%–31.3%)) (Table 6). Of special note is that the inter-individual variability in the percentage of completed repetitions with respect to the XRM always decreased with the increment in the number of repetitions (Figure 1).This study was designed to examine the effects of different grip widths on the XRM and XVLT during the Smith machine BP exercise. The main findings of this study were that: (I) XRM, XVLTs, MVslowest, and MVfastest were not significantly affected by the grip width; (II) there were positive correlations between the grip widths for both XRM and XVLTs and between XRM and XVLTs across the grip widths; (III) relative strength and anthropometric characteristics did not consistently present a significant correlation with the XRM or XVLT; and (IV) a high and comparable inter-individual variability was observed using both FNR and VTL. When considered collectively, these results suggest that the training volume is not influenced by Smith machine BP grip width and that using VLTs do not allow for a more homogeneous prescription of the set volume with respect to the XRM when compared to using the traditional FNR methodology.Supporting our first hypothesis, the Smith machine BP grip width did not affect the XRM and XVTL completed against the same absolute load (75% of the narrow grip width 1RM). These findings are in line with those previously demonstrated by Wilk et al. [25], who did not find significant differences between the narrow (95% of the biacromial width) and wide (200% of the biacromial width) grip widths for the XRM completed against the grip-specific 75% 1RM loads during the free-weight BP exercise. It has previously been suggested that changes in BP grip width can affect 1RM performance [19,20,21] and kinetic and kinematic outputs [22,23,24]. However, based on the findings by Wilk et al. [25], and the results of the present study, it seems that the grip width might not affect the total repetition volume or the repetition volume before exceeding a pre-determined VLT. Even more important is the fact that the equipment (machine-based vs. free-weight movement) used to perform the BP exercise does not appear to influence the reported findings, but caution should be taken due to the methodological differences between the studies. Furthermore, García-Ramos et al. [30] has recently shown that MVfastest was predominantly observed during the 1st repetition (53%) and 2nd repetition (32%) and that individuals sometimes produced a velocity output above a VLT once this threshold is exceeded for the first time (on 0 to 4 occasions for 15% and 30% VLTs, and on 0 to 2 occasions for 45% VLT). In the present study, and in agreement with the findings by García-Ramos et al. [30], MVfastest was also predominantly observed during the 1st repetition (63%) and 2nd repetition (28%). In addition, subjects in the present study frequently produced velocity outputs above 15% (13 occasions), 30% (4 occasions), and 45% (4 occasions) VLTs once these thresholds were exceeded for the first time. It is also worth noting that the MVslowest was always observed in the last repetition of the set to failure. In that regard, the present study provides additional evidence on the importance of considering the reference repetition when implementing VLTs to prescribe and monitor training volume during resistance training. Collectively, these findings suggest that XRM, VLTs, MVslowest, and MVfastest are not affected by the grip width during the Smith machine BP exercise. Therefore, the self-selected grip width could be the simplest and most ecologically valid strategy of Smith machine BP execution during resistance training.Previous studies have explored the influence of relative strength and certain anthropometric characteristics on XRM [11,27,28,29], but no previous study has looked at the effects of these factors on XVLTs. Supporting our third hypothesis, neither the relative strength nor the anthropometric characteristics were systematically related to the XRM and XVLTs. These results are in line with previous studies that failed to find significant relationships between XRMs and relative strength in the BP exercise against a range of relative loads (50–90% 1RM) [11,27]. Our results are also in agreement with previous studies that did not find a clear relationship between the XRM and anthropometric characteristics such as body height (r range = −0.50 to 0.16), body mass (r range = −0.44 to 0.21), or total arm length (r range = −0.46 to 0.16) against a range of relative loads (40–85% 1RM) [11,28]. However, our results are in disagreement with other investigations that have reported high correlations between the XRM and biacromial width (r range = −0.60 to −0.50) and XRM and chest girth (r range = 0.56 to 0.60) against different relative loads (40–100% 1RM) [28,29]. The discrepancy between the findings could likely be explained by the differences in study methodologies, technical execution of the BP exercise, and the sample size. Nevertheless, it is important to note that the present study expands the current knowledge while examining not only XRM, but also certain XVLTs in the BP exercise with different grip widths. Based on available evidence, the large inter-individual variability observed for the XRM might be mainly caused by other factors such as training background or specific muscle characteristics of the individuals [36] rather than relative strength or anthropometric characteristics.Resistance training volume has been commonly prescribed using a predetermined FNR to be completed in each exercise set [2]. However, since there is a large inter-individual variability in the XRM completed against a given relative load [11,26], requiring all individuals to perform the same FNR will likely result in a different training stimulus for each athlete. In this regard, one of the greatest challenges facing coaches, strength, and conditioning professionals is how to accurately prescribe training volume to elicit specific adaptations. As a potential solution, previous research has proposed using different VLTs [7,8,9] and VLT-based equations to estimate with a low inter-individual variability (CV = 2.7%–12.1%) the number of repetitions left in reserve in a set during the BP exercise [11,37]. However, our fourth hypothesis was rejected because a high and comparable inter-individual variability in the percentage of completed repetitions with respect to the XRM was observed for both FNR and VLT. These findings are in contrast to a previous study conducted by Gonzalez-Badillo et al. [11], who observed a considerably lower inter-individual variability for the percentage of completed repetitions with respect to the XRM before exceeding a predetermined VLT in the same exercise against the 75% 1RM load (VLT-15%: 31.3% vs. 10.9%; VTL-30%: 23.5% vs. 8.5%; VLT-45%: 15.8% vs. 7.1%). However, our findings are in agreement with a more recent study by García-Ramos et al. [30], which showed a comparable inter-individual variability of the completed repetitions with respect to the XRM before exceeding a predetermined VLT against multiple short (≤12 XRM) or long (>12 XRM) training sets in the Smith machine BP exercise. Such discrepancies could be attributed in part to methodological differences between the studies (e.g., velocity variable (mean velocity vs. mean propulsive velocity) or execution mode (touch-and-go technique vs. concentric-only technique from the bar holders or chest)). Furthermore, in agreement with both González-Badillo et al. [11] and García-Ramos et al. [30], the inter-individual variability in the percentage of completed repetitions with respect to the XRM for both VLT and FNR tended to progressively decrease as the number of repetitions completed increased. Nevertheless, due to the high inter-individual variability in the percentage of the completed repetitions with respect to XRM for both VLT and FNR, caution should be practiced when prescribing resistance training since neither of the strategies guarantee that all individuals are experiencing similar levels of exertion after completing each set of the Smith machine BP exercise.The grip width during the Smith machine BP exercise did not affect XRM or XVLTs. In addition, the fastest and the slowest repetitions were also not affected by the grip width during the BP exercise against the same relative load in the set to muscular failure. Therefore, coaches, strength, and conditioning professionals are encouraged to implement the self-selected grip width as the simplest and most ecologically valid strategy of Smith machine BP execution during resistance training. Furthermore, while XRM and XVLTs were not affected by the relative strength or anthropometric characteristics of the individuals, a large inter-individual variability was observed for the percentage of completed repetitions with respect to the XRM, further suggesting that neither FNR nor VLT guarantee that all individuals are experiencing a similar level of exertion after completing each training set in the Smith machine BP exercise. In that regard, individual determination of the XRM and XVLTs is therefore recommended for more accurate and objective monitoring of repetition volume during each set of the BP exercise. However, since our sample size was relatively small and consisted of exclusively males with moderate resistance training experience and variable maximal strength values (BP 1RM = 0.97 ± 0.19 kg⋅body mass−1), future studies should investigate whether the results of the present study can be extrapolated to females or athletes with higher resistance training experience.Conceptualization, A.P.-C. and A.G.-R.; methodology, A.P.-C., I.J., and A.G.-R.; formal analysis, A.P.-C., I.J., G.G.H., and A.G.-R.; data curation, A.P.-C.; writing—original draft preparation, A.P.-C.; writing—review and editing, A.P.-C., I.J., G.G.H., and A.G.-R.; visualization, A.P.-C., I.J., G.G.H., and A.G.-R.; supervision, A.G.-R. 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 University of Granada Institutional Review Board (approval: 491/CEIH/2018).Informed consent was obtained from all subjects involved in the study.We would like to thank all the authors who participated in the data collection.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.Coefficients of variation for the percentage of completed repetitions with respect to the maximal number of repetitions performed to failure when reaching a fixed number of repetitions ((a) FNR; upper panel) or a given velocity loss threshold ((b) VLT; lower panel) for each bench press grip width. The grand mean was calculated as the average value of the four bench press grip widths. The selected FNR were based on the average number of repetitions completed by the participants when reaching the VLT.Participants’ characteristics.SD, standard deviation; 1RM, one-repetition maximum.Comparison of the fastest mean velocity (MVfastest), slowest mean velocity (MVslowest), number of repetitions performed to failure (XRM), and number of repetitions performed before reaching certain velocity loss thresholds (XVLT) between the four bench press grip widths.Data are mean ± standard deviation (range). χ2, Chi-square.Correlations between the four bench press grip widths for the number of repetitions performed to failure (XRM) and number of repetitions performed before reaching certain velocity loss thresholds (XVLT).The grand mean was calculated as the average value of the XRM and XVLTs. *, significant correlation (p < 0.05).Correlations between the number of repetitions performed to failure (XRM) and the number of repetitions performed before reaching certain velocity loss thresholds (XVLT) separately for each bench press grip width.The grand mean was calculated as the average value of the four bench press grip widths. *, significant correlation (p < 0.05).Correlations of the number of repetitions performed to failure (XRM) and before reaching certain velocity loss thresholds (XVLT) with the relative strength and anthropometric characteristics for each bench press grip width.The relative strength is calculated as the one-repetition maximum divided by the subject’s body mass. *, significant correlation (p < 0.05).Percentage of completed repetitions with respect to the maximal number of repetitions performed to failure when reaching a fixed number of repetitions (FNR) or a given velocity loss threshold (VLT) for each bench press grip width.Data are mean ± standard deviation (range). The selected FNR were based on the average number of repetitions completed by the participants when reaching the VLT.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01058.txt

@@ -0,0 +1 @@
+The Development of Robust and Innovative Vaccine Effectiveness (DRIVE) project is a public–private partnership aiming to build capacity in Europe for yearly estimation of brand-specific influenza vaccine effectiveness (IVE). DRIVE is a five-year project funded by IMI (Innovative Medicines Initiative). It was initiated as a response to the guidance on influenza vaccines by EMA (European Medicines Agency), which advised vaccine manufacturers to work with public health institutes to set up a joint IVE study platform. The COVID-19 pandemic reached Europe in February 2020 and overlapped with the 2019/2020 influenza season only in the last weeks. However, several elements of the DRIVE study network were impacted. The pandemic specifically affected the study sites’ routines and the subsequent assessment of the 2019/20 influenza season. Moreover, the current social distancing measures and lockdown policies across Europe are expected to also limit the circulation of influenza for the 2020/21 season, and therefore the impact of COVID-19 will be higher than in the season 2019/20. Consequently, DRIVE has planned to adapt its study platform to the COVID-19 challenge, encompassing several COVID-19 particularities in the study procedures, data collection and IVE analysis for the 2020/21 season. DRIVE will study the feasibility of implementing these COVID-19 components and establish the foundations of future COVID-19 vaccine effectiveness studies.The ‘Development of Robust and Innovative Vaccine Effectiveness’ (DRIVE) (https://www.drive-eu.org/) project was launched as a five-year public–private partnership in order to respond to a guideline issued in 2017 by the European Medicines Agency (EMA), requesting vaccine manufacturers to provide a yearly brand-specific effectiveness evaluation for all influenza vaccines as part of their post-licensure regulatory requirements [1].For this purpose, the European Commission, through the Innovative Medicines Initiative (IMI), supported the DRIVE project, a pioneer public–private partnership (PPP) that advances European cooperation in influenza vaccine effectiveness (IVE) studies. The data generated through DRIVE are expected to increase the understanding of influenza vaccine effectiveness, lead to enhanced monitoring of influenza vaccine performance by public health institutes and allow manufacturers to fulfil regulatory requirements. DRIVE’s main objective in its first pilot season (2017/2018) was to establish and test the feasibility of a new multi-country platform using a limited number of sites (five study sites from four different countries, see Table 1) [2]. Although the influenza season was severe, precise IVE estimates were not obtained due in part to the limited number of participating study sites and the resultant limited sample size. However, DRIVE succeeded in setting up the IVE study platform and implementing DRIVE generic protocols and standard operating procedures across the different sites. For the 2018/2019 season, the network expanded from 5 to 13 sites from seven different European countries (Table 1). DRIVE protocols were further harmonized, the statistical analysis plan was improved, and age- and setting-stratified IVE estimates were calculated [2,3]. Furthermore, a post hoc analysis of the 2018/2019 data allowed for the simplification of the confounders that were considered for the following season, 2019/2020, permitting the participation of study sites that had limited data on confounders and avoiding potential over-adjustment [4,5].In 2019/2020, DRIVE continued its expansion and included 14 sites from eight different European countries. For the 2019/2020 season, four primary-care-based test-negative design (TND) studies (Austria, England, Italy and the UK), eight hospital-based TND studies (Finland, France, Italy, Romania, Spain) and one register-based cohort study (Finland) were conducted (Table 1, Figure 1). The COVID-19 outbreak impacted influenza surveillance; thus, the study period for the main analysis was truncated. The pandemic and the subsequent lockdown measures interfered with and capped an already mild influenza circulation and impacted data collection within DRIVE study sites. Despite all these challenges, precise brand-specific estimates were obtained from the pooled TND studies in DRIVE, in addition to those from the population-based cohort study [2]. COVID-19 Europe-wide circulation started in February 2020 and overlapped with the 2019/2020 influenza season only for a few weeks between the end of February and beginning of March [6]. Thus, there was not an actual co-epidemic of influenza and SARS-CoV-2 in the 2019/2020 season in Europe; however, COVID-19 and influenza are both respiratory infectious diseases with similar clinical presentation. As a consequence, the COVID-19 pandemic directly impacted influenza surveillance systems and data collection at the DRIVE study sites (this will be described in detail in the next sections). First, the COVID-19 health crisis has pushed healthcare systems to the limit of their capacities or beyond and has changed health-care–seeking behaviors for people presenting influenza like illness (ILI) symptoms.Second, measures taken by public health authorities and governments to reduce virus spread, such as containment, social distancing and lockdown, for COVID-19 dramatically affected the transmission of other infectious diseases such as influenza. Finally, the COVID-19 pandemic forced several sites to stop the inclusion of influenza cases in early March, mostly due to lockdown measures, safety measures within the hospitals, COVID-19–related work overload and clinical staff being diverted to COVID-19– related tasks (e.g., SARS-CoV-2 testing and clinical management of COVID-19 patients). Overall, data collection for DRIVE and testing for influenza was de-prioritized, and as a consequence, the last influenza swabs of the season were collected in mid-March of 2020. The DRIVE study period for the 2019/2020 season effectively ended in week 12 of 2020 in all study sites (as the study period end was defined as the week prior to the first of two consecutive weeks when no influenza viruses were detected) [2]. Altogether, the previous points highlighted DRIVE’s necessity to account for the COVID-19 pandemic in the IVE study platform.DRIVE updated both the Statistical Analysis Plan (SAP) and the IVE results report for the 2019/2020 season to account for the altered COVID-19 case management and influenza surveillance among study sites (i.e., study period and sensitivity analysis).At the site level, the end of the study period was defined as either the week prior to the first of two consecutive weeks when no influenza viruses were detected, or the previously set end of the study period (30 April 2020), whichever occurred first. As illustrated in Figure 2 and described in Table 2, several sites stopped collecting data by the beginning of March. Moreover, the influenza season was shorter than expected due to the lockdown, social distancing and other COVID-19 protective measures; thus, no influenza-positive swabs were collected beyond week 12 of 2020. Therefore, the end of the study period for the main analysis was established as 29 February 2020 (two months earlier than the originally established date of 30 April 2020).A sensitivity analysis that included data up to 30 April was performed. All local study periods effectively ended before 30 April (Figure 2), but no more influenza positive tests were reported after week 12. Precise IVE estimates in the main analysis were similar to the estimations obtained from the sensitivity analysis with the extended study period, and two additional estimates with a confidence interval (CI) width of <40% were obtained [2,7].DRIVE study sites did not exclude COVID-19 cases from the DRIVE study for the 2019/2020 season. This is in line with previous DRIVE seasons, in which test results for respiratory viruses other than influenza did not impact the inclusion of subjects. Thus, COVID-19 patients were only excluded from DRIVE study if they met the regular DRIVE exclusion criteria. However, data on the COVID-19 status of subjects recruited was not collected as this was not foreseen in the DRIVE protocol for the 2019/2020 season. Several DRIVE study sites implemented a different triage protocol in response to SARS-CoV-2 emergence, and all SARI cases arriving to the hospital were tested first for SARS-CoV-2; if results were negative, later tests were performed for other respiratory viruses such as influenza (Table 2). This new triage strategy was not expected to significantly reduce the number of influenza cases captured in the DRIVE dataset in 2019/2020, as very few cases of co-infection of influenza/SARS-CoV-2 were reported at the DRIVE sites that did test swabs for both viruses; this was due the minimal overlap between the influenza season and the emergence of the SARS-CoV-2 pandemic. This section provides more details on the experiences of two DRIVE study sites in 2019/2020 in terms of management of the DRIVE study during the COVID-19 pandemic as well as the challenges faced and solutions designed to overcome these challenges. Each of the study sites below approached the DRIVE study during the COVID-19 outbreak in different and unique ways: CIRI-IT, in Italy, had to deal with an intense COVID-19 outbreak, whereas HUS Jorvi Hospital in Finland faced a moderate COVID-19 outbreak.CIRI-IT (Interuniversity Research Center on Influenza and other Transmissible Infections) participates in DRIVE: (1) through a hospital network composed of five Italian hospitals and (2) through a network of 35 physicians (25 GPs and 10 pediatricians) in two Italian regions, Liguria and Veneto. The first local case of COVID-19 in Italy was confirmed on 22 February, and over the following days, cases were reported in several other regions. This emergency led to a national lockdown on 9 March 2020.Until the first week of March 2020, the data collection for DRIVE was not affected by the COVID-19 emergency. However, later in March, the COVID-19 emergency affected the enrollment of SARI cases in all CIRI-IT hospitals and the enrollment of ILI cases by the CIRI-IT physicians and it resulted in an end to data collection for DRIVE (last SARI swab collected for DRIVE was 15 March 2020 in Siena hospital and last ILI swab was collected on 13 March 2020).During the initial stages of the COVID-19 pandemic, in order to minimize the possibility of contagion, ILI cases enrolled by the physicians’ network and the patients admitted with SARI in the hospitals were regarded as “suspected COVID-19 cases”. These subjects followed a new clinical pathway, which made it difficult to (a) collect useful information for filling in the questionnaire required by the DRIVE project, and (b) carry out PCR tests for the presence of influenza. These patients were initially subjected to laboratory tests for SARS-CoV-2 only. However, after testing for SARS-CoV-2, the swab sample from each patient was frozen, and testing for the detection of influenza was performed on the frozen samples later on. Accordingly, the collection date of the last swab positive for influenza was actually 13 March 2020 for the hospital study and 4 March 2020 for the primary care study, but the testing for influenza of those frozen samples was performed in April. In addition, patients and physicians were also contacted at a later stage to obtain the information necessary for the DRIVE study.In Finland, the COVID-19 pandemic remained moderate in comparison with countries such as Italy and Spain. DRIVE research collaborators in HUS were able to keep their staff working for DRIVE and did not stop collecting data for the DRIVE study, finishing data collection on 30 April, as initially scheduled. However, after the emergence of COVID-19 and the subsequent lockdown measures, influenza cases nearly disappeared (and, previously, the influenza season in Finland was also considered very mild). As a result, the number of influenza patients fell short of what was expected (33 vs. 74 last season). The influenza cases recruited for the DRIVE study in the 2019/2020 season were also included in a prospective observational study comparing the clinical characteristics and outcomes of hospitalized adult COVID-19 and influenza patients [8].Suspected COVID-19 patients were recruited in the DRIVE study, as they fulfilled the SARI criteria, and the study site protocol accounted for SARS-CoV-2 testing. However, as SARS-CoV-2 testing was prioritized over influenza, sample logistics were disturbed and, in many cases, only SARS-CoV-2 PCR testing was performed. Most of the samples tested for SARS-CoV-2 were stored for later re-analysis for influenza. Unfortunately, a small percentage of the stored samples were lost and consequently not re-analyzed for influenza.The HUS study site faced several other challenges due to the COVID-19 pandemic. For instance, patients could not be visited due to strict isolation, and thus it was difficult to obtain their informed consent. To overcome this problem, HUS implemented the strategy of oral witnessed consent. Oral consent was taken on the phone (on speaker, so the witness heard the conversation) or from outside the isolation area. The witness had to be a member of the hospital staff who was not included in the study team. The study nurse and the witness signed the consent form. Oral informed consent given by the next of kin was used in the same manner.Oral witnessed consent was a form of consent inbuilt to the HUS study protocol and was approved by their ethics committee. This form of consent was initially included as there was the possibility of influenza patients in the ICU who might not be able to give their written informed consent due to being too ill to write, but who could still give their oral consent. As it turned out, it was mostly needed for including seriously ill or isolated COVID-19 patients. During the COVID-19 epidemic, witnessed oral consent was very practical both for avoiding the risk of transmission and for saving PPE sorely needed by hospital staff.In the current COVID-19 pandemic context, it is expected that influenza and SARS-CoV-2 viruses will co-circulate partially or entirely during the 2020/2021 season [9], so the impact on the DRIVE study will likely be higher than in the 2019/2020 season. The change in health-seeking behavior and further lockdown policies due to subsequent COVID-19 pandemic waves in Europe, differences in vaccine recommendations and distribution, the standard of care and access to flu vaccination [10], and further elements yet to be explored, will undoubtedly generate bias in IVE estimates and/or affect the operational execution of the IVE study platform. After consultation with EMA and IMI in April 2020, the DRIVE consortium decided to account for COVID-19 in its study documents (protocols, SAP, etc.) and operational procedures for the 2020/2021 season. Since then, DRIVE is closely tracking the evolution of the pandemic and liaising with the study sites to rapidly adapt to this ever-changing situation.Consequently, DRIVE will use the 2020/2021 season as a feasibility study for discussion on COVID-19 components that can be implemented in the study platform in order to better understand how COVID-19 would impact the brand-specific IVE. A few points should be considered:SARS-CoV-2 virus infection could potentially be a confounder for the association between influenza vaccination and influenza disease [11].In potential co-circulation of influenza and SARS-CoV-2, efforts to increase rates of influenza vaccination are important to reduce the additional influenza-related burden, particularly among groups at high risk of influenza complications and severe COVID-19 disease. Thus, COVID-19 might influence the influenza vaccine recommendations or its acceptance, increasing the differences between the vaccinated and unvaccinated [12].There may also be differences in healthcare-seeking behavior for influenza vs. COVID-19 (e.g., at-risk individuals and severity of the new disease) leading to under- or over-estimation of IVE.Co-infection may affect clinical outcomes and individual response to vaccine-derived protection for influenza.The standard of care and access to influenza vaccination might be impacted in the case of SARS-CoV-2 virus circulation.DRIVE has updated its TND protocol for the 2020/2021 season in order to account for several of the COVID-19 particularities mentioned above. First, two COVID-19 related objectives were added to the protocol. A secondary objective is to estimate the COVID-19 impact on IVE within the adults/older adults in a hospital setting, given the COVID-19 epidemiology. As exploratory objective is to describe clinical signs and symptoms as well as laboratory features, around the point of admission, among hospitalized COVID-19 cases as compared to influenza cases.Moreover, the following variables will be collected across the study sites:COVID-19 positivity in the current season (RT-PCR test)Date of COVID-19 testCOVID-19 positivity in the previous seasonUse of COVID-19 treatments and type (e.g., antivirals)Clinical symptoms to distinguish between influenza and COVID-19, such as anosmia, ageusia, etc.Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19The collection of these variables is not mandatory for DRIVE study sites, but preferrable. Table 3 indicates which study sites will be able to collect COVID-19–related variables in the 2020/2021 season.Triage strategies will differ among sites in the 2020/2021 season (Figure 3, Table 3). The majority of the DRIVE study sites have established a common triage for all SARI/ILI cases, which will be tested simultaneously for both influenza and SARS-CoV-2 in multiplex RT-PCR. However, other sites will prioritize SARS-CoV-2 testing, and only those swabs testing negative for SARS-CoV-2 will be stored for subsequent influenza testing.DRIVE has envisioned since its very beginning a vaccine effectiveness (VE) platform beyond influenza, which will eventually be able to evaluate the effectiveness of vaccines for other respiratory diseases. The urgency of the COVID-19 pandemic and the imminence of the new COVID-19 vaccines has pushed DRIVE to move quickly and to leverage DRIVE assets to set up a COVID-19 VE platform as soon as possible. Because of the uncertainty of the allocation modalities, distribution and timing of early COVID-19 vaccine candidates in Europe, it will be very challenging to predict the level of uptake and ensure sufficient a sample size for VE studies [13,14]. This will necessitate the set-up of large study network with wide geographical coverage. In addition, COVID-19 and influenza are both respiratory infectious diseases with similar clinical symptoms, respiratory specimens and laboratory tests as well as potentially having time periods leading to an overlap of surveillance [15,16]. Biases may also be expected if they are not jointly considered. As presented in Figure 4, DRIVE has developed several assets that can be leveraged to readily encompass SARS-CoV-2 in addition to influenza. DRIVE’s existing brand-specific study platform, governance, tools and methods could be readily adapted to add COVID-19–specific components, allowing for limited start-up cost and increased efficiency. Thus, within the current vaccines’ post-marketing landscape, DRIVE could potentially provide the base of a new study platform. leading the COVID-19 vaccine effectiveness monitoring field.DRIVE was able to obtain precise brand-specific IVE estimates in the 2019/2020 season despite the start of the COVID-19 pandemic during the influenza season [2]. This achievement reflects the soundness of the DRIVE study network and how well DRIVE study sites managed the challenges posed by the COVID-19 pandemic, which have been reflected in the present paper. The pandemic and the subsequent lockdown measures interfered with and capped an already mild influenza circulation and impacted data collection within DRIVE study sites. COVID-19 will also have a significant impact on the DRIVE study platform for the 2020/2021 season, both in terms of epidemiology and data collection at DRIVE study sites. In order to better interpret IVE, the DRIVE generic protocol for TND studies and SAP has been adapted to encompass several COVID-19 components regarding the operational aspects of data collection and analysis. These adaptations aim to estimate the COVID-19 impact on IVE and to compare clinical and laboratory features of COVID-19 and influenza cases at the time of hospital admission. The COVID-19 and influenza testing strategy at the study sites has yet to be fully understood, as healthcare-seeking behaviors, triage strategies and testing pathways have been adapted in many European countries. A good understanding of all the COVID-19 adaptations that DRIVE sites have implemented will be important to accurately describe the study population. Consequently, DRIVE is currently studying the feasibility of implementing COVID-19 surveillance elements in the DRIVE study platform. In the future, DRIVE proposes to leverage the COVID-19 experience during the 2019/2020 and 2020/2021 seasons to design its proposal for a post-DRIVE study platform for influenza and other potential respiratory diseases (e.g., COVID-19, respiratory syncytial virus (RSV)) by developing a proof of concept. This proof of concept will consider the feasibility of adapting the DRIVE influenza study platform for the evaluation of future COVID-19 vaccines and will propose a PPP model to create a common study platform to monitor the vaccine effectiveness of several respiratory infectious diseases.Contact with study sites and collection of the data, A.C. and C.M.-Q.; writing—original draft preparation, A.C.; writing—review and editing, A.S., A.D., C.M.-Q., A.M.-I., J.D.-D. and L.T.-P.; supervision, J.D.-D. All authors have read and agreed to the published version of the manuscript.The DRIVE project received funding from the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement No 777363. This Joint Undertaking receives support from the European Union’s Horizon 2020 research and innovation programme and EFPIA.Not applicable.Informed consent was obtained from all subjects involved in the study.The results from the DRIVE IVE studies (seasons 2017/2018, 2018/2019 and 2019/2020) are available in the DRIVE website via the following link: https://www.drive-eu.org/index.php/results/. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. DRIVE study sites 2019/2020 map. Fourteen study sites from a total of eight different European countries collected data for DRIVE during the 2019/2020 influenza season. In total, there were 19 hospitals and 388 GPs for the TND studies and more than 1.5 million subjects for the register-based cohort study.New study period adjusted due to COVID-19 pandemic. Study period for main analysis was considered until 29 February 2020; sensitivity analysis considered data collected until 30 April 2020. However, the actual study period end (defined as first week of two subsequent weeks of no influenza cases detected) was in week 12 of 2020 at all DRIVE sites, indicating there was no more influenza circulating even at the sites that did still collect data beyond end of February. Only showing TND studies.DRIVE study sites for the 2020/2021 season: 13 study sites from 7 different European countries.DRIVE key assets and achievements. Since its inception in 2017, DRIVE has achieved several milestones and developed key assets that make it a unique study platform for brand-specific VE studies in Europe. DRIVE has developed methods to pool brand-specific IVE data.Evolution of DRIVE study platform in the years since its start in 2017. Specified information includes the number of participant study sites/European countries; total number of general practitioners (GP)/hospitals participating in the test-negative design (TND) studies; number of influenza-like illness (ILI) cases, severe acute respiratory infection (SARI) cases and laboratory-confirmed influenza (LCI) recruited in the TND; figures for the register-based cohort studies and influenza vaccine brands captured in DRIVE dataset.DRIVE study sites: COVID-19 impact on data collection, triage and testing strategies during the 2019/20 season. * Children: 6 months to 17 years old recruited in 2019/2020 season; ** Adults: 17 to 65 years old recruited in 2019/2020 season; *** Older adults: >65 years old recruited in 2019/2020 season.DRIVE study sites: triage and testing strategies as well as ability to collect information on COVID-19 variables for the 2020/21 season.COVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the previous seasonUse of COVID-19 treatments: dexamethasone, remdesivir, favpiravir, lopinavir/ritonavir and non-steroidal anti-inflammatory drugsClinical symptoms to distinguish between influenza and COVID-19Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19 (pneumonia, etc.)COVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the previous seasonUse of COVID-19 treatments: according to hospital’s protocolsClinical symptoms to distinguish between influenza and COVID-19 Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19 (pneumonia, etc.)COVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the previous seasonCOVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testClinical symptoms to distinguish between influenza and COVID-19Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19 (pneumonia, etc.)Clinical symptoms to distinguish between influenza and COVID-19Co-morbidities to identify risk-specific groups for COVID-19COVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the previous seasonUse of COVID-19 treatments: antivirals, monoclonal antibodies/IL-6 blockers, chloroquine, hydroxychloroquine, antibiotics, corticoidsClinical symptoms to distinguish between influenza and COVID-19: anosmia, ageusia, etc.Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19 (pneumonia, etc.)COVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the previous seasonUse of COVID-19 treatments: all administeredClinical symptoms to distinguish between influenza and COVID-19Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19 (pneumonia, etc.)COVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the previous seasonUse of COVID-19 treatmentsClinical symptoms to distinguish between influenza and COVID-19Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19 (pneumonia, etc.)COVID-19 positivity in the current season (RT-PCR test)Time of COVID-19 testCOVID-19 positivity in the previous seasonUse of COVID-19 treatments: dexamethasone, remdesivirClinical symptoms to distinguish between influenza and COVID-19Co-morbidities to identify risk-specific groups for COVID-19Morbidity related to COVID-19 (pneumonia, etc.)Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01059.txt

@@ -0,0 +1,19 @@
+Demographic evolution is resulting on an aged population increment in Spain. This growth has been more relevant in rural areas, where the population has traditionally lived under hard socio-economic conditions and leveraging the natural resources such as food from family orchards to survive. Studies that have investigated the possibilities and uses of these traditional family orchards today in relation to health-related quality of life in the elderly are scarce. Based on a previous ethnography, this mixed research aims to describe a protocol that will evaluate the effects of the use of traditional family orchards as a daily resource on fitness and quality of life of the elderly population in Las Hurdes (Spain). Body composition, fitness, mental health, health-related quality of life, and activity-related behaviors of participants will be assessed. The outcomes of this study might enable us to design further tailored physical exercise-based interventions using family orchards as an adequate resource to improve the health-related quality of life and fitness of the elderly in rural areas. In addition, the study detailed here might also be applied to other similar rural areas in Spain and worldwide.Population aging is a challenge to modern societies [1,2]. Every country in the world is experiencing an increase in the number and proportion of elderly people in its population [3]. This situation is due to different causes, including improvement in the quality of life and access to health resources that result in lower mortality in old age, and a significant decrease in the birth rate [4]. The phenomenon falls within what has been called “the second demographic transition” [5]. In this context, in Spain, according to the National Statistics Institute (INE), the Ageing Index (AI) has increased from 37.35 to 120.46 in the last 30 years (1978–2018), with a percentage of older people over the total population of 19.1%, one of the largest in the world [6]. Projections put this percentage at 29.4% in 2068 [7]. Moreover, the impact of this aging is greater in the rural areas where, since the 1960s, internal migratory movements have determined a negative demographic balance [8]. Thus, there are some regions in Spain where a very old population is living without support of younger people, together with the usual problems in rural areas such as more precarious social and health services.The Spanish autonomous community of Extremadura is a paradigm in this sense. Located in the southwest of the country, which borders Portugal (Figure 1), it is a region with a very old population (AI: 140.84) [9] that is also historically poorly communicated, sparsely populated, and impoverished, with a fragile economy. Within this region, the area of Las Hurdes is a particularly sensitive area due to its social and geographical conditions. Located in the northwest of Extremadura, it has been considered as a place that is outside of history [10] (p. 42). It has been the paradigmatic expression of misery in Spain for a long time, representing hunger and the most extreme poverty, that has impacted heavily on the inhabitants’ fitness and quality of life [11]. The diet of this population was essentially based on the only foods available, which were mainly produced in their own family orchards, essential for the economy of the area throughout its history [12]. Given this past, today the zone presents a very old population, with villages like Casares de las Hurdes, representing an extreme situation, with 55.4% of people over 65 years of age in 2019 [6].In this context, the consequences of the combination of an aging rural population, depopulation, precarious health and social services, and an economically depressed area where hunger was present until a few years ago—about the 1950s—has led the population to suffer hard socio-economic conditions over the last 50 years; thus, health-related problems could be expected nowadays in the elderly. A group of anthropologists has been carrying out some field work in this area for two years, doing qualitative research linked to the traditional tools of ethnographic fieldwork interested in the health conditions and perceptions of the elderly [13]. When we asked non-dependent old persons about their “perception of health” or “quality of life”, most of the answers were positive, and when we asked for the reasons, most of the answers referred to physical activity: “Here we’re fine”; was the most common answer to our questions, even if the questions were asked to very old people. In this sense, the relationship between health perception/quality of life and physical exercise is widely accepted in the scientific literature [14,15]. However, what kind of physical exercise do the aging people do in an area like this without social services?Routinely the answers to this question referred to the activities carried out in the family orchards, the same orchards that have been used to obtain food and survive in the area throughout all their lives: “We are fine because we move a lot in the orchards!” Family orchards are defined as a traditional agricultural production system managed with family labor [16]. In addition, the owners usually maintain the family orchards, as they received them from their parents because they represent an ethnological heritage of the first order with traditional knowledge [17], passed down from generation to generation, from parent to children [18]. Moreover, they are sustainable and small agroecosystems developed over generations [19], where ecological, agronomic, cultural, social, and physical processes take place [20,21]. As we have already mentioned, this type of family orchards has traditionally been the only mechanism that has enabled low-income families to overcome the socio-economic problems in the area of Las Hurdes, since the family tended to take advantage of the resources obtained from the land. So, they have been an essential support for the region’s economy, to the point that, as indicates Catani [22], “the only thing that feels comfortable to a “Hurdano” is to stay near his family orchard.”However, at present, these small pieces of land, which are part of the heritage, have important implications not only to provide the population with food—and therefore sustainability—but also in affecting their health and quality of life as the elderly people have confirmed. Today, the role of these natural spaces includes leisure and occupational aspects that may have a positive influence on the mental health of the elderly people, as well as encouraging physical activity and sustainability of a distinctly aging population.Previous scientific literature shows that tasks performed in the family orchards may have a positive impact on older adults’ behaviors, well-being, and quality of life, including decreased falls and a reduction in the use of psychotropic medications, improved memory, enhanced social interaction, the achievement of meaningful goals, and enhanced interpersonal intimacy [23,24,25,26]. However, there is a lack of research that analyzes the specific effects of work in the orchards on the fitness and quality of life of elderly people in rural environments. Previous research has shown that similar activities such as gardening could be classified as low-to-moderate intensity physical activity, which require whole-body exercise, which helps owners to maintain a healthy fitness level [27]. De Keijzer et al. (2019) informed that the proximity to natural green and blue areas is related to slower decreases in walking speed and grip strength in people aged between 50 and 74 years from urban areas [28]. Other studies showed that living near natural environments increases physical activity levels [29,30,31], decreases psychological stress and improves perceived health among their dwellers [32,33,34,35]. Moreover, participation in agricultural activities leads people to increase their physical activity level, performing both fine motor activities (such as cutting or grafting) and gross motor activities (such as digging or ploughing) [36,37] and an association has been found between the implementation of gardening and a reduced risk of obesity and cardiovascular disease [36]. Therefore, scientific evidence tends to show that natural environments and family orchards may impact on health status, fitness and daily life in rural dwellers. However, the complexity of rural nature and the daily living conditions cannot be easily synthesized [38], considering their rooted features, habits, past conditions and traditions, and other context parameters that should be considered at different levels.For all these reasons, from an anthropological ethnographical viewpoint, a second line of investigation has been derived, trying to discover to what extent family orchards are positive elements for the health of the elderly in Las Hurdes. To answer this question, we decided to create a multidisciplinary working group between anthropologist and specialists in Sports Sciences to do mixed research. The joint multidisciplinary study aspires to link previous research with such contemporary topics as health. Thus, we present here a study protocol which will evaluate the effects on health and quality of life of the use of traditional family orchards as a daily resource for the elderly of Las Hurdes. The results, in the case that they are positive, could be very interesting, since it is a very easy activity to encourage in other rural areas.This research is based on a previous ethnographic work that allowed us to know the potentiality of this approach and to have a good access to the object of study. Based on this information, a new and more targeted ethnography will allow us to obtain the sample of informants and participants in order to carry out a mixed-type research. From there a qualitative study (phase 1) and a quantitative study will be carried out (phase 2). Semi-structured interviews, observations and a field-diary will be used to collect qualitative data during the phase 1 and body composition, fitness, mental status, health-related quality of life, and activity behaviors will be assessed using different validated instruments during the phase 2. The study protocol sequence is shown in Figure 2.The present research was approved by Bioethics and Biosafety Committee at the University of Extremadura (approval number: 61/2020. Date 13 July 2020).To make the sample calculation, the Visual Analogic Scale (VAS) of the EQ-5D-5L has been taken as a reference. Accepting an alpha risk of 0.05 and a beta risk of 0.2 in a two-sided test, 22 subjects are necessary in first group and 22 in the second to recognize as statistically significant a difference greater than or equal to 16.44 units [39]. The common standard deviation is assumed to be 18.43 [40].A purposeful sampling design will be used to systematically represent a variety of perspectives on the topic under study [41]. Participant recruitment will take place in the selected geographic area marked on Figure 3 where there is a higher concentration of family orchards. Based on previous experiences, Casares de las Hurdes and its surroundings will be the preferred area since the previously developed work will provide an easier access to informers and participants. A total of 44 participants will be recruited from the ethnographic fieldwork and assigned to two groups (22 with family orchards and 22 without family orchards). We believe that it could be a representative sample size as the study will be conducted in a depopulated environment, since several rural and depopulated contexts exist in other areas of Spain. The 44 participants selected according to the inclusion criteria will be part of both phase 1 and phase 2 of the study.To be included in the sample with family orchards, participants will must comply the following eligibility criteria for both stages:Age > 65 years old.To be living at present and during the last 12 months in Las Hurdes, Extremadura, Spain.Not to be institutionalized (i.e., not to be subject to the living arrangements of institutions such as nursing homes, day centers, or care homes).Not to suffer any cognitive impairment considering the Mini-Mental State Examination score. This is a 30-point test whose thresholds for cognitive impairments are ≤9 points, “severe”; 10–18 points, “moderate”; 19–23 points, “mild”; and ≥24 points, “normal cognition” [42]. The test will be applied during semi-structured interviews. If at this time the score is not sufficient, the participant will be released for both phases of the study.Not to present health contraindications or medical conditions such as suffer from severe cardiovascular risk or severe back injuries that prevent performing the assessment tests. Participants will be required to present a health certificate from their primary health physician.Age > 65 years old.To be living at present and during the last 12 months in Las Hurdes, Extremadura, Spain.Not to be institutionalized (i.e., not to be subject to the living arrangements of institutions such as nursing homes, day centers, or care homes).Not to suffer any cognitive impairment considering the Mini-Mental State Examination score. This is a 30-point test whose thresholds for cognitive impairments are ≤9 points, “severe”; 10–18 points, “moderate”; 19–23 points, “mild”; and ≥24 points, “normal cognition” [42]. The test will be applied during semi-structured interviews. If at this time the score is not sufficient, the participant will be released for both phases of the study.Not to present health contraindications or medical conditions such as suffer from severe cardiovascular risk or severe back injuries that prevent performing the assessment tests. Participants will be required to present a health certificate from their primary health physician.For the second stage of the study, participants will additionally need to meet the following inclusion criteria:To carry out activity in family gardens frequently, at least three times a week.Not to engage in any significant physical activity other than tending family orchards.To have accepted voluntary participation in the study and signed an informed consent.To carry out activity in family gardens frequently, at least three times a week.Not to engage in any significant physical activity other than tending family orchards.To have accepted voluntary participation in the study and signed an informed consent.To be considered into the sample without orchards, participants had to comply the same eligibility criteria as previously described, except to frequently carrying out activity in family orchards.Based on knowledge of the field from a previous work, we will carry out a specific ethnography to select the study participants. Data and empirical materials regarding the ethnographic fieldwork (phase 1) will be collected through in-depth semi-structured interviews, and participant observation in the field. A field diary will allow to build up a more comprehensive image of subjects through providing additional information on specific contexts and issues [43,44]. All research techniques are shown in Table 1.Semi-structured interviews will be conducted with informers, meaning that predesignated core topics, but not specific questions, will be covered in each session, to allow the elaboration of areas whose significance emerges as the interviews proceed. This approach will ensure that the specific areas of interest will be covered and allow unanticipated topics to emerge. Core topics for interviews will include (1) well-being and health perception; (2) quality of life sense; and (3) type of activity in traditional orchards. Privacy will be protected by conducting the interviews in locations where the conversation cannot be overheard. Interviews will be usually conducted in the participant’s own native language, probably in Spanish. Interviews will be audio and video-recorded with the previous consent and permission of the interviewee and have an average length of about 40 min, considering that these are older people who may feel tired during the interview.Participants will be asked about some specific points during the first session: age, income, pathology presence, educational level, marital status, and different questions about orchard management (e.g., growing type, daily and weekly time dedicated to orchard management, type of activity, covered distance during a working day, means of transport to go to the orchard, etc.).Observations are a hallmark of ethnographic data collection that involve the presence of the researcher in the naturalistic environment to record events and activities that could be interesting to the research inquiry [45]. Participant observation will allow the researcher to have a direct view of events under study to complement the information collected in the interviews. The following activities will be observed by researchers in the focus area: (1) local people’s daily routine; (2) type of physical activity carried out in the family orchards; and (3) daily activities in line with the sense of well-being. Participant observation sessions will last throughout all the field works in the area.Once the data collection procedures are finished, the ethnographic fieldwork information will be prepared and analyzed. The Dédalo Platform that is specialized in management of the Intangible Cultural Heritage and oral sources will be used to ensure the correct use and treatment of the obtained information, since we are aware that the collected information belongs to the interviewees. Dédalo Platform Intagible Heritage Management is a web-format digital private access repository, where the informers’ testimonies will rest in the “cloud”, both their original video/audio and transcript formats. Likewise, the obtained information from questionnaires will be processed through specific office packages for this purpose.Considering this field work and the selection of participants who met the specific inclusion criteria, the second research line of the project will be developed (Phase 2). It will assess the impact that the use of orchards currently has on physical activity levels and health-related quality of life of participants. The following assessment will be conducted:
+            Balance
+          “Time Up and Go (TUG) test.” After getting up from a chair, participants will walk in a straight line a distance of 2.40 m [46], 3 m [47], and 10 m [48] (depending on the variant of the test), make a 180° turn and return to the chair to sit on it again. The time will start at the “go” signal and stop when the participant touches the back of the chair with his/her back. Each test will be performed twice, with a one-minute recovery time between each test, and a familiarization test will be performed before recording these two measurements. The best attempt will be considered for each variant.“One-leg stance.” Participants will be timed to one-leg balance with crossed-arms over their chest. The time will count from when the foot is lifted off the ground and will end when: (1) the arms are not crossed, (2) lifted foot touches the ground, (3) movement of supported foot, and (4) maintain position over than 45 s. The assessment will be performed both open and close eyes, and in the latter case, the time will also end when the eyes are opened. Three attempts will be made with each leg both with eyes open and closed and the best value will be recorded in each case [49,50].
