Flow-chart of literature search.
General characteristics of systematic reviews or meta-analyses
| Study | No. of publications included (year) | Meta-analysis | AMSTAR / PRISMA | Target age (y) | Type of intervention | Intervention duration | Intervention setting | Outcome | |
|---|---|---|---|---|---|---|---|---|---|
| Primary | Intermediate | ||||||||
| Wang Y et al. (2015)35) | 139 | Meta | High/22 | 2–18 | PA, N, C | 1 y≤ | S, H, Cm, Pc, | BMI, BMI Z-score, % | Intermediate behav- |
| (1985–2014) | 6 mo≤ | Cc, CHI | BMI, WC, %BF, | ioral outcomes (i.e. dietary intakes, PA and sedentary behaviors) and obesity-related clinical outcomes (e.g. blood pressure and blood lipid levels) | |||||
| (school-ba sed) | SFT, P | ||||||||
| Peirson L et al. (2015)30) | 90 (1998–2013) | Meta | High/25 | 0–18 | PA, N, Ed | 12 wk≤ | S, H, Cm, Cc | BMI, BMI Z-score, P | |
| Vasques C et al. (2014)34) | 52 (2000–2011) | Meta | Mod/23 | ≤ 19 | PA, N, Ed | 6 wk≤ | S, after-school | BMI, BMI Z-score, %BMI, | |
| %over-weight/obesity, %BF | |||||||||
| Marsh et al. (2014)13) | 17 | SR | Mod/21 | 2–18 | SB | 6 wk-2 y | S, H, Cm, Pc | Sedentary time | Diet, PA, body |
| (2001–2011) | composition, Bwt | ||||||||
| Langford et | 67 | Meta | High/26 | 4–18 | PA, N, C | 8 wk-6 y | S, H, Cm, En | BMI, BMI Z-score, | School attendance, |
| al. (2014)38) | (1998–2013) | PA, physical fitness, fruit and vegetable intake | Non-academic school, process, curriculum, school environment, engagement with families or communities or both | ||||||
| Chen and | 14 | SR | Mod/21 | 12–18 | Technolog | 10 wk-2 y | NA | BMI, BMI Z-score | Diet intake, phys- |
| Wilkosz (2014)17) | (1990–2014) | y-based (internet, video-game, social media, mobile) | ical fitness, physical activity, psychosocial variable (self-esteem, self-competence) | ||||||
| Williams et al. (2013)18) | 26 | Meta | High/24 | 4–11 | School | 8 mo-9 y | S | BMI, BMI Z-score, | |
| (2003–2012) | policy related to PA or N | %BMI, %BF, BMI healthy fitness zone | |||||||
| Sobol-Gold- | 32 | Meta | Mod/26 | 5–18 | N, PA, Ed, | NA | S | BMI | |
| berg et al. (2013)33) | (2006–2012) | environmental, paren-tal/social support | |||||||
| Silveira et al. (2013)32) | 8 (2004–2010) | Meta | High/27 | 5–18 | N, Ed | 4 mo-3 y | S | BMI | |
| Showell et al. (2013)40) | 6 (2001–2013) | SR | Mod/23 | 3–17 | PA, N, C | 52–104 wk | H±S, Cm, Pc, CHI | BMI, BMI Z-score, %BF, Bwt, P | Diet intake (FV, energy), PA, SB |
| Sbruzzi et al. (2013)31) | 26 (1989–2012) | Meta | Mod/26 | 6–12 | PA, N, Ed | 6–72 mo | S, H, Cm En | BMI, BMI Z-score, WC, BP, lipid | |
| Bleich et al. (2013)39) | 9 (2003–2010) | SR | Mod/21 | 2–18 | PA, C | 12–48 mo | C±S, H, Pc, Cc | BMI, BMI Z-score, fat mass, P | Diet intake (FV, fatty food, sugar sweetened beverages, totalE), PA (MVA, TV watching time, active commuting to school, accelerometer) |
| von Grieken et al. (2012)15) | 34 (1999–2010) | Meta | Mod/24 | 1.9–18.2 | SB, PA, N, Ed | 7 d-4 y | S, H, Pc, Cm, nursery | SB | BMI, BMI Z-score, %BF, SFT, WC, %overweight |
| Osei-Assibey et al. (2012)19) | 35 (1978–2011) | SR | Mod/21 | birth-13 | N | 4 d-4 y | En | Bwt, BMI, BMI Z-score, dietary intake, food choice, nutrition knowledge | |
| Niemeier et al. (2012)29) | 36 (2004–2010) | Meta | Low/23 | 4–18 | PA, N, Ed, C, B | 9 wk-4 y | S, H, clinic | BMI | |
| Luckner et al. (2012)12) | 68 (1982–2008) | Meta | Mod/26 | <18 | PA, N, Ed, TV viewing | 1 mo-7 y < | NA | BMI, %BF | |
| Lavelle et al. (2012)11) | 43 (1991–2010) | Meta | Mod/23 | ≤18 | PA, N, Ed, SB, C, B | 1 mo-6 y | S | BMI | |
| Friedrich et al. (2012)42) | 23 (1998–2010) | Meta | Mod/21 | 7–17 | PA, N | 3–72 mo | NA | BMI | |
