Background: Urban design may affect children's habitual physical activity by influencing active commuting and neighborhood play.
Purpose: Our objective was to examine associations between neighborhood built-environment features near children's homes and objectively measured physical activity.
Methods: We used geographical information system (GIS) protocols to select 2016 households from 48 low- and high-walkability neighborhoods within four New Zealand cities. Children (n = 227; mean age ± standard deviation [SD] 9.3 ± 2.1 years) from the selected households wore accelerometers that recorded physical activity in the period 2008-2010. We used multilevel linear models to examine the associations of GIS and street-audit measures, using the systematic pedestrian and cycling environmental scan (SPACES), of the residential environment (ranked into tertiles) on children's hourly step counts and proportions of time spent at moderate-to-vigorous intensity on school and non-school days.
Results: During school-travel times (8:00-8:59 a.m. and 15:00-15:59 p.m.), children in the mid-tertile distance from school (~1 to 2 km) were more active than children with shorter or longer commute distances (1290 vs. 1130 and 1140 steps·h(-1); true between-child SD 440). After school (16:00-17:59 p.m.), children residing closest to school were more active (890 vs. 800 and 790 steps·h(-1); SD 310). Neighborhoods with more green space, attractive streets, or low-walkability streets showed a moderate positive association on non-school day moderate-to-vigorous steps, whereas neighborhoods with additional pedestrian infrastructure or more food outlets showed moderate negative associations. Other associations of residential neighborhoods were unclear but, at most, small.
Conclusions: Designing the urban environment to promote safe child-pedestrian roaming may increase children's moderate-to-vigorous physical activity.