+            Fear of Falling
+          The Falls Effectiveness Scale-International (FES-I) questionnaire will be used to assess the fear of falling, which has been validated by the Fall Prevention Network Europe (ProFaNE) and is a widely accepted tool to assess the fear of falling [51]. Furthermore, it is valid and reliable in different languages [52], and its Spanish version will be used for this study [53]. This questionnaire evaluates the concern about falls in a series of activities of daily life and consists of 16 items with a four-point scale (1 = not very concerned to 4 = very concerned); that is, 16 would be the best possible value and 64 the worst. The fear of falling will also be assessed through the Visual Analogical Scale for Fear of Falling (VAS-FOF) [54].
+            Body Composition
+          A stadiometer (Seca 22, Hamburg, Germany) will be used for measuring height and weight. Waist and hip circumferences will be also recorded using an anthropometric tape (Harpenden Anthropometric Tape, Holtain Ltd., Crymych, United Kingdom) and with participants on a standing position. In addition, the body composition will be evaluated with a bioimpedance meter (TANITA MC 780MA).
+            Physical condition
+          “The 6-min walk test.” Participants must walk a 45.7 m rectangular course for 6 min. The maximum distance walked by each participant will be recorded [47].“Lower body strength.” The 30-s standing test will measure how many times participants can, from a sitting position with their back supported, sit on and get up from a chair for 30 s without arm propulsion [55].“Upper body strength.” This test will measure how many times participants are able to lift a weight of 2.3 kg in their hand by flexing their arm for 30 s [55]. A grip test with a digital dynamometer will also be performed (TKK 5101 Grip-D; Takey, Tokyo, Japan) [56]. Two repetitions will be performed with each hand alternately. The average of both attempts of each hand will be used for the analysis.“Muscular strength and resistance of the trunk.” For this evaluation a test of abdominal and trunk muscular resistance will be carried out. To assess the strength of the trunk flexor muscles, participants will be asked to lie supine on a mat and to rise with 90 degrees of flexion on the hip and knee. To assess the strength of the extensor muscles, participants will lie prone and lift the shoulders off the floor by bending or extending the trunk. Both positions must be kept as long as possible, not exceeding 5 min [57].“Upper-Limb Flexibility.” This will be evaluated by the back-scratch test, which measures the full shoulders range of motion of the shoulder. It consists of determining the distance between (or overlap of) middle hand fingers behind the back using a ruler or tape. Two attempts will be made and the best of both will be registered for both arms. The average of the two previous measurements will be used for analysis [55].“Flexibility of the lower extremities.” The sit and reach test will be used to evaluate this parameter. Participants will sit with one leg extended and lean forward by sliding their hands across their leg until touching or overpassing their toes. The measurement will be taken in cm [58]. Two attempts will be made for each leg and the best value will be recorded for each one. For the analysis, the best measurement of each leg will be averaged.“Velocity.” This will be evaluated by the Brisk Walking Test. Participants will walk 30 m and the time will be recorded. Two reps with 1-min rest between them will be performed. The best trial will be used for computations.“Functional reach.” The Functional Reach Test will be applied [59], in which participants will stand in front of a wall and must reach the maximum frontal distance with their arms raised to 90 degrees from the trunk, maintaining this position for a few seconds. The maximum distance reached perpendicularly to the wall will be recorded.“Short physical performance battery (SPPB).” This battery will evaluate walking speed, balance, and time used to get up from a chair five times by direct observation [60].“Self-perception of physical fitness.” For this evaluation the International Scale of Physical Fitness (IFIS) [61] will be used. Participants must answer different questions about their self-perception of general and cardio-respiratory fitness, strength, speed-agility and flexibility compared to their friends using a 5-level Likert scale. Answer options are: “very poor”, “poor”, “average”, “good”, and “very good”.
+            Health-Related Quality of Life (HRQoL)
+          EQ-5D-5L. This questionnaire evaluates the state of health, first in levels of severity by dimensions (descriptive system) and later through a visual analogical scale (VAS) [62]. The descriptive system contains five health dimensions (mobility, personal care, daily activities, pain/discomfort and anxiety/depression) and each one of them presents five severity levels (no problems, slight problems, moderate problems, serious problems and extreme problems/impossibility). Furthermore, participants should complete the VAS referencing the best state of health imaginable. The EQ-5D-5L questionnaire was developed from its preliminary Spanish version [62] proving to be valid and reliable.15-D. This instrument has a total of 15 dimensions (mobility, vision, hearing, breathing, sleep, eating, speech, elimination, habitual activities, mental function, discomfort and symptoms, depression, anguish, vitality, and sexual activity) with 5 levels of response each, obtaining a single score ranging from 0 to 1, where 0 corresponds to the worst possible quality of life and 1 to the best [63,64]. The Spanish version of 15-D has been developed by the author of the questionnaire, with a rigorous process of translation and cultural adaptation. This version is available upon request from the author of the questionnaire; however, it has not been published.
+            Depression
+          To evaluate depressive symptoms, the Geriatric Depression Scale (GDS) questionnaire will be used. This scale consists of 15 questions about how the participant has felt in the last 14 days, with “yes” or “no” answers [65]. The GDS questionnaire has proven to be reliable and valid in the Spanish population [66].
+            Happiness
+          It will be evaluated through the Subjective Happiness Scale and the Satisfaction with Life Scale (SWLS).This measure will be evaluated with the General Happiness Scale [67] which consists of 4 items where participants must respond between a score of 1 to 7, where 1 would imply less happiness and 7 more happiness. In addition, the Life Satisfaction Scale (SWLS) will also be used [68], which consists of 5 items using a 7-point Likert where 1 corresponds to strongly disagree and 7 to strongly agree. The Subjective Happiness Scale presented adequate reliability and validity in the Spanish population [69] has proven to be valid and reliable in the Spanish population [70].
+            Physical activity and sedentary behaviors
+          Physical activity in leisure time. It will be assessed through the Leisure Time Physical Activity Instrument (LTPAI). This questionnaire includes 4 items with 3 activity levels: light, moderate, and vigorous. Participants will be asked about the time per week that they had engaged in physical activity and at what activity level during the last 4 weeks. The scale will be simplified into three levels: (1) 0.5 to 1.5 h per week, (2) 2 to 4 h per week, and (3) more than 4 h per week. For the first two activity levels, the average number of hours (1 and 3 h, respectively) will be considered to calculate the total score. If no level is selected for a category, the number of hours will be equal to 0. Leisure-time physical activity level for one week will be obtained by the sum of the number of hours indicated by participants for every category of intensity [25].
+            Physical Activity at Home and Work
+          It will be measured using the Physical Activity at Home and Work (PAHWI) tool. It consists of seven items with 3 work levels performed at home (light, moderate, and heavy activity) and four levels for employment (sedentary, light, moderate, and heavy activity). The items present a brief explanation and participants should report on the time they spend on each. The result will be obtained by adding up the hours in each category to obtain the total score [25].Initially, to draw a clear profile of participants’ characteristics, descriptive analyses will be conducted. Data will be presented as Mean (M) and standard deviation (SD) for continuous variables and; as frequencies and percentages for categorical variables. Normality will be determined by the Kolmogorov–Smirnov test and Lilliefors correction). Between-group differences on participants’ characteristics will be analyzed using independent sample t-tests for continuous variables and Pearson’s chi square test will be performed for categorical data if data are normally distributed.Between-groups differences will be checked applying the Mann–Whitney U test. Moreover, correlations between the main study variables, physical activity, fitness and HRQoL will be calculated using Pearson’s correlation coefficient with Bonferroni adjustment. Finally, a stepwise linear regression analysis will be carried out to comprehend which variables may influence fitness, physical activity, and participant’s HRQoL, including those parameters that show statistically significant associations with these parameters. Overall, statistical significance level will be set at p ≤ 0.05 for the most of test and regression. However, specifically for correlations, significance level will be fixed at p ≤ 0.005 due the number of comparisons.This procedure will be conducted by two members of the research team, (an anthropologist and a sports science specialist), who will examine, analyze and interpret all the materials obtained in the end, collaboration with a third expert in as a referee in case of discrepancies, thus completing the triangulation. The confluence of professionals with different origins, backgrounds, and skills has been proposed as an essential aspect to achieve success in an investigation such as the one proposed here. This interdisciplinary research allows the introduction of the anthropological view in the ways of thinking of specialists in physical activity. As opposed to the uniqueness of the laboratory, the diversity of the field and the concrete social life. Thus, the final outline of the study protocol is shown in Figure 4.Interdisciplinary studies have proved to be an excellent research strategy to address the complexity of the real world [71]. In fact, previous research has shown this trend [72]. Scientists with different disciplinary backgrounds often have different views on what count as good data, good evidence, a good model, or a good explanation [73]. With this research method, collaborations have emerged that were once unthinkable [74,75]. The social sciences in general and anthropology in particular have also contributed with their way of seeing the world. From ethnography it is possible to take into account social factors such as age, rurality index or gender [76]. Contributions to the field of health have been commonplace [77,78], and related to this is a growing collaboration with Sports Science professionals [79,80]. The present Study Protocol is a good example in this regard, identifying from field work the possible benefits of family orchards.Family orchards have traditionally been understood as a piece of land that provides the population with several benefits, especially in rural contexts, as these natural spaces were essentially used to obtain food and represented the base of the local economy in rural areas [81]. In this respect, less economically developed rural areas in Extremadura (Spain) like Las Hurdes have survived during the last 50 years due to the work and efforts of the local population in these traditional orchards [12]. However, besides this economical support for the families, these orchards involve a wide variety of cultural, social, physical, and psychological processes that often are developed in relation to them [20,21].Previous research has shown the beneficial effects of activities that are developed in nature such as gardening, farming or horticultural activities in older adults, since these activities contribute to improving physical health (such as, functional ability, greater physical activity, reduced weight gain and body mass index, decreased falls episodes, total cholesterol, blood pressure, etc.) [27,82,83]. Likewise, these activities contribute to an increase in well-being and quality of life [26,82], improve self-esteem and provide opportunities for social interaction [84] and psychological benefits, such as personal life satisfaction [33] and the reduction in depression and anxiety levels [85]. However, there is a lack of research regarding the effects on HRQOL and fitness in older people of a variety of similar activities also performed in nature such as tasks related with agriculture in family orchards, especially for those living in rural contexts. In this respect, despite traditional orchards having represented a way of living for the rural population throughout history, as previously stated, many other effects on their owners may arise from the work performed on these pieces of land.This mixed study details the protocol and the methodology that will analyze the effects of working in family orchards on fitness status and health-related quality of life of elderly people in a rural population located in Las Hurdes (Spain). The assessment instruments that will be used in this study are standardized and validated into Spanish for the targeted population. The assessment procedures will be also adapted for participants, as this evaluation will be performed in the participants’ environment. In order to complement the outcomes and information obtained using the fitness tests previously described, other evaluation instruments will be employed such as semi-structured interviews and validated questionnaires. Thus, it is considered that the methodology of this study is robust and adequate to evaluate the effects and the importance of the family orchards for this population.The outcomes of this study might enable us to design further physical exercise-based interventions using family orchards as an adequate resource to improve the health-related quality of life and fitness of older people in rural areas. Despite the relatively small sample size could present a limitation on the generalization of the results of fitness and HRQoL to other contexts. It is mainly caused by this being a depopulated area like most rural areas and implies an important handicap. Therefore, future studies in this context might require specific adjustments in this regard. However, physical activity to perform the basic and instrumental daily activities such as gardening, work in orchards or feeding livestock are culturally, and socially settled in rural environment, therefore, the study detailed here might also be applied to other similar rural areas in Spain and worldwide.From a previous ethnography, we identified the possibility that family orchards were beneficial for the health of the elderly in rural areas. An interdisciplinary work team was then created to design this protocol which aims to analyze the effects of the use of these traditional family orchards as a daily resource on health-related quality of life, well-being, and fitness status in elderly from a rural environment of Las Hurdes (Spain). The findings of this mixed study will help—from the possibilities offered by interdisciplinary studies—to discover how the orchard activities influence health, well-being, and fitness, and if these activities cause a positive effect on population’s health-related quality of life. Thus, they could help to design different strategies and apply physical exercise-based interventions using family orchards as a main resource to improve the health-related quality of life, well-being, and fitness of older people in rural areas. Moreover, the possible developed strategies could be applied to other rural areas with similar characteristics.This work was funded by the Diputación Provincial of Cáceres (Spain). Project “An ethnography of Hurdana resistance. Heritage and sustainability”. Grant number [2019/00455/001].The authors declare no conflict of interest.Situation of the region of Las Hurdes in the autonomous community of Extremadura (Spain).Study protocol Implementation Sequence.Planned fieldwork area for the project.Final outline of proposed study protocol.Research techniques.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01060.txt

@@ -0,0 +1 @@
+The prevalence of childhood obesity is increasing worldwide. Some obese children can go on to develop metabolic syndrome (MetS), but exactly who among them remains to be determined. The aim of this study was to indicate predisposing factors for metabolic syndrome, especially those that can be modified. The study comprised 591 obese children aged 10–12 years. They were all Caucasian residents of Gdańsk, Poland, with similar demographic backgrounds. Clinical examination, anthropometry, biometric impedance analysis, blood tests (including oral glucose tolerance tests (OGTT) and insulinemia), and dietary and physical activity evaluation were conducted. The results of our study show that the risk factors for MetS or any of its components include male sex, parental (especially paternal) obesity, low body mass at birth, as well as omitting breakfast or dinner. There are few risk factors for metabolic syndrome both in obese adults and children. Some of these predictors can be modified, especially those in relation to lifestyle. Identifying and then influencing these factors may help to reduce the development of metabolic syndrome and consequently improve health and quality of life.The prevalence of obesity is rapidly increasing worldwide. According to the World Health Organization (WHO), in 2016 obesity affected as many as 650 million people, and two billion adults were considered overweight (i.e., at a high risk of becoming obese); 41 million children under 5 years of age were overweight or obese and over 340 million children and adolescents aged 5–19 were overweight or obese [1]. The number of children with obesity continues to increase and it is therefore expected that complications from obesity in these age groups will also increase.Obesity, defined as the excess of fat tissue, is wide spread all over the world and it affects all age groups [2]. Obesity, among other factors, is well known risk factor for metabolic syndrome (MetS) which comprises cardiovascular diseases (CVDs), diabetes mellitus (DM), hypertension, and atherosclerosis, as well as other complications [3]. Although the etiology of obesity is complex, genetic predisposition is permissive and actually interacts with environmental agents including physical activity and diet. Heritable factors seem to make a 40–85% contribution to obesity’s etiology. Apart from genetic predisposition, other recognized constituents such as the metabolom and metabolic programming during both the gestational and post-gestational periods can be modified to some extent [4]. Consequently, development of obesity and its complications might be reduced [5]. Metabolic syndrome is a recognized consequence of obesity and it can occur as early as in adolescence. Certainly, not all obese children will develop all or any of the complications of obesity. Which children living with obesity are most prone to MetS remains to be fully elucidated. In a systematic review of 85 studies in children, the median prevalence of metabolic syndrome in all populations was 3.3% (range 0–19.2%), in overweight children it was 11.9% (range 2.8–29.3%), and in obese subjects it was29.2% (range 10–66%). For non-obese, non-overweight populations, the range was 0–1% [6]. Almost 90% of obese children and adolescents have at least one feature of metabolic syndrome [7]. On the basis of National Health and Nutrition Examination Survey (NHANES) 1999 to 2002 data, the prevalence of metabolic syndrome in adolescents 12 to 19 years old ranged from 0 to 9.4%; variation in this estimate was the result of different criteria used to define metabolic syndrome [8]. In a report from (Biobank Standardisation and Harmonisation for Research Excellence in the European Union (BIOSHARE-EU), prevalence of metabolic syndrome in obese subjects ranged from 24 to 65% in females and from ≈43 to ≈78% in males and substantially exceeded the prevalence in metabolically healthy obese subjects [9]. These divergences depended on the country.In the light of children’s dynamic growth and maturation, the population of children is unique. Obesity is diagnosed on the basis of centile charts, when BMI ≥95th centile. Moreover, according to The International Diabetes Federation (IDF) consensus—commonly used by most of the authors—metabolic syndrome may be recognized in children not younger than 10 years old [10]. It takes some time before metabolic syndrome complicates obesity. Nevertheless, the younger the child becomes obese, the earlier in life he or she might suffer from its complications [11]. However, one must remember that not all children suffering from obesity will develop MetS. The aim of our study was to identify factors favoring the presence of one or all compounds of metabolic syndrome in obese children. Secondly, we analyzed the results in order to indicate which of them are modifiable. This study was a part of the “6–10–14 for Health” integrated weight management program for children with overweight and obesity from Gdansk, Poland. The detailed design is available in previous publications and on request [12,13,14]. The analyzed data included children aged 10–12 years attending the intervention program in 2011–2015. Children were screened in primary schools and if overweight or obesity was diagnosed they were invited to the multidisciplinary, 12-month-long program. Patient flow is presented in Figure 1. The demographic background of the participants was similar in the terms of ethnicity and socioeconomic status, however some minor discrepancies were present. The population of the City of Gdansk is over 99% Caucasian.All the children were examined by pediatrician during the first intervention visit (including weight, height, and waist circumference measurement) and bioelectric impedance analysis (BIA) was performed. All the children were referred for blood tests within 4 weeks of the first visit.Medical history was taken and collected data included body mass at birth (below 2.5 kg was assumed as hypotrophy, more than 4.0 kg—macrosomy), parents’ BMI, any metabolic disease in family members, sleeping disorders, and gastrointestinal complaints. Body mass, BMI centile, waist circumference centiles, and blood pressure centiles were assessed using Polish centile charts, as recommended by the WHO [15,16,17]. Overweight was diagnosed if the BMI centile was ≥85th and obesity was diagnosed when BMI centile was ≥95th on the recommended centile charts [15]. Waist circumference (WC) over the 90th percentile and waist–hip ratio (WHR) > 0.8 for girls and >0.9 for boys were interpreted as abnormal. For children younger than 10 years old, waist–hip ratio (WhtR) seems to be more accurate than WC. Abnormal WhtR was defined when WhtR ≥0.5.In our study, although all the participants were at least 10 years old, WhtR was also calculated.The results of biometric impedance analysis were assessed according to standard values [18]. Lifestyle was evaluated on the basis of information obtained during the visit. Physical activity was evaluated by means of a Kash Pulse Recovery Test [19,20,21] and classified as excellent, very good, good, moderate, poor, and very poor.Nutritional habits were evaluated by a dietitian on the basis of recall data given by children and parents. Specific questions regarding quantity and quality of meals constituted the original questionnaire. Special attention was paid to breakfast (first course within 2 h after waking up) and dinner (last meal eaten 2 h before sleep), including the time of day.Blood tests comprised aminotransferases, lipid fractions, thyroid stimulating hormone, tyrosine, Hb1c, and oral glucose tolerance test (OGTTs), along with insulin levels at the same time points, and the results were compared to standard values for appropriate age and sex.Metabolic syndrome (MetS) was diagnosed according to IDF [22] in children with WC > 90th centile and at least two out of the following metabolic features: HDL < 40 mg/dL, TG > 150 mg/dL, glycemia > 100 mg/dL, and blood pressure ≥ 130/85 mmHg.Statistical analyses included normal distribution of continuous variables, which was verified with the Shapiro–Wilk test [23]. Descriptive statistics are presented as the mean or median and standard deviation from the mean. Between-group comparisons were carried out using the Mann–Whitney U test and ANOVA Kruskal–Wallis test [23]. Nonparametric tests were chosen because of the large number of significant Shapiro tests, which were used for normality assumption assessment. All statistical tests were 2-tailed and performed at the 5% level of significance. Statistical analysis was performed using Statistica 10 software (TIBCO Software Inc., Tulsa, OK, USA 2014). This study was accepted by the Independent Bioethics Committee for Scientific Research at the Medical University of Gdańsk (NKBBN/228/2012) on 25 June 2012. The study is registered in clinicaltrials.gov (NCT number: NCT04143074).591 children aged 10–12 years who entered the program, completed the questionnaire and had blood tests performed. None of the children had any chronic disease that could influence investigated parameters as well as no infection on the day of examination.The girls were younger (p = 0.031), shorter (p = 0.028), and had lower WC (p < 0.0001), lower WHR (p < 0.0001), lower WhtR (p = 0.002), and lower DBP (p = 0.044) compared to the boys. Moreover, higher BMI centile (p < 0.001) and higher fat tissue content were characteristic for girls. Elevated systolic blood pressure (SBP) was found in 10% of children, with no difference between girls and boys (Table 1).Based on BMI percentile criterion there were 401 (67.9%) overweight (BMI 85–95 centile) and 190 (32.1%) obese children (BMI ≥ 95 centile). Obesity was diagnosed in 85 (30.9%) girls and in 105 (33.2%) boys, whereas 190 (69.1%) girls and 211 (66.8%) boys were overweight. Girls had higher BMI centiles compared to boys (p < 0.0001).Among 76 (12.9%) children with metabolic syndrome there were 30 girls and 46 boys (Table 2 and Table 3).73% of children with MetS had elevated blood pressure in comparison with 1.1% of children from groups II and III (p = 0.005).Reduced levels of HDL (<40 mg/dL) were found in all the children with MetS and in 27.6% of the remaining children (i.e., groups II and III), which was significantly different (p = 0.029). More than half of participants with MetS (55.3%) had increased TG concentration compared to 4% of the remaining children (p = 0.001).To achieve this aim we compared children constituting group I (MetS+) to the rest of the children i.e., groups II and III (Met− and MetS+/−).According to medical history, hypotrophy at birth was three times more prevalent in boys with MetS than in boys without MetS. Discerning gender showed that low birth weight was more prevalent in boys with MetS (p = 0.034). All the girls with MetS were eutrophic at birth.Further analyses revealed that 47.3% of fathers of children with MetS were obese and 14.5% fathers had normal body mass. Maternal BMI did not differ between the groups.There was no difference between the subgroups in terms of analyzed metabolic diseases (DM2, hypothyroidism, MetS, or dyslipidemia) in children’s parents. Similarly, prevalence of cardiovascular diseases in parents was not significant.The correlation between metabolic syndrome and its components was significant forPaternal obesity (p = 0.023);Obesity in at least one parent (p = 0.046);Low body mass at birth in boys (p = 0.046).The results provided that children with MetS (44.7%) omitted breakfast more often than children from group II (35%) and group III (28.7%) (p = 0.03). This phenomenon was especially distinctive in boys. Therefore, not eating breakfast was a risk factor not only for MetS (p = 0.027, OR = 1.74; 95%CI: 1.06–2.87) but also for at least one of its components (p = 0.036) (OR = 1.46; 95% CI: 1.02–2.09).Omitting dinner was more prevalent in children from groups I and III (OR = 1.63; 95% CI: 1.13–2.35) as well as in boys (OR = 2.66; 95% CI: 1.56–4.55) (see Supplementary Table S3, Additional File S1). Not eating dinner seems to be another risk factor for MetS.On the other hand, there were no differences in eating fruit or vegetables (in terms of quantity and frequency) between children with MetS and the rest of the study population.The physical performance results (evaluated by the Kash Pulse Recovery Test) are presented in Supplementary Table S4, Additional File S1. This shows that children with MetS had very poor physical performance, three times more often than the rest of the children. However, boys with MetS had significantly worse physical performance than the rest of the males (p = 0.15). The correlation between metabolic syndrome and its components was significant forOmitting dinner: in the whole population (OR = 1.63; 95% CI: 1.13–2.35), especially in boys (OR = 2.66; 95%CI: 1.56–4.55);Not eating breakfast (p = 0.036) (OR = 1.46; 95% CI: 1.02–2.09).Biochemical assessmentElevated levels of insulin were most common in children with MetS; twice as much as in children from group II and fourfold than in group III. Significantly, more obese children without MetS had normal insulin levels (p = 0.001) (Table 4).Insulin resistance was also more common in children with MetS compared to the remaining children (Table 5).Five children with MetS had fasting glucose >126 mg/dl and were consulted by a diabetologist.Children without MetS had normal fasting glucose concentration, whereas children from groups I and III were pathological in 27.6% cases and 9.9% cases, respectively.Abnormal Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) values were significantly more common in children with metabolic syndrome (p = 0.005) (Table 6).Although children with MetS had elevated aminotransferase activity more often, the difference was significant for Alanine transaminase (ALT) only (p = 0.011). Hypertriglyceridemia, Hyperglycemia and lower HDL concentration were similar in boys and girls. (Table S1). Nevertheless, the mean values of these parameters differed: boys had higher concentration of glucose and HDL and lower TG concentration (Table S2).The prevalence of obesity is high worldwide and some describe it as pandemic. This disease is known to be one of the so-called “lifestyle diseases” and it develops mostly in adolescence and adulthood. However, the incidence of obesity in children is growing fast and the burdens of its complications must be considered not only from a medical but also a socioeconomic point of view. Obesity increases the risk of other diseases of affluence such as hypertension, dyslipidemia, and glucose intolerance, and at the same time it is a well-known risk factor for CVD and MetS [24,25,26,27]. Many authors have reported that children with a BMI over the 75th centile have higher morbidity and mortality due to DM2 and CVD in adulthood [28,29,30,31]. Thus, quality and expectancy of life certainly is and will be affected by obesity. Metabolic syndrome has become the epitome of obesity’s complications with a high impact on human wellbeing.Despite various definitions of MetS in children, apparently not all teenagers with obesity will develop metabolic syndrome or even one of its components. Which obese children (and in fact which of the whole obese population) are especially predisposed to MetS is still not fully understood. However, being aware of the risk factors for metabolic syndrome might allow prevention of MetS or at least minimize its prevalence and consequences.In our study, 12.9% of obese children 9–12 years old, participating in “6–10–14 for Health” had metabolic syndrome diagnosed—10.9% of girls and 14.6% of boys. These results are similar to those in other publications [32]. It seems that prevalence of obesity and metabolic syndrome in children is more or less the same all over the world. Taking into consideration sociodemographic variables, differences in religion, economic status, etc., there must be some stand-alone background.According to Abdullah et al., young age at the onset of obesity, as well as the time period for which obesity lasts, are essential factors for MetS in adolescence [33]. Clearly, appropriate prophylaxis should be undertaken as soon as possible to stop this process.Our main aim was to identify children who were at the highest risk of developing MetS. Recognition of predisposing agents which can be modified may be crucial, since it is our deep belief that not undertaking any prevention will eventually lead to MetS.The results show that obese children with metabolic syndrome are characterized by poor physical performance, bad nutrition habits, and glucose intolerance with insulin resistance.It is hard to say whether poor physical performance is secondary to obesity and MetS, or rather the reason for these. Contemporary lifestyle with ubiquitous hi-technology, and fast, mechanized transportation, do not mobilize children to undertake physical activity. A sedentary lifestyle also favors certain nutritional choices, including fast-foods, snacks, and sweet beverages-all high-sugar, high-fat products.Similarly to Mazur et.al., we also found that pathological WhtR predisposes children to both MetS and certain components of the syndrome [34].In our study, we found that boys with low body mass at birth more often had MetS. Obviously this fact is irreversible, but nevertheless it is known that metabolic programming, which begins in fetal life, carries on after birth. There is a slight opportunity to influence this process by means of a healthy lifestyle for pregnant woman and this can be achieved by educating not only doctors and health providers but also mothers. Promoting breastfeeding exclusively with recommended weaning time is an easy way to influence metabolic programming and weight gain in the first 2–3 years of life.Both or either parent’s obesity—especially the father’s—might suggest genetic predisposition to excess body mass in a child, but, on the other hand, it could illustrate the family’s lifestyle. It is already known that obese children skip breakfast more often than healthy children [35]. Contrary to some authors, we also found that omitting dinner is related to metabolic syndrome or to components of it.This feeding pattern fits into the modern model of life, characterized additionally by little physical activity and much sedentary time (at school, work, and home). Similar observations have already been published [36].The recall data collection is undoubtedly the limitation of this study. Both interviews and questionnaires have flaws. Respondents may have not given accurate answers. Moreover, the recall data most probably are not precise and are rather “approximate”; some information may also have been omitted.In some cases the prospective dietary records were ordered, but could be verified only after the first interventional visit (i.e., during the 2nd or even 3rd visits, when the recommended actions should have been implemented). The strengths of this study, however, were the number of subjects and detailed lifestyle and laboratory evaluation, as well the performance of a test to assess physical fitness—these are rarely performed in population studies in children.One cannot precisely assess what is a result of genetic involvement and what is a result of a family’s habits in the development of obesity. Nevertheless, the latter can be modified. However, if genetic predisposition runs in the family, this is a red flag and preventive measurements should be undertaken in order to minimize the unavoidable consequences of obesity and MetS. Sound knowledge and education on energy balance (via proper diet and physical activity) should be promoted as a driving force against obesity. Changes implemented for a child’s benefit will benefit the whole family.To our knowledge this is the first study on the factors predisposing obese children and teenagers to metabolic syndrome that has been carried out in such a large, homogenous population (age, residency).Taking into account genetic predisposition and environmental influences, we tried to indicate modifiable risk factors.The results of our study could be used as warning signals for subjects who are genetically predisposed to obesity. In these children certain preventive measures should be taken. Although, unfortunately, it has become a catchphrase, physical activity and a healthy diet are the best way to lessen the risk of severe complications from obesity and metabolic syndrome itself.Among the many risk factors of metabolic syndrome, besides those that are irreversible (such as body mass at birth, gender, genetic predisposition, etc.) there are many factors that are dependent on lifestyle. Proper, increased physical activity and rational nutrition (regular healthy meals) can be modified and, by this, the risk of metabolic syndrome in obese children can be diminished in an inexpensive way.Education and preventive companies addressing health providers and parents are required in order to lessen morbidity in relation to obesity and metabolic syndrome or at least minimize their prevalence and consequences.Intervention programs addressed to all overweight and obese children, especially those who are at risk of metabolic syndrome, should be designed (preferably on an international level).The following are available online at https://www.mdpi.com/1660-4601/18/3/1060/s1, Additional file S1. Table S1: Mean glucose, triglycerides (TG) and high density lipids (HDL) concentration in study population. Table S2: Comparison of glycemia, TG and HDL abnormalities between boys and girls. Table S3. Dinner consumption in the study population; Table S4 Physical performance analysis.A.J.: design of the work, patient recruitment, data collection and analysis, paper drafting, revision and approval of the manuscript, M.B.: design of the work, substantive revision, approval of the manuscript, A.R.-S.: design of the work; analysis and interpretation of data, patient recruitment, data collection and analysis, approval of the manuscript, A.S.-S.: design work, revision and approval of the manuscript. All authors have read and agreed to the published version of the manuscript.The City of Gdansk authorities. The City of Gdansk had no financial input into the presented manuscript, nor did they have any impact on the design of the study or collection, analysis, and interpretation of data, or in the writing of the manuscript.The study was conducted with the approval of an independent bioethics committee at the Medical University of Gdańsk (NKBBN/228/2012.), 25th June 2012 in accordance with the requirements of the Helsinki Declaration. Each parent/legal guardian had to express written consent to the child’s participation in the programme.Informed consent was obtained from all subjects involved in the study.Availability of data and materials: the data that support the findings of this study are available] but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are, however, available from the authors upon reasonable request and with permission of “6–10–14 for Health” integrated weight management program for children with overweight and obesity from Gdansk, Poland.We thank our colleagues from “6–10–14 for Health” for their hard work and engagement. Most of all we would like to thank our patients and their parents for their trust in our work.The authors declare no conflicts of interest.Patient screening and qualification for the study-study flow.Results of anthropometry, blood pressure, and fat tissue content in the study population.p < 0.05 Mann–Whitney U test analysis.Number of components of metabolic syndrome in girls and boys.p < 0.05 Chi-square test analysis by sex.Number of girls and boys in Groups I, II and III.p < 0.05 Chi-square test analysis.Fasting insulin level (Ins), one and two hours after glucose intake (oral glucose tolerance test (OGTT)) in study population.p < 0.05 Kruskal–Wallis test analysis.Glucose concentration in study population during oral glucose test.p < 0.05 Kruskal–Wallis test analysis.HOMA-IR values (normal and elevated) in study subgroups.p < 0.05 Kruskal-Wallis test analysis.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01061.txt

@@ -0,0 +1 @@
+Working in the Arctic increases the risk of occupational diseases, which is especially important in the context of acute shortage of manpower in the region. The purpose of the study was to comparatively evaluate the working conditions and occupational pathology in Nenets Autonomous Okrug (NAO) and Chukotka Autonomous Okrug (ChAO) of Russia. We analyzed the results of socio-hygienic monitoring “Working Conditions and Occupational Morbidity” in 2008–2018. Despite similar climatic and socio-economic conditions, significant differences exist in the health risks of the working populations of the two regions. In NAO two-thirds of workers were employed at facilities with satisfactory sanitary and epidemiological well-being, while in ChAO only 13% of workers had such conditions. In NAO, almost all occupational diseases (93.2%) were due to exposure to noise among civil aviation workers. In ChAO, health problems mainly occurred among miners (81.5%). The most common of these were noise effects on the inner ear (35.2%), chronic bronchitis (23.1%), and mono- and polyneuropathies (12.5%). In 2008–2018, the occupational pathology risk in ChAO was higher than in NAO: RR = 2.79; CI 2.09–3.71. Thus, specificity of technological processes and forms of labor organization create significant differences in health risks for workers. It is necessary to use modern mining equipment to decrease the occupational morbidity in ChAO. In NAO, this effect can be achieved by updating the fleet of civil aviation.The Russian Arctic (Decree of the President of the Russian Federation dated 2 May 2014 No. 296 (revised 5 March 2020)) is the main source of raw materials for the country in the short and long term. Therefore, a powerful industry has been created in the region, and the scale of economic activity significantly exceeds the level of other polar countries. One of the problems of the Russian Arctic is the shortage of labor resources due to the actions of several factors. Firstly, there is a small number of the indigenous population employed mainly in traditional types of economic activities (reindeer husbandry, fishing, and sea animal hunting). Secondly, there is an insufficient number of people who have come from other regions of the country with professional skills to work in mining and processing enterprises. In addition, a significant number of nonindigenous inhabitants of the Arctic left the region due to the economic difficulties of the 1990s and early 2000s. Currently, the Russian Arctic, which occupies about 18% of the country’s territory, is home to no more than 1.5% of its 145 million population.In full measure, the problem of a shortage of labor resources is relevant for the Nenets Autonomous Okrug (NAO) and Chukotka Autonomous Okrug (ChAO) (Figure 1), the populations of which are 44.1 and 50.7 thousand people, and the densities are 0.25 and 0.07 people/km2, respectively (as of 1 January 2020). In addition, with extreme climatic conditions, underdeveloped social and economic infrastructures are common to both regions. Their difference lies in the leading type of economic activity. In NAO, the economy is based on the extraction of oil and natural gas, whereas the extraction of hard and brown coal, alluvial and ore gold, and other nonferrous metal ores prevail in ChAO. Despite the noted difficulties in terms of the size of the gross regional product per capita in 2018, NAO ranked first (6288.5 thousand rubles) and ChAO ranked fifth (1.386.1 thousand rubles) in Russia [1,2].It is known that the climatic conditions of the Arctic modify the action of harmful occupational factors, increasing the risk and reducing the time of formation of occupational diseases [3,4,5]. In 2008–2018, the level of occupational morbidity in the Arctic, in contrast to the Russian Federation as a whole, tended to increase [6,7]. Previous studies were mainly devoted to the problems of populous and economically more important regions of the Arctic: Murmansk oblast, Yamalo-Nenets Autonomous Okrug, and Krasnoyarsk Krai [4,5,8]. At the same time, due attention was not paid to the NAO and ChAO, which have less access to the leading scientific and medical centers of the country. It is important to note that the early termination of labor activities due to occupational diseases is an additional factor that aggravates the shortage of labor resources in the northern regions [9,10].In Russia, monitoring of working conditions, occupational health, and morbidity falls within the competence of the Russian Federal Service for Surveillance on Consumer Rights Protection and Human Wellbeing (Rospotrebnadzor). All standards for working conditions, regulations for periodic medical examinations of workers exposed to harmful occupational factors, criteria for occupational diseases, and their list and registration rules are the same for the whole country and are regulated by Federal laws and orders of the Ministry of Health of Russia. Only teams of doctors specially created for this purpose, trained in occupational safety and health, have the right to diagnose an occupational disease. Data on working conditions and indicators of occupational morbidity in all 85 regions of Russia and at the federal level are updated annually and presented on the Internet in the form of government reports.Taking into account the especially difficult demographic situation in the northern regions of the country [11], studying the influence of harmful environmental factors, including the working environment, and creating a set of measures for maintaining the health of the working population are two of the priorities of the state policy of the Russian Federation in the Arctic (On the fundamentals of the state policy of the Russian Federation in the Arctic for the period until 2020 and beyond. Rossiyskaya Gazeta, No. 4877 (18 September 2008)).The aim of this research was a comparative study of causes, structures, and incidence of occupational pathology in the two Russian Arctic regions: Nenets Autonomous Okrug and Chukotka Autonomous Okrug.The study used information and data from the Russian Federal Information Fund for Social and Hygienic monitoring (section “Working conditions and occupational disease incidence”) in NAO and in ChAO in 2008–2018 (Access to the materials is provided only to public health professionals. However, part of the information is published annually in state reports on sanitary and epidemiological well-being in Russia. Available at: https://rospn.gov.ru/documents/documents.php (accessed: 2021/01/06). (in Russian)).The analysis included data on the annual number of employees of enterprises who had contact with ambient factors in workplaces (hazardous substances, fibrogenic aerosols, electric and magnetic fields, heat and cold, noise, vibration) as well as information about heavy physical work and psychosocial hazards and risks (working arrangements, routine work, intellectual, sensory, and emotional stress). We determined the number of jobs in enterprises with satisfactory, unsatisfactory, and extremely unsatisfactory working conditions (classified according to the data of a comprehensive assessment of the action of all factors of the production environment and the labor process). The analysis also dealt with all newly registered cases of health disorders of occupational etiology (diseases, intoxications, and their nosological forms, acute or chronic course of the process). In addition, we assessed data on gender, age, profession, length of service of sick persons, the nature of harmful production factors, and the exposure circumstances, the circumstances of the initial detection of health disorders of occupational etiology.Microsoft Excel 2010 and IBM SPSS Statistics v. 22 programs were used to process the study findings. We used a Kolmogorov–Smirnov test to determine the normal distribution in the “Age” and “Employment duration” samples in NAO and ChAO, a Mann–Whitney criterion (U-test) (also for these samples), agreement criterion χ2 (for unpaired nominal data, i.e., to compare NAO and ChAO) or Fisher exact (one-tailed) criterion (if the number of observations in the sample was less than 5), and McNemar’s test (for paired nominal data, i.e., to compare indicators in one of the territories in 2018 and 2008). We used the polynomial trend (n = 4) to assess the dynamics of occupational morbidity indicators, the forecast for 1 year was given, and the coefficient of determination of the model (R2) was calculated.We calculated relative risk (RR) and 95% confidence interval (CI). RR was calculated as the incidence rate of the outcome in the tested group, divided by the outcome of the control group. CI was determined by using a two-step procedure: CI was generated for Ln(RR), and then the antilogs of the upper and lower limits of CI for Ln(RR) were computed to give the upper and lower limits of Cl for the RR [12].Numerical data are presented as absolute values, percentages, and a mean with a standard deviation or a median with the interquartile range (IQR, Q1, and Q3). The critical level of significance for the null hypothesis was 0.01.The study showed that the most prevalent occupational hazards (percentage of each being over 10% of cases) to which NAO enterprise employees were exposed in 2008–2018 included nonionizing electromagnetic fields and radiations, cooling workplace microclimate, heavy physical work, noise, and chemical hazards. Working conditions changed insignificantly over the eleven years. A decreased percentage of workers exposed to noise and increased percentage of those exposed to chemical hazards (p < 0.001) in 2008 and 2018 were the only reported changes. The share of other occupational hazards in their total spectrum has not changed significantly.Cooling workplace microclimate, noise, and the combined action of several occupational hazards were most prevalent at ChAO enterprises. Since 2008, the percentage of workers exposed to cooling workplace microclimates, heavy physical work, fibrogenic aerosols, and chemical hazards (p < 0.001) decreased in 2018. At the same time, the combined action of occupational hazards, nonionizing electromagnetic fields and radiations, and hand–arm vibration increased (p < 0.001). No changes in the prevalence of whole-body vibration or psychosocial hazards and risks were detected. According to the average annual indices in 2008–2018, at the enterprises of NAO, in comparison with ChAO, a greater share of employees was exposed to the cooling microclimate, heavy physical work, psychosocial hazards and risks, nonionizing electromagnetic fields and radiations, and chemical hazards (p < 0.001). A larger number of workers at ChAO enterprises as compared to NAO were exposed to a combined effect of occupational hazards, noise, whole-body and hand–arm vibrations, and fibrogenic aerosols. If we assess working conditions by the dynamics of particular occupational hazards throughout 2008–2018, an absence of significant changes in NAO and tendency to improve in ChAO can be noticed (Table 1).A comprehensive assessment of working conditions based on the percentage of workers employed at the production facilities of three groups of sanitary-epidemiological well-being showed that in NAO in 2008–2018, almost two-thirds of employees worked in satisfactory working conditions (the first group). Over eleven years, the number of persons with satisfactory working conditions increased (p < 0.001), and the number of workers with unsatisfactory conditions (the second and third groups) decreased (p < 0.001). In 2018, there was no enterprise left in NAO, characterized by extremely unsatisfactory sanitary-epidemiological well-being (the third group). In 2008–2018, only 13% of workers in ChAO were employed at enterprises with satisfactory working conditions, and almost one third of workers were reported to have extremely unsatisfactory sanitary-epidemiological well-being. For eleven years, there has been an increase in the proportion of workers at the facilities of the first and second groups (satisfactory and unsatisfactory conditions), while at the facilities of the third group (extremely unsatisfactory conditions) it decreased. According to a complex assessment during 2008–2018, working conditions at the enterprises of NAO were more favorable than in ChAO, which was manifested by a larger share of workers employed at the facilities of the first group and a smaller share of workers engaged at the second and third group facilities (p < 0.001). It is also important that the improvement of working conditions that took place in 2008–2018 was more significant in NAO than in ChAO (Table 2).In 2008–2018, 59 and 216 occupational diseases were detected for the first time in NAO and ChAO, respectively. Almost all cases were men. The seniority at the time the disease was established was higher among workers in NAO, who were employed in air transport and in oil and gas production. In ChAO the percentage of mining enterprise workers (coal, lignite, ore raw materials, and alluvial gold) was higher and the percentages of air transport employees was lower than in NAO. Of the 176 occupational diseases identified in miners in ChAO, 106 cases were related to the extraction of coal and lignite and another 70 cases were related to the extraction of ore raw materials (Table 3).Almost all occupational diseases among NAO enterprise employees resulted from exposure to industrial noise and only two cases were related to heavy physical work. The formation of occupational pathology among the employees of enterprises in ChAO was caused by seven occupational hazards. The most common of them were noise and fibrogenic aerosols. Physical factors (noise, hand–arm and whole-body vibrations) predominated (55.1%) in the structure of harmful production factors that caused occupational diseases among employees in ChAO. In NAO, occupational pathology formation was in most cases (89.8%) due to equipment design defects. In ChAO, occupational diseases were associated with imperfection of technological processes and, to a lesser extent, with design flaws and malfunction of machines, mechanisms, equipment, devices, and tools (Table 4).Occupational diseases in NAO workers included only two nosological forms: noise effects of inner ear and radiculopathy. Occupational pathology structure of ChAO workers was much more diverse. Among the diagnosed health problems were diseases of ear, musculoskeletal and nervous systems, injuries and other consequences of exposure to external causes, respiratory diseases, and malignant neoplasm. However, as with NAO workers, noise effects of the inner ear were the most prevalent health problem. Respiratory diseases (chronic bronchitis, pneumoconiosis) were detected mainly in coal and lignite miners. All occupational health disorders had a chronic course (Table 5).Occupational pathology in NAO was first detected after mandatory periodical medical examinations (53 cases or 89.8%), and only six diseases (10.2%) were identified as a result of workers’ self-requests for medical help because of poor health. In ChAO, there was the opposite ratio of occupational diseases diagnosed during medical examinations and due to workers´ self-requests for medical assistance: 57 (26.4%) and 159 (73.6%) cases, respectively.The annual number of occupational diseases first diagnosed in NAO ranged between 2 (in 2008 and 2011) and 9 (in 2018) cases. The level of occupational morbidity in the okrug did not significantly differ from the all-Russian one. The eleven-year