Association of neighborhood economic status with adults' physical inactivity in Los Angeles County

Background: Physical inactivity is a major risk factor for obesity and diabetes. We hypothesized that adjusting for other predictors including individual-level income, neighborhood economic status (NES), measured as median income per capita within the neighborhood is independently associated with physical inactivity in adults.

Methods: We used three cross-sectional Los Angeles County Health Survey samples of adults (ages 18+): 2002 (n=8050), 2005 (n=8484) and 2007 (n=7140), linked with 2000 census data by zip code. We dichotomized the self-reported physical inactivity measure: 1= minimal and no activity in a usual week and 0= some vigorous or moderate activity. NES was categorized into three strata; low (<$20,683), median ($20,683-$41,366) and high (>$41,366). To achieve adequate statistical power, we applied ArcGIS to merge small zipcode areas into large regions. Multi-level logistic regression was used in each cross-sectional dataset.

Results The odds ratios with median NES level for the three datasets (2002, 2005 and 2007 respectively) after adjusting for individual and neighborhood-level variables were the following: 0.92 (95% confidence interval 0.77 to 1.09), 0.95 (0.81 to 1.12) and 0.90 (0.75 to 1.09). The odds ratios with high NES level were 0.79 (0.60 to 1.04), 0.81 (0.63 to 1.04) and 0.76 (0.57 to 1.01). The P-values with continuous NES measure were statistically significant (P<0.01) except for the 2007 dataset (P=0.086).

Conclusions We found a moderately negative association of neighborhood economic status with individuals' physical inactivity in LA County. Interventions promoting physical activity should target low-income neighborhoods.