School-based physical activity interventions: which intervention characteristics are associated with participation and retention? A meta-analysis

OBJECTIVE: Many school-based intervention studies are conducted to increase students' physical activity (PA). Recruitment and retention problems potentially impact the robustness of RCT findings. We conducted a meta-analysis to summarize recruitment and retention rates in long-term secondary school-based PA intervention studies and examined associated participant and intervention characteristics. METHODS: Web of Science, Pubmed, Medline, and PsychInfo were searched until March 20th 2023. We included studies on secondary school-based PA interventions =12 weeks, aimed at typically developing adolescents. We abstracted number of schools and students invited, randomized, and participating at follow-up to calculate pooled recruitment and retention rates; participant and intervention characteristics were abstracted to execute subgroup or meta-regression analyses. RESULTS: Recruitment rates were 51% for invited schools and 80% for invited students, the retention for schools was almost 100% and for students 91%. Interventions with fixed and flexible components, executed in Asia and South America, and from later publication years had higher student recruitment rates. Students' retention rates were lower for interventions which had flexible components, were theory/model-based, used an accelerometer, had a longer intervention duration, and included more females. CONCLUSION: Recruitment and retention rates in school-based PA interventions are high. Some participant and intervention characteristics influence these rates: flexibility of the intervention, theory/model-based intervention, accelerometer use, intervention duration, continent, and number of females. Researchers should consider these characteristics in intervention development to achieve optimal balance between intervention effectiveness, recruitment, and retention.