Effective strategies for childhood obesity prevention via school based, family involved interventions: a critical review for the development of the Feel4Diabetes-study school based component

Background Although there are many interventions targeting childhood obesity prevention, only few have demonstrated positive results. The current review aimed to gather and evaluate available school-based intervention studies with family involvement targeting dietary, physical activity and sedentary behaviors among primary schoolchildren and their families, in order to identify the most effective strategies. Methods Studies published between 2000 and January 2015 were retrieved from scientific electronic databases and grey literature. The databases used included MEDLINE/PubMed, Web-of-Science, CINAHL and Scopus. Included studies had to be experimental controlled studies and had duration over 1 school year, had family involvement, combined PA and dietary behaviors and were implemented in school setting. A complementary search was executed to update the review to cover the period from February 2015 to January 2019. Results From the studies examined (n = 425), 27 intervention programs (33 publications) fulfilled the inclusion criteria. Among these, 15 presented significant effect on weight status and/ or overweight/ obesity or clinical indices, 3 presented significant effect on most energy balance-related behaviors (EBRBs) while 9 presented significant effect on some/few EBRBs or determinants. Strategies implemented in effective interventions were: teachers acting as role-models and being actively involved in the delivery of the intervention, school policies supporting the availability of healthy food and beverage choices and limiting unhealthy snacks, changes in the schoolyard, in the recess rules and in the physical education classes to increase physical activity, and involving parents in the intervention via assignments, meetings, informative material and encouraging them to improve the home environment. Use of incentives for children, social marketing techniques, collaboration with local stakeholders were found to increase effectiveness. Programs that focused only on educational sessions and material for parents, without promoting relevant environmental and policy changes, were found to be less effective. Cultural adaptations have been suggested to increase the intervention’s acceptance in specific or vulnerable population groups. Conclusions Several effective strategies were identified in the reviewed programs. Outcomes of the current review were taken into account in developing the Feel4Diabetes-intervention and summed up as recommendations in the current work in order to facilitate other researchers designing similar childhood obesity prevention initiatives.


