Clinical and school-based intervention strategies for youth obesity prevention: A systematic review

Introduction In the last couple of decades, numerous intervention strategies (ISs) have been formulated in school/community or clinical sectors using physical activity (PA) in order to prevent youth obesity because they have been highly effective in addressing this issue. These two sectors have revealed some interesting information in terms of efficient results and best practice mechanisms, but comparisons between them to learn about their functioning have been rare. Methods Therefore, the aim of this systematic review was to analyze and synthesize PA ISs from school/community or clinical domains for the period 2013-2017, in French or English, targeting youths aged 5-19 years old through primary, secondary, and tertiary prevention. Results In total, 68 full articles were reserved for data extraction and synthesis and 617 were excluded because they did not meet eligibility criteria (61 of 68 were kept for the final analysis). The results identified a number of differences between the studies of the various IS sectors and also a third type of IS, mixed sector. Mixed ISs (clinical and school-community) have a special advantage because they can benefit from the strengths of both school/community-based and clinical-based ISs. Mixed ISs showed the most promising results. This review also highlighted the differences between sectors and their ISs in terms of intervention teams, prevention objectives, duration, materials, and efficiency. Conclusion Future studies should focus on establishing a prevention program in a given geographical area involving all stakeholders with their respective skills/knowledge, in the area of decision-making and in the development of ISs, to ensure that the program is the most efficient and best adapted to its environment.


Introduction
Nowadays, it is well known that obesity is a chronic pathology with multifactorial origins, defined by the World Health Organization (WHO) as an abnormal or excessive fat accumulation that presents a risk to health (e.g., vascular, endocrine) (1,2). A study reported that in the last four decades, the number of young obese children (aged [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] multiplied by 10 worldwide (3). Another study reported that the rate of youth obesity prevalence increased by 47% in the last three decades (4). To help control this disease condition, the WHO made some recommendations [e.g., physical activity (PA); nutrition] (5,6) to be followed, in order to adopt a healthy lifestyle and reduce health problems and the risk of obesity.
In parallel, it is also known that PA has many health benefits (physical, mental, and social, among others) that can help reduce obesity and maintain a certain weight when combined with nutrition (7). Moreover, the adoption of a healthy lifestyle during childhood or adolescence tends to spill over into adulthood. Nevertheless, 80% of youths aged 13-15 do not follow those PA recommendations (8) and tend to lead a sedentary lifestyle (9). This sedentariness and inactivity could lead to overweight or obesity (10). To address this issue and curb this tendency, intervention strategies (ISs) and programs integrating PA have been developed and implemented in different settings to prevent overweight and obesity (11). According to Gadais (11), ISs are initiatives and programs with thematic content and events that directly or indirectly aim to facilitate people to adopt healthy lifestyles for the benefit of their immediate and future health. When we look closely at the literature on obesity prevention (12), a few different settings emerge, and two of them have been widely studied (13,14): the clinical setting on how to manage childhood obesity and the school/community setting on how to deal with obesity prevention and the role of the school in such prevention. It is from this perspective that we decided to focus our work on these two promising intervention sectors.
As we suggested earlier and to quote Lydecker et al. (15), "prevention assumes that individuals have some degree of susceptibility to obesity and would benefit from medical and psychosocial interventions to counter that susceptibility" (15). If the degree of susceptibility to obesity varies from one individual to another, prevention must also take place at different levels: primary, secondary, and tertiary. Primary prevention targets every individual without any distinction, for example, advertisements on television that invite people to be active and eat better. Secondary prevention targets subsamples of the population: people at risk of becoming overweight or obese, for example, children in the upper BMI range or who engage in very little physical activity. Tertiary prevention targets specific individuals who are already overweight or obese with complications, in this case, interventions that aim to help people obtain sufficient weight loss to reduce comorbidities (15)(16)(17). These three types of interventions are generally implemented in two major sectors: clinical or school and community sectors.
A clinical setting is a place where people are treated (e.g., hospital, health center). In this context, clinical ISs seem to contain better financial resources (e.g., exergaming) (18) and human resources/expertise to act as a source of quality and reliable information (19). Clinical ISs have also shown good results in the fight against childhood obesity (20-22), making it an important contributor in the management of obesity. Nevertheless, not all studies show good results, as prevention does not involve only treatment, which is mostly the last step of prevention, thus making the ISs of the other sector also useful.
The school and community sector can be seen as a place dedicated to learning where children develop their knowledge and skills (e.g., physical, social, cognitive skills). Many authors agree that school is a privileged place for the prevention of obesity (23-25). Indeed, children and adolescents spend most of their time at school and it is "possible to globally reach the population of interest without stigmatizing or discriminating and without being primarily dependent on families" (26). According to these authors, the school/community sector could assume an important role to promote positive change in children's lifestyles, in order to make them adopt a healthy way of life (11,27). Yet, some studies have demonstrated that school-based obesity prevention interventions with children have produced limited efficacy (28,29), generally lacking in financial or human resources, among others.
According to the literature, school/community-based ISs and clinical ISs seem to be different because they do not employ the intervention on the same level of childhood obesity prevention. Interestingly, both seem to show promising results in preventing obesity. Therefore, a question arises: Could it be possible to consider a global prevention strategy (primary, secondary, and tertiary levels) to reduce youth obesity prevalence and incidence in the coming years by integrating the best practices from one sector into another? We, therefore, sought to know if there were relevant elements in the ISs from these two sectors that would help formulate effective strategies for the prevention of obesity among young people through mutual enrichment.
The aims of this study were to (1) prepare an extensive inventory on the recent literature regarding programs and ISs that aimed at preventing youth obesity, from clinical or school-community perspectives; (2) extract information in order to identify the mechanisms that make programs effective in a clinical or school/community sector; (3) propose some recommendations from the point of view of both sectors (clinical and school/community) and improve the current ISs for future studies.

