Ecology of poor health, good health: Constraints to developing a small area database for health disparities analysis

Many states, including Tennessee, have identified what they regard as the leading causes of death within their geo-space. However, collecting data on these leading causes at levels of geography below the county appears problematic. Despite evidence that the census track level of geography presents the best demonstration of the effect of socioeconomic inequality on health status, there exists no census track database for these leading causes in Tennessee or any state. Such a database would be a matrix with a column vector of heath status measures as the dependent variables and SES variables as the horizontal vector of independent variables. The independent variable set comes from US census data and the hierarchical structure of reporting these data makes them readily available. The problem is with the dependent variable vector which is the responsibility of state and local public health agencies, and sometimes, public health researchers in the respective public health communities throughout the nation. A superior-subordinate hierarchical structure for collecting and reporting these dependent variables would facilitate the development of small area databases. The 2006 National Health Disparities Report encourages that by suggesting “… neighborhood solutions” as “the key to achieving the elimination of health disparities” and calling for “…focused community-based projects that are supported by detailed local data”. The objective of this study is to discuss some of the structural constraints to this small area approach, based on a survey of Tennessee state and Davidson county public health departments.