Disparities in mortality across geographic scales in Texas

Public health researchers rely on statistics developed by multiple parties to understand the health of their target populations. Unfortunately, at the local level geographies, small sample sizes and large variations can render drawing conclusions extremely difficult. This presentation examines the impacts of aggregation at multiple levels of geography and examines differences in findings in terms of racial and ethnic disparities in all cause and cause-specific mortality in Texas based on different geographic levels of analysis.

Data for this study were obtained from the Census Summary Files (SF) 2 and 4 for the year 2000, the Rural Health Research Center Rural-Urban Commuting Area Codes (RUCA), Texas Department of State Health Services Border and Public Health Region definitions, and Texas Vital Statistics for Multiple Causes of Death. Data were aggregated by public health regions, counties, and census tracts. Death counts were analyzed using hierarchical modeling controlling for individual level demographics and geographic level factors including poverty, ethnic density, proximity to the border, and rural-urban status, as well as visually using geographic information science (GIS).  

Findings from thematic mapping indicate that non-Hispanic Whites have some of the highest rates of mortality for all causes across all levels. Non-Hispanic Blacks were largely at lower risk, overall, than non-Hispanic Whites, and Hispanics had the lowest rates across all groups. When using smaller-scale geospatial identifiers, results show greater variation in the needs and risk of different areas. This is especially true with census tracts where differences in rates can vacillate wildly in a small area.