Methods
To conduct this systematic review, we followed the six steps of the PRISMA (30) for preparing a flow chart ( Figure 1).

Phase 1: Identification of studies
In the first phase, an exhaustive search through computerized databases was performed to identify scientific publications targeting youth obesity prevention. Particularly, we used specific search equations to conduct the first overall research. Relevant articles were identified by means of a computerized search through three databases (i.e., SPORTDiscus, PubMed, and ScienceDirect) with different combinations of keywords (e.g., program; obese; physical activity, child) (Appendix A) and Equations/Meshterms (Appendix B).

Phase 2: Screening
In phase two, 40 duplicates were eliminated and 685 papers were identified. We started looking for non-systematic reviews and prioritized peer-reviewed articles. Some reports from credible organizations such as government agencies, international agencies, or academic centers were also included in our research (e.g., WHO), but none of them were finally considered.

Phase 3: Eligibility
The selection of the abstract was done on the basis of inclusion criteria. To dispel any doubts on this process, the third author was consulted. Six inclusion criteria were applied for article eligibility: (1) articles published between 2013 and 2017, (2) those written in French or in English, (3) full text available online, (4) youths 5-19 years' old targeted by studies, (5) PA as a part of ISs, (6) ISs that targeted obesity prevention and not only comorbidities or physiological risk factors (e.g., diabetes). In total, 68 full articles were reserved for data extraction and synthesis and 617 were excluded because they did not meet the eligibility criteria.

Phase 4: Inclusion
The fourth phase of the literature search involved obtaining copies of the articles previously identified in phase 3. After collecting the articles, the overall sample was reassessed. A complete reading was made in order to eliminate non-relevant articles (according to inclusion criteria). Finally, 61 articles were retained for data extraction by meeting all inclusion criteria and focusing on the topic of this study.

Phase 5: Data extraction
A data sheet was used to extract information concerning the date of the study; intervention sector (school/community or clinical); design (collaborative ISs or not); variables studied; data collection (equipment, hardware); participants (non-obese, overweight, obese, age); synthesized results; synthesized ISs. The first and the second author performed data extraction. To dispel any doubts on this process, data synthesis tables were discussed until agreement was reached on the presentation form and what should be extracted. The completion of phase 5 marked the start of the content analysis of the 61 articles.

Phase 6: Data analysis
To identify the differences, the results will be presented in five different categories (i.e., intervention team, prevention objective, duration, material, and efficiency). As previously mentioned, the intervention sectors (clinical and school/community) are different from each other but show interesting results (10)(11)(12). Therefore, it is interesting to see who conducted the intervention, its purpose, the means available, how long the ISs lasted, and the results.
The first author carried out the initial selection in the literature search on the basis of the abstract and title (n = 725). He also performed the initial analysis of the data and wrote the first draft of the manuscript. The second and third authors helped with data analysis and result presentation. They also revised the entire manuscript.

