Determinants of County-level Stroke Mortality in the Southeastern United States, 1999--2002

• Activity Number: 521
• Type: Contributed
• Date/Time: Thursday, August 11, 2005 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Eric Tassone*+ and Lance A. Waller
• Companies: Emory University and Emory University
• Keywords: Bayesian disease mapping ; determinants of stroke ; hierarchical models ; small area analysis

Abstract

We estimate county-level rates of stroke mortality and associations with determinants using a hierarchical Bayesian disease mapping model that permits inclusion of county-level covariates of stroke, something not possible with many spatial smoothing methods. The study population was 35-64 year olds in the southeastern United States during 1999--2002. Stroke deaths were defined as those for which the underlying cause of death was classified as I60-I69, according to the International Classification of Disease, 10th Revision. County-level maps of median stroke death rates per 100,000 person-years of risk are presented for men and women. Variability estimates associated with the county-level rates are presented separately. Results of the Poisson regression modeling inherent in this Bayesian approach indicate both similarities and differences in the covariates of stroke mortality by geography and gender. The findings highlight geographic disparities across counties in the magnitude of stroke burden and quantify the magnitude and statistical significance of the selected covariates while allowing for state-level and gender differences in the impacts of the covariates.