Relationships Between Objective Measures Of The Built Environment And Children’s Active Transportation And Physical Activity

PURPOSE: Previous research on the relationship between objective measures of the built environment (BE) and children’s physical activity (PA) has yielded inconclusive findings. One potential explanation is that the effect of the BE varies between domains of PA. METHODS: Canadian children aged 9-11 years (N=484; 56.6% female) self-reported their primary mode of transportation to/from school, and they were asked to wear an ActiGraph accelerometer for 7 consecutive days, using a 24-hour wear-time protocol. ArcGIS software was used to measure home-school distance, the distance between home and the nearest park, and the number of parks and the average slope of the area within a 500 meter buffer of the child’s home. Publicly-available WalkScore®, BikeScore®, and TransitScore® indicators were obtained for participants’ home address. Generalized linear mixed models examined the relationships between BE features and measures of active school transportation (AST) and moderate-to-vigorous PA (expressed as minutes/day) while controlling for gender, parental education, and household income. RESULTS: Each unit increase in the logarithm of home-school distance (OR=0.26; 95% CI=0.18-0.37) and in the WalkScore® of the home address (OR=0.99; 95% CI=0.98-1.00) were associated with lower odds of AST. A 10-point increase in WalkScore®, which ranges from 0 to 100, was associated with a 13% decrease in the odds of AST. No built environment feature was associated with moderate-to-vigorous PA. CONCLUSIONS: Consistent with previous research, longer home-school distance was associated with lower odds of AST. The negative relationship between WalkScore® and AST suggests either that walkability does not increase the likelihood of AST in children of this age, or that the WalkScore® scoring algorithm (which was developed for adults) puts excessive emphasis on access to destinations that are not relevant for children. Our results also suggest that the relationship between the BE and PA is domain-specific. Finally, it is worth noting that our sample provided limited variability in access to park and average slope of the home neighborhood, underscoring a need for future multi-site studies. Supported by The Coca-Cola Company