Prevention objective
Of all the studies aimed at preventing childhood obesity through school/community ISs, nine dealt with secondary or tertiary obesity prevention, since they were only interested in young people who were already overweight or obese (18,25,37,38,47,48,51,59,60). For example, Larsen et al. (48) selected "One hundred fifteen 11-13 year-old children with overweight and obesity" and Wright et al. (25) mentioned, "Students (n = 251) were English or Spanish speaking, had a BMI > the 85th percentile". The majority of the studies (23, 24, 27, 31-36, 39-46, 49, 50, 52-58, 61, 62) directed their attention toward a relatively primary prevention of obesity and did not necessarily target an overweight or obese population but rather an entire population. For example, Lau et al. (49) mentioned, "the average BMI was 17, which was within a healthy range." These studies tended to include children "at risk" as their participants because of the absence of selection criteria. For example, Smith et al.'s (57) intervention group included 2 underweight children, 110 normal weight, 39 overweight, and 30 obese.

Material
Studies from the school/community sector did not automatically consider BMI with age and sex in their anthropometric measurement (e.g., BMI z-score, BMI percentile). Two studies did not consider anthropometric measurement in their outcomes (40,46) and two others considered only BMI (not for age and sex) (49,59). Of the 34 studies with anthropometric measurements, 23 were based on a measurement related to BMI or abdominal circumferences (e.g., BMI score, waist circumference). Other studies (indicating anthropometric measurements) systematically considered at least a second objective measure such as dual energy x-ray absorptiometry (DXA) (one study), skinfold thickness (three studies), or impedance (seven studies). Ning et al. (37) mentioned that "body composition was assessed by bioelectrical impedance analysis (…) body composition was also estimated at baseline and 6 months using dual x-ray absorptiometry" and Johnston et al. (47) mentioned that "body composition was assessed using triceps skinfold thickness." For studies that assessed PA measurement (32/36), 20 used one objective measurement (not a self-reported). Among these, the most common instruments were accelerometer (9 studies), pedometer (5 studies), and various fitness tests (13 studies) (e.g., Fitnessgram). For example, Larsen et al. (48) used a "progressive bicycle ergometer protocol (…) Actigraph GT3X + for ten consecutive days" and Grao-Cruces et al. (32) reported that "a pedometer was used for evaluation and follow-up purposes." Finally, regarding nutrition, 19 studies collected data. Most of them were carried out through three types of instruments: survey (15 studies), recall/diary (6 studies), and interview (1 study). For example, Ning et al. (37) mentioned that it was "assessed by a 48-hr diet recall."

Efficacy
In the school/community sector, the major objective tended to be the prevention of obesity, and interestingly, "lose weight" was not the first goal.

Prevention objective
None of the studies targeted the primary prevention of obesity and 18 of them worked on secondary or tertiary prevention of obesity, because they only targeted participants with a BMI >85th percentile.  (19) used a BMI between the 75th and the 95th percentiles as an inclusion criterion. Nevertheless, participants were judged at risk of weight gain due to their BMI, based on their last medical consultation.

Efficacy
In the context of secondary or tertiary prevention of obesity, one of the main objectives remained BMI decrease and fat loss in favor of lean mass. Out of 19, 14 (73.68%) studies showed significant effects on BMI or participant body fat (19, 63-65, 68, 71-75, 77-80). For example, Marild et al. (71) "reported a significant reduction in BMI and BMI-SD in the Nurse-Dietician-Physiotherapist managed treatment compared to the control group with obesity (p = 0.0007 and p = 0.0005 respectively)." Nevertheless, some studies showed significant effects only on some of their participants. For example, Martín-García et al. (72) "found that, in the intervention group, boys decreased their whole-body fat mass (p < 0.04) and reduced their percentage of body fat (p < 0.001); moreover, boys' body lean mass increased significantly (p = 0.003)." Of the five other studies that did not have a direct effect on BMI, four had at least a significant influence on physical, psychological, or nutritional health factors (66,69,70,76). Furthermore, in these studies, the intervention group was compared with a control group performing a "less advanced" intervention. Serra-Paya et al. (76) mentioned that "the intervention group received organized physical activity sessions, theoretical and practical sessions for parents, behavior counselling for children and parents, 3 weekend activities organized outside the family for children; unlike the counselling group that received only the behavior counselling sessions." Only one article showed no effects (67).

Intervention team
Of the six studies we identified ( Table 3) and that were carried out in a "mixed" setting (school/community and clinical), all of them used at least two types of contributors, with one (or more)

Prevention objective
All these studies were part of a secondary or a tertiary prevention of obesity, because they were interested only in children who were already overweight or obese. Four of them were based on data on overweight and obesity for age and sex of CDC (BMI > 85th percentile). For example, Morano et al. (83) selected participants "with a BMI ≥ 85th percentile for age and sex according to the CDC growth reference" and Sanders et al.  (84) selected "overweight or obese based on the CDC growth chart." The two remaining studies were based on different data from the WHO: in Maatoug et al. (82), "Z-score were derived using the world health organization references" or from English references; in Kokkvoll et al. (85), "≥98th percentile according to the UK references." None of these studies were directly concerned with the primary prevention of the pathology.

