Disparities in the Receipt of Local Treatment of Node-positive Prostate Cancer.

Abstract

We sought to determine whether any sociodemographic disparities are present in the receipt of local treatment for node-positive prostate cancer. We identified 9771 patients with clinical N1M0 prostate cancer diagnosed from 1998 to 2012 using the National Cancer Database. We used multivariable logistic regression modeling to identify patient-specific factors that were associated with a reduced likelihood of receiving prostate or pelvic radiation or radical prostatectomy. We also used Cox regression modeling to estimate the differences in overall survival (OS) using these factors. The factors associated with a reduced likelihood of receiving local therapy included black race versus white race (43.6% vs. 49.4%; adjusted odds ratio [AOR], 0.76; P=.001), bottom income quartile versus top quartile (44.7% vs. 52.7%; AOR, 0.69; P= .001), age > 66 years versus≤ 66 years (40.8% vs. 55.1%; AOR, 0.48; P< .001), diagnosis before 2005 versus after 2005 (30.5% vs. 61.7%; AOR, 0.66; P< .001), and Medicaid or no insurance versus private insurance (41.0% vs. 49.4%; AOR, 0.41; P< .001). Although patients had reduced 5-year OS if they were from lower income quartiles (adjusted hazard ratios [AHRs], 1.18-1.22; P< .05), were older (AHR, 1.82; P< .001), or had Medicaid or no insurance (AHR, 1.24; P= .032), these disparities were no longer present or were smaller in magnitude after adjustment for receipt of local treatment. Significant treatment disparities exist in the receipt of local therapy for node-positive prostate cancer. Given the accumulating evidence supporting this practice, the factors underlying these disparities should be studied and addressed.