| Branscum et al. (2012)21) | 25 (2006–2011) | SR | Low/14 | 7.3–14 | PA, N | 4 wk-3 y | After-school | Bwt, antecedent of behavior (self-efficacy), B (PA, D), body composition, aerobic fitness, BP | |
| Waters et al. (2011)36) | 55 (1993–2009) | Meta | High/26 | 0–18 | PA, N, C | 12 wk≤ | S, H, Cm, Cc | Bwt, Ht, %BF, BMI, ponderal index, SFT, P | |
| Wahi et al. (2011)16) | 13 (1995–2010) | SR | Mod/26 | 3.9–11.7 | PA, N, SB | 1–24 mo | S, Cc, Cm, clinic | BMI | Screen time |
| Zenzen et al. (2009)37) | 16 (2001–2006) | SR | Low/16 | 5 y-12th grade | PA, N, C, Ed | 5 wk-8 y (mean 16.8 mo) | S, H | BMI, SFT, %BF, WC, Bwt, knowledge, B, fitness, dietary intake, psychological measure | |
| Harris et al. (2009)25) | 18 (1993–2008) | Meta | Mod/27 | 1–12th grade | PA | 6 mo-3 y | S | BMI, %BF, WC, WHR, SFT | |
| Gonzalez-Suarez et al. (2009)24) | 19 (1996–2007) | Meta | Mod/21 | 11≤ | PA, N, behavior, combination | 6 mo, 1 y, 2 y | S | BMI, %BF, WC, SFT, WHR, P | |
| Cook-Cot-tone et al. (2009)9) | 40 (1997–2008) | Meta | Mod/20 | Prescho ol-12th grade | PA, N, SB, psycho-educa- tion | 0–12, 13–27, 28–32, 32 wk ＜ | S | BMI, BMI Z-score, %BMI, SFT, %obesity, %BF | |
| Brown et al. (2009)20) | 38 (1993–2007) | SR | Low/19 | 4–18 | PA, N, C | 12 wk-22 y | S | Bwt, BMI, BMI Z-score, %BF, SFT, %overweight | |
| Katz et al. (2008)10) | 8 (1985–2004) | Meta | Mod/21 | 3–18 | PA, N, SB, B, C | 6 mo-5 y | S, H, En | BMI, standardized BMI, Bwt, SFT, %BF, %obesity | |
| Li et al. (2008)41) | 22 (1993–2006) | SR | Mod/21 | 3–21 y | PA, N, Ed, B, school environment, acupuncture | 10 wk-3 y | S | P, Bwt, SFT, BMI Z-score, biochemical marker, changes in knowledge and behavior | |
| Kropski et al. (2008)27) | 14 (1993–2005) | SR | Mod/17 | 4–14 | PA, N, C | 6 mo-6 y | S±En | BMI, BF, %obesity | Dietary intake, PA (fitness level, accelerometer), SB, serum cholesterol |
| Kamath et al. (2008)26) | 36 (1988–2006) | Meta | Mod/23 | 2–18 | PA, N | <3, 3–6, >6 mo | 6 S, H, Cm, clinic | BMI, PA, SB, dietary habit | |
| Lissau (2007)28) | 14 (1995–2005) | SR | Low/13 | 5–15 | PA, N, PA+N | 12 wk-3 y | S | Bwt, BMI, SKT, WC, %BF | |
| Connelly et al. (2007)22) | 28 (1986–2005) | SR | Mod/18 | 9 m-17 | 7 PA, N, Ed | 12 wk≤ | S, H, Cm | BMI, weight-for-height-Z | |
| –score, WC, WHR, %BF, SFT, %obesity, dietary | |||||||||
| intake, PA, fitness level, BP, cholesterol, insulin | |||||||||
| Stice et al. (2006)14) | 46 (1983–2006) | Meta | Mod/21 | 0–22 | PA, N, psycoe-duca- tional, SB | 7–140 wk | NA | BMI | |
| Doak et al. (2006)23) | 25 (1977–2004) | SR | Low/12 | 4–16 | PA, N | 8–260 wk | S, H, En, Cm | m BMI, %obesity, WC, WHR, SFT, %BF | |
| Budd and Volpe (2006)8) | 12 (1988–2005) | SR | Mod/11 | 1–10th grade | h PA, N, SB Ed | , 8 wk-3 y | S | BMI, %BF, SFT, dietary habit, nutritional knowledge, PA, SB, fitness level, BP, cholesterol | |
Abbreviations: AMSTAR, assessment of multiple systematic reviews; PRISMA, preferred reporting items for systematic reviews an meta-analysis; PA, physical activity; N, nutrition; C, combined nutrition and physical activity; y, year; mo, month; S, school; H, home; Cm community; Pc, primary care; Cc, child care; CHI, consumer health informatics; BMI, body mass index; WC, waist circumference; BF, bod fat; SFT, skin fold thickness; P, prevalence of obese or overweight; Ed, educational intervention; wk, week; Mod, moderate; SR, systematic re view; SB, sedentary behavior; Bwt, body weight; En, environment; NA, not assessed; FV, fruit and vegetables; MVA, moderate to vigorous a tivity; B, behavior; Ht, height; WHR, waist/hip ratio.
