Preventing childhood obesity and diabetes: is it time to move out of the school?


  • Joel Gittelsohn,

    Corresponding author
    1. Center for Human Nutrition, Department of International Health, Bloomberg School of Public Health, Baltimore, MD, USA
      Joel Gittelsohn, MS, PhD
      Room 2041A
      Center for Human Nutrition
      Department of International Health
      Bloomberg School of Public Health
      Baltimore MD 21205
      Tel: +1 410 955 3927;
      fax: +1 410 955 0196;
    Search for more papers by this author
  • Mohan B Kumar

    1. Center for Human Nutrition, Department of International Health, Bloomberg School of Public Health, Baltimore, MD, USA
    Search for more papers by this author

Joel Gittelsohn, MS, PhD
Room 2041A
Center for Human Nutrition
Department of International Health
Bloomberg School of Public Health
Baltimore MD 21205
Tel: +1 410 955 3927;
fax: +1 410 955 0196;


Abstract:  Childhood obesity interventions in the USA and Europe have predominantly focused on the school environment for over two decades with mixed or modest success. The focus on school – while intuitive, apparently efficient, and convenient – does not address larger upstream environmental factors, which affect obesity among youth. In this article, we examine potential drawbacks and limitations of previous school-based obesity and diabetes prevention programs. The future of school-based obesity and diabetes interventions and potential strategies for improvement is explored. Increased use and reporting of diversified theoretical frameworks, formative research to inform the interventions, and process evaluations to improve programs are recommended. More importantly, addressing the broader issue of the overall food environment and its impact on children’s diet with intensified involvement of key stakeholders, including families, supermarkets, and corner stores is essential. We discuss the development of healthy eating zones around schools as a potential tool in the fight to reduce childhood obesity.

It is predicted that US life expectancy will decline, for the first time in recent history, as a result of increasing childhood obesity (1). Obesity is a primary risk factor for cardiovascular disease, cancer, and diabetes (2–4). Using Centers for Disease Control-defined cutoffs to describe childhood obesity among US youth between 6 and 19 yr of age, 31% are at risk of overweight and 16% are overweight, with 35% or more of African-American (AA) youth overweight or at risk of overweight (5, 6). Obese children have worse lipid profiles (7, 8) and obesity increases the risk for many chronic conditions (9, 10). Rates of pediatric diabetes are particularly high among native North Americans (11) and other low-income ethnic minority populations in the USA. However, the obesity and diabetes epidemic is also a worldwide phenomenon (12).

The burgeoning epidemic of child obesity and diabetes has drawn the attention of researchers for decades. Most obesity and chronic disease prevention research in children has focused on the school as the primary venue for intervention. This approach would seem to make great sense. The vast majority of children in the USA and Europe are educated in schools for seven or more hours a day, creating a captive audience for intervention and a heightened sense of ‘apparent efficiency’– you can reach many children in a relatively short period of time in schools. Schools would also appear to be a path of least (or lesser) resistance. Providing health education and a healthy environment are part of the mission of most schools and have been legislated at both the state and federal levels. Once key school administration has been convinced to adopt a program, it can be institutionalized and maintained for extended periods.

Despite these apparent strengths, after two or more decades of intensive work within schools for obesity and chronic disease prevention, the impact of school-based studies for obesity prevention has been mixed or relatively modest (13, 14). In order to understand these findings, this paper first summarizes the findings and conclusions of 15 review articles published since 2000 on school-based obesity prevention, examines the available literature on school-based diabetes prevention efforts, and attempts to answer the following questions:

  • (i) In what dimensions have school-based obesity prevention programs achieved their greatest successes and greatest failures?
  • (ii) What is the future of school-based obesity prevention? How can we improve school-based programs as a venue for intervention?
  • (iii) What is the current state of the art in school-based diabetes prevention?
  • (iv) Are we focusing too many resources on chronic disease prevention in schools? Should we move outside schools and into the community and household?
  • (v) What are the most promising ways to move outside schools?


