Variation in Use of Prostate Biopsy Following Changes in Prostate Cancer Screening Guidelines.

Abstract

Prostate biopsy rates have paralleled decreasing prostate specific antigen screening rates since 2012. We hypothesized that biopsy rates and the change in rates since 2012 would vary considerably across hospital referral regions. Using Medicare data from 2012 through 2014 we identified prostate biopsies performed by physicians who performed 11 or more biopsies annually. We calculated annual biopsy rates and changes in rates from 2012 to 2014 across 306 hospital referral regions. We performed multivariable regression adjusting for factors associated with annual biopsy rates (eg percent of patients older than 75 who were screened with prostate specific antigen and percent of the population that was African American). We also estimated adjusted prostate biopsy rates and changes with time across regions. We identified 395,993 biopsies. The overall rates decreased from 11.68 biopsies per 1,000 men in 2012 to 10.23 per 1,000 in 2014 (-12.4%, p = 0.11). Biopsy rates were higher in regions in which a greater percentage of the population was African American (β = 0.810, 95% CI 0.235-1.384, p = 0.006), ambulatory surgical centers were available where biopsy could be performed (β = 0.892, 95% CI 0.108-1.676, p = 0.026) and prostate specific antigen testing occurred more frequently (β = 2.462, 95% CI 1.153-3.771, p <0.001). There was marked geographic variation in the adjusted average biopsy rate (median adjusted rate 9.08 biopsies per 1,000 men, IQR 7.65-10.76) and in the change in biopsy rates with time (median adjusted rate change -1.49 biopsies per 1,000 men, IQR -1.94--1.22 per 1,000). Since 2012 there has been considerable geographic variation in the performance of prostate biopsies as well as changes with time after prostate specific antigen recommendations changed. Characterizing the role of unmeasured patient and physician level factors is crucial to optimize the use and minimize the harms of prostate biopsy.