A hierarchy of sociodemographic and environmental correlates of walking and obesity.

Objectives: Initial studies demonstrate the need for further investigation of how the association of built environment with physical activity and BMI may differ by sociodemographic subgroups. The aim of this study was to use a novel statistical technique to identify possible subgroups.

Methods: Data from the 2002 Strategies for Metro Atlanta's Regional Transportation and Air Quality (SMARTRAQ) study were analyzed to explore relationships between measures of residential density, street connectivity, land use mix, and sociodemographic characteristics of individuals in predicting walking, overweight and obesity status. Chi-squared Automatic Interaction Detector (CHAID) analyses were used to partition the population into subgroups (N=13,065).

Results: Subgroups, were more likely to walk if they lived in neighborhoods with greater residential density, greater street connectivity and greater land use mix. A similar relationship was seen in men for the outcomes of obesity and overweight. Male residents of more walkable neighborhoods were less likely to be obese or overweight. In contrast, features of walkability were related to higher rates of obesity and overweight in women and non whites.

Conclusions: These analyses reveal that gender and ethnic subgroups display substantially different weight outcomes across different levels of walkability. In contrast, walking was consistently higher for all groups in the more walkable neighborhoods. This information can contribute to better targeting of interventions, and calls for more detailed investigation of the moderators that affect weight and physical activity across subgroups. This information supports a more efficient use of scarce resources to promote physical activity and healthy body weight.