Area characteristics, community resources and spatial dynamics of self-care disability among older adults: An analysis of county-level variation

Older adults remain at the highest risk of disability compared to all other age groups. Geography-specific features can affect the disability process. In addition, these factors can impose constraints which may lead to spatial clustering of people with common characteristics. I conducted both standard and spatial linear regression models to assess the association between area characteristics and self-care disability. Data was merged from the 2000 US Census, the 2000 Area Resources File and the 2000 Uniform Crime Data. The sample consisted of the 3,109 counties in the contiguous United States. My outcome of interest was the county-specific proportion of adults 65 and older who report a self-care disability, defined as difficulty “dressing, bathing, or getting around inside the home”. I used principal component analysis to group 51 different area-level characteristics into such domains as environmental stressor, economic environment, and area resources measures. Exploratory spatial analysis indicated clear areas of spatial clustering such as the Southeast. According to the multivariable linear regression model, areas with environmental stressors, such as high poverty and crime, were associated with increased county-level of self-care disability. The general model fit improved when the spatial distribution was accounted for in a spatial error model. Estimates for the regression coefficients remained virtually the same (change <10%). Results indicate local area resource deprivation contributes to self-care disability, an effect that is stronger when such communities are spatially proximate. These findings may help us target interventions to better meet the needs of community-dwelling older adults with self-care disabilities.