Summary of the positive results of the systematic reviews or meta-analysis for childhood obesity prevention
| Study | Main findings | Strengths | Limitations | Implications for practice | Implications for research |
|---|---|---|---|---|---|
| Wang et al. (2015)35) | School-based interventions were significantly effective (mean BMI SMD=-0.30; 95% CI −0.45 to −0.15; |
Comprehensive analysis was performed with large numbers of studies. Assessed the strength of evidence for each study. Identified some important implications for clinical decision and policy making. | Great heterogeneity in the included studies. Did not perform stratified analyses based on intervention types or settings. | Combined nutrition plus physical activity intervention based on school with family or parent involvement are effective in childhood obesity prevention. | Future research is needed to evaluate interventions conducted in other environments than in school, and the impact of policy and CHI. Research based on established behavioural theories and novel methodologies is needed. |
| Peirson et al. (2015)30) | Behavioural prevention interventions showed a small but significant effect on BMI and BMI Z-score (SMD=-0.07; 95% CI=-0.10 to −0.03; a reduction in BMI (mean difference −0.09 kg/m2, 95% CI −0.16 to −0.03); and a reduced prevalence of overweight and obesity (RR=0.94; 95% CI=0.89–0.99). | Detailed description of research objective, search strategy, the risk of bias in individual studies and analysis plan. | Did not assess the risk of bias in present study. Great heterogeneity in the included studies. | Trials with larger sample sizes that are powered to detect small differences across subgroups are needed. Future researches involving normal-weight children and very young children are required. | |
| Vasques et al. (2014)34) | Interventions had a small but significant positive effect in prevention and decreasing childhood obesity. ( |
Present the effect size of weight related outcomes by quantitative analysis. Detailed description about classification and definition of moderator variables. | BMI may not accurately reflect a child's fat mass loss. No detailed description of physical activities including types, intensity and frequency. Did not examine socio-economic status of subjects and the risk of bias in individual studies. | Programs have to consider the characteristics of each participant, such as gender, age. Interventions lasting 1 year, with physical activity and nutritional education, with parental involvement are effective. | A detailed description of the methodologies used in the measurement is important for further research. Reviews should be conducted using several anthropometric measurements and evaluating their impact on the metabolic profile of children. |
| Marsh et al. (2014)13) | Studies including a parental component of medium-to-high intensity were associated with significant changes in sedentary behaviours. TV exposure appeared to be related to changes in energy intake rather than physical activity. | Detailed description of study characteristics and the risk of bias in individual studies. | Great heterogeneity of the included studies. Included studies with small sample sizes and short follow up. Inclusion of studies that did not have change in sedentary time as a primary endpoint. | Parental involvement is more important than environmental component. If it is difficult to manage actual screen time itself, parents can focus on associated dietary behaviours. | Future research needs to assess whether targeting of parents considered to be at high risk for low intervention compliance may help improve outcomes, and the mechanism(s) underlying the relationship between screen time and body weight in children. |
| Langford et al. (2014)38) | Interventions had positive effects for BMI, physical activity, physical fitness, fruit and vegetable intaketobacco use, and being bullied. Physical activity (mean BMI=-0.38, 95% CI 0.73–0.03) and physical activity plus nutrition intervention are effective in obesity prevention (mean BMI=-0.11, 95% CI 0.24–0.02). | Included cluster-RCT that addressed all points in the HPS framework. Assessed the quality of evidence and the risk of bias in the individual studies. Categorized timing of outcome assessment as short, medium or long term. Detailed description of study objective, data collection, data extraction and analysis plan. | Great heterogeneity of included studies. Difficulty in assessing complicated interventions. Limited generalizability due to inclusion of standardized interventions. | School-based intervention, like the HPS framework, can be effective at improving a number of health outcomes in students including BMI. Despite the inextricable links between health and education, there are structural barriers in reality. Cross-departmental working between health and education is required to allow the HPS policy to achieve its potential | More evaluations are required that target older children (over 12 years of age). Future research should use outcome measures including academic achievement and behaviours. Studies should evaluate cost effectiveness of the interventions. |
| Williams et al. (2013)18) | Unlike the results of National School Lunch Program, the results of school breakfast program showed a significantly reduced BMI-SDS (ES=-0.080, 95% CI −0.143 to −0.017). | Analysis of school policies according to PICOS format. Detailed description of the quality and risk of individual studies. Inclusion of the variety of databases. | Poor description of the risk of bias across studies | Obesity prevention interventions should focus on multiple factors, such as diet, physical activity, sedentary behaviour, self-esteem, and environment. | Natural experiments could be used to evaluate new policies. The policy would need to be multidimensional and to extend outside of schools. |