Identification of the relevant literature for school-based obesity prevention reviews began by performing a literature search for the years 2000–2007 in Medline, PubMed, PsycINFO, and the Cochrane Database of Systematic Reviews. Initially, key word searches included ‘childhood obesity’, ‘obesity’, ‘prevention’, ‘children’, ‘intervention’, and ‘schools’. We retained only review articles that considered multiple studies with obesity as the primary outcome and those that examined key dietary and physical activity determinants of obesity. Fifteen peer-reviewed articles were found matching these criteria.

Because of the limited published literature, identification of the relevant literature for school-based diabetes prevention studies utilized the PubMed database for the years 1985–2007. Key word searches included ‘diabetes’, ‘prevention’, and ‘schools’. We retained any study that reported on the outcomes of a school-based program for diabetes prevention. Seven peer-reviewed articles were found matching these criteria.

Summary of review articles of school-based programs for obesity prevention

We identified 12 review articles that have been published since 2000 on school-based obesity prevention programs (Table 1). Most of these articles overlap considerably in terms of the studies included, but differ in terms of criteria for inclusion, outcomes considered, and their conclusions about the state of the art and recommendations for the future. However, some common themes emerge.

Table 1.  Summary of review articles on school-based childhood obesity prevention, 2000–2006
Review articleNumber of studies reviewed and primary outcomesTypes of interventions consideredMain inclusion criteriaSuccess rateMain conclusions/recommendations
  1. BMI, body mass index; RCT, randomized controlled trial; SB, sedentary behavior; WHR, waist to hip ratio.