curve of occupational disease rate in NAO had a “sawtooth” shape, with rises and falls, and there was a slight general upward trend. In 2018, the risk of occupational pathology development in NAO was not higher than in 2008: RR = 2.99; CI 0.65–13.8; χ2 = 2.17; p = 0.141.In ChAO, the annual number of newly diagnosed occupational diseases ranged from 6 (2008 and 2017) to 37 (2015), which caused significant changes in the level of occupational morbidity. In 2008–2015, it was growing steadily. In 2016–2017, there was a decrease in indicators, followed by their rise in 2018. Over eleven years, occupational disease level in ChAO exceeded the all-Russia indicators. In 2018, occupational pathology risk in ChAO significantly exceeded the level of 2008: RR = 3.46; CI 1.34–8.91; χ2 = 7.48; p = 0.006 (Figure 2). It was not possible to identify changes in working conditions with which the subsequent dynamics of occupational morbidity indicators could be associated, both in NAO and ChAO. In general, in 2008–2018, the risk of occupational pathology formation among employees of enterprises in ChAO was higher than in NAO: RR = 2.79; CI 2.09–3.71; χ2 = 53.6; p < 0.001.This study showed that, despite the similarities of climatic, demographic, and social indicators in NAO and ChAO, the working conditions, structure, and incidence of occupational pathology in these two Arctic regions have significant differences. In 2008–2018, more favorable working conditions were seen in NAO, as almost two-thirds of employees were engaged in facilities with satisfactory indicators of sanitary-epidemiological well-being. This can be also evidenced by a longer employment duration period before occupational pathology development as compared with workers at ChAO enterprises. The data obtained contradict the information on the increased risk of developing diseases of the musculoskeletal and nervous systems and circulatory and respiratory organs in oil and gas industry workers in the subarctic and arctic regions, summarized in the report of the International Labor Organization [13].In NAO, the vast majority of occupational pathology cases were formed not in the oil and gas industry workers, as one might have expected, but among civilian flight personnel exposed to noise. In addition to more favorable working conditions, one of the possible explanations for rare cases of occupational pathology in oil and gas production may be the widespread use of the shift method of work. First, the detection and registration of diseases is known to be extremely difficult in shift workers [14,15]. Secondly, shift workers are people with initially better health indicators than the population of the region and the country as a whole [16]. One should also consider a tendency of some employees to hide the true state of their health in order to maintain highly paid jobs in the oil and gas industry in the Arctic [17,18]. In general, the data obtained are consistent with the level of occupational morbidity in Russia in the extraction of natural resources. In 2017, in oil and gas production, it amounted to 2.12 cases per 10,000 employees, and in the extraction of all types of raw materials—26.87 cases per 10,000 employees [7].In 2008–2018, only 13% of ChAO workers were employed at facilities of the first group of sanitary-epidemiologic well-being, which explains the higher level of occupational disease rate and its tendency to grow over the past eleven years. Occupational pathology structure in ChAO is typical for mining enterprises. It includes musculoskeletal, nervous and respiratory system diseases, noise effects of the inner ear, injuries, and other external effect outcomes [19,20,21,22].The dynamics of occupational morbidity in Chukotka is characterized by a pronounced rise from 2008 to 2015, a sharp decline in 2016–2017, and a subsequent upward trend from 2018. As we have already noted, we were unable to identify changes in working conditions that could cause such dynamics. Based on the data of the Russian Federal Information Fund, we can state a significant improvement in working conditions only at energy facilities. However, these objects initially made little contribution to occupational morbidity. We assume that the sharp decline in morbidity may be associated with a deterioration in the quality of medical examinations of employees. In 2017, 100% of cases of occupational diseases were detected when employees independently sought medical help, while in 2015 the share of such cases was 59.26%.A single case of occupational pathology in women, while in Russia their number is 14.2–42.2% [23], is explained by a sharp restriction on the use of female labor in the mining industry and as pilots in the civil aviation. Only two cases of occupational pathology in extreme climatic conditions of NAO and ChAO were related to a cooling workplace microclimate. In this situation there was probably an incomplete account of the effect of exposure to cold on the health of workers. This fact may be due to the peculiarities of the special assessment of working conditions currently used in Russia. In ChAO, it is noteworthy that most occupational diseases were detected because of self-request of workers for help, which may be a consequence of the low quality of periodic medical examinations [24,25].It is logical to compare the prevalence and structure of occupational pathology in NAO and especially in ChAO with similar indicators in the geographically close American state of Alaska. Problems of occupational pathology, as well as cases of insect and animal bites, various injuries and poisoning associated with industrial activities, are within the competence of the Alaska Occupational Disease and Injury Surveillance System. The literature provides data on work-related nonfatal health disorders, assessed by the number of compensation claims filed by injured workers [26,27] and by the number of accidents [28,29]. Therefore, in 2014–1015 there were 44 such requirements for all sectors of the economy per 1000 employees. The highest rates were observed for seafood processors (63 cases per 1000 employees). Less frequently, compensation was demanded by builders (34 per 1000), food processors (37 per 1000), car drivers (38 per 1000), medical workers (45 per 1000), and other professionals. Unlike ChAO and NAO, there were no miners, oil workers, or pilots among them. Among health disorders, injuries prevailed, which most often occurred among firefighters (162 cases per 1000 workers) and law enforcement officers (121 cases per 1000 workers) and less often among loggers, aviation personnel, and other workers.The analyzed articles mainly discuss the causes and frequency of occupational injuries. Only among seafood processors have diseases of the musculoskeletal system, caused by physical overstrain and frequent repetitive movements, been described [26]. There are no data on diseases associated with exposure to industrial vibration, noise, and dust aerosols. However, it is these factors, along with heavy physical work, which cause the development of most cases of occupational pathology in NAO and ChAO. Unfortunately, we have to admit that differences in the interpretation and registration of occupational health disorders in Russia and the United States make it impossible to compare them in the Arctic regions of the two countries.The technological specifics of the extraction of natural resources and the organization of work in the Arctic create significant differences in health risks for workers at enterprises in NAO and ChAO. In the Arctic region, there is a higher risk of occupational pathology in workers employed in the mining industry than in the extraction of oil and gas. To decrease the occupational morbidity rate in Chukotka miners, it is necessary, first of all, to improve technological equipment and technological processes, aimed at reducing levels of noise, whole-body, and hand–arm vibration, concentration of fibrogenic aerosols, and heavy physical work. A decrease in occupational morbidity rates in NAO can be achieved by updating the civil aviation fleet with aircraft with reduced noise characteristics and by the use of more effective individual antinoise equipment. Special attention should be paid to improving the quality of periodical medical examinations of industrial workers in ChAO.Conceptualization, S.G. and S.S.; methodology, S.S. and A.K.; software, A.K.; validation, S.G., S.S. and A.K.; formal analysis, S.G.; investigation, S.S..; resources, S.S. and A.K.; writing—original draft preparation, S.S.; writing—review and editing, S.G. and A.K. All authors have read and agreed to the published version of the manuscript.This research received no external funding.Not applicable.Not applicable.Restrictions apply to the availability of the data presented in this study. Data was obtained from the Russian Federal Service for Surveillance on Consumer Rights Protection and Human Wellbeing (Rospotrebnadzor) and are available on request from the corresponding author (kovshov@s-znc) with the permission of Rospotrebnadzor.The authors declare no conflict of interest.Political map of Russian Federation. (Available at: https://www.maps-of-the-world.ru/europe/russia/large-scale-administrative-divisions-map-of-russia-2009).Occupational morbidity rates in Chukotka Autonomous Okrug, Nenets Autonomous Okrug, and Russian Federation in 2008–2018 (per 10,000 employees).Share of employees exposed to major occupational hazards (%).Notes: upper line—indicators in Nenets Autonomous Okrug (NAO), lower line—indicators in Chukotka Autonomous Okrug (ChAO).Share of workers at enterprises of three groups of sanitary-epidemiological well-being (%).Notes: Upper line—indicators in NAO, lower line—indicators in ChAO.General characteristics of workers with newly diagnosed occupational diseases.Factors and circumstances associated with occupational disease development (cases and %).Clinical characteristics of occupational pathology (cases and percentages).Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01062.txt

@@ -0,0 +1 @@
+Mitigating the adverse physical health risks associated with COVID-19 has been a priority of public health incentives. Less attention has been placed on understanding the psychological factors related to the global pandemic, especially among vulnerable populations. This qualitative study sought to understand the experiences of children and adolescents during COVID-19. This study interviewed 48 families during the COVID-19 pandemic restrictions, and a national lockdown, to understand its impacts. The study used an Interpretative Phenomenological Analysis (IPA) methodology. Parents and children discussed the negative impact of the restrictions on young people’s wellbeing. Children and adolescents experienced adverse mental health effects, including feelings of social isolation, depression, anxiety, and increases in maladaptive behaviour. Families with children with Autism Spectrum Disorders reported increased mental health difficulties during this period mostly due to changes to routine. The findings highlight the impact of severe restrictions on vulnerable populations’ wellbeing and mental health outcomes, including children, adolescents, and those with Autism spectrum disorder (ASD).Discourse and policies around COVID-19 disproportionately focus on the adverse effects of the public health crisis on adults. Research shows that depression, anxiety, and post-traumatic stress are the most common psychological reactions to the pandemic in adults [1]; however, the impact of COVID-19 on young people is not fully understood. Though children and adolescents are not at the forefront of the pandemic, the United Nations Sustainable Development Group (UNSDG) labels them a vulnerable group that is at-risk for becoming its biggest victims and for some, the impact will be lifelong [2]. From a medical perspective, evidence suggests that young people are less likely to experience severe symptoms from contracting the disease [3] but this does not mean they are equally able or supported to cope with the psychological, economic, and social effects. Furthermore, increasing inequalities in parents’ resources and shifting conditions of home environments may exacerbate disparities in opportunities for advantaged and disadvantaged youth [4]. The short-term policy strategies of governments have focused on mitigating the physical risks of COVID-19 by placing limits on social interactions and freedom of movement. These policies have transformed how children aged 4–12 years, and adolescents aged 13–18 years behave in their lives. There has been an interruption to care arrangements, youth education and leisure services, as well as to schools and other organisations. Disruptions have been shown to disproportionately affect women [5], while reductions in families’ incomes, from a decrease in work hours, have negatively impacted households’ quality of life. These factors have had significant impacts on children’s health and wellbeing [6]. Reports by the UNSDG show that as of April 2020, more than 188 countries had some form of school closures, placing over 1.5 billion children and adolescents out of schools [2]. The impact of these policies on the mental health of young people is not yet clear. Though the longitudinal data is not yet available, the assumption is that there will be mental health consequences resulting from the pandemic [7]. Research shows that social isolation and loneliness are impacting young people [8]. There have been increases in adolescent anxiety [9,10], which has been due in some part to pandemic related restrictions [11,12]. The extent of the impact will likely depend on factors like age, gender, socioeconomic indicators, and country of origin [13]. The short- and long-term impacts are unclear, as well as if it will cause an increase in mental health disorders in younger people. As many mental health disorders begin in childhood, it is essential to consider children and adolescents’ psychological welfare during crises. This consideration is especially vital as young people are more susceptible to experiencing long-term consequences of mental health [14]. To date, there has been little systematic research exploring the psychological impact of the pandemic on children and adolescents in Ireland. This study, alongside relevant international research, analyses and presents essential data that informs the impact of COVID-19 on young people living in Ireland. Some groups may be more vulnerable to the psychosocial effects of the COVID-19 pandemic than others. Because they are in a critical period of development, with half of all mental health disordered developing before the age of 14 [15], children and adolescents must be provided with adequate supports. Factors associated with mitigation measures such as social distancing, family discord, school closures, fears about the future, and quarantine are disrupting young people’s lives [4,16]. These disruptions include changes in routine, break in the continuity of learning with the closure of schools, break in health care missed significant life events and the loss of a sense of security and safety [17]. Early research has indicated an increase in adverse mental health outcomes for young people. These include increased levels of depression and anxiety [9,10] as well as a greater likelihood of post-traumatic stress symptoms [14].In Ireland, the landscape of youth mental health has deteriorated over recent years. There is a notable increase in anxiety and depression among young people and higher rates of self-harm [18]. Furthermore, Ireland has the fourth-highest suicide rate in Europe among young people [19]. With the assumption that the psychosocial effects of COVID-19 will disproportionately affect young people, the immediate and long-term consequences must be understood. This understanding will aid the development of supports and provide insight into how best to respond to crises like those seen in recent times. The expected impacts of COVID-19 on young people will likely differ depending on age and the family’s social and demographic qualities. Parental stress caused by the pandemic-related factors may have a more significant impact on younger children, which can result in behavioural problems and can have a substantial effect on children with developmental disabilities [20]. Research also shows that increases in parental stress levels during a pandemic directly interfere with a child’s quality of life [21]. This systematic review of the impact of pandemics reveals that quality and duration of sleep and decreased levels of physical and outdoor activities can, and does, prevent child development from reaching its full potential in times of crisis.Children and adolescents with pre-existing mental health conditions may be more susceptible to new mental health conditions during the pandemic [22]. The closure of schools has led to many children losing access to mental health services received directly through their schools [23]. A study in the United States (US) found that 13.2% of adolescents were receiving some form of mental health service from their school before the pandemic [24]. By closing schools, they may have lost their support network during a time where they may be suffering from social isolation, loneliness, and anxiety as a result of COVID-19 [14]. Furthermore, research suggests that parental mental health during COVID-19 can influence child mental health [25]. Increases in parental stress during quarantine is associated with their children’s negative emotions and behaviour; the more stressed a child’s parent is, the more stressed the child themselves are [26]. These parental mental health difficulties may lead to negative parenting behaviours related to nurturing and monitoring, decreased attention to a child’s needs and more significant dysfunction within the home [27,28]. Increased family stress exacerbated by home-schooling and reduced psychological supports through school means that young people might be more vulnerable [29].Mental health problems emerge in early childhood, with almost half of the acute mental illnesses occurring by the age of 14, and 75 per cent manifest by the mid-20s [4]. It is well documented that children and adolescents’ psychological wellbeing is critical for good mental health later in life. Mental Health America reported that the mental health impacts of COVID-19 are more pronounced in age groups under 25 and are placing a generation at risk of long-term effects if these issues are not addressed. Early evidence suggests that around 9 in 10 young people are screening with moderate-to-severe depression and 8 in 10 with moderate-to-severe anxiety, as a result of the crisis [30]. This study extends this agenda by exploring the impact that the crisis has had on the family’s wellbeing, focusing specifically on young people’s mental health. The potential short and long-term impact of COVID-19 on children and adolescents’ psychological wellbeing is unclear, making this study imperative to understanding these prospective impacts and informing needed interventions. This research uses a qualitative approach to examine children and parents’ perceptions and experiences of COVID-19; it explores how the restrictions imposed on families have impacted young people’s psychological wellbeing in Ireland.Forty-eight families (N = 94) were recruited for this study. Parents, guardians, caregivers, and their children were recruited in this study to share their experiences of COVID-19. One hundred and nine participants agreed to participate, and 94 in total completed interviews. Twenty per cent were parents on their own, and 3% were couples. Families were recruited through a convenience sampling method and snowballing sampling techniques. Participants were invited to participate through social media and other networking platforms. The final sample demographics include two-parent families, single-parent families, those from different socioeconomic backgrounds, and rural and urban settings. Though the researchers sought an inclusive sample set, participation was based on people’s willingness to participate in the study, and 48 families agreed to participate. Additionally, more female parents or caregivers were willing to participate (see Table 1) This study received ethical approval from The Social Research Ethics Subcommittee, Maynooth University (2407411).This qualitative study features an Interpretative Phenomenological Analysis (IPA) approach [31]. This study was conducted and reported in accordance with the Consolidated Criteria for Reporting Qualitative Research (COREQ). The design featured online interviews with families, and the researcher asked open-ended questions through a semistructured interview format. This design aimed to explore individual experiences of COVID-19, exploring children and adolescent mental health during this period.The IPA interview schedule was developed based on academic expertise. The semistructured interview employed “a funnelling technique” in which the methodology allows discussion to flow from general topics to specific themes [31]. This design was administered due to its allowance to provide structure and yield rich data while maintaining consistency across interviews.Participants were recruited from social media and various networking platforms. Participants signed up online and gave their contact details for the researcher to schedule an initial briefing and to send the relevant information forms. In order to prevent the spread of COVID-19, hard copy consent forms were not used. Consent was achieved verbally through the recruitment phone call and during the interview. Obtaining consent from the participant was recorded in the interview. The interviews were conducted through Microsoft Teams, recorded, and transcribed, based on participant consent. Microsoft Teams was the chosen method for the interviews as the research team has a professional Microsoft Office account, ensuring data protection and ownership of the data. The researcher’s Microsoft Office (Redmond, WA, US) account was used to store participant details and the recorded audio from each interview.The interviews were conducted during the first set of pandemic restrictions in Ireland. The interviews lasted for approximately 30 min and were led by the researcher. At the beginning of each interview, families were informed of their right to leave the research at any time without suffering any consequences. Each interview took place with families as a group. Following the initial briefing and obtaining consent, the researcher facilitated open-ended questions through a semistructured interview. This allowed for flexibility of questions during the interview. Families’ “funnelled” through their experiences of COVID-19 and discussed their experience of government restrictions and their shifted family dynamic. Following each interview, the researcher debriefed the family and asked if they had any questions. When the interview was finished, the researcher saved the audio file and transcribed the interview verbatim. The participants were also contacted to ensure that they were happy with their interview and if they had any queries about the research process.Families reflected on their situation during the pandemic and the first lockdown in Ireland. At this time, people were only allowed to leave home to shop for food and exercise briefly within 2 km of their homes. All schools were closed, and people were asked to work from home where possible. Additionally, public or private gatherings were banned, as were visits to hospitals and prisons, with some exceptions on compassionate grounds. People aged over 70 and vulnerable groups were told to “cocoon”, and travel to offshore islands was limited to residents only.A thematic approach to IPA was used for the data analysis using the NVIVO software package (QSR, Melboure, Ausrtalia). The analysis we employed followed the six steps outlined by Smith and Shinbourne [31]:Reading/Re-reading—The research team familiarise themselves with the interview concept by immersing themselves with the transcript.Coding—The research team identify codes and organise them into initial themes.Clustering—Themes are emerged by common themes and subthemes.Iteration—The iterative process involves several revisions, including checking themes, subthemes, and quotes.Narration—The research theme develops a narrative based on the findings. The narration process involves describing the themes and using quotes to illustrate them.Contextualisation—The researchers interpret the findings within the context of existing literature.Reading/Re-reading—The research team familiarise themselves with the interview concept by immersing themselves with the transcript.Coding—The research team identify codes and organise them into initial themes.Clustering—Themes are emerged by common themes and subthemes.Iteration—The iterative process involves several revisions, including checking themes, subthemes, and quotes.Narration—The research theme develops a narrative based on the findings. The narration process involves describing the themes and using quotes to illustrate them.Contextualisation—The researchers interpret the findings within the context of existing literature.IPA thematic analysis moves from descriptive to interpretive. The process consists of first categorising each transcript into broad themes and through continued review translating the data into more specific themes [32]. In line with the IPA approach, focuses on the subjective lived experiences of people [33], and, therefore, the researchers demonstrate reflexivity throughout the study [34]. To elaborate, reflexivity encourages the researchers to consider how their subjective worldviews may impact the research process. This approach benefits this research’s quality, particularly concerning the data analysis as it buffers against personal experiences and biases impacting the research findings [35].The interviews were conducted by Amy McGrane, MSc, who was provided with interview training by the research’s principal investigator, Katriona O’Sullivan. Once interviews were transcribed, to ensure procedural consistency across interviews, all researchers initially separately coded the interviews. The themes that emerged were then discussed and analysed together to determine the final themes. The research also counted the frequency of occurrences of reference to each code, and these are presented in Table 2.