Duration
With regard to these mixed studies, two had a duration above or equal to 1 year (82, 85)

Material
Anthropometric measurements, in these mixed studies, systematically considered BMI for age and sex (e.g., BMI z-score, BMI percentile). Nevertheless, it was not the only measurement, because in three studies, this was combined with at least one other measurement of body composition [WC (three studies), skinfold thickness (two studies), impedance (one study)]. For example, Kokkvoll et al. (85) used "bioelectrical impedance," in Morano et al. (83), "Skinfold thickness was determined (…) with a skinfold caliper," and in Rito et al. (81), "waist circumference was obtained for every child." With regard to PA, three studies clearly used an objective measurement coupled with a second selfreported one. This helped avoid over/underestimating the results. Sanders et al. (84) mentioned that "preand post-intervention fitness tests were administered to participants (…) program participants and their parents completed a physical activity and nutrition behavior questionnaire." The other studies were based only on self-reported data [Maatoug et al. (82) mentioned that participants "responded to a 24 h food and physical activity recall questionnaire"] or no PA measurement was done in them (85). Three tools were mainly used to measure nutrition: questionnaire (two studies), 24 h recall (two studies), and a 7-day dietary diary (one study). For example, Morano et al. (83) reported, "dietary habits were assessed with a 7-day food diary." Only in one study, nutritional measurement (85) was not performed.

Efficacy
As mentioned previously, all mixed studies focused on secondary or tertiary prevention of obesity. One of the major objectives was therefore to influence downward BMI and weight in order to reduce the fat mass of the participants. Of the six programs, five had direct effects on BMI (81-85). One article did not show significant BMI reduction, even though it indicated a tendency to slow its growth. Rieder et al. (86) mentioned that percentile BMI measurements taken before and after the intervention indicated a general upward trend (p = 0.0003). Nevertheless, during the intervention period, the slope of the BMI percentile showed a downward trend (p = −0.0001). Moreover, a comparison of the results of the preintervention phase and the intervention phase showed significant variations (p = 0.003). "For intervals T12 to T0 vs. T0 to T9, there were significant decreases in rates of gain in BMI (0.13 vs. 0.04, p < 0.01, BMI percentile [0.0002 vs. −0.0001, p < 0.01]." Each of these studies also presented at least one positive variation on one of the various health, physical, psychological, or nutritional factors. For example, Maatoug et al. (82) showed positive effects on PA, p-value (pre-post) = 0.001, and reduction of caloric intake; p-value (pre-post) < 0.001; Rito et al. (81) mentioned "vigorous physical activity (day/week), CI 95% 0.1-0.5, p-value = 0.008" and Morano et al. (83) showed that "Actual (p < 0.001) and perceived (p < 0.03) physical abilities, physical activity enjoyment (p = 0.03), and psychosocial HRQoL (p < 0.05) also improved from pre-to post-intervention."

Summary of the findings
This study identified a number of differences between the studies of the various IS settings ( Table 4); these differences could be grouped into five elements (intervention team, prevention objective, duration, material, and results).
Intervention team: The intervention team in the school/ community was mostly composed of a single stakeholder that was often the teacher (54.5% of cases). In the clinical and mixed sectors, the ISs largely depended on a multidisciplinary team with various members specialized in health. In both sectors, stakeholders contributed to the success of the ISs as key actors.
Prevention objective: The school/community ISs mainly targeted primary prevention because there was no selection of participants, while participants in the other sector were chosen by targeted criteria such as overweight or obese (according to their BMI for age and sex), and the same was the case for mixed studies. Duration: One interesting point in the school/community setting was that it allowed for a relatively long intervention duration, and ISs aimed at preventing obesity needed time to embed and develop before being evaluated (60). On the other hand, clinical and mixed ISs tended to last for a shorter period of time. More than threefourths of the studies done in the clinical setting and two-third in the mixed setting lasted less than a year.

Findings
: What can be understood and learned?

Material
Clinical and mixed ISs tended to use more objective instruments, requiring more skills and knowledge. This allowed them to associate and combine certain measurements to achieve more accurate results and not over/underestimate their results.

Efficacy
Many authors agreed (23-25) that school was a privileged place for prevention. Our results seem to confirm this tendency, because school/community ISs showed significant and promising results both on anthropometric measurements relative to obesity (72%) and on health-related factors (90%). The clinical setting was also a beneficial location for the treatment of obesity (secondary/tertiary preventions). This setting was seen as a source of quality and reliable information (17) and it provided important results on both obesity (74% of the studies) and health-related factors (83% of the studies) (87).

