Youth physical activity and the neighbourhood environment: examining correlates and the role of neighbourhood definition.

The primary objective of this study was to examine relationships between neighbourhood built and social environment characteristics and moderate to vigorous physical activity (MVPA) in a sample of children aged 8-11 in Vancouver, British Columbia and the surrounding lower mainland region (n = 366). A secondary objective was to assess how neighbourhood definition influences these relationships, by using measures calculated at multiple buffer sizes: 200, 400, 800 and 1600 m (1 mile). Geographic information systems -software was used to create a broad set of measures of neighbourhood environments. Physical activity was measured objectively using accelerometers. Relationships between MVPA and neighborhood characteristics were assessed using generalized estimating equations to account for the clustering of children within schools. Sex specific relationships were assessed through sex stratified models. When controlling for child age, sex and ethnicity, MVPA was positively associated with commercial density, residential density, number of parks and intersection density; and negatively associated with distance to school and recreation sites. When entered as a composite index, these measures accounted for 4.4% in the variation in MVPA for the full sample (boys and girls). Sex stratified models better explained the relationships between neighbourhood environment and physical activity. For boys, built and social environment characteristics of neighbourhoods accounted for 8.7% of the variation in MVPA, and for girls, neighborhood factors explained 7.2% of the variation. Sex stratified models also point towards distinct differences in factors associated with physical activity, with MVPA of boys associated with wider ranging neighborhood characteristics than MVPA of girls. For girls, two safety-related neighbourhood features were found to be significantly associated with MVPA: cul-de-sac density and proportion of low speed limit streets. In all models, larger buffer sizes, and predominantly the largest buffer size, best explained environment-physical activity relationships.