Background: Responding to the obesity epidemic requires robust evidence to help prioritize the allocation of scarce resources to preventive interventions. The aim of this study was to evaluate interventions that promote healthy weight [defined as reduction in body mass index (BMI) or percentage body fat] in general populations (unselected by weight) using a comprehensive meta-analysis. Interventions with both single and multiple components were considered.
Methods: Studies were first identified through well-conducted systematic reviews complemented by a search for single studies in five large medical databases up to 6 November 2008. Sixty-eight controlled studies were included. For each intervention type and age group, all relevant studies were pooled in a random effects meta-analysis.
Results: In children, the highest reductions in mean BMI were achieved through promoting reduced television viewing [-0.27 kg/m(2) (95% CI -0.4 to -0.13 kg/m(2))]. Programmes combining physical activity, specifically themed or general health education and nutrition achieved a lower reduction [-0.1 kg/m(2) (95% CI -0.17 to -0.04 kg/m(2))]. Other interventions had high heterogeneity or showed no statistically significant reduction in outcomes. In adults, single component interventions were found to reduce both outcome measures. Their mean percentage body fat was reduced through education by -1.22% (95% CI -1.92 to -0.52).
Conclusion: The evidence for the effectiveness of promoting healthy weight in general populations is limited, though multi-component interventions in schools and encouraging reduced children's television viewing are promising strategies. Improving the reporting of outcomes is vital, as imputation of inadequately reported measures may have contributed to the observed heterogeneity. Longer follow-up is essential for understanding policy relevance.