Variation in Use of Active Surveillance among Men Undergoing Expectant Treatment for Early Stage Prostate Cancer

Purpose: We examined variation in active surveillance use in Medicare eligible men undergoing expectant treatment for early stage prostate cancer. Materials and Methods: Using SEER (Surveillance, Epidemiology and End Results) and Medicare data we identified 49,192 men diagnosed with localized prostate cancer from 2004 through 2007. Of 7,347 patients who did not receive treatment (ie expectant management) within 12 months of diagnosis we assessed the prevalence of active surveillance (ie repeat prostate biopsy and prostate specific antigen measurement) vs watchful waiting across health care markets. We fit multivariable logistic regression models to examine associations of active surveillance with patient demographics, cancer severity and health care market characteristics. Results: During the study interval use of active surveillance vs watchful waiting increased significantly in patients treated expectantly from 9.7% in 2004 to 15.3% in 2007 (p < 0.001). Active surveillance was less common in older patients, those with high risk tumors and those with more comorbidities (each p < 0.001). Patients who were white and had higher socioeconomic status were more likely to receive active surveillance (each p < 0.05). After adjusting for patient and tumor characteristics significant differences in the predicted probability of active surveillance persisted across health care markets (range 2.4% to 30.1%). No significant variation in active surveillance use was associated with specific health care market characteristics, including intensity of end of life care, Medicare reimbursement or provider density. Conclusions: Active surveillance has been relatively uncommon in Medicare beneficiaries with localized prostate cancer. Its use relative to watchful waiting varies based on patient demographics, tumor severity and geographic location.