| Sobol-Goldberg et al. (2013)33) | School-based obesity prevention intervention were significantly, but mildly effective in reducing BMI, primarily in children but not teenagers (SMD= −0.076; 95% CI −0.123 to −0.028; |
Present the effect size of BMI by quantitative analysis. Good description of search strategy, study selection, data extraction, data collection process, data items. | Poor presentation of the study characteristics Did not report the study limitation. | Interventions for teenagers are required. Future researches should clearly identify the theoretical model guiding their intervention so that more precise data would be available regarding what interventions work and for which populations. | |
| Silveira et al. (2013)32) | School-based nutrition education interventions were effective in reducing BMI (mean SMD=-0.33; 95% CI −0.55 to −0.11). | 84% of included studies assessed as high quality studies. Low risk of publication bias. Good external validity due to inclusion of various countries. | Small number of included study due to the limited number of available RCTs. | Nutrition education intervention needs to be longer than one school year. BMI Z score standardized to age and sex may be an ideal outcome measurer. Result should be interpreted considering different gender characteristics and stages of sexual maturation | Future researches need to identify which approaches, considering the theoretical framework and intervention components, are most effective in obtaining the expected effect over medium- and long-term periods. |
| Bleich et al. (2013)39) | Significant changes in BMI or BMI Z-score found in 4 of the 9 included studies. | A range of community-based childhood obesity prevention interventions from various countries were included. | Many included studies have suboptimal study designs, which may lead to biased results. | Combination interventions implemented in multiple settings may be more effective at childhood obesity prevention. | More research and more consistent methods are needed to understand the comparative effectiveness of these intervention programs. |
| von Grieken et al. (2012)15) | Results showed significant decreases for the amount of sedentary behavior (mean SMD=-17.95 min/D; 95% CI −30.69 to –10.20) and BMI (mean SMD=-0.25; 95% CI −0.40 to −0.09). | Researchers included many studies and were able to estimate an effect based on all interventions combined. | Did not include unpublished studies. Included studies reported several distinct types of sedentary behavior. | Future researches need to provide details on the intervention and the types of outcome measures taken. Studies with longer follow-up time are required. | |
| Osei-Assibey et al. (2012)19) | There was moderately strong evidence to support interventions on food promotion, large portion sizes and sugar-sweetened soft drinks. These interventions would support individual and family-level bahaviour change. | This study is the first focus on the influence of the food environment on overweight and obesity in younger children. Used experts' and practitioners' perceptions about food environment. | The majority of the intervention studies were short term. Not all the evidence outcomes in this review were reported in anthropometric indices. | Reducing food promotion to young children, increasing the availability of smaller portions and providing alternatives to sugar-sweetened soft drinks should be considered in obesity prevention programs aimed at younger children. | Future researches are needed to identify the optimal design and delivery of the interventions, and impact on body weight and BMI rather than food intake. |
| Niemeier et al. (2012)29) | Longer interventions that include parental involvement appear to have greater success. Interventions that require parent participation are likely to reduce child and adolescent participants' BMIs roughly 1.2 kg/m2 relative to controls. | Focus on the influence of the parental participant on childhood obesity. Detailed description of study characteristics in individual studies. | The lack of ability to clearly distinguish between participant age groups. Great heterogeneity of included studies. Poor description of the risk of bias in the individual studies and across studies. | Childhood obesity prevention interventions should include parental involvement and have longer duration. | This study supports the development and testing of interventions that focus primarily on parents to aid them in helping their children develop positive weight-related health behaviors. |
| Luckner et al. (2012)12) | In children, the reductions in mean BMI were achieved through promoting reduced television viewing (MD=-0.27; 95% CI −0.4 to −0.13) and programmes combining physical activity, specifically themed or general health education and nutrition (MD=-0.1; 95% CI −0.17 to −0.04). | Researchers analyzed the effectiveness of intervention according to intervention type and outcome measure. Detailed description of study characteristics in individual studies. | A potential risk of bias due to including controlled but non-randomized studies. Great heterogeneity of included studies. BMI does not fully capture changes in body composition. Body fat was measured differently across the studies that reported it. | Interventions with physical activity and nutritional education are effective. | Future studies should evaluate the effect on both body mass index and percentage of body fat and should report confidence intervals around all outcome estimates. |
