Family, school and individual characteristics associated with adolescents' physical activity at school in Hong Kong: the iHealt(H) study.

Background: Adolescents (11-18-year-olds) are at risk of physical inactivity. There is limited knowledge of physical activity (PA) levels among adolescents in the school setting in Hong Kong. We developed and tested a novel theoretical model of how household/family characteristics, school-level social and physical environmental factors and individual adolescent's characteristics impact on their school-based PA during and after school hours.

Methods: Cross-sectional study participants were Hong Kong adolescents attending secondary school, paired with their parent/caregiver (n = 1299 dyads). Parents survey-reported on household/family characteristics, parental PA and rules related to PA. Adolescents survey-reported on school PA-friendly policy, PA equipment at school (combined to create PA-friendly index), social support for PA from peers, athletic ability, attitude to and enjoyment of PA. Adolescents self-reported their school-based PA during school hours (physical education; recess) and after school (sports teams/classes). Objectively-measured moderate-to-vigorous PA (MVPA) was measured using accelerometers for a sub-sample of adolescents (n = 588). Generalized additive mixed models were used to estimate how household/family-level and school-level factors, and adolescents' individual characteristics were related to adolescents' school-based PA in Hong Kong, and to identify potential mediators of these associations.

Results: A complex network of potential pathways of influence on adolescents' school-based PA was identified. Overall, most of the significant effects were indirect ones. However, there were far fewer significant pathways between household/family characteristics and objectively-measured MVPA at school than there were for self-reported PA at school. In fact, there were no indirect pathways between these variables and MVPA at school. Gender disparities among pathways were identified. For example, school PA-friendly index was significantly associated with MVPA after school only among girls (e = 1.06, 95%CI (1.02,1.12)).

Conclusions: Key points of intervention identified by our study may be in the re-design of PE classes so that adolescents spend more time being physically active during these classes, and promotion of active play during recess. Further research measuring amount, intensity and location of adolescents' PA using accelerometer and Global Positioning Systems is required in Hong Kong, as well as observational studies of PA during PE classes and in the schoolyard during recess, to guide the design of PA interventions.