(Continued from previous page)
Results: From the studies examined (n = 425), 27 intervention programs (33 publications) fulfilled the inclusion criteria. Among these, 15 presented significant effect on weight status and/ or overweight/ obesity or clinical indices, 3 presented significant effect on most energy balance-related behaviors (EBRBs) while 9 presented significant effect on some/few EBRBs or determinants. Strategies implemented in effective interventions were: teachers acting as role-models and being actively involved in the delivery of the intervention, school policies supporting the availability of healthy food and beverage choices and limiting unhealthy snacks, changes in the schoolyard, in the recess rules and in the physical education classes to increase physical activity, and involving parents in the intervention via assignments, meetings, informative material and encouraging them to improve the home environment. Use of incentives for children, social marketing techniques, collaboration with local stakeholders were found to increase effectiveness. Programs that focused only on educational sessions and material for parents, without promoting relevant environmental and policy changes, were found to be less effective. Cultural adaptations have been suggested to increase the intervention's acceptance in specific or vulnerable population groups.
Conclusions: Several effective strategies were identified in the reviewed programs. Outcomes of the current review were taken into account in developing the Feel4Diabetes-intervention and summed up as recommendations in the current work in order to facilitate other researchers designing similar childhood obesity prevention initiatives.
Keywords: Obesity prevention, Type 2 diabetes prevention, Primary school children, Families, School based intervention Background Childhood obesity is a growing world-wide health problem. It is estimated that about 170 million children are currently with overweight globally [1]. In US children, obesity prevalence has increased from 6% in 1980 to 17% in 2008 [2] and to 18.5% in 2016 [3]. In Europe, the trend of childhood obesity is following the same pattern, with its prevalence increasing across most European countries [4], especially in low-to-middle-income countries (LMICs) and in vulnerable groups [5,6].
Being overweight or obese has serious health consequences, especially for children. Having a high body mass index (BMI) is a major risk factor for diseases such as cardiovascular disease, type 2 diabetes and cancer later in life [7,8]. These diseases, also known as noncommunicable diseases (NCDs), can cause premature mortality as well as long-term morbidity. Due to the significant increase in the prevalence of obesity and the serious public health consequences, obesity is considered one of the most important public health challenges of the twenty-first century [9,10].
Early childhood is considered to provide a unique opportunity to establish lifestyle behaviors such as healthy eating habits, physical activity and limited sedentary time that will promote health and minimize the risk of obesity. This is further supported by the fact that, as a results of these behaviours, childhood obesity tracks into adulthood with overweight preschool children being more likely to become overweight adults in comparison to their normal-weight peers [11][12][13][14][15]. Thus, interventions promoting such healthy behaviors in childhood provide a key strategy in the prevention of obesity, since it has been shown that treating obesity in adulthood poses more difficulties than changing lifestyle habits earlier on [16].
To date, many programs have been developed in order to prevent obesity in children. The vast majority of these programs use schools for the implementation of interventions [17]. Schools are considered an important setting for intervening in children's obesity related behavior, for a number of reasons: (i) primary school education is compulsory for all children in most countries and reaches all children with different backgrounds (ii) children spend a significant part of their daily life at school, usually consuming one or two meals per day; (iii) schools offer physical education classes and provide opportunities for physical activity during recess; (iv) schools offer a structured environment where interventions can be easily applied/ fit, (v) implementers can reach many children in a relatively short time via schools; (vi) teaching staff can significantly facilitate and contribute to the delivery of the intervention, thus increasing the intervention's sustainability [18]. Despite the aforementioned advantages, the overall impact of school-based interventions is questionable and generalizable recommendations cannot be easily extracted [2].
For the needs of the European multicenter Feel4Diabetes-study the current study examined interventions that have been implemented in the school setting and focused on the promotion of healthy eating and physical activity and the reduction/interruption of sedentary behavior aiming to prevent childhood obesity. The Feel4Diabetes-study was implemented in the overall population in low/middle-income countries (Bulgaria, Hungary), in low socio-economic areas in high-income countries (Belgium, Finland) and in countries under austerity measures (Greece, Spain). More information on this project can be found in the literature [19]. The aim of the current review was to identify the most effective strategies in improving health behaviors and tackling obesity in primary school-aged children with special emphasis given to low socioeconomic status and vulnerable groups that Feel4Diabete-study specifically targeted, taking into account their increased risk for obesity [9,20] and type 2 diabetes [21]. The Feel4Diabetes project was developed using the PRECEDE-PROCEED model as the theoretical basis. The current review served as a part of the PRECEDE phase of this specific theoretical model and its outcomes were utilized in the development of the Feel4Diabetes school-and community-based intervention.

Search strategy
The bibliographic search strategy focused on articles published in peer-reviewed, English language journals, published from 2000 to January 2015, when the Feel4Diabetes-intervention was developed. A complementary search was executed to update the results for the publication to cover the period from February 2015 to January 2019 following the same methodology. The databases used included MEDLINE/PubMed, Web-of-Science, CINAHL and Scopus. Citations in reviews were also used. The PICO key terms used for conducting the literature search are presented in Supplementary  Table 1.

Selection criteria
To be included, studies had to be original experimental controlled studies with school-based interventions aiming to prevent obesity in primary school-aged children from any country, published between 2000 and January 2019, in English. Additional inclusion criteria were applied based on recent literature [18,22]. These criteria were, namely: a) duration of > 1 school year (i.e. at least 6 months), b) family involvement, either by contacting parents via their children with the use of newsletters, etc. or by having meetings with them, c) combination of physical activity and dietary behaviors, d) implementation of the school-based intervention by school personnel (i.e. healthcare professionals working at schools or teachers). These criteria were applied in order to avoid repetition of already proven to be effective strategies and take the existing literature a step further.

Exclusion criteria
Exclusion criteria included interventions implemented in preschools, early childcare programs, adolescents or after-school programs, descriptive or analytic studies, reviews on surgical or pharmaceutical treatments, literature reviews, opinions or editorials, reports published as meeting abstracts only, reviews of causal relationships between obesity and related factors, articles reporting study design and/or process evaluation only as well as papers focused on physiological, molecular or genetic research and papers focused on morbidities (such as kidney disease and diabetes) in which obesity is a comorbidity.

