Measuring and Mapping Small Area Racial Disparities in Heart Disease

In light of the increasing interest in documenting and eliminating racial and ethnic disparities in health, a pressing need exists for appropriate methods to measure these disparities at the local level. The authors measured county-level disparity in heart disease mortality between African Americans and whites in South Carolina during 1996– 2000, modelled via an extension of a spatial hierarchical Bayesian disease mapping model derived for incidence proportions. County-level maps of median disparity appear with figures displaying associated variability estimates. Race-specific maps of the burden of heart disease enable the reader to evaluate race-specific contributions to the disparity measure for each county. The findings highlight geographic variations across counties in magnitude of excess burden of heart disease mortality borne by African Americans relative to whites and indicate the largest racial disparities in heart disease mortality often occur in counties containing metropolitan or economically developed areas. Other maps and figures illustrate the versatility of the disparity measure. The authors discuss the merits and potential liabilities of the disparity measure and provide suggestions for wider application of such Bayesian techniques to the measurement of small area racial disparities, such as identifying community-level covariates of the racial disparities in heart disease mortality. Bayesian smoothing; health disparities; hierarchical Bayesian disease mapping; rate standardization; small area analysis; standardized mortality ratio (SMR). (Word counts: Abstract, 193 words; Text, 3911 words.) Abbreviations: SMR, Standardized Mortality Ratio; MCMC, Markov chain Monte Carlo.