“My conclusion is that children have borne the brunt of this.”Throughout the interviews, we observed the stresses and strains placed on families in the COVID-19 crisis. One common theme that emerged was the direct impact that the restrictions were having on young people. Throughout the interviews, both parents and children referred to the extra strain placed on children during the pandemic. Many parents argued that their children struggled the most with the lockdown measures. For example, one mother described the impact on her young children, and how there was nowhere for them to go during the crisis, they did not have the benefit of being able to connect through technology:“My conclusion is that the children have taken, have borne the brunt of this. Really, you know they didn’t…have much to go on and my biggest point and the reason I really wanted to participate in this study is because a 9-year-old and a 7-year-old are not able to talk like we’re able to talk over a zoom call. It’s just it’s just not. I don’t think they’re neurologically mature enough, I know they’re digital natives, but I don’t think they have the capacity to learn or communicate across digital platforms.”Children were socially withdrawn and socially isolated—and parents said that children felt it much more than adults due to their lack of fluidity with digital means of communication. Additionally, parents were concerned that children were not supported to stay connected to their friends. There was limited peer interaction throughout the lockdown: “Yeah, I think that kills them the most with not having any social time, not seeing their schoolmates.” (Parent) “And you know, there’s a lot of stuff that they need to talk about, and I think even nothing to do with education, but even seeing their, their classmates on a screen and being able to share experiences would have gone a long way towards alleviating any source of negative mental health.”A recurring theme was how the young people were missing out the most through the pandemic and that their life was changed for the worse. Especially when it came to schooling and those transitions that young people usually get to experience.One parent describes this perfectly: “It was really sad, like they were really upset because they lost out on all the typical rites of passage that they were looking forward to, like the graduation their end of year ceremony” (Parent). Children and young people were seen as having the worst experiences in the COVID-19 crisis because they could not be children.According to both the parents and children in this study, children’s adverse mental outcomes were provoked by experiences of loneliness and social isolation. As young as five, children stated that, “I hated everything about COVID” (Child aged 5). when discussing being separated from friends and school. All the parents interviewed illustrated deep concern over the effects of the prolonged social isolation for their children:“Because I even found my son, he’s only gone seven since June, he was six at the time. He was a really outgoing child but because with the lock down. The hassle I had to get him back out on the road just, you know when the restrictions were lifted a bit. He went really into himself- really shy, now that has all completely turned back around, he is back to his old self now. But for a few weeks, I was actually getting worried about him like his mental health at six years of age. He didn’t want to go out and he got very into himself because he wasn’t out with friends.”A common understanding among parents was that their children would catch up on what they missed out academically. However, the children were most negatively affected by the lack of social interaction with their peers:“It was the lack of social interaction with his peers. I wasn’t even worried in the beginning. I was a bit like Oh God, his maths, his whatever. But after a while I was like I don’t actually care about that. That can be picked back up well, interacting with your peers is so important you know and just weeks and months of just sitting in front of a screen.”Another parent described the subtle changes in their child’s behaviour and how they felt powerless to help:“I’d say he would miss the social aspect…we have observed certain things, certain behavioural changes, not big things, but just there’s more frustration from being home all the time. I think that you just can’t. You can’t just replicate the craic a bunch of five- and six-year-olds have together and the engagement that they have…”Another catalyst of negative mental health outcomes for children was the stress placed on them to complete their home-schooling. One child stated that:“There were so many projects, nearly there was one project every week and then there was like every single subject on the thing and like she’s getting us to do like so much. And then I just got like really like it was too much for me. ‘Cause I’m used to like smaller work and she would give us more time to finish it.”This child found distance learning difficult and struggled to work independently. Like many other children, she missed the collaborative working style of the classroom and found home-school expectations stressful. This theme was recurrent in this study. Many families reported that to salvage the wellbeing within their family unit, they would not inflict home-schooling requirements onto their children:“She really missed her friends and she’s at that age where she was just, she used to get quite emotional about, and anxious about getting it done and stuff like that. So there were days where I just kind of told her look. We don’t have to do anything today. You know we can just. You know, go upstairs and read a book or something, or chill out. Or watch a movie or something like that…. We will just we just put on a movie and chill out and forget about school for today.”Several parents reported that their children experienced anxiety throughout the lockdown imposed by COVID-19. For example, one parent noticed the emotional changes in her children after a few weeks of restrictions:“In the beginning it was a huge novelty. Of course, this is great. You can do half an hour work and watch TV and play Xbox, but as the weeks went on, I felt all of us really. The morale was very low. Our moods were very low. He was very tearful. And even the teenager who was living the dream of doing a tiny bit of work and being online all day…he was low and crying and tearful.”Other parents articulated that their children’s anxiety manifested in attachment issues and frustration:“My daughter really struggled a lot, and she was like she wasn’t really able. She wanted a lot of input, even though she didn’t really need it. So, she was very, very anxious and so basically, she’d want to be sitting here. You know, when you’re doing, the work, this is the bedroom she might be sitting on the bed and you know she’d be interrupting. Yeah, and then she’d be huffing and puffing, you know what I mean.”Parents described having to control the information being given to their children about COVID-19 and how watching the news caused some anxiety.“She kept asking to watch the news, I said no, you’re not watching the news. I’ll explain to you what’s going on afterwards.”Another parent described the anxiety, which emerged when her child was considering the restrictions being placed on the family:“she asked…if we go on a bit further because the lady and the radio says that you’re not going to go more than two kilometers? And I’m gonna go out, I’m going to be. Are we going out about 2 km zone here? We ought not be doing this, you know, yeah, so there’s no, you can’t sugarcoat if the child is able to be picking up on public health messaging.”Overall families described their children as anxious and high strung during the pandemic, and that this was facilitated by the media coverage of the pandemic and the lack of access to friends and family.Interestingly, many parents of children younger than 10, noticed that their children developed increased negative or maladaptive behavioural changes following the restrictions. For example, one single parent noticed that her son reverted to an old bedwetting behaviour, possibly correlated with not seeing his father for an extended period:“For a good few weeks, he didn’t see the kids (their father), so that did affect my youngest son. One of the twins, his behaviour went all over the place and he was acting up and everything and went backwards and he was starting to bed wet again and still is, so I had all like to deal with as well.”Moreover, parents reported increased frustration among their children:“You know he’s an absolute nightmare, or beating up his sister, one or the other. I was on many calls where, like my boss would say, do you want to go and sort that out because you could hear the fighting from two rooms away.”However, the most reported maladaptive behaviour among young children was the decreased attachment security, and “clinginess” children demonstrated towards their parents:“Because of the lock down he didn’t see even his grandparents. Or you know anyone else for about three months, so he got really, really clingy with my wife, like his mam. And then he couldn’t like, she has to sleep in the same room with him, and if she is going to the toilet, he would follow her. I think the lock down really did play; I mean he was pretty attached but it got way worse. We had gotten used to the behaviour, and now that were coming out of the lock down, we can see already, he’s a little bit less, because he’s getting to see his grandparents once a week and he gets to go out on more trips without being confined. So, he’s coming out of it a little bit, but there were definitely behavioural issues because of the lockdown. It was it was blowing over like the you know, but through this winter again I’d say he’d regress all over again.”While adolescents experienced similar negative mental health outcomes to younger children, their mental health experiences manifested differently. Depression and anxiety were commonly reported experiences during the lockdown for adolescents, and many parents sought mental health services for their teenage children:“He was getting very depressed. At one stage he even asked me to get a counsellor. That’s how depressed he was getting. So, I was getting really worried about it. He seems to be coming back to himself but for a while, yeah. It had more impact on him then it did the little ones.”Negative mental health outcomes were reported following the mourning of milestones that were cancelled due to COVID-19:“He spends a lot of time in my mother’s house so yeah, he was pretty much locked in and he was getting very down some days. So, we were glad when they lifted a little bit that you could start going out and mixing with his friends and that a bit more. But I know that things like his graduation. He was very down about that like I think he was more down about that then missing his Leaving Cert I’ll be honest with you.”Another parent said:“She was robbed of a rite of passage I suppose. And she does feel upset over that and disappointed.”In addition, many parents reported that teenagers experience increased anxiety over future transitions, such as the transition into secondary school:“So she had months on her own dwelling on that. So that wasn’t a good year for her, and that’s a funny age as well, because you’re kind of like maturing and then puberty and everything. Yeah, full on hormones and she doesn’t like change. The biggest change of her life is coming, and then she just sat at home for months to think about it. Yeah, so that that was. I had to send her to a counsellor then after that ‘cause she just was struggling with anxiety for it”.The confined restrictions associated with COVID-19 had adverse mental health outcomes for all adolescents interviewed in this study. The lack of freedom forced teenagers to be confined to their household, which provoked various challenges. For example, one teenager stated that:“we stressed out so much because my parents are actually in the middle of a divorce, but they live in the same house. So that alone is stressful, then lockdown came, and my dad was always at home. Obviously, we couldn’t go anywhere we can’t get out of the house and it really added to the stress ‘cause you can’t walk into a room where my two parents are because you’re like stepping on egg-shells. It’s not really a relationship. I know for me it really effects like my mental health and everything. I hate being stuck inside. I don’t mind being inside, but you know, I just wanted my one time where I was like. I really want to go out. I really wanna get outta here. I even missed getting the bus.”For this teenager, she missed everyday life outlets that would typically distract her from a difficult family situation. Many teenagers suggested that they struggled with the confinements associated with lockdown despite their family dynamic. Moreover, adolescent parents exhibited concern regarding their lack of routine within the family household:“it was a struggle to get my 14-year-old out of the bed, because, you know, they’re not tired enough, so they’re not going to sleep. First of all, it was midnight. Then it was 1 AM like the average now its 3 AM for the two older ones. And of course, they don’t get up till 11 or 12.”This disruption in routine has been heavily correlated with an increase in screen time, including staying up to play video games or the influx of social media usage:“then for the past three months she’s lay on the sofa on her phone, and there’s nothing I can do that will motivate her to do anything. You know, should I take the phone off her? Should I ban her?”The removal of structure, routine, and outlets of support for teenagers during COVID-19 has been associated with the accentuation of their modern-day challenges. Parents articulated that their concerns for their teenage children have multiplied since they lost their standard outlets of support. For example, one parent discusses an increased awareness of her daughter’s body image issues and anxiety during this time:“We were on holiday last year with family that live in the UK and their daughter sent her some pictures that had come up, like they were looking at from last year and my daughter was just like Oh my gosh like, I hate myself. I hate myself. I hate my body and it was purely body image and she’s hitting that age.”There was reference made to some of the strategies that families used to mitigate the stress and anxiety associated with the pandemic restrictions. Families described spending time outside in nature together, while baking and household chores were described as calming. Many talked about the importance of routine with some saying it helped focus them.“So, we had to go and buy a printer so he could print out worksheets and stuff for him to do and he painted a lot of furniture. He painted the decking. He did gardening… You know anything just to keep him busy.”A theme that emerged from the analysis was the impact of COVID-19 restrictions on children with developmental disorders. Six parents, two special needs teachers, described the effect that the restrictions had on children’s mental health outcomes and students with Autism spectrum disorder (ASD) during the COVID-19 restrictions. While the children and adolescents with Autism vary in their developmental trajectories, they described similarly concerning responses to the lockdown. All participants in this study discussed how the drastic change in routine had provoked anxiety among the children with Autism:“Especially my younger daughter is on the ASD spectrum and she will have to integrate again. She does it every time she’s in school and they’re off on Friday and back on Monday, it’s a big stress for her, so you can imagine how 6 months off is going to be for her. It’s going to be like the first day of school again.”Anxiety was a prominent theme among these children as all families articulated its presence in the household. A parent of an ASD child with high needs discussed that his child’s anxiety manifested in relation to his attachment difficulties:“Like he used to freak out if my wife would leave the house at all without him, because she hadn’t left the house without him in months. He lost his routine and then the new routine again was really unhealthy in a way, because there was no, kind of external anything for him.”A parent and teacher who works with a class of high- functioning adolescents on the ASD spectrum articulated that his students had a comprehensive understanding of the virus which provoked fear of COVID-19:“You know it’s funny because some of the kids that I have like they would be very tuned in to, um, just you know they, they have a special subject like most kids with Autism have special interests…. But yeah, some of them are really aware of, of these viruses and how they could lead to other viruses and they were really stressed.”Moreover, this awareness of COVID-19 provoked maladaptive existential anxiety categorised by an overwhelming fear of the virus:“He said that like even if there was a vaccination, he wouldn’t get it. And, because he’s afraid of needles, and he’s afraid. Basically, he started to think an awful lot about you know what will happen to him when his parents are dead and when he’s like an individual living in the world with Autism and not having the support he’s had all his life, so he’s gone really into existential anxiety and. You know this is what COVID has done. Like you know, it’s a psychological disease as well.”This research explores Irish children and adolescents’ experiences during COVID-19, specifically the pandemic’s psychological impact on this age group. The study uses an inductive qualitative research design and documents the psychological effects on young people primarily through parents and caregivers’ observation. The key findings suggest that public health restrictions had adverse implications on children and adolescents’ mental health. Parents and caregivers reported higher levels of stress, depression, and anxiety among their dependents resulting from social isolation. These conditions were exacerbated in children and adolescents with developmental disorders. These findings are concerning to healthy development because adverse psychological experiences in childhood are associated with an increased risk of anxiety later in life [36]. The themes found in Table 2 are discussed below and often intertwine.Research has shown that social isolation and loneliness are impacting young people [8], and there have been increases in adolescent anxiety [9,10], due in some part to pandemic related restrictions [11,12]. Families in this study also described increases in adolescent anxiety, depression, and despair. At the same time, many mourned the cancellation of milestone events and were negatively affected by the loss of routine and guidance. These findings are in line with previous research [14,29,37], which indicates that the restrictions associated with the COVID-19 pandemic have impacted the psychosocial wellbeing of young people. The research found that the use of social distancing and stay-at-home orders to mitigate the spread of COVID-19 is negatively affecting the mental health of young people as has been seen elsewhere [8]. There is an increased risk of developing psychiatric disorders, such as depression and anxiety during these formable years with half of all mental health disordered developing before the age of 14 [15]. For adolescents, this risk is complicated by how hormonal and neurobiological changes correspond to increased emotional reactivity and continual development of stress regulation and coping strategies [38]. Additionally, social interactions with peers contribute to adolescents’ social health, creating a sense of belonging and reducing feelings of burdensomeness on others. These factors are essential for interpersonal needs [38]. It is apparent from this research that as the pandemic subsides, and there is a lifting of restrictions, the emotional consequences of COVID-19 on young people needs consideration and support systems emplaced to mitigate any long-lasting psychosocial effects [9,10].This research shows that the stress over home-schooling was an added burden to families. The closure of schools shifted education from the classroom to the home. This change was unprecedented, and families had little or no experience of protracted home-schooling [39]. School routines are crucial coping mechanisms, particularly for young people with mental health issues [15]. Prolonged quarantine, fear of infections, boredom, insufficient personal space, and separation from classmates and teachers causes stress in children and adolescents [12]. With home-schooling increasing family stress [29] especially in already burdened families [40], this research shows a need to support young people to engage with education in ways that are not stressful, which suit their family situation. As discussed, social isolation is damaging to the psychological wellbeing of young people. In the United States, epidemiological data indicated that 35% of adolescents who used mental health services between 2012 and 2015 availed their school’s mental health services [10]. With schools closed and the decrease in services offered by them, home-schooling stresses may be acting to exacerbate the long-term psychological impacts on young people; young people’s experience in this study point towards this.The restrictions associated with COVID-19 had negative mental health outcomes for adolescents in this study. Increased levels of anxiety and heightened fear were observed. These findings matched those from previous research, where increases in adolescent anxiety have been seen [9,10]. The lack of freedom forced upon young people through the national restrictions provoked various challenges; with the closure of schools and home-quarantine increasing anxiety and loneliness among adolescents and increasing negative behaviour in younger children. Moreover, reported elsewhere, this study supports the emerging view that pandemic related restrictions [11,12] have negatively impacted the psychological wellbeing of young people.The research also observed that children and adolescents with ASD had specific mental health challenges related to the disruption to routine, which emerged as a direct result of the restrictions. These results highlight children’s mental and emotional vulnerability during this pandemic, especially those with developmental disabilities. Previous research shows that these children are at increased risk as the pandemic dissipates and that they require specific interventions to minimise having disproportionate consequences on their psychosocial and emotional wellbeing [41]. The outcomes of this research support this and indicate in the relatively small sample of families with a child with ASD; they required support to reinstate routine and support a transition back to pre-COVID times.Though this study produced valuable data, there are some limitations. First, the researchers sought a sample set that would be equally represented across specific demographics, such as single-parent families and two-parent families, socioeconomic backgrounds, and rural and urban settings. However, the sample set was dependent on the willingness of families to participate. As a result, the sample set was not equally distributed across these demographics. For instance, because Dublin has the largest population size in Ireland, many families resided in Dublin, an urban setting. Gender also played a role, as mothers were more likely than fathers to be interviewed. A second limitation is the use of the remote interview technique. Though this was an essential tool because of COVID-19, this technique limits nonverbal communication between the interviewer and interviewee and nonverbal communication can provide important insights into the interview.This study, alongside international research, illustrates the mental health challenges arising from COVID-19 pandemic related restrictions. Young people are at risk of suffering psychological consequences, and we are not yet sure of the long-term impact that this will have. In addressing these difficulties, there is a growing need to implement policies that will help children and adolescents cope with the short-term and long-term psychological effects of the pandemic, especially those who are deemed more vulnerable in this group. We need to be considering the impact of further restrictions and ensure that mental health services for young people are easily accessible if we are to prevent longer-term mental health impacts [16]. With the increasing anxiety and depression levels reported among young people and the parents’ observations in this study that children are struggling to cope, it is crucial to ensure these support services are made available. Young people need support to develop healthy coping mechanisms as they begin to process the potentially adverse effects of COVID-19.Semistructured Interview ScheduleCan you tell me a bit about your day-to-day family life during the COVID-19 pandemic?Can you tell me a little bit about what your average school day looked like before the COVID-19 pandemic?Did you do homework?Favourite subjects?What was fun in school?Do parents work—how does school life and home-life meet?Can you talk about how the COVID-19 has affected your schoolwork?Did the school talk to you about what was happening at the start?Did you (the child) know what was happening?Was your work life affected?Were there any emotional effects?Were you able to engage with school subjects and content?Who in the household manages the home-schooling?Many families have found home-schooling difficult or strange—some have done it, some have not—can you tell me if you have been able to keep up with school during the crisis? Tell me how your day-to-day school engagement looks (even if there is none it is fine).Is there anything that has been particularly helpful to you over the last few months in doing home-school work?Is there anything that has been particularly challenging?How have you had contact with school?Would you have preferred different contact?How about you (child) did you see your teacher in the last while?Have you any ideas on how we could have made this better for children?Do you guys use computers tablets and/or the Internet—tell me what you like to do on these?Do you use these for school-work? Is there any apps or sites you have used for home schooling?Can you describe what that has been like?Any challenges?Any good points?Does this feel the same as school learning?Have you engaged with Raidió Teilifís Éireann (RTE) school or any specific things that support school work?What was it like? Can you talk me through your answer a little?If this pandemic happened again what do you think schools could do to help you?Is there anything that we could do online to help with schooling in the future?Would you like to tell me anything important about your home school experience that you have not had a chance to say or has just occurred to you?Are you worried about going back to school?What is the biggest thing you both/all miss about going to school? What don’t you miss?Can you talk about how the COVID-19 has affected your family?Can you tell me what you think the future will look like for you and your family?Conceptualization, K.O. and K.M.; methodology, K.O., K.M. and A.M.; formal analysis, S.C., A.M. and N.B., investigation, K.O., A.M., S.C.; resources, N.B., N.J.; writing—original draft preparation, N.R., N.J., L.B., A.M., S.C., K.O.; writing—review and editing K.O., S.C., A.M., supervision, K.O. and K.M., S.C.; project administration, S.C., A.M., K.O., N.B.; funding acquisition, K.O. and K.M. All authors have read and agreed to the published version of the manuscript.SFI/EI/IDA COVID-19 Rapid Response (Award ID: 20/COV/0151).The study received ethical approval from The Social Research Ethics Subcommittee, National University Ireland, Maynooth, (2407411. 15/05/2020).Informed consent was obtained from all subjects involved in the study.The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.The authors declare no conflict of interest.Demographics of the sample set by gender.The referral and percentage of qualitative themes in this data set.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01063.txt

@@ -0,0 +1 @@