Campbell et al. (2001) (61)Seven studiesSchool- and community-based1. RCT, Non-RCT with concurrent control group66.7% of long-term studies reported decrease in prevalence of obesity1. Mixed results on effectiveness of obesity prevention interventions.
2. Minimum observation period 3 months
2. Range of interventions addressing multiple obesity casual factors is required.
3. Ages <18 yr
4. All interventions excluding drug or surgical interventions50% of short-term studies reported decrease in obesity prevalence
BMI, triceps skin fold, dietary intake, physical activity levels, WHR, parental support5. Must include one or more of the following primary outcomes: BMI, % body fat, ponderal index, skin fold thickness
6. Report outcome data at baseline and postintervention or baseline and postintervention change
Reilly and McDowell (2003) (62)Four studiesSchool- and family-based (television viewing, nutrition and physical activity components)1. RCT75% of studies reported changes in BMI1. Promising targets for obesity prevention emerging [sedentary behavior (SB)]
2. 12-month follow up
Obesity prevalence and remission, serum cholesterol, % overweight, aerobic capacity, dietary intake, activity2. Generalizability and clinical relevance unclear
Bautista-Castano et al. (2004) (63)Fourteen studiesSchool and community-based1. 0–18 yr14.3% studies reported reduction in overweight indices in both genders; 21.4% in one gender1. Longer-term interventions (>6 months) more effective
2. Published 1993–2003
2. Nutritional education and physical activity components combined with behavior change necessary
 BMI, skin folds and % of body fat mass 3. Deal with the effects on ponderal status 3. Family/parent involvement critical
4. RCT and non-randomized intervention studies with control groups
4. School cafeteria component not decisive in increasing effectiveness
5. Minimum of 12 wk follow-up
5. Improvement in dietary habits seen in most programs with potential implications on chronic disease risk factors
Caballero (2004) (64)Seven studiesSchool- and family-based1. US & Canada locations42.9% of studies reported reduction in BMI1. Lack of carefully controlled obesity prevention interventions for children
2. Aimed at childhood obesity prevention
BMI, % overweight, triceps skin fold, waist circumference, % body fat
2. School-based interventions successful in decreasing energy and fat intake, sedentary activities
Summerbell et al. (2005) (65)Twenty-two studiesSchool-, community- and family-based interventions1. RCT and controlled clinical trials20% of long-term studies reported reduction in overweight status1. Most studies were short term
2. >12 wk duration2. Combined dietary and physical activity programs did not improve BMI
16.7% of short-term studies reported modest changes in overweight status
3. Interventions focusing on dietary or physical activity component produced small BM reductions
4. Attention to appropriateness, duration, design, intensity and duration necessary
5. Short term, behavior change focus unsustainable or ineffective
6. Sustainability and environment change focus necessary with involvement of all stakeholders
Doak et al. (2006) (16)Twenty-five studiesTelevision watching, physical education, nutrition education (school and home)1. Focus on school-aged children (6–19 yr of age)68% of studies reported reduction in BMI or skin folds1. The majority of prevention programs effective
2. Body weight or adiposity measured at baseline and follow-up2. Family component: Mixed results for family involvement (more non-effective programs target diet and activity or outside school activity or physical environment or family level factors, but difference in effectiveness between effective and non-effective programs are small and not statistically significant)
BMI or skin folds3. Must incorporate diet and/or physical activity-related behavior
Flynn et al. (2006) (17)One hundred and forty-seven studiesSchool-, community- or clinic-based1. Top tertile of any of the four predetermined appraisal types1. 73% school-based programs associated with reduction in ‘fatness’1. Only 14.3% of studies incorporated high methodological rigor
1. Indices of obesity (anthropometry, BMI, body fat distribution, growth rates)2. No single program was a model of best practice, but studies present innovative ideas in dynamic settings
 2. Risk factors for obesity  2. 68% of studies associated with improvements in physical fitness3. Lack of focus on subgroups of children and youth
3. Chronic disease risk factors
2. High appraisal scores in program development and evaluation4. Limited interventions in home and community setting
4. Chronic disease associated with obesity
3. 70% of school-based programs non-effective in changing psychosocial outcomes
5. Physical activity component critical
5. Adverse effects (eating disorders, smoking)
6. Involvement of stakeholders in program design, implementation and evaluation necessary
Pyle et al. (2006) (66)Nine studiesSchool-based (nutritional, physical activity and family involvement interventions)1. School-basedN/A (not indicated: seven articles measured overweight indices of which four were reported no change, two indicated significant reduction and one reported mixed results). Most reported changes in psychosocial outcomes1. Early targeting of nutritional behavior (preschool) and continuation into high school necessary
2. Obesity prevention
BMI, skin fold thickness, % overweight, % body fat, body image, depression, food intake, diet and physical activity behavior2. Increased role of school psychologists
Flodmark et al. (2006) (67)Twenty-four studiesSchool-based (nutrition education and/or physical activity) or family-oriented1. Controlled studies41% of studies reported reduction in overweight indices1. Obesity in children and adolescents preventable through limited school-based programs incorporating healthy dietary habits and physical activity
2. 12 months or more follow-up
3. Outcome should include BMI, skin fold thickness or % of overweight or obesity
4. Children recruited from normal or high-risk populations
BMI, skin fold thickness, % overweight/obesity
2. Interventions targeting high-risk activities are few, poorly tested, and with unconfirmed benefits
3. Risk of adverse effects needs to be addressed
Cole et al. (2006) (68)Ten studiesSchool-based (nutritional education, lifestyle and behavior changes and physical activity)1. School-based70% of studies reported reduction in BMI/weight outcome, 30% indicated mixed results (in girls only)1. Influence of children’s social context on treatment should be taken into consideration
2. Published before Feb. 5, 2005
BMI, % overweight, weight loss3. Ages 4–14 yr
4. Healthy lifestyle education, dietary habit and/or physical activity intervention
2. Role of healthy education lifestyle in modifying behavior at a cognitive level is significant and needs to be incorporated
5. Significant decrease in BMI or weight
3. Support of friends and family and providing education in a familiar, non-threatening setting important
Sharma (2006) (13)Eleven studiesSchool-based curricular1. English publication, 1999–200445.5% of studies reported BMI reduction (some in girls, some in both sexes)1. Modest changes in behaviors
2. Mixed results with obesity indicators
3. Focus on general population
3. Most modifiable behaviors: TV watching followed by physical activity and nutrition behaviors
BMI4. Explicit school-based obesity prevention curriculumOthers showed increase in nutrition knowledge or decrease in TV watching
4. Interventions should be based on behavioral theories
5. Family component: 1 study involved family participation (outcome: no change in body fat, but increase in knowledge attitudes and behaviors)
Thomas (2006) (14)Fifty-seven studiesSchool-based1. Elementary or secondary school interventions, with or without parents and/or the community involvement.7.0% studies reported statistically and clinically significant reduction in BMI1. Need for addressing statistical vs. clinical significance
2. Focus on rigorous methodology (sample size and selection, masking, and sub-analysis)
Nutrition, Inactivity, physical activity measures2. Comparison group required57.9% reported mixed/modest effects3. Monitoring of program integrity and dosage
3. Multiple outcomes21.1% reported increase in physical activity4. Successful programs included family/parent involvement
Budd and Volpe (2006) (15)Twelve studiesSchool-based cluster RCT1. RCTs must include BMI for age and gender as an outcome33.3% studies reported reduction in BMITwo types of successful programs for reducing child obesity (classroom instruction and physical education effective in improving physical activity both in and out of school, behavior change programs to target SB
2. In US schools during the school day (elementary, middle, or high school)
3. Peer-reviewed publication
BMITwo studies included families with no change in BMI
Policies required for reducing energy intake and increasing expenditure
Sharma (2006) (69)Twenty-one studiesMostly international elementary School-based (nutrition and/or physical activity)1. English publication, 1999–200554.5% studies reported BMI reduction1. Family/parental involvement critical
2. Non-US
BMI, skin fold thickness3. General population of children in school (3–18 yr)75.0% with parental involvement reported BMI change2. Targeting physical activity, nutrition and TV watching necessary
Demattia et al. (2007) (70)Twelve studiesClinic-, school-based populations and population-based1. Controlled interventions100% of studies decreased SB and improved weight indices1. Targeting SB is an effective intervention to control weight in children and adolescents
2. Children or adolescents
3. Designed to reduce SB in a natural setting
Measure of SB, BMI, % overweight2. Studies should incorporate family-centered components

