Is the relationship between the built environment and physical activity moderated by perceptions of crime and safety?

Background Direct relationships between safety concerns and physical activity have been inconsistently patterned in the literature. To tease out these relationships, crime, pedestrian, and traffic safety were examined as moderators of built environment associations with physical activity. Methods Exploratory analyses used two cross-sectional studies of 2068 adults ages 20–65 and 718 seniors ages 66+ with similar designs and measures. The studies were conducted in the Baltimore, Maryland-Washington, DC and Seattle-King County, Washington regions during 2001–2005 (adults) and 2005–2008 (seniors). Participants were recruited from areas selected to sample high- and low- income and walkability. Independent variables perceived crime, traffic, and pedestrian safety were measured using scales from validated instruments. A GIS-based walkability index was calculated for a street-network buffer around each participant’s home address. Outcomes were total physical activity measured using accelerometers and transportation and leisure walking measured with validated self-reports (IPAQ-long). Mixed effects regression models were conducted separately for each sample. Results Of 36 interactions evaluated across both studies, only 5 were significant ( p < .05). Significant interactions did not consistently support a pattern of highest physical activity when safety was rated high and environments were favorable. There was not consistent evidence that safety concerns reduced the beneficial effects of favorable environments on physical activity. Only pedestrian safety showed evidence of a consistent main effect with physical activity outcomes, possibly because pedestrian safety items (e.g., crosswalks, sidewalks) were not as subjective as those on the crime and traffic safety scales. Conclusions Clear relationships between crime, pedestrian, and traffic safety with physical activity levels remain elusive. The development of more precise safety variables and the use of neighborhood-specific physical activity outcomes may help to elucidate these relationships.