Adolescent physical activity profiles as determinants of emerging adults' physical activity.

Background: Although adolescent physical activity (PA) tracks into adulthood, it is unclear if the context of PA practiced during adolescence predicts adult PA. We previously identified five adolescent PA profiles and aimed to assess associations between these profiles and levels of PA in emerging adulthood.

Methods: Using data from the first 8 years of the MATCH study, when participants were 11 to 18 years, we identified five adolescent PA profiles: "non-participants" (9% of the sample), "dropouts" (30%), "active in unorganized PA" (19%), "active in organized PA" (27%), and "active through a variety of PA" (15%). The same participants reported their PA level (IPAQ) 2.5, 3.5, 4.5, and 5.5 years later when they were emerging adults (20, 21, 22, and 23 years). The adolescent PA profiles were included in a mixed-distribution two parts model as predictors of i) the probability of reporting any PA during emerging adulthood, and ii) the PA level of emerging adults reporting PA.

Results: Being categorized as "non-participant" or "dropout" during adolescence was associated with similar likelihoods of reporting PA and PA level during emerging adulthood. In contrast with "non-participants", those in the "active-unorganized" (OR, 95% CI: 2.8, 2.1-3.8), "active-organized" (2.4, 1.7-3.2), and "active-variety" (3.7, 3.3-5.1) were considerably more likely to report any PA during emerging adulthood. Among emerging adults who reported some PA, those with an adolescent profile of "active-unorganized", "active-organized" and "active-variety" had higher PA levels than "non-participants" (all p < 0.01).

Conclusion: Profiles of PA participation during adolescence represent strong predictors of PA in emerging adulthood. Promoting participation in various types of PA during adolescence is key to preventing low PA among emerging adults.