Prospects
In light of these findings, it is necessary that mixed studies should be prioritized, with a combination of school/communitybased and clinical-based strengths. Indeed, our study found that mixed ISs provided the most promising results; 83% of the studies showed a positive influence on obesity and 100% on health-related factors. Nevertheless, those we considered sought to apply a relatively clinical model to the school/community setting (82) but did not participate in an exchange relationship, and therefore, the strengths of school/community-based ISs were "left behind." To enhance global obesity prevention and in line with health recommendations to prevent childhood obesity, Table 5) proposes recommendations for future studies to be more effective.
Clinical: For the purposes of synthesis (Figure 2), we recommend the implementation of a transparent local council involving the entire local community (e.g., school children, representatives, clinical specialists, stakeholders, parents, associative representatives) (25) responsible for the development, improvement, and implementation of prevention programs at the local level. The most important point here is that each intervention sector should have its own prerogatives. Nevertheless, to achieve effective obesity prevention, the different settings need to function in a more transparent manner without ignoring the three different aspects of prevention (primary secondary = school/community; secondary tertiary = clinical). Another interesting aspect is the use of new technologies in prevention. On this point, further studies are needed to evaluate the potential added value of technological tools in obesity prevention.

Limitations
Many studies use BMI as a measurement variable or other variables related to it. However, although it is easy to use, BMI alone is not a representative indicator of the benefits of a program. Morano et al. (83) showed that the use of multiple body composition measurements provide better indications of changes in body fat, which is more representative of expected changes. Furthermore, weight-related measurements can skew the results (72). In PA-oriented programs, which lead to positive body composition change (e.g., lean mass gain and body fat loss), the participant weight can increase (lean mass is heavier), which can induce an increase in BMI measurements. This lack of precision relative to these measurements can also distort the results of the studies that use only BMI as the measurement variable or variables relative to the weight of the participants (22).
A second limitation could be the lack of information about programs in some studies to assess the quality of the study. This sometimes makes it difficult to identify the equipment used, the staff involved, and their skills/knowledge. The duration of the program is another limitation. Indeed, all the identified studies do not last the same amount of time (e.g., more than 1 year; less than 6 months; a few weeks). Furthermore, they do not use the same evaluation time (e.g., pre-post; pre-post + follow-up). It is, therefore, difficult to evaluate the effectiveness of one IS in relation to another over a short period of time. Moreover, to assess the efficacy of ISs, more time is required to embed them (60). The number of participants in each study is also highly variable and therefore can make a generalization or a comparison with other studies impossible. The last limitation pertains to the number of studies selected for each setting. Indeed, the number of studies being relatively low in the mixed setting can lead to an over/ underestimation of the results. Nevertheless, this work seems to yield promising results, and future studies must continue to move to mixed setting, to nested ISs.

Conclusion
The main objective of this study was to propose a first combination and comparison of obesity prevention intervention programs from the clinical and the school or community sectors. However, our review showed that comparisons are difficult to make since the standards and units used for measurements are different and vary according to the protocols and areas of application. Nevertheless, we believe that this study offers an initial proposal for bridging the gap between the clinical and the school/community sectors, the two most promising sectors in terms of outcomes for obesity prevention in youth in particular.
Future studies should focus on establishing a prevention program in a given geographical area (e.g., town, county), involving all stakeholders with their respective skills/knowledge, in the decision-making process and in the development of ISs (e.g., parent association, professors, doctors, local representatives, sports association), so that it becomes the most efficient and best adapted to its environment. Although this study focused on physical activity interventions, it would be relevant to also look at nutrition interventions, since nutrition is a major theme in obesity prevention. The main objective of this study was to propose a first combination of obesity prevention intervention programs with the clinical and the school or community sectors. Our review showed that comparisons are difficult since the standards and units used to measure are different and vary according to the protocols and areas of application. However, we believe that this study offers an initial proposal for bridging the gap between the clinical and school/community sectors, the two Modelling of physical activity ISs.
Caron et al. 10.3389/fspor.2022.906857 Frontiers in Sports and Active Living most promising sectors in terms of outcomes for obesity prevention in youth in particular.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

Author contributions
Conceptualization was done by TC and TG; the methodology of the study was prepared by TC and TG; software was provided by PB; validation was done by TC, PB, and TG; formal analysis was done by TC; investigation was performed by TC; resources were provided by TG; data curation was done by PB and TG; writing-original draft preparation-was done by TC; writingreview and editing-was done by PB and TG; supervision was carried out by TG; project administration was looked after by TC. All authors contributed to the article and approved the submitted version.