Selection strategy and data extraction procedures
The flowchart of the publication identification process is presented in Fig. 1. Data from the included studies were extracted using a standardized form. Extracted data included: author, year of publication, project name, country, participant data, intervention duration, theory used and strategies implemented as described in the methodology of the relevant publications, outcomes as well as follow-up period if available. In addition, the setting in which the intervention was implemented was also extracted in order to identify strategies that proved to be effective in low-SES areas or in vulnerable groups. A summary table of the selected studies was constructed and the clinical significance of the results was evaluated for each study, to facilitate interpretation of the effectiveness of intervention. The scoring is marked in the summary table as follows: (+++) significant effect on weight status and/ or overweight/ obesity or clinical indices. (++) significant improvement in most important target behaviors. (+) significant improvement in some/few secondary target behaviors. (−) no effect.

Results
The studies included in the current analysis have been summarized in Table 1 outlining the target population, intervention strategies and design, measures, outcomes, and findings relevant to obesity prevention. The interventions in table have been arranged alphabetically based on first author's name but with clustering the publications of same projects together.

General description of findings
Using the aforementioned search strategy 1148 publications were identified. After removing duplicates and a first title screening 425 articles remained. These 425 articles were screened based on their abstract, in order to discard studies without yet published results, those targeting specific populations (e.g. only obese children or patients), preschoolers or adolescents. Of the 122 articles that remained, full-texts were extracted and reviewed based on the inclusion/ exclusion criteria. Finally, 27 intervention programs (33 publications) fulfilled the criteria and were included in the current review.
Most of the included interventions (n = 19) were implemented in high income countries (HICs), however a few of them (n = 8) were specifically focused on vulnerable population groups or low-to-middle income countries (LMICs), while two of the studies implemented in HICs took vulnerable groups under consideration using extra analysis [23]. The sample size varied across the studies, ranging from 108 to 7406 participants.

Outcome and effectiveness of school-based interventions on weight status and energy balance related behaviors
Out of the 27 intervention studies (33 publications) included in the current review, 15 (55% of included studies) presented significant effect on weight status and/ or overweight/ obesity or clinical indices, 3 (11% of included studies) presented significant effect on most target behaviors while 9 (33% of included studies) presented significant effect on some/few target behaviors. Only a few interventions (n = 4) followed-up the participants post-intervention, showing promising longterm effectiveness especially in high-socioeconomic status (SES) groups.

Intervention strategies
Key strategies have been identified from the effective intervention programs. In studies including various intervention strategies and several outcomes, no direct cause-effect link could be identified, however the strategies or the combination of strategies used in successful interventions were found. Regarding teachers' involvement, having the teachers trained by health professionals to deliver the intervention, coordinate school-based activities and promote healthy energy-balance related behaviors (EBRBs) during school hours instead of having researchers or other personnel implementing the intervention has been suggested as an effective strategy. Inclass material (workbooks, posters, manuals) was used to facilitate the process. Moreover, teachers acting as rolemodels and getting actively involved in all targeted EBRBs has been highlighted as another effective strategy.
Regarding school policies, several strategies such as the increased water accessibility, the free provision of           fruits, the availability of only healthy options in the school cafeteria and the removal of vending machines seem to be efficacious strategies. The formation of a wellness council (by school staff) and a written wellness policy were reported as some more drivers of the intervention's effectiveness. The involvement of the family was a prerequisite for the inclusion of studies in the current review. Parents were approached via several methods, i.e. family "homework" assignments, educational/ informative material (newsletters) including healthy recipes and school meetings/ events. In all cases the aim was to encourage changes at the home environment as well in order to further promote the targeted behaviors (e.g. availability of fruit and vegetables) and ensure the continuity of the intervention after school hours.
Several other strategies were also implemented in the childhood obesity prevention interventions that were found to be effective. More specifically, physical activity promotion during recess and the development of an active playground as well as non-competitive, enjoyable activities, promoting whole class participation during physical education (PE) classes and the provision of additional hours to physical activity were shown to be effective strategies. Motivational incentives (e.g. stickers, t-shirts) as well as social marketing techniques targeting children, e.g. slogans, characters, raffle tickets have also been shown to contribute to effectiveness of the interventions. The collaboration with community stakeholders, local community and school policy and media in every aspect of a health promotion project targeting physical activity, healthier diet or both has been also highlighted as an important contributor to the implementation fidelity, and thus effectiveness, of the intervention, also providing the basis for a sustainable approach.
Regarding intervention programs delivered in vulnerable groups and/ or low-SES areas, effective strategies focused on school policy and environmental changes, parental engagement, incentives and interactive activities such as cooking lessons and cultural dance classes, with less emphasis given on the educational part of the intervention. Furthermore, an intervention program particularly targeting Jewish population offered gender separated physical activity classes for pupils, adapting the intervention to the needs of the specific population group thus increasing adherence and effectiveness of the intervention.