School-based obesity programs are often successful in improving psychosocial factors (knowledge, self-efficacy, intentions) relating to diet and physical activity. School-based programs are also associated with statistically significant improvements in diet and physical activity, although the clinical/public health significance of the size of these changes is not great. There is considerable variation between reviewers on the proportion of studies that show a statistically significant impact on obesity ranging from about 20 to 75%. This variability is strongly related to differences in inclusion–exclusion criteria in the selection of studies. Overall, when there is an effect, it is quite modest, and of questionable clinical significance at the individual level.

Social Cognitive Theory (SCT) is the most common conceptual framework driving both the intervention and evaluation strategies of the programs. This makes sense, as SCT is based on learning theory and it is a comfortable framework for educators and education researchers, and, perhaps as a result, successful programs have often targeted older children who are thought to be in a better position to internalize and act upon new concepts and behaviors (15). However, many programs reviewed are not based on any underlying theoretical framework (14). In addition, interventions addressing the environment and cultural factors explicitly are lacking in the literature (14).

Most of the reviewers note that the primary focus of the interventions are in terms of changing individual behavior, specifically physical activity and nutrition behaviors, as a means of reducing obesity. Few programs have attempted environmental change, and most of those that did, focused their effort on change within the school (e.g., changes to the school food service). There was little emphasis, impact, or evaluation of factors outside the school.

A great concern is the lack of publication of how the program was developed (formative research) and of process evaluation findings. Lack of publication in these areas makes it very difficult to know whether failure or modest success of the intervention was because of a problem in the development or content of the intervention approach or with its implementation. Other concerns include insufficient duration of follow-up and recidivism (15), inadequate methodological rigor (14, 15), generalizability (14), and gender-specific effects (15, 16). Also, the effect of culture, socioeconomic status, and risk level on the outcomes is often not addressed (14).