| Lavelle et al. (2012)11) | School-based interventions were effective in reducing BMI (ES=-0.17 kg/m2; 95% CI 0.08–0.26; |
This review was conducted in accordance with the PRISMA guideline. Good description of the risk of bias across studies. | BMI may not be the best measure of childhood adiposity. A potential risk of bias due to including non-randomized studies. Poor description of data collection and extraction. | The interventions examined to date appear to be less effective in boys than girls and further work is required to explore the reasons and whether they require modifications to the school-based interventions or an alternative approach. | Further research is required to determine whether the effect of study is maintained after 6 years. Further research is required to determine the ideal type of intervention, taking cognisance of cost-effectiveness as well as clinical effectiveness. |
| Friedrich et al. (2012)42) | Interventions that combine physical activity and nutritional education present better effects on the reduction of BMI among students, (SMD −0.37; 95% CI −0.63 to −0.12) than if applied in an isolated manner. | Researchers analyzed the effectiveness of intervention according to intervention type. | The majority of included studies was performed with a small sample and was considered of low quality. This review is subject to publication bias. | The most challenging aspect for health promotion strategies is adherence outside of schools, since health is negatively impacted by the food industry through advertisements and commercials for calorie-dense foods. | There is a need for randomized controlled studies with well-designed methodologic criteria in order to evaluate the effect of interventions. |
| Waters et al. (2011)36) | Interventions were significantly effective in reducing BMI or BMI Z-score (SMD= −0.15 kg/m2; 95% CI −0.21 to −0.09). | Good description of study objective, methods, analysis, results. Researchers attempted to provide a synthesis of a variety of “implementation factors”, such as age, intervention type, setting, duration, the risk of bias. | Great heterogeneity of included studies. A potential risk of publication bias. | Curriculum on healthy eating, physical activity and body image integrated into regular curriculum Creating an environment and culture that support children eating nutritious foods and being active throughout each day Engaging with parents to support activities in the home setting to encourage children to be more active, eat more nutritious foods and spend less time in screen-based activities | Future trials should be larger, longer term and include assessments of costs, harm, equity impacts, implementation factors and sustainability. |
| Cook-Cottone et al. (2009)9) | Results indicated a small but significant effect for school-based interventions to reduce obesity in children ( |
Researchers analyzed the effectiveness of intervention according to moderating factors. | There is a distinction between a lower weight and actual physical fitness and health. This study did not examine socioeconomic status, parental weight and follow-up periods, which may be moderating factors. | Interventions must be carefully planned and suited to each school's population, risk, and needs. Additional intervention goals should include the following: improved nutrition and health knowledge through psychoeducation, encouragement of nutritional change, reduction of sedentary behaviors, and a high level of parental involvement. | Future research should address the efficacy of integrating a holistic body and mind approach to obesity prevention that integrates an attention to the causes of binge eating and eating disorder risk and prevention. |
| Brown and Summerbell (2009)20) | Combined diet and physical activity interventions prevent children from becoming obese in the long term. | Detailed description of the individual study result. Some interventions appear to vary in effectiveness according to gender, age or weight status of the children. | Some studies pilot studies and have low statistical power. Some of the interventions were of insufficient length or intensity to produce change weight or BMI. The findings are inconsistent. | Dietary interventions providing breakfast for adolescents and PA interventions particularly in girls may help to prevent becoming overweight in the short term. | Studies using quantitative and qualitative outcomes and focusing on study population characteristics that may impact on effectiveness were needed. Study to view behavior change within the context of an obesogenic environment was needed. |
| Katz et al. | Nutrition plus physical | Standardized data | Did not report quality | If we want more | Future studies need to |
| (2008)10) | activity intervention was | extraction according to | assessment of | evidence-based | focus on intermediate |
| significantly effective in | CDC community guide | individual study and | practice, we need more | outcome such as | |
| reducing weight | data abstraction form | limitation of the study. | practice-based | attitude, knowledge as | |
| (SMD=-0.29; 95% CI | evidence. Because the | early measures of | |||
| –0.45 to −0.14). | primary mission of | success in obesity | |||
| Comprehensive | schools is to educate not | control. | |||
| interventions including | to promote health, a | ||||
| family or parent | priori evidence should | ||||
| involvement were | be presented. | ||||
