Addressing childhood obesity in low-income, ethnically diverse families: outcomes and peer effects of MEND 7–13 when delivered at scale in US communities

Objectives: Implementation of a large-scale, child weight management program in underserved communities provided an important opportunity to evaluate its effectiveness under service level conditions. Methods: MEND 7–13 is a community-based, multicomponent, childhood obesity intervention designed to improve dietary, physical activity and sedentary behaviors. It comprises twice weekly sessions for 10 consecutive weeks (total of 35 contact hours) and is delivered to groups of children and their accompanying parents/caregivers. The current evaluation used an uncontrolled, repeated measures design. 4,324 children attended 415 MEND 7–13 programs in seven USA states, of which 2,738 (63%) had complete data for change in zBMI. The intervention targeted underserved families (70% with an income <$40,000 per year; 85.6% Hispanic or African American). Changes in anthropometric, fitness and psychological outcomes were evaluated. A longitudinal multivariate imputation model was used to impute missing data. Peer effects analysis was conducted using the instrumental variables approach and group fixed effects. Results: Mean changes in BMI, and zBMI at 10 weeks were -0.42 kg/m2 (95%CI: -0.49, -0.35) and -0.06 (95%CI: -0.08, -0.04) respectively. Benefits were observed for all other study outcomes. Mean peer reduction in zBMI was associated with a reduction in participant zBMI in the instrumental variables model (B=0.78, p=0.04, 95%CI: 0.03, 1.53). Mean program attendance and retention were 70.8% and 84.7% respectively. Conclusion: Similar to recently published efficacy trial results, implementing MEND 7–13 under service level conditions was associated with short-term improvements in anthropometric, fitness and psychological indices in a large sample of underserved overweight and obese children. A peer effect was quantified showing that benefits for an individual child were enhanced if peers in the same group also performed well. To our knowledge, this is the first study to show positive peer effects associated with participation in a childhood obesity intervention.