RESUMEN
AIM: This study aimed to determine over a three-year period what the PA levels and patterns of adolescents will be after participating in a physical activity (PA) intervention. METHODS: African adolescents (158 boys, 211 girls, N.=309) from a disadvantaged environment in South Africa with a mean age of 14.9±1.4, boys and 14.3±1.4 girls) participated in the study as a PA intervention group (N.=282) and a control group (N.=87). Multi-level modelling (repeated measures analysis of covariance (ANCOVA) adjusted for attendance percentage and gender to compare the intervention and control groups and repeated measures analysis of variance (ANOVA) over time, with a Bonferroni post hoc test were used to determine how the variables differ across the various testing opportunities. Week and weekend PA levels and patterns were analysed separately. RESULTS: Similar mean PA levels were found among the boys during week and weekend baseline measurements, with a significantly higher mean PA-level after three years in the intervention group. Baseline measurements for girls indicated statistically significant mean differences during the week and the weekend, although these mean differences were not significant during the end measurements. Boys and girls (excluding the intervention group of boys), showed a tendency of declining mean PA over the three years. The PA intervention showed a stabilising effect on PA levels and contributed to more active choices among the intervention group of boys, especially during weekends. CONCLUSION: Participation in similar PA interventions is recommended for adolescent boys, with adaptation to the content for girls.