Discussion
The current review aimed to gather and evaluate available school-based, family-involved obesity prevention interventions targeting dietary, physical activity and sedentary behaviors among primary schoolchildren and their families. Moreover, it aimed to identify the most effective strategies in improving those EBRBs and tackling obesity in primary school-aged children with special emphasis given on low-SES and vulnerable groups.
Previous reviews on the same field were used as a basis for the current review and the inclusion criteria applied [18,22,24]. In order to avoid repetition of already proven to be effective strategies and in order to extract new meaningful results from the existing literature, the current review goes one step further in the identification of effective strategies by incorporating those already known to be effective as inclusion criteria. From the total number of studies selected 27 intervention studies (33 publications) fulfilled the inclusion criteria and were included in the current analysis.
As previously described, primary schools seem to be the ideal setting for childhood obesity prevention interventions since it offers many opportunities for physical activity promotion and nutrition education and reduction of sedentary behavior through practice, policy, and a supportive environment [18]. This approach requires the active involvement and participation of teachers. Since schools mainly focus on academic achievement, many teachers find it difficult to include extracurricular modules into the already tight schedule. Furthermore, not all teachers are willing to act as champions for a health promotion intervention, especially if their own lifestyle habits do not follow/ agree with the healthy lifestyle messages delivered via the intervention [25]. Strong engagement by the school headmaster/ leadership is crucial in order to achieve engagement of families and teachers [26] and ensure the delivery of the intervention with the highest possible fidelity and thus, effectiveness as well as teachers' behavior as role-models.
Parental influence on children's nutrition and physical activity habits is a well-known determinant of childhood obesity [27]. Based on research indicating that the involvement of parents in addressing nutrition, physical activity and sedentary behavior in children is essential, the current review only included interventions that involved parents [18,22]. Several effective strategies such as including parents via family "homework" assignments, educational/ informative material (newsletters) and school meetings/ events have been used to involve parents in the intervention. Meetings with parents at school seemed to lead to increased effectiveness in more cases in comparison to written material, i.e. newsletters.
Given that vulnerable groups and/ or people living in low-SES areas are more likely to be overweight or obese, have poorer health outcomes, and tend to benefit less from interventions in comparison with their counterparts from more affluent backgrounds, both developed and developing countries should also target low-income, high-risk groups [28]. Tailoring interventions to fit the needs of a specific population group, as done in the study by Benjamins et al. where gender separated physical activity classes were offered for Jewish pupils, has been shown to be imperative for the intervention's acceptability and effectiveness [28]. A thorough situational analysis and formative research should be included in the time planning for the development of interventions, especially when targeting schools in low-SES areas and vulnerable groups [28]. The results of the current review are in agreement with previous work highlighting that intervention programs delivered in vulnerable groups and/ or low-SES areas should focus on school policy and environmental changes and interactive activities with less emphasis on the educational part of the intervention [28,29].
Using a solid theoretical framework as a basis in an intervention's design is vital to its chances of success, as evidence suggests that such interventions are more effective compared to those that are not based on a theory [30]. Various theoretical frameworks have been utilized by the included interventions, while almost half of the projects did not mention the use of a theoretical model, a fact that might have limited their effectiveness. CDC's ecological model-based Coordinated School Health Program which is commonly used in the US was one of the two more popular theories used in the development of health promotion programs in the current review. Its significant effect might derive from the fact that among several targets and in contrast to most frameworks it also aims at improving employee wellness. This characteristic might have made a difference for its popularity and effectiveness. One the other hand, the Social Cognitive Theory is a commonly used basis for interventions design. In the current review, even though it was popular in the current review, its effectiveness varied. Last but not least, integrating policies and/ or environmental changes, such as increased availability and accessibility of healthy choices and removal of unhealthy choices but also the collaboration with community stakeholders, have been shown to be a high impact intervention strategies [31,32], especially when targeting lower SES groups.
Several outcomes have been considered to assess the interventions' effectiveness in various studies. Both behaviors and clinical outcomes where considered in the included studies, however the interventions did not always succeed in having an effect on clinical outcomes. The most commonly used and assessed behavioral outcome was the consumption of fruits and vegetables and many interventions succeeded in increasing their consumption followed by physical activity, which proved to be more difficult for the interventions to significantly improve. On the other hand, BMI was the most commonly used clinical outcome assessed. Several interventions succeeded in improving BMI, while some that did not showed an effect on other important adiposity measures, i.e. waist circumference of skinfold thickness. Still, having a significant effect on clinical outcomes is not always feasible even in well-designed interventions, especially in this age group.