Many reviewers expressed concern about the modest amount of parental involvement in most school programs. The duration and intensity of parent activities vary widely between the different programs (14). Most efforts to reach parents have been through children (newsletters, take-home assignments, etc.) and very occasional school events where parents are invited. While some effort has been made in selected schools to modify the within-school food environment, very little effort has been made to influence the home food environment (15–17). School-based intervention seems to be a weak method for influencing family food patterns in the rare instances when they are measured (18). On the other hand, family-based obesity programs are estimated to yield greater impact in obesity and need to be incorporated aggressively into school programs (13, 19).

Summary of school-based programs for diabetes prevention

Table 2 summarizes the limited literature available on school-based diabetes prevention programs. Many school diabetes prevention programs focus on behavior change (diet and physical activity) and obesity as outcomes, rather than on more proximal indicators of diabetes status.

Table 2.  Summary of school-based diabetes prevention programs
AuthorProgram namePopulation, age groupStudy designInterventionMain results
  1. AA, African-American; CST, coping skills training.

Ritenbaugh et al. (2003) (25)Zuni high school diabetes prevention programZuni high school students, 16–18 yrOne school, 3 yr intervention, no comparison groupFitness center with instructors in schoolReduced soft drink consumption
Water made readily available
Replaced regular soda with diet soda in machinesNo change in fasting glucose levels
Diabetes prevention curriculum in science classSignificantly decreased fasting insulin levels
Grey et al. (2004) (28)No nameHigh risk (obese, family history) multi-ethnic middle school students, 10–14 yrTwo schools, 12 month interventionFamily-centered interactive nutrition education curriculumIntervention school children showed trends in improved usual food choices, increased dietary knowledge, lower glucose and insulin levels
School 1 (control) (nutrition education and physical activity)After school physical activity for 2 d/wk for 16 wk
School 2 (intervention) (same as 1 plus CST)CST including culturally sensitive weight management material
Telephone support
Paradis et al. (2005) (21)Kahnawake School Diabetes Prevention ProgramMohawk elementary school studentsThree schools (two intervention, one comparison), 8 yr school and community intervention, independent samplesHealth education curriculum for grades 1–6Initial decreased rate of skin fold thicknesses in intervention children, but not in BMI, PA, fitness or diet
Community media
Community walking/biking pathsDecreased gym class in intervention children
Cross-sectional measures from 1994 to 2002 showed increased BMI
School food policy changes
Saksvig et al. (2005) (20)Sandy Lake School Diabetes Prevention ProgramOji-Cree elementary school childrenOne school, 9 month intervention, no separate comparison groupHealth curricula for third, fourth, fifth gradesSignificant increases in diet-related psychosocial factors, and dietary fiber intake
After school cooking club
School food policy changes
School breakfast/lunch program initiated
Rosenbaum et al. (2007) (27)El Camino Diabetes Prevention ProjectHispanic junior high school studentsOne school, 4 month intervention, control group (same school)Diabetes prevention/research curriculum in science classSignificantly reduced body fatness, insulin resistance and C-reactive protein
Exercise program (kick-boxing)
Shaw-Perry et al (2007) (26)NEEMAAA elementary school childrenSix schools, 4.5 month intervention, no comparison groupClassroom health curriculumSignificantly decreased fasting glucose and percentage body fat
After school health club
Family health fair
School cafeteria changes
Ho et al. (2007) (25); Ho et al. (2006) (71)Zhiwaapenewin Akino’maagewinOjibway elementary schools childrenFour schools, 9 month intervention, (two intervention, two comparison)Health curricula for third and fourth gradeIn adults: significantly increased knowledge and healthy food purchasing among adults
Food stores stocked healthier foodsImpact on children not assessed
Community events and promotions

All the programs reviewed were conducted in populations at exceptionally high risk for diabetes. This includes several programs targeting native North American children, including the Sandy Lake School Diabetes Prevention Program (20), the Kahnawake School Diabetes Prevention Program (21–24), the Zhiwaapenewin Akino’maagewin: Teaching About Diabetes Program (Ho et al., unpublished data), and the Zuni High School Diabetes Prevention program (25). Other programs focus on Hispanics or AA children (26–28).