| effective (SMD=-0.20; | |||||
| 95% CI −0.41 to 0.00). | |||||
| Lissau (2007)28) | ) Half of the 14 included | Good description of | The included studies | The barriers of | Further studies need to be |
| studies had an effect on | study objective and | differred greatly in | school-based | evaluated using various | |
| overweight. | search strategy. | regards to age group, | interventions are (1) | outcome measure as | |
| type and length of | healthy eating has a low | well as BMI. A | |||
| intervention, type and | priority, (2) lack of | prevention project must | |||
| amount of actions and | support at the school | be theory-based. | |||
| statistical power. | for healthy food and | ||||
| meals, (3) the school | |||||
| staff are not motivated | |||||
| or are too overloaded | |||||
| with work to give | |||||
| attention to nutrition, | |||||
| and (4) poor or lack of | |||||
| supervision of the | |||||
| school meals. | |||||
| Connelly et al. | The main factor | Comparative analysis | Did not validate intensity | Compulsory aerobic | Further research is |
| (2007)22) | distinguishing effective | between studies | score as elements of an | physical activity may be | required to identify |
| from ineffective trials | reporting effective and | effective intervention. | related to a decrease in | how compulsory | |
| was the provision of | ineffective outcomes | adiposity in children. | physical activity can be | ||
| moderate to vigorous | was done. | Nutritional education | sustained and | ||
| aerobic physical activity | and skills training may | transformed into a | |||
| in the former on a | reasonably be | personally chosen | |||
| relatively ‘compulsory' | considered useful in a | behaviour by children | |||
| rather than ‘voluntary' | general health | and over the life course. | |||
| basis. | promotion sense. | ||||
| Stice et al. | 21% of the 64 included | Researchers evaluated | Poor description of the | Larger effect sizes tended | Future studies are needed |
| (2006)14) | studies had a significant | putative moderators of | risk of bias in the | to emerge in trials | to conduct follow-up |
| weight gain prevention | obesity intervention | individual studies and | involving children and | trials of enhanced | |
| effect (average effect size | effects. | across studies. Great | adolescents, in | versions of the | |
| heterogeneity of | female-only trials, in | programs and to design | |||
| 0.5). | included studies. | interventions below the | new programs. Need to | ||
| median of 16 weeks, in | determine how to better | ||||
| interventions that | design obesity | ||||
| targeted only weight | prevention programs | ||||
| change. | for preadolescents and | ||||
| males. Future trials | |||||
| should include | |||||
| multi-year follow-ups. | |||||
| Need to evaluate the | |||||
| mediators that | |||||
| putatively account for | |||||
| any weight gain | |||||
| prevention effects. | |||||
| Doak et al. (2006)23) | 68% the interventions, or 17 of the 25, were ‘effective' based on a statistically significant reduction in BMI or skin-folds for the intervention group. | Researchers evaluated potential adverse effect. Inclusion criteria were kept broad in order to include interventions focusing on ‘health promotion' as well as prevention' of obesity and obesity-related behaviours. | A potential risk of publication bias. Difficulty in comparing outcomes that are reported in different ways, including height for weight as well as skin-fold measures. | Future interventions should take body composition measures such as skin-folds as well as height and weight to better assess body composition changes. More attention should be given to improving the participation rates of interventions. Health promotion messages should be tailored appropriately according to ethnicity, gender and age. | Future studies targeting a broader age range could test whether the 8–10-year-old age group requires a specific approach. Additional studies are needed to measure the costs and benefits of interventions, as well as potential adverse effects. |
| Budd and Volpe (2006)8) | The use of a multicomponent, comprehensive, and detailed nutrition and physical activity curricula for the students in higher grades greatly contributed to the success of programs. | Detailed description of study characteristics and implication for practice. | The randomization was either by school or by classroom while the findings were reported on individuals. When classrooms were randomized, it was likely that the control group was aware of the study objectives, which may have weakened the findings. | Strategies might include using behavior modification techniques with younger students to reduce sedentary behavior, increase physical activity, and encourage proper nutrition and instituting a schedule of physical education classes with longer and more vigorous exercise. | The new participatory action models of community-centered and community-partnered research are required. Future studies can use the framework that the Prevention Group of the International Obesity Task Force presented. |
Abbreviations: BMI, body mass index; SMD, standardized mean difference; CI, confidence interval; CHI, consumer health informatics; RR, relative risk; RCT, randomized controlled trial; HPS, health-promoting school; SDS, standard deviation score; MD, mean difference; ES, effect size; PICOS, population, interventions, comparators, main outcome, study design; PRISMA, preferred reporting items for systematic reviews and meta-analysis; PA, physical activity; CDC, Centers for Disease Control and Prevention.
