Residential racial segregation, or the relative degree to which members of two or more racial groups are spatially distributed throughout a defined geographical area, is associated with the health status of individuals living in racially segregated communities. Despite pronounced racial disparities in prostate cancer incidence and mortality, no known study has investigated the association between racial residential segregation and prostate cancer post-diagnosis treatment. The objective of this study is to evaluate the independent effect of residential racial segregation as measured by the isolation index on prostate cancer treatment decisions. Incident prostate cancer cases among men living in metropolitan statistical areas that were reported to the SEER-Medicare linked database between 1995 and 2002 were identified. Differences between individual and area-level characteristics at varying degrees of residential segregation were statistically evaluated using Chi Square and Student's t tests and ANOVA. Hierarchical logistic regression models with MSA- level random effects were used to characterize the relationship between level of residential segregation and post-diagnosis prostate cancer treatment received. African American men living in metropolitan areas with high segregation were statistically significantly more likely to receive expectant management (OR=1.42, 95% CI= 1.09, 2.23) and were statistically significantly less likely to receive radiation therapy (OR=0.56, 95% CI=0.34, 0.93) compared to white men living in areas of low segregation. Black men living in metropolitan areas with high segregation were more likely to receive radical prostatectomy (OR=1.59, 95% CI= 0.57, 4.45) and hormone therapy (OR= 2.08, 95% CI=0.91, 4.77) than white men living in areas of low segregation, although these differences were not statistically significant. Researchers should continue to investigate the implications of access to care and healthcare utilization by various populations and their effect on prostate cancer treatment decisions.

Citation Information: Cancer Epidemiol Biomarkers Prev 2010;19(10 Suppl):B71.