How to get children moving? The effectiveness of school-based interventions promoting physical activity in children and adolescents – A systematic review and meta-analysis of randomized controlled- and controlled studies

Abstract

This systematic review and meta-analysis summarizes the effectiveness of school-based physical activity (PA) interventions on children's and adolescents' PA. As no robust empirical evidence exists regarding what seems to characterize the school-based PA interventions that are most effective, we further aimed to map key factors of particular importance when trying to increase PA in early stages of life through school-based strategies.

Intervention effects were calculated as standardized between-group (i.e., intervention vs. control) mean differences (SMD) in PA from baseline to follow-up. In total, 189 publications were included. Few studies (7%) were of high quality. Our results demonstrate that school-based interventions only have a small positive effect on children's and adolescents' PA levels. Compared to the effect observed during total day (SMD = 0.27, p < 0.001), a slightly larger effect was observed during school hours (SMD = 0.37, p < 0.001), while no intervention effect was observed during leisure time (SMD = 0.07, p = 0.20). There was a tendency for interventions to be more effective if theoretical frameworks for behavior changes were used in the design phase. The largest effect size was observed when experts from outside school delivered the program (SMD = 0.56, p = 0.01), but training of personnel involved in delivery was the determining factor for program effectiveness as no effect was observed if interventions were delivered primarily by schools' untrained staff (SMD = 0.06, p = 0.61). Intervention effects where larger if parents were involved in the intervention program (parents involved: SMD = 0.35, p < 0.001; parents not involved: SMD = 0.16, p = 0.02). Small positive intervention effects were sustained at long-term follow-up after end of intervention. Overall, the certainty of the evidence of the findings is rated as low.