Summary of the negative results of the systematic reviews or meta-analysis for childhood obesity prevention
| Study | Main findings | Strengths | Limitations | Implications for practice | Implications for research |
|---|---|---|---|---|---|
| Chen and | There is no clear evidence | Researchers evaluated the | Did not assess study | Technology-based | Future study should |
| Wilkosz | that technology-based | intervention effects | limitations. | intervention for weight | include evaluation of |
| (2014)17) | interventions decreased | according to | management needs to | cost-effectiveness, the | |
| obesity in adolescents. | technology. | be continued in order to | mediating and | ||
| see sustainability. | moderating factors | ||||
| Depending on the age | associated with | ||||
| of the participants, | effective | ||||
| different modalities | technology-based | ||||
| might be more | interventions, more | ||||
| attractive than others. | long-term follow-up, | ||||
| and assessment of | |||||
| weight-related health | |||||
| outcomes, such as | |||||
| physical activity, | |||||
| sedentary activity, | |||||
| dietary behaviors, | |||||
| self-efficacy, and | |||||
| quality of life. | |||||
| Showell | None of the 6 included | Good description of | Researchers limited | Studies with larger | More research is needed |
| et al. | studies reported a | study characteristics | review to studies with at | sample sizes, testing the | to evaluate the impact |
| (2013)40) | significant effect on | and the risk of bias in | least 1 year of | intervention effect in | of home- and |
| weight outcomes. | individual studies. | follow-up and only | the various settings, and | family-based | |
| Combined interventions | Evaluated the effects of | included those from | targeted entire families | interventions with | |
| had beneficial effects on | the interventions on | high-income countries. | in households showed | larger sample sizes, | |
| fruit/vegetable intake | multiple outcomes | favorable diet or PA | greater intervention | ||
| and sedentary behaviors. | including | outcomes. | duration and intensity, | ||
| weight-related | and adequate | ||||
| outcomes and | participant follow up to | ||||
| behavioral outcomes. | improve statistical | ||||
| power of studies. | |||||
| Sbruzzi | There were no differences | The focused review | Most of the studies | The use of BMI and waist | Future studies should be |
| et al. | in outcomes assessed in | questions | included were of low | circumference for the | carried out with a larger |
| (2013)31) | prevention studies. | A comprehensive and | methodological quality. | prediction of risk factor | number of participants. |
| Educational | systematic literature | clustering among | |||
| interventions were | search with no language | children and | |||
| associated with a | restrictions | adolescents is useful for | |||
| significant reduction in | The collaboration of a | clinical practice. | |||
| waist circumference, | multidisciplinary team | Short-term | |||
| BMI and diastolic blood | of cardiologists, | interventions may be | |||
| pressure in treatment | endocrinologist, | more intensive and have | |||
| studies. | healthcare researchers | higher frequency and | |||
| and methodologists | adherence of the | ||||
| participants, which may | |||||
| have contributed to | |||||
| better results. | |||||
| Branscum | Interventions resulted in | Good description of | Poor description of study | Obesity prevention | Future study is needed to |
| and | modest changes in | study characteristics | method. | interventions should | be based on behavioral |
| Sharma | behaviors and behavioral | such as outcome | target both physical | theories and to | |
| (2012)21) | antecedents, and results | measures, sample size | activity and nutrition | implement more than | |
| were mixed and | calculation, process | behaviors. | one type of process | ||
| generally unfavorable | evaluation, setting. | evaluation. | |||
| with regards to | |||||
| indicators of obesity. | |||||
| Wahi et al. (2011)16) | Interventions aimed at reducing screen time had no overall effect on the reduction of BMI. A subgroup analysis of preshcool children showed a difference in mean change in screen time (unadjusted difference in mean=-3.72; 95% CI −7.23 to −0.20). | Explored the methodological quality and quality of evidence using the GRADE criteria | Parent reporting of screen time might have biased the measurement. Great heterogeneity of included studies | Future study should include a report of potential adverse effects. Future study is required to evaluate pragmatic interventions that could feasibly be implemented in fewer sessions, over shorter periods of time, with longer follow-up, and focused on key age groups where behavior change may be sustainable, such as the preschool age group. | |