Most school diabetes prevention programs include extensive references to diabetes, its causes, and how to avoid the condition as part of the curriculum. Programs for Zuni (25), middle-school Hispanic (27), and high-risk multi-ethnic children (28) were successful in improving insulin sensitivity. Other programs demonstrated improvements in dietary behavior (20, 25) and psychosocial factors (20, 25, 28).

All the studies reviewed must be considered pilot studies. Most were conducted in just one or two schools (none in more than four), and many had no or unmatched comparison groups or non-random assignment to treatment condition. Most of the school-based diabetes prevention programs took place entirely within the schools and did not take advantage of the school (e.g., food service) or community environment as part of the intervention, with two exceptions (21, 25).

Potential for community settings for obesity and chronic disease prevention among children

There is a growing body of evidence arguing for the implementation of interventions in community and other environmental settings (29, 30). Environmental factors play a significant role in the development of obesity (31, 32). Individual eating habits are directly related to the availability of and access to healthy food options, especially in disadvantaged populations (33–35). The higher cost of healthier foods can also limit access, particularly to low-income neighborhoods (36). The presence of supermarkets is associated with intake of fruits and vegetables (37). Their migration away from low-income urban neighborhoods, combined with inadequate transportation, is speculated to contribute to the obesity epidemic in low-income residents (38, 39).

Children’s diets, in particular, are heavily influenced by their immediate food environment. Where do children obtain the food they eat? In school-aged children, food sources include the home, school cafeteria, local stores, fast-food restaurants, and vending machines (40). Among the barriers to healthy food consumption are ease of access to cheap snacks at local stores and fast food, and lack of availability of healthy food choices at home (41). As a consequence, most minority children have low intakes of fruits and vegetables, usually lower than Caucasian children (42–45), with consequences for childhood obesity.

Children from low-income ethnic minority populations appear to be especially vulnerable. One source of unhealthy foods are fast-food restaurants, which are often clustered around schools in inner-city neighborhoods, often within a 0.5 km radius, providing children with easy access to less nutritious food environments (46). One-third of the children and adolescents consume fast food each day (47), contributing over 19% of total energy intake (40). For adolescents, fast-food use is significantly associated with increased fat intake (43, 48). Food store-derived snacks contribute 11% of total energy intake in adolescents (40).

These dietary patterns emphasize the importance of exploring settings in addition to the school for obesity prevention among children. There are many advantages to taking a community-based approach to address the child obesity problem. Communities are constituted by multiple institutions and individuals representing food stores, restaurants, faith-based groups, youth organizations, and many other stakeholders (49). Bringing together these diverse viewpoints and letting them all contribute to the development of intervention strategies would likely enhance the probability of success and long-term sustainability.

Community-based interventions aimed at chronic disease prevention in adults had some early success, but recent review shows that this is not consistent (50). There is a dearth of childhood obesity prevention efforts focused at the community level. A notable exception is the recently completed Shape Up Sommerfield (SUS) study (51), which was designed to change the environment to prevent obesity in elementary school-aged children. SUS took a multilevel social ecological perspective, combined with community-based participatory research principles. Intervention components were varied and included changes in the before, during, and after school environments – coupled with changes in home and community settings. Early results of this study showed that children in the two intervention communities decreased their body mass index in comparison with children in the control community (51).