+Almost all European citizens rank patient safety as very or fairly important in their country. However, few patient safety initiatives have been undertaken or implemented in Poland. The aim was to identify patient safety strategies perceived as important in Poland and compare them with those identified in an earlier Dutch study. A web-based survey was conducted among primary healthcare providers in Poland. The findings were compared with those obtained from eight other countries. The strategies regarded as most important in Poland included the use of integrated medical records for communication with specialists and others, patient-held medical records, acceptable workload in general practice, and availability of information technology. However, despite being seen as important, these strategies have not been widely implemented in Poland. This is the first study to identify strategies considered by primary care physicians in Poland to be important for improving patient safety. These strategies differed significantly from those indicated in other countries.The 2005 Council of Europe definition of patient safety is “freedom from accidental injuries during the course of medical care; activities to avoid, prevent, or correct adverse outcomes which may result from the delivery of healthcare” [1]. The European Commission Directorate General Health and Consumer Protection (SANCO) Eurobarometer survey found that 78% of European citizens ranked medical errors as very or fairly important in their country, and 98% felt that having national political support for patient safety was of high importance [2]. In addition, the World Alliance for Patient Safety has urged member states to develop a coherent strategy for improving patient safety [3], resulting in the development of various safety strategies and validated instruments to promote a culture of patient safety [4,5,6,7,8,9] and research on its implementation in various countries [10,11].In Poland, most activities related to improving the safety of medical procedures have been local initiatives focused primarily on medication errors. No systematic monitoring of sentinel events (any unanticipated events in a healthcare setting resulting in death or serious physical or psychological injury to a patient, not related to the natural course of the patient’s illness), circumstances, near misses (unplanned events that have the potential to cause, but do not actually result in human injury), or preventable events as described by The Joint Commission (https://www.jointcommission.org/resources/patient-safety-topics/sentinel-event/) has been implemented in general practice, nor has any detailed data been acquired and any root cause analysis of incidents been performed [12,13,14,15]. Furthermore, the health service in Poland suffers from lack of public funding and shortages of medical personnel; these have negative effects on access to healthcare services [8] and, potentially, safety. The aim was (1) to identify and map the patient safety strategies perceived as important in Poland, and (2) to compare the views on importance of these strategies in Poland with those of healthcare professionals in a previous Dutch study [16].A convenient sample of randomly selected consecutive primary care physicians in Poland were surveyed using an electronic questionnaire. Participants were recruited from several conferences and seminars in Poland where the strategies were presented; most respondents were physicians potentially interested in patient safety. Further contacts were recruited through a snowball sampling procedure. In order to obtain the reliability of the answers, context descriptions were provided. Participation in the study was voluntary. The initial survey was developed in the Netherlands and used in eight countries: Austria, Denmark, France, Germany, the Netherlands, New Zealand, Slovenia, and the United Kingdom [16].The questionnaire was translated and validated for a Polish context by the study team. It covered five categories: facilities in the practice, patient safety management, communication and collaboration, generic conditions for patient safety, and education regarding patient safety, as well as 37 patient safety strategies (items) including incident reporting, medication alerts, patient safety indicators, periodic medication review, training on patient safety, or culture conditions [16]. A full list of strategies is given in Table 1. A professional research and consulting company collected the questionnaires from the Polish respondents.The responses regarding the presence of a strategy were “no”, “no, but planned”, “yes, <50% present in the country”, and “yes, >50% present in the country”. Responses concerning the importance of the strategies were “no”, certainly not”, “no, probably not”, “partly yes, partly no”, “yes, to some extent”, and “yes, very much”. The respondent could add a comment to each item. Respondents were also asked to list any other strategies that were not mentioned in the questionnaire.The percentage of respondents indicating the variants “yes, >50% present in the country” and “yes, very much” were calculated.The results were compared with those obtained in Dutch study using the two-proportion z-test. The following statistical hypotheses were formulated: (H0) the proportion from the studied countries is equal to the proportion from Poland, or (H1) that it is not equal. This was a two-tailed test. The null hypothesis was that no difference existed between the two population proportions or, more formally, that the difference was zero [17]. The z-value and associated p-value were calculated. The null hypothesis (H0) was rejected if z ≥ 1.96 or if z ≤ −1.96 or p < α = 0.05. The statistical analysis was performed using SPSS 20 (Statistical Package for the Social Sciences, International Business Machines Corporation (IBM), New York, USA).During March 2019, 1300 questionnaires were sent to Polish respondents, and 251 replies were received. Out of these 251 replies, three were not completed and, hence, were discarded. Therefore, 248 individuals were included in the study. Of these, 53.2% were male, 56.9% were general practitioners (GPs), 41.1% were in internal medicine, 20% were physicians with just general medical training, and 2% were professionals with other specialties (gynecologist, surgeon, or endocrinologist). Eight percent of the respondents were also involved in teaching and research. Most respondents (66.1%) worked in health centers, 21.4% worked in group practice, and 24.6% ran an individual practice. The physicians differed significantly with regard to the number of patients under medical care of a physician, ranging from 0 to 4000 (mean = 1.763; SD = 873) and the size of their place of work, with the patients in their facility ranging from 300 to 16,220 (mean = 5091; SD = 3096). In addition, 49.6% of respondents conducted their medical practice in cities with over 100,000 inhabitants (Table 2).Regarding the “very important for patient safety” category, in the Polish study, the percentage of affirmative responses was much lower than in the Dutch study (the population in the Dutch study by Gaal et al. [16] was almost five times smaller (58 from eight different countries compared to 248 respondents in the present study)) for four strategies: “telephone facilities that allow quick access to the practice, particularly for urgent health problems”, “planned checks of safety of equipment, medication, and other facilities in the practice”, “forms for reporting incidents available”, and “computerized decision support regarding medication safety in daily practice. The percentage of affirmative responses for the remaining strategies was similar to that in the Dutch study.The Polish respondents were much less likely to indicate the top score (yes: >50% present in the country) than in the Dutch study; this response constituted fewer than 10% of all responses. The response was given in 27% of responses for “planned checks of safety of equipment, medication, and other facilities in the practice”, 15.3% for “forms for reporting incidents available”, 13.7% for “computerized medical record system, which is adequately kept”, and 10.9% for “working agreements with pharmacists when problems arise with delivering medication, e.g., alerts, interaction”. This highlights the rudimentary implementation of these strategies in Poland (Table 1).The Polish and Dutch studies returned similar percentages of responses “very important for patient safety” for the strategies “nationwide or regional incident reporting weeks” and “campaigns to increase patients’ and public awareness of patient safety in general practice”, while significantly higher percentage responses were given by Poland for “surveys and other types of consultations of patients regarding safety incidents”.The response “>50% present in the country” was less common in the Polish study than the Dutch study for five strategies, but similar results were observed for the other five (Table 2).In this category, the response “very important for patient safety” was significantly more common in the Polish survey than the Dutch survey for the strategies “integrated medical records for communication with specialists and others” and “patient-held medical records”.The response “>50% present in the country” was recorded at similar frequencies in the Polish and Dutch studies for the following strategies: “structured formats for information on referral of patients”, “the pharmacist conducted periodic reviews of the patient’s medications for potential interactions”, and “patient-held medical records”. For the remaining strategies, it was recorded much less frequently in the Polish study than the Dutch study (Table 1).In Poland, this category was the most important. The Dutch and Polish studies returned similar percentages of “very important for patient safety” for the strategies “understanding of patient safety in health professionals, particularly regarding how it differs from complications of treatment” and “adequate procedures for identifying and managing burnout in health professionals”. However, a significantly higher percentage was noted for “culture and mentality which facilitates learning from incidents” in the Dutch study than the Polish study, while a significantly higher percentage was returned in the Polish study for “workload is perceived as acceptable in general practice” and “availability of information technology in general practice, and skills to use these adequately” than the Dutch study.The Polish respondents were significantly less likely to indicate “>50% present in the country” for “availability of information technology in general practice, and skills to use these adequately”. However, similar responses were observed between the Polish and the Dutch studies for perception of presence in the country (Table 1).Education was seen as the most important factor for improving patient safety in both studies. In this category, the response “very important for patient safety” was observed more frequently in the Dutch study than the Polish study.Similar frequencies of “>50% present in the country” were observed between the Polish and Dutch studies for the strategies “education on patient safety in the vocational training of practice nurses” and “postgraduate education on patient safety of practice nurses”. However, it was observed much less frequently in the Polish study for the remaining strategies (Table 1).This study is one of the first in Poland to identify strategies considered to improve patient safety. It compares the perception of strategies needed to improve patient safety in Poland with those identified in an earlier study carried out in a number of other European countries [13,14,15,16]. Many differences appear to exist between Poland and other European countries with regard to the perceived importance of patient safety strategies. Although 14 of the 37 strategies included in the survey were regarded as being similarly important in the Polish and Dutch studies, very few in Poland perceived them as being implemented in daily practice. This is one of key differences with those obtained by Gaal et al. [16].In Poland, the most important strategies included “the use of integrated medical records for communication with specialists and others” and “patient-held medical records, acceptable workload in general practice and availability of information technology”. However, despite being seen as important, these strategies have not been widely implemented in Poland; similar results have been obtained in previous studies [15,16]. In the present study, the highest indications were given for “generic conditions for patient safety” and the lowest (all less than 5%) for “patient safety management”; these strategies tended to be less frequently implemented. There is little correlation between the intention of a healthcare worker and the subsequent (improvement) behavior [12,16].The respondents in Poland ranked all educational items similarly to those in the other countries. However, the implementation of this strategy was found to be lower than in other countries. Hence, including further education on patient safety in vocational training and postgraduate programs in Poland would be desirable. Moreover, a patient safety program could be valuable in education for practices (such as root cause analysis), as noted previously [14,15,16].The response rate for this study was acceptable. However, most of the respondents were practicing GPs (56.9%), which can be seen as a potential bias. Earlier studies [15] found that, although “regular” practicing GPs found patient safety highly relevant, they tended to have a very broad idea about patient safety. The method used requires a random sample of each population group to compare categorical data and a number sample greater than 100. The total sample size was 248, and the population was not truly definable. Their expertise to speak on the subject was uncertain and may not actually be comparable to the Dutch study, with only 58 respondents [16]. The respondents of the survey showed that the implementation of patient safety strategies is low, probably due to the very low number of respondents (19%), which can be seen as a potential bias. Moreover, among specialties, most of the respondents were practicing GPs which could direct their responses on the basis of their daily work and, therefore, can be seen as another potential bias. Furthermore, the survey only obtained general opinions from the respondent group, and it is difficult to extrapolate system-level changes and recommendations for strategies on the basis of such limited qualitative data.The two populations did appear to be significantly different with regard to their opinions. Polish respondents more often declared that “availability of information technology in general practice, and skills to use these adequately”, “acceptable workload in general practice”, and “patient-held medical records” play the most important roles in patient safety. More worryingly, our findings suggest that patient safety management strategies are not perceived as being implemented in Poland at all. The findings may be used to focus the attention of healthcare authorities and professionals on specific safety strategies considered essential by professionals in Poland.Key points are the following:The perception of patient safety by GPs varies between present studies.For Polish respondents, the availability of information technology and the skills to use it were most important for patient safety.The differences between various countries regarding attitudes toward patient safety should be addressed in the international regulations of medical practice.The conclusions are realistic, demonstrating a low implementation of safety management strategies in Polish primary care.The study design was conceptualized by K.K., M.G.-C., M.W. and A.K., who also prepared the Polish version of the questionnaire and organized data collection. Analyses were performed by I.S. and A.D. (SPSS 20) and they prepared the first draft of the manuscript. All authors have read and agreed to the published version of the manuscript.The research leading to these results received funding from the European Community’s Seventh Framework Program FP7/2008-2012 under grant agreement n°223424, as well as from the Ministry of Science on Higher Education in Poland and the Medical University of Lodz.The internal group of experts, consisting of a GP and public health specialist, advised against external ethical assessment since the research was restricted to collecting feedback related to information and communication artifacts with exclusion of personal data. This study was performed as part of the LINNEAUS PC (“Learning from International Networks about Errors and Understanding Safety in Primary Care”) project http://www.linneaus-pc.eu, funded by the European Union Framework 7 program. LINNEAUS EURO-(2008-12) within theme 1: Co-operation and Health (Grant Agreement number 223424).Informed consent was obtained from all subjects involved in the study.The dataset supporting the conclusions of this article is included within the article. A copy of the questionnaire can be obtained from the first authors.The authors kindly thank Michel Wensing for his permission to adapt his previously implemented questionnaire for use in the present study. The participation of general practitioners is gratefully acknowledged. The authors specially thank Edward Lowczowski for English language assistance.The authors no conflicts of interest.Views on the importance and implementation of patient safety interventions. GP, general practitioner.1 consecutive number; * p < 0.05, ** p < 0.01, *** p < 0.001.Demographic characteristics in Polish study.* Some respondents were specialists in more than one discipline.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Med-MDPI/ijerph_6/ijerph-18-03-01064.txt

@@ -0,0 +1,7 @@
+Background: Drawing on the impact of the COVID-19 global pandemic and its sanitary measures on coping strategies for preserving health, it is also necessary to add exposure to certain work stressors, such as burnout. The aim of the study was to assess the influence of the confinement situation caused by COVID-19 on the levels of self-perceived health and psychological capital in a sample of workers, as well as to analyze whether exposure to burnout before social isolation would help to explain the levels of health and psychological capital. Methods: Data were collected in a longitudinal design. Time 1 surveys (December 2019) were sent to a sample of 354 Spanish workers while in Time 2 (April 2020) the employees completed 235 questionnaires. Results: Our findings indicate a significant worsening of employees’ health perception (t = −4.13; p < 0.01) and psychological capital (4.10, p < 0.01) levels during mandatory confinement in Spain. Our results also revealed that emotional exhaustion is the only burnout dimension capable of explaining the variance of health while self-efficacy does regarding psychological capital. Conclusion: We conclude a significant reduction in self-perceived health and psychological capital during COVID-19 mandatory confinement, and that burnout acts as a predictor variable in both health and psychological capital variance.The 12 December 2019 marked, in the Chinese city of Wuhan, the appearance of a new human coronavirus with acute respiratory syndrome. Provisionally called 2019-nCoV or SARS-Cov2, it is the origin of a global pandemic, declared by the World Health Organization (WHO) on 11 March 2020. According to the latest data from the WHO [1], the case fatality rate (CFR; the ratio between confirmed deaths and confirmed cases) varies widely between countries. The case fatality rate has reached 4.9% in China, whereas in European countries differs from 0.5% in Iceland to 3.5% in Italy. In Spain, the case fatality rate had increased up to 12.2% in the worst moment of the pandemic. Nowadays, this rate is 2.5%. It is important to note that this information should be treated with caution due to the difficulties in collecting and interpreting the data. Currently, we finally can say there is a specific antiviral treatment. After the discovery of the vaccine against the virus, distribution to all countries has recently begun. In most of them, the vaccination process in the general population has already started.Unfortunately, this pandemic not only has had effects on population health but also on the labor market. According to the latest International Labor Office report [2] in the third quarter of 2020 it has been estimated a decline in global working hours of 12.1%, equivalent to 345 million full-time jobs. In terms of work, younger and older people and especially women have been hit particularly hard by the COVID-19 economic crisis.Considerable research attention has been directed toward physical and psychological effects caused by the COVID-19 virus on general population health and well-being [3,4,5]. Nevertheless, we consider a lack of research about how previous work stressors as burnout could increase the pandemic effects on employees’ health. Hence, the main goal of this study was, on the one hand, to examine the influence of the confinement situation caused by COVID-19 on employees’ self-perceived health and psychological capital levels. On the other hand, analyze whether exposure to burnout before social isolation would help to explain the levels of self-perceived health and psychological capital.One of the main strategies to contain the pandemic is confinement that is, legally forcing someone to stay in a certain place, usually the home. In Spain, for example, the confinement period has been prolonged for 97 days. During this time, it has been pointed out that due to the separation from loved ones, the feeling of loss of freedom and the uncertainty associated with the state of the disease, loss of job or professional status, this situation could have had dramatic long-term effects on general population psychological health [3,6]. Hence, these effects were already showed during the investigations carried out in the course of the previous epidemic associated with another SARS virus [7,8].According to Hobfoll’s Conservation of Resources Theory [9] (p. 339), resources are defined as: “those objects, personal characteristics, conditions, or energies that are valued in their own right, or that are valued because of the act as conduits to the achievement of protection of valued resources”. The basic tenet of the Conservation of Resources Theory is that individuals strive to obtain, retain, protect, and foster those things that they have valued [9]. Hence, having a higher level of resources is favorable, especially in situations of high psychosocial stress [9].Psychological capital is one of the resources that could be framed in Conservation of Resources Theory [10,11]. It has been proposed as a useful personal resource for coping with work stress [12,13], mobbing [14], burnout [15], or even unemployment [16,17]. Luthans, Youssef, and Avolio [18] propose psychological capital as a second-order construct that brings together four resources (optimism, resilience, hope, and self-efficacy), and is defined as a state of positive psychological development characterized by having the confidence to face challenges and difficult tasks (self-efficacy); make positive attributions about the present and future triumphs (optimism); visualize and persevere in the goals, as well as redirect the objectives when necessary to achieve success (hope); recover and even emerge stronger from adversity (resilience). The Conservation of Resources Theory by Hobfoll [9] defends that resources can be treated independently or integrated into more complex models. With this, the author pointed out the fact that some psychological concepts are better understood as the representation of a global multidimensional factor. That is, this theory defends the positive synergy between certain resources when they are part of a second-order construct that works better than each of its components separately [19]. Given the above, psychological capital research indicates that its four components have an underlying common bond that makes it a higher-order construct. This means that, if we take into account the four components of psychological capital as a whole, instead of focusing on them individually, their effects will have a greater impact than each of the four facets separately [18,19].Even though psychological capital could help in high-stress situations, some authors have emphasized that the continued use of individual resources (tangible and intangible) to achieve vital objectives and face stressful situations, could result in resource deterioration [20]. When someone needs to face a high-caliber stressor, such as the confinement situation caused by COVID-19, two scenarios can be observed: (a) the resources are not adequate to cope with a certain stressor; (b) in certain situations, the stressor is so overwhelming that the resources available to a person, even if they are adequate, are not enough. In both cases, an irreparable deterioration of resources would occur [9]. Likewise, we consider that the confinement situation caused by COVID-19 is a serious stressor [3,4,5] before which, people’s psychological capital levels, could be deteriorated.Drawing on the impact of the COVID-19 global pandemic and its sanitary measures’ impact on health [3,4,5] and psychological capital [9], it is also necessary to add the exposure to certain work stressors, such as burnout [21].Current literature supports the idea that burnout is a chronic work stressor, consisting of three symptoms: emotional exhaustion, cynicism, and low professional efficacy. Burnout syndrome is not an automatic process originated exclusively by work stressors, but the result of an interaction between personal characteristics and organizational features [22].Many studies have reported a positive association between burnout syndrome and worsening of workers’ health [23], alcohol abuse consumption [24], sleep disorders [25], depression [26], sedentary lifestyle [27] or musculoskeletal pain [28]. Similarly, numerous investigations showed a negative association between burnout and psychological capital [15,29]. In line with this approach, psychological capital has clarified the relationship between burnout and job satisfaction [30], stress and burnout [31], as well as between work–family conflict and burnout [32].For explaining the association between burnout syndrome, self-perceived health, and psychological capital during COVID-19 social isolation, we apply the loss cycle put forward in Hobfoll’s Conservation of Resources Theory [9]. This theory has been previously proposed as a valid theoretical framework to explain burnout outcomes [9]. As explained above, people strive to protect, conserve, promote, and obtain more resources [33]. In line with this argument, stressful situations appear when there is a real loss of resources or when they are threatened. Hence, when it is not possible to avoid the loss or recover from it, higher stress levels would give way to the loss cycle. This will, in turn, promote a predisposition towards the protection of the resources that are already owned but not always towards the creation of new resources [34].The loss cycle hypothesis could be aligned with the idea that workers under high levels of stress perceive and create more work demands over time [35,36]. Apparently, this could be related to self-undermining behaviors, understood as certain behaviors that create various obstacles and end up undermining one’s performance [36]. Empirical evidence has shown that those workers who tend to self-weaken behaviors are more likely to experience higher levels of chronic stress, emotional demands, work pressure, and burnout [36,37]. Likewise, employees under self-undermining behaviors perceive a more stressful, confusing, and conflictive work environment, which ends up leading to an increase in labor demands [35,37]. All in all, this self-weakening conduct could be the fuel necessary to activate the loss cycle associated with burnout according to Hobfoll’s Conservation of Resources Theory [9]. Hence, those employees with higher levels of work stress are more likely to show self-debilitating behaviors, which will lead to higher levels of work demands and, consequently, will have fewer resources to deal with them, thus increasing their stress levels again [36]. That is, it is possible that the fact of being exposed to burnout before the start of the pandemic and its associated containment measures, makes workers fall into a loss cycle where they are not able to mobilize their resources to face this situation, thus affecting their health levels and weakening their levels of psychological capital [38]. Following Hobfoll [9] it is possible that when a worker has already seen deteriorated their personal psychological resources (in this case, psychological capital) trying to cope with burnout before confinement social isolation when they have to face another seriously stressful situation such as the one caused due to the COVID-19 virus, they experienced a lack of resources to deal with it, and therefore their health is affected. According to Hobfoll [9] when some resources have been invested in a previously stressful situation, they may not be available to meet future demands, leaving people exposed to the impact of stress.According to the arguments exposed above, we hypothesize that:
+            The confinement situation caused by the spread of the COVID-19 virus will hurt the perception of health.
