The built environment and physical activity levels: the Harvard Alumni Health Study

Physical activity is associated with better health, but many individuals are insufficiently active. Modifying the built environment may be an approach capable of influencing population-wide levels of physical activity, but few data exist from longitudinal studies that can minimize bias from active people choosing activity-friendly neighborhoods.

This study aims to examine cross-sectional and longitudinal associations between the built environment and physical activity on a large scale.

This study examined cross-sectional associations between urban sprawl (mapping addresses to corresponding counties) and physical activity (self-reported) among men throughout the U.S. in 1993 and in 1988, and longitudinal associations between changes in exposure to urban sprawl for movers and physical activity, 1988-1993. Included were 4997 men (mean age, 70 years) in the 1993 cross-sectional study; 4918 men in the 1988 cross-sectional study; and 3448 men in the longitudinal study, 1988-1993. Data were collected prospectively in 1988 and 1993, and analyses were performed in 2007-2008.

In cross-sectional analyses, less sprawl was significantly associated with more walking OR, comparing least with most sprawling areas, for meeting physical activity recommendations by walking=1.38 [95% CI=1.09, 1.76] in 1993 and 1.53 [1.19, 1.96] in 1988). Less sprawl also was associated with lower prevalence of overweight (corresponding OR=0.79 [0.64, 0.98] in 1993 and 0.81 [0.66, 1.00] in 1988). However, longitudinal analyses assessing change did not show that decreasing sprawl was associated with increased physical activity or decreased BMI.

These findings suggest that the cross-sectional results may reflect self-selection, rather than indicating that the built environment--as measured by urban sprawl--increases physical activity. However, the longitudinal findings were limited by small numbers of men changing residence and associated sprawl levels.