Our review of the literature on child obesity and chronic disease prevention in schools leads us to four main conclusions. First, school-based obesity prevention researchers need to improve their theoretical frameworks and reporting of formative research and process findings. Second, the field needs to move to larger randomized clinical trials (RCTs) for child diabetes prevention. Third, it is time for investigators to reduce their overly heavy emphasis on school-centered approaches. Finally, environmental-based and community-based approaches that focus on families hold great promise as more effective strategies for the prevention of childhood obesity and (probably) diabetes.

In the reviews presented, a common concern raised is the lack of information or inconsistent information presented by investigators, making comparisons between studies difficult. Investigators need to publish more details about their intervention trials, in order to permit adequate understanding and synthesis of their findings. A single paper on impact is insufficient basis on which to make a judgment of efficacy. We recommend at least two additional types of papers on formative research and process evaluation be published for each intervention trial. Formative research is an essential part of the development of culturally appropriate and effective interventions (52). There are many examples of formative research for school-based interventions in the literature (52–55). Yet, for most of the school-based intervention trials published, the formative work is only an occasional mention.

Process evaluation papers are also relatively rare in the school health promotion literature. Their absence leaves a crucial question unanswered, as has been noted by other reviewers (14): was lack of success or modest success because of a failure in the overall intervention approach or because of weak implementation? Further, a strong process evaluation can help assess which components of the intervention are associated with success. Addressing these issues is essential in order to improve school-based interventions in the future. In smaller trials, formative research and process evaluation could be combined into a single paper (56).

The authors argue that new directions beyond the central focus on schools must be explored in order to effectively combat the burgeoning obesity and diabetes epidemics among children. The mixed and modest results of school-centered trials to date argue that this approach alone is not effective in making a real dent in the problem. An approach worth trying, and one that is seeing increased support, are community-based and/or household-centered trials.

The rationale for working in community settings as the future focus for addressing the childhood obesity epidemic is strong. Adolescents obtain more than 90% of their total calories from outside the school setting, the majority from within their homes (60.5%) or from restaurants (19.3%) (40). They also spend more than half their waking hours outside the school, with multiple opportunities to engage in physical activity. Efforts to modify the nutritional or physical activity environments within the school may be undermined by opposing influences outside the school (15, 57). School-based programs have had limited success in engaging parents to change household behavior, and thereby reducing the obesigenic environment at home, with some exceptions (18).

Parents and other child caregivers generally have very limited interaction with schools – except for a subset of involved parents. On the other hand, all parents are most definitely community members and engage with institutions in the community on a regular basis, such as food stores, churches, restaurants, and so on. Changes and outreach in these settings has a strong potential to change adult behavior and lead to positive change in the household. It is well recognized that family eating patterns affect the nutritional status of their children (58, 59). Targeted intervention programs aimed at parents of obese children have already shown success in reducing obesity (60).

It is a great challenge to intervene in an entire community environment (49), with its diversity of settings, institutions, and multiple stakeholders. Interventions of this nature require different theoretical frameworks, including multilevel social ecological approaches. The recent SUS study mentioned earlier offers an example of a successful program to change the community environment for school-age child obesity prevention (51).

Where to begin? One compromise approach to consider would involve developing healthy eating zones around schools. By working in food stores, carryouts, and fast-food restaurants within three to four blocks of schools, we have the potential to influence the food environment of both children and their parents who use those food sources. This approach could also include work within schools, using approaches that have proven effective in previous studies. Other potential ways of impacting on children’s food environment could involve increasing the number of supermarkets in low-income neighborhoods and/or reducing the amount of food-related media focused on children. Future efforts to improve the food environment should employ more than one strategy.

The recent surge in pilot diabetes prevention trials in schools that show positive effects is reassuring and demonstrates how saliency of a health condition like diabetes can motivate and drive behavioral improvements. We are not advocating dropping school-based approaches to chronic disease prevention wholesale. Nonetheless, we feel that the same challenges that limit the effectiveness of school-based programs for obesity prevention, may ultimately also limit the effectiveness of diabetes prevention programs. Therefore, the future of childhood diabetes prevention most likely also lies outside schools.

Conflicts of interest

The authors have declared no conflicts of interest.