Multilevel determinants of physical activity in children and adolescents: a meta-analysis guided by social ecological model

Abstract

Purpose: We conducted a systematic review and meta-analysis to identify factors influencing physical activity (PA) in children and adolescents aged 5–18 years through a social-ecological model (SEM) perspective. Additionally, we examined the combined effects of factors across SEM levels on PA behaviors. Methods: Following a comprehensive search of Chinese and English databases, two researchers independently screened studies, assessed quality, and selected 15 articles meeting inclusion criteria. A dual analytic approach was employed: semi-quantitative evaluation Method (systematic scoring of study findings to synthesize directional associations) and meta-analysis using R4.4.2 to pool effect sizes (Cohen’s d) with random-effects models. Subgroup analyses addressed heterogeneity sources. Results: Semi-quantitative synthesis revealed consistent positive associations for interpersonal-level factors, including parental co-participation in PA (83.3%), parental rules (66.7%), peer support (66.7%), and team sports participation (100%). At the individual level, self-efficacy (84.6%) and enjoyment (62.5%) showed the strongest effects. Meta-analysis confirmed multilevel SEM influences (SMD = 0.443,95%CI [0.784,0.885]), Due to variations in statistical methods (I2 = 45.9–99.8%), survey setting scope (I2 = 0–99.6%), geography (I2 = 0.86–99.7%), and Research design (I2 = 67.5–99.8%), there was a considerable degree of heterogeneity among the included studies. Conclusions: Key findings indicated that at the individual level, PA engagement was significantly associated with perceived barriers, self-efficacy, motivation, and enjoyment. At the interpersonal level, parental co-participation, rule-setting, peer support, and team sports emerged as critical facilitators. However, meso-level factors (e.g., school physical education curricula, accessibility of community sports facilities, extracurricular programming) and macro-level structural determinants (e.g., national physical education policy alignment, urban built environment planning, sociocultural sports values) remain underexplored, with current research disproportionately focused on micro-level individual traits (e.g., motivation) and interpersonal dynamics (e.g., family support). The study recommends: (1) individual-level interventions prioritizing self-efficacy and intrinsic motivation enhancement; (2) interpersonal strategies fostering family-peer collaborative networks; (3) meso-level improvements to school PA curricula and equitable community infrastructure; and (4) macro-level research on policy coherence and sociocultural value systems. Future investigations should integrate longitudinal designs within social-ecological frameworks to elucidate dynamic interactions across multilevel factors, thereby advancing theoretical and policy frameworks for youth PA promotion. These insights provide empirical foundations for optimizing school-based PA initiatives, urban planning for active spaces, and culturally responsive community interventions. Trial registration: PROSPERO CRD420250651231.