UCSF

Purpose

Veterans Affairs (VA) health care system investigators perform large clinical trials in prostate cancer treatment but potential differences between VA and other patient cohorts have not been explored systematically.

Materials and methods

Cancer of the Prostate Strategic Urologic Research Endeavor is an ongoing observational database of men with prostate cancer, comprising 7,202 patients treated at 35 sites across the United States. Three sites that together contribute 241 patients are VA medical centers. Demographic and clinical characteristics were compared between all VA and nonVA patients in the database and a multivariate model was used to explore the interactions between ethnicity and VA status for predicting clinical characteristics.

Results

VA patients were 4 times as likely as nonVA patients to be black. They had lower income, less education and more co-morbidity at presentation (all comparisons p <0.0001). VA patients had higher risk disease. Mean serum prostate specific antigen at diagnosis was 20.1 vs 15.3 ng/ml for nonVA patients (p = 0.003). Mean Gleason score was 6.4 for VA patients vs 6.0 for nonVA patients (p <0.0001). Differing ethnic distributions explained the differences in prostate specific antigen between VA and nonVA patients. However, VA status, socioeconomic level and ethnicity independently predicted Gleason score. VA patients were more likely to undergo watchful waiting or primary hormonal therapy and less likely to receive definitive local treatment (p <0.0001).

Conclusions

Significant sociodemographic and clinical differences exist between VA and nonVA patients, which should be borne in mind when extrapolating the results of VA clinical trials to the general population. These observations require validation in larger patient cohorts.