+          
+            The confinement situation due to the spread of the COVID-19 virus will harm psychological capital levels.
+          
+            Previous COVID-19 confinement burnout workers’ level will be able to predict their psychological capital (a) and self-perceived health (b) during COVID-19 social isolation measures.
+          A longitudinal non probabilistic design was followed. Data were collected in two-time points. Burnout, psychological capital, and health were assessed at Time 1 (December 2019), whereas psychological capital and health were also assessed four months later at Time 2 (April 2020), during the social isolation confinement caused by COVID-19 in Spain. The study was carried out under the recommendations of the ethics committee of Universidad de Murcia. All subjects gave written informed consent following the Declaration of Helsinki. Participants received an email with a direct link to the questionnaire hosted on the Universidad de Murcia inquiry platform to ensure confidentiality and anonymity. The email also contained the research goal, the instructions for the correct completion of the questionnaire, and a reminder of voluntary participation in the study.Time 1 surveys were sent to a sample of 354 workers from different Spanish socioeconomic sectors. A total of 235 usable questionnaires were received, resulting in a response rate of 66.4%. About half of the respondent were men (59.1%; N = 139) with an average age of 36.62 years (SD = 13.43; range = 18–62 years). In terms of education, 37.8% (N = 89) had university studies, 26.8% (N = 63) followed vocational training studies, 18.7% (N = 44) had primary studies, and 16.6% (N = 39) had completed high school. Regarding job position, 35.3% (N = 83) held positions as a qualified worker, 21.7% (N = 51) were technicians, 18.3% (N = 43) worked as an unskilled worker, 12.8% (30) were a middle manager, and 9.4% (22) executives. Most respondents were permanent employees (53.2%; N = 125). Lastly, the average professional experience was 9.37 years (SD = 10.62, range = 1 month to 42 years).Time 2 surveys were sent to a sample of 235 workers. A total of 198 usable questionnaires were received, resulting in a response rate of 84.3%. About half of the respondents were women (50.7%; N = 100) with an average age of 39.73 years (SD = 12.78; range = 18–58 years). In terms of education, 48.6% (N = 95) had university studies, 22.8% (N = 46) followed vocational training studies, 18.6% (N = 37) had completed high school, and 10% (N = 20) had primary studies. Regarding job position, 36.8% (N = 73) held positions as a qualified worker, 23.5% (N = 47) were technicians, 11.8% (N = 23) worked as an unskilled worker, 16.2% (32) were a middle manager, and 11.8% (23) executives. Most respondents were permanent employees (61.4%; N = 122). Lastly, the average professional experience was 11.74 years (SD = 10.57, range = 2 months to 38 years).We assessed Burnout by the Maslach Burnout Inventory-General Survey (MBI-GS), developed by Maslach, Jackson, Leiter and Schaufeli [39] in the Spanish version by Salanova, Schaufeli, Llorens, Peiró, and Grau [40]. This scale is a 15-item self-report measure of job burnout that includes three dimensions, namely, emotional exhaustion, cynicism, and job self-efficacy. The items are rated from 1 (never) to 7 (every day). Some items are “I have become less enthusiastic about my work” (item 9) and “I have become more cynical about whether my work contributes anything” (item 13). Regarding internal consistency, Cronbach’s alpha coefficients were recorded for burnout at T1: emotional exhaustion α = 0.87; cynicism α = 0.82 and job self-efficacy α = 0.72. The test–retest reliability was r = 0.38 (p < 0.001) for the exhaustion subscale, r = 0.34 (p < 0.001) for the cynicism subscale, and r = 0.37 (p < 0.001) for the efficacy subscale.The OREA questionnaire developed by Meseguer-de Pedro, Soler, Fernández-Valera, and García-Izquierdo [41] was used to measure psychological capital. It consists of 12 items, three items for each psychological capital dimension. The OREA questionnaire items are rated on a four-point scale, that is, going from 1 (strongly disagree) to 4 (strongly agree). Sample items per dimension are: “In difficult times I usually expect the best” (Optimism); “I usually reach my goals even if there are obstacles” (Resilience); “I think my life is worth it” (Hope); “I am confident about I could effectively handle unexpected events” (Self-efficacy). The coefficient alpha for this scale was 0.82. Regarding internal consistency, the following internal consistency coefficients (Cronbach’s alpha) were found: α = 0.88 at T1 and α = 0.88 in T2; with a test–retest reliability of r = 0.60 (p < 0.001).As an indicator of the perception of psychological health, the GHQ-12 questionnaire by Goldberg and Williams [42] was used in the version validated by Rocha, Pérez, Rodríguez-Sánz, Borrel, and Obiols [43]. This instrument strives to estimate two kinds of phenomena: first the unfitness to experience and carry out activities in a functional, healthy way, and secondly the appearance of stressful phenomena. It contains 12 items referring to problems with well-being suffered in recent weeks (such as item 5: ‘Have you constantly felt overwhelmed and tense?’). It is evaluated via a four-point Likert scale ranging from 1 (‘not at all’) to 4 (‘much more than usual’), and therefore high scores indicate a lower level of health. Regarding internal consistency, the following internal consistency coefficients (Cronbach’s alpha) were found: α = 0.83 at T1 and α = 0.88 at T2, with a test–retest reliability of r = 0.60 (p < 0.001).The statistical analyses were performed using the IBM SPSS Statistics Versión 24.0 computer program. Bearing in mind the longitudinal design of the study, we proceeded as follows: first, internal consistency and reliability analyses (test–retest) were performed, followed by descriptive and bivariate correlations. Subsequently, mean comparisons were made between the two measurement moments to check if there were changes in the socioeconomic and demographic variables and to test the temporal persistence of the main study variables. Finally, for testing means differences, Student’s t-test was used. Finally, to analyze the longitudinal relationships, different multiple-step hierarchical regression analyses were carried out.We first inspected the descriptive statistics. Table 1 displays the means, standard deviations, and correlations of variables included in this study.Secondly, we tested the existence of significant differences between self-perceived health and psychological capital in Time 1 and Time 2. For fulfilling this purpose, we applied the T Student proof. We found significant differences both in self-perceived health (t = −4.13; p < 0.01) and in psychological capital (t = 4.10, p < 0.01) between Time 1 and Time 2, which resulted in a confirmation of self-perceived health and psychological capital deterioration during home confinement. In a more detailed analysis, the items that contribute the most to a worsening of self-perceived health are: item 2 “has felt overwhelmed and stressed?” (t = −2.37, p = 0.02), item 10 “Have you been able to enjoy your daily activities?” (t = −3.70, p < 0.01), and item 12 “do you feel reasonably happy?” (t = −5.67, p < 0.01). Similarly, in psychological capital, significant differences were found in the four psychological capital dimensions: optimism (t = 2.88, p < 0.01), resilience (t = 3.47, p < 0.01), hope (t = 2.33, p < 0.05), and self-efficacy (4.37, p < 0.01).Thirdly, we examined Time 1 predictors with the ability to commit to self-perceived health and psychological capital worsening in Time 2. To accomplish this goal, we carried out two hierarchical regression analyses (one for self-perceived health and another for psychological capital as output variables). We decided to introduce as input variables the three burnout dimensions (emotional exhaustion, cynicism, and self-efficacy) and age as a control variable, following the recommendations of previous health (Pinquart, 2001; Robert, 1999) and psychological capital research (Baron, Franklin & Hmieleski, 2013).The results show that age and emotional exhaustion explain 27% of self-perceived health total variance in Time 2 (see Table 2). While, age and self-efficacy are responsible for explaining 24% psychological capital variance in Time 2 (see Table 3).The uncommon scenario caused by COVID-19 has meant a great change in the way of life of the population. It is not surprising the negative impact that such a situation can have on mental health in the short and long term. The present study examined the influence of the confinement situation caused by COVID-19 on the levels of self-perceived health and psychological capital in a sample of workers, as well as to analyze whether exposure to burnout before social isolation would help to explain the levels of health and psychological capital behind it.Interestingly, the findings of this study provide support for the three proposed hypotheses: the first one stated that the confinement situation caused by the spread of the COVID-19 virus would have a negative influence on the perception of health; the second argued that the confinement situation caused by the spread of the COVID-19 virus would have a negative influence on psychological capital; finally, the third consisted of analyzing whether workers exposed to burnout before the confinement caused by COVID-19, would be more likely to enter a cycle of losses that would lead to a loss of their levels of self-perceived health and psychological capital.Regarding the first hypothesis, our findings indicate a significant worsening of employees’ health perception during mandatory confinement in Spain. These results strengthen the research carried out about the psychological impact of extreme emergencies on health [44], as well as those that have specifically focused on the analysis of health and wellbeing of the general population during COVID-19 social isolation measures [3,4,5,6].Respecting the second hypothesis, our results have also revealed a significant reduction in employees’ psychological capital levels during mandatory confinement. One possible explanation could be the idea of resource depletion framed in Hobfoll’s Conservation of Resources Theory [9]. According to previous research [20,45] the continued use of resources, such as psychological capital, in highly stressful situations, as could be the current one, can result in resources deterioration. In this regard, future research could try to elucidate whether the psychological capital decline is because perhaps its components (optimism, resilience, hope, and self-efficacy) are not the most appropriate to deal with this kind of substantial stressors, or that perhaps the stressor is so overwhelming that, although the resources are adequate, their levels are insufficient to cope with it.Concerning the third hypothesis, our findings revealed, on the one hand, how age becomes a total variance predictor of both perceived health and psychological capital levels. What this seems to indicate is certain age-related changes affecting personal resources reserves would help to explain variations in physical health and psychological well-being [46]. To illustrate the relationship between psychological resources and age, one of the first studies carried out on this subject showed that as people aged, their resources were adapted to the demands they had to face [46]. Furthermore, from a cognitive perspective, it is considered that, as people get older, a greater need appears to allocate unused resources to compensate for the large number of resource losses that have occurred over the years. Therefore, although adults’ resource loss is practically inevitable, it is crucial to conserve and successfully mobilize the greatest number of resources to preserve health to a greater extent [45,46].On the other hand, regarding burnout dimensions when explaining the variance of health and psychological capital levels, there is a significant and negative association between employees’ burnout in Time 1 with health and psychological capital in Time 2. These results corroborate previous research that has repeatedly shown the adverse burnout consequences on worker’s health [23,25,26], as well as those that have related burnout negatively and significantly to psychological capital [5,29,30,32]. It may be speculated that this is probably so because, applying the idea of loss resources cycle [9] together with the self-weakening concept (Bakker & Demerouti, 2017), those employees who perceive work-related stress are more likely to have self-debilitating behaviors, which will lead to higher levels of job demands and will have fewer resources to deal with these stressors, thus increasing their stress levels [36,37]. Hence, not being able to mobilize adequate resources to face the new situation, would significantly affect their health levels [9]. This argument could be illustrated with a situation in which a worker has invested his psychological resources (in this case, psychological capital) to cope with exposure to burnout before confinement so that when he has to face another serious stressor (in this case, mandatory social isolation measures) would not have sufficient levels of resources to deal with it.Given the previous explanation, when analyzing our results in detail, it could be observed, first, that emotional exhaustion is the only burnout dimension capable of explaining the variance of health. In this regard, there is a consensus when it comes to affirming that emotional exhaustion is the central core of burnout, and as such the variable that would have the most influence on people’s health levels [47]. Traditionally, it has been thought that burnout (physical or mental) is a legitimate label that is used for problems that can occur both inside and outside the work environment since in any context people can feel exhausted. However, the other two remaining dimensions of burnout, cynicism, and professional inefficacy, would be reduced to the work context. Perhaps, in this case, emotional exhaustion is the only variable that contributes to explain the variance in health perception because, in addition to the argument exposed above, we are facing a non-work stressor (COVID-19 mandatory confinement). Second, the lack of professional efficacy is the only dimension of burnout that explains part of the total variance of psychological capital. If we understand job self-efficacy as a resource, following Ross and Mirowsky [48], when resources substitute for each other, the presence of one makes the absence of another less damaging. That is, the effect of having a specific resource is greater for those who have fewer alternative resources. In this case, workers with lower levels of psychological capital (alternative resource) may be more affected by the loss of a specific resource such as professional efficacy. Besides, following Youssef-Morgan and Luthans [49], and taking into account that one of the sources of self-efficacy (one of the components of psychological capital) is the experience of success in past situations, since one of the characteristics of burnout is the perception of professional ineffectiveness and therefore, the repeated feeling of failure when facing work tasks, this could prevent the necessary process for the attention, interpretation, and retention of positive and constructive memories, characteristic of psychological capital.Firstly, regarding the theoretical implications, the results of this study help to reinforce the application of the Conservation of Resources Theory for studying employee’s burnout and its consequent coping strategies. Secondly, two possible practical approaches emerge from this study: on the one hand, taking into account the substantial change in working conditions caused by COVID-19 and its influence on workers’ health, a jobs psychosocial risks reevaluation is needed both in face-to-face and teleworking modality. On the other hand, from the point of view of human resources management, it is necessary to build up employees’ personal psychological resources through training interventions, as well as to provide them with additional resources, such as social support from colleagues and supervisors.Several limitations of the present study should be acknowledged. First, the information was obtained through self-report questionnaires, which can produce response biases, exacerbating the common variance, and artificially increasing the correlations between variables [50].Second, the presence of social desirability in the responses to the questionnaire could be another limitation. To try to detect the effects of this phenomenon on the results, it would be beneficial to include additional measurement scales that estimate its presence in subsequent studies [51]. An example is the Marlowe-Crowne Social Desirability Scale (MCSDS) [52].Further research is needed to focus on the prevalence of the disease in different settings. For example, future studies could examine if our results could differ according to the essential and nonessential workers distinction, between different labor market strata or even between labor layers.Overall, the results of this study have shown a significant reduction in self-perceived health and psychological capital between the two periods analyzed (before and during the COVID-19 mandatory confinement). Specifically, burnout together with age act as predictor variables in both health and psychological capital variance between the two times included in this study. However, not all dimensions of burnout contribute to the same extent. Thus, regarding health effects, emotional exhaustion explains 27% of the total variance, while, for psychological capital, the dimension that contributes the most is job self-efficacy, explaining 24% of its variance.Conceptualization, M.M.F.-V. and M.G.-I.; methodology, M.M.d.P.; software, M.M.F.-V. and M.M.d.P.; formal analysis, M.I.S.S.; investigation, M.M.F.-V. and M.M.d.P.; resources, M.I.S.S.; writing—original draft preparation, M.M.F.-V.; writing—review and editing, M.M.F.-V.; supervision, M.I.S.S. and M.G.-I. 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 Murcia.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 restrictions.The authors declare no conflict of interest.Means, standard deviations, and correlations between the study variables.T1: Time 1; T2: Time 2. * p < 0.05 (bilateral); ** p < 0.001 (bilateral). Note: Regarding self-perceived health, high scores indicate a lower level.Hierarchical regression analysis with self-perceived health as output variable.* p < 0.05, ** p < 0.01.Hierarchical regression analysis with psychological capital as output variable.* p < 0.05, ** p < 0.01.Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.