Modeling Longitudinal Changes in Physical Activity Levels: Oporto Growth Health and Performance Study

Adolescence is often viewed as a critical window for the development of healthy behaviors which can prevent the development of metabolic risk factors and related comorbidities. Although there is evidence of the independent associations of physical activity with subject and contextual factors, few studies examined the joint longitudinal links of these factors. PURPOSE: We aim to model longitudinal changes in youth total physical activity (TPA) as a function of their biological maturation (BM), weight status, socioeconomic status (SES), physical fitness (PF), sleep habits, fruits and vegetables consumption and built environment. METHODS: This is a mixed longitudinal study and the data comprised 7315 adolescents (3621 girls) divided into four age cohorts (10 to 12, 12 to 14, 14 to 16 and 16 to 18 years) measured annually for three consecutive years. TPA was estimated with the Baecke questionnaire; BM was indirectly assessed by the maturity offset; weight status was defined with cut-off points for age and sex defined by the International Obesity Task Force; SES was determined by the Portuguese schools’ social support system; PF was assessed with the Fitnessgram test battery; sleep habits, fruits and vegetables consumption and built environment information were obtained by questionnaire. Random effects mixed-models were used and computations done in SuperMix v.1. RESULTS: Girls’ TPA at baseline (10 years) was 7.27 points (possible range: from 3 to 15), being boys systematically more active (β=0.85, p<0.001); with increasing age, TPA decreases, independently of sex (β=-0.07±0.03, p=0.022); and the fittest teenagers were those more active (β=0.29±0.039, p<0.01); favorable neighborhood environments to healthy lifestyles were positively linked to TPA over the age (built facilities: β=0.06±0.02, p<0.01; sociocultural: β=0.09±0.02, p<0.01; and economic/political: β=0.07±0.02, p<0.01); although marginally, BM was positively associated with TPA changes (β=0.07±0.04, p=0.064). The others predictors did not show any significant association with TPA trajectories either sex. CONCLUSIONS: These results showed the decline of TPA during adolescence, emphasize the relevant role of PF, as well as favorable neighborhood environments to provide increments in TPA among adolescents.