Disparities in prostate cancer treatment associated with population density of the county of residence.

117 Background: We sought to assess disparities in primary treatment of prostate cancer by examining the effect of population density of the county of residence on treatment for clinically localized prostate cancer and quantify variation in primary treatment attributable to county and SEER site. Methods: Participants included 138,226 men with clinically localized prostate cancer who underwent surgery, radiation, or conservative therapy in the Surveillance, Epidemiology, and End Result (SEER) database diagnosed in 2005 through 2008. Mixed effects hierarchical logit models were used to examine the effect of population density of the county of residence on treatment while accounting for the random effects of counties nested within SEER sites. In order to quantify the effect of county and SEER site on individual treatment, the percent of total variance in treatment attributable to county of residence and SEER site was estimated with residual intraclass correlation coefficients. Results: Men in metropolitan counties had 23% higher odds of being treated compared with men in rural counties, controlling for number of urologists per county as well as patient clinical and sociodemographic characteristics. Age, race/ethnicity, PSA at diagnosis, and tumor characteristics were also significant predictors of treatment in the multivariable model. The percent of variation in treatment attributable to the county of residence was consistently higher than that attributable to SEER site. 3% (95% CI: 1.2%-6.2%) of the total variation in treatment was attributable to SEER site, while 6% (95% CI: 4.3%-9.0%) of variation was attributable to county of residence, controlling for clinical and sociodemographic characteristics. This variation at the county and SEER site increased to 4% and 10%, respectively, when specific type of treatment (surgery or radiation) was assessed. Conclusions: Variation in treatment for localized prostate cancer exists for men living in different population dense counties of the country. These findings highlight the need for comparative effectiveness research and health policy initiatives to advance prostate cancer research and reduce disparities in current clinical practice