Strengths and limitations
The results of the current work should be interpreted in consideration of the study's strengths and limitations. In the current review studies in languages other than English were omitted, which could limit the generalization of the results in other countries, especially low-to-middle income countries (LMICs). Studies without published outcomes were excluded as it was not possible to evaluate their efficacy. Although the benefits of single health behavior change interventions vs. multiple health behavior change interventions remain unclear [33], the present review only included interventions targeting dietary, physical activity and sedentary behaviors simultaneously. On the other hand, the current review goes beyond the existing literature by applying specific inclusion criteria that have been linked to increased effectiveness and thus, focusing on the most effective programs to extract the best intervention strategies, contributing to the updating of the literature.

Conclusions
School-based interventions are vital in the prevention of the globally rising childhood obesity. Many interventions have shown promising results, which were supported by a number of effective and high-impact strategies. Multiple strategies are used in effective interventions, highlighting the fact that a one-size-fits-all approach is not applicable in childhood obesity prevention intervention programs development and that many different strategies can be effective. However, future school-based obesity prevention interventions should build on already successful intervention strategies while also addressing and integrating culture specific strategies. Including long-term follow-up measurements to assess the efficacy of school-based interventions will facilitate the identification of the most effective strategies in the long-term.
Further studies are needed to elucidate the effectiveness of specific strategies aiming at long duration interventions. Both the development and the implementation methodology of the research, as well as the barriers, challenges and possible facilitators should be encouraged to be thoroughly recorded and published in order to inform the scientific community on the feasibility and sustainability of implementing interventions in real life situations.

Recommendations for future school-based obesity prevention interventions
Apart from the already mentioned recommendations that have been used as inclusion criteria in the present review, several strategies have been proven efficacious, i.e.: The use of a suitable theoretical framework should be considered in the intervention design Interventions should be adapted to the needs of the targeted population group(s) Collaboration with community stakeholders, local policy and media should be targeted in order to increase intervention's acceptability and sustainability School teachers should act as role-models and get actively involved in all targeted behaviors The teaching staff and school personnel should promote the availability and accessibility of healthy snacks and water and monitor the provision only of healthy food options by the school cafeteria/canteen and attend to the removal of unhealthy choices and vending machines, if any Schools should be encouraged to form a wellness council (school staff and/or parents' association) and write a wellness policy Families should be targeted via educational/ informative material (newsletters), family "homework" assignments or school meetings/ events for the children and their families in order to change the home environment to promote the targeted behaviors (e.g. availability of fruit and vegetable, as well as other healthy food items/snacks) School yards should be accessible after school hours and physical activity during recess should be promoted PE-instructor/school teachers should develop and promote non-competitive, enjoyable activities, promoting whole class participation Social marketing techniques, interactive activities and motivational incentives can be utilized for children. In low-SES areas or in migrant population groups, interventions should focus on school policy and environmental changes, parental engagement and interactive team activities, rather than the educational part of the intervention.