Does Car Availability Moderate The Relation Between Neighborhood Walkability And Physical Activity?

PURPOSE Neighborhood walkability, characterized by mixed land use, street connectivity, and residential density, has been found to be related to overall physical activity. In previous analyses, this association was significant in high income adults, but not in low income adults. The purpose of the present study was to test the hypothesis that car availability moderated the relation between neighborhood walkability and physical activity. METHODS A random sample of adults was recruited from neighborhoods in Seattle and King County, WA that varied by walkability and income. Participants reported the number of operating motor vehicles in their households. Physical activity was measured by 7 days of monitoring with Actigraph accelerometers, and mean daily minutes of moderate-to-vigorous physical activity (MVPA) were computed. A neighborhood walkability index was calculated for a 1-km radius around each participant's home based on the street network using Geographic Information Systems. Multiple regressions were conducted separately for low- and high-income participants to explain MVPA, controlling for age and sex. RESULTS Of the 624 low-income participants, 8.5% had no car. Of the 659 high-income participants, 1.8% had no car. The median of cars per household for each group was 2.0. Interactions between walkability and car availability were not significant for low- (p<.15) or high-income (p<.46) participants. In regressions without interactions for the high-income group, walkability explained 3.6% of the variance (p<.001), but car ownership explained only .003% (p<.90). Results for the low-income group showed walkability explained 1.2% of the variance (p<.01), and there was a trend for car ownership to be negatively related to MVPA (.006% of variance, p<.06). CONCLUSIONS The hypothesized moderating effect of car availability on the relation between walkability and MVPA was not supported for low- or high-income adults. However, when car availability was included in the regression model for lowincome adults, it was marginally significant, indicating further study of this variable is warranted. In contrast to previous analyses, when car availiability was in the model for low-income adults, walkability was significantly related to MVPA. Different patterns of associations between environmental variables and MVPA may be operating for different population subgroups. Supported by NIH grant HL67350.