| Zenzen and Kridli (2009)37) | This review points to no specific intervention or combination of interventions as the most beneficial. | Good description of study characteristics such as level of evidence and theoretical framework. | A potential risk of publication bias. Study duration did not appear to be adequate, especially in studies looking for outcomes related to changes in BMI. | School-based obesity intervention programs should be guided by behavioral theoretical frameworks. Intervention components should include physical activity, diet, healthy lifestyle education and parental involvement. | The program should be of a duration long enough to give the participants ample time to exhibit the desired outcome. Intervention need to have aspects specifically tailored to each community. |
| Harris et al. (2009)25) | Results showed that BMI did not improve with school-based physical activity interventions. | Researchers conducted sensitivity analysis according to study characteristics, such as duration of intervention, quality of studies, genders of participants. | It is possible that school-based physical activity could increase lean muscle mass and decrease fat mass with no overall change in BMI. There was a variation among the studies. Did not assess the adherence to study protocols and the “dose” of physical activity. | Current population-based policies that mandate increased physical activity in schools are unlikely to have a significant effect on the increasing prevalence of childhood obesity. | Having a study that is appropriately powered is critical. In addition to collecting appropriate anthropometric data, future studies should also assess the impact of such interventions on the metabolic profile of children. |
| Gonzalez-Suarez et al. (2009)24) | The risk of overweight and obesity was significantly lower in intervention program (OR=0.74; 95% CI 0.60–0.92), and programs with duration of 1≤ years are more effective. The results showed no significant differences in decreasing BMI (WMD= −0.62; 95% CI −1.39 to 0.14). | Detailed description of search strategy, data extraction and analysis plan. Systematic presentation of study according to PICOS framework. | Poor description of individual study characteristics. | Multidimensional programs with longer in duration were more effective. | Intervention with larger scales, longer than 2 years in duration and with better reporting of methodology are needed. |
| Li et al. (2008)41) | Most studies reported a beneficial effect of the intervention with one or more of the study outcomes, but most of the intervention studies suffered from weak methodology. | The kindergartens and schools that participated in these interventions are likely to be largely representative of Chinese educational establishments in urban areas. | All of the studies had serious, or moderate, methodological weaknesses. Limited external validity for countries other than China. | The interventions that focused on health education and/or lifestyle behavioural changes are effective in knowledge improvement and in the prevention of overweight and obesity. | Future studies should address the methodological weakness of previous interventions. |
| Kropski et al. (2008)27) | Quantity and quality of evidence were insufficient. Twelve of 14 studies reported significant improvement in at least one measure of dietary intake, physical activity and/or sedentary behaviour. | Good description of the risk of bias and the quality of evidence in individual studies. | A potential risk of publication bias. Great heterogeneity of included studies. No flow chart/description of selection process. | Programs grounded in social learning may be more appropriate for girls, while structural and environmental interventions enabling physical activity may be more effective for boys. | Future studies should include a well-designed evaluation protocol that assesses appropriate outcome measures, appropriate target age, optimal duration of intervention and cost-effectiveness. |
| Kamath et al. (2008)26) | Included interventions caused small changes on their respective target behaviors and no significant effect on BMI compared with control (ES =-0.02; 95% CI −0.06 to 0.02). | Good description of the quality of evidence in individual studies. Researchers conducted subgroup analyses by grouping studies. | Included studies are prevention studies that mostly included a mixed group, a proportion of which was already overweight. Included studies have important methodological shortcomings (loss of follow-up, lack of blinding). | . Strategies attempting to reduce unhealthy behaviors (i.e. decreasing sedentary behaviors and dietary fat) seem to be more effective than those promoting positive behaviors (i.e. increasing physical activity and consumption of fruits and vegetables). | The long-term impact of behavioral interventions on maintenance of target behaviors needs further exploration along with methodological rigor in the definition and measurement of the target behaviors. Future studies should evaluate the adverse effects of behavioral interventions or the targeted bahaviors. |
Abbreviations: PA, physical activity; BMI, body mass index; CI, consumer health informatics; GRADE, the grading of recommendations assessments, developments and evaluation; OR, odds ratio; WMD, weighted mean difference; PICOS, population, interventions, comparators, main outcome, study design; ES, effect size.