The over-diagnosis and over-treatment of prostate cancer is a major public health concern, and in 2012 the United States Preventive Services Taskforce (USPSTF) recommended against prostate cancer screening. Prostate cancer is usually detected by performing a prostate biopsy. Previously, many men received a biopsy at the first sign of an elevated cancer risk identified by screening. Currently, physicians have more tools are their disposal to select men for biopsy who are likely to have clinically significant cancers, including repeat prostate specific antigen (PSA) testing, PSA density, PSA velocity, PSA free/total ratio, and age-specific cutoffs. These tests allow physicians to reduce the number of unnecessary biopsies performed on lower risk patients. One would expect that the use of these tests, in addition to more selective screening, would decrease the incidence of prostate biopsies in the population.
I hypothesized that in the last 10 years in Ontario: 1) the incidence of prostate biopsy has decreased, 2) the proportion of biopsies that are malignant has increased, and 3) patients receiving biopsies are healthier.
I performed a secondary analysis of population-based administrative databases. I validated the prostate biopsy procedure code in the Ontario Health Insurance Plan (OHIP) then used this code to create a cohort of Ontario men who received their first prostate biopsy between 1992 and 2012. Crude and age standardized incidence rates of prostate biopsy were determined for each study year. Era-specific inter-censal population estimates from Statistics Canada were used to establish the number of men at risk of biopsy each year. Changes over time in prostate biopsy incidence were examined using negative binomial regression by comparing the biopsy incidence of each year to a referent year expressed as incident density ratios. Similar analyses were performed to examine changes over time in the proportion of biopsies that are malignant and the health status of patients receiving biopsy. Health status was determined by calculating the Aggregated Diagnosis Group (ADG) score for each patient.
The sensitivity of the OHIP prostate biopsy code improved during the study period and was approximately 90% in recent years. The specificity for identifying the first prostate biopsy a patient received was estimated to exceed 95%. The crude and age standardized incidence of prostate biopsy in Ontario gradually increased between 1992 and 2007 and then dropped sharply in 2008 and 2012. Overall, 39% of biopsies were malignant but this proportion increased during the study period. The health status of patients receiving biopsy, as measured by the ADG score, improved over the study period.
This is the first study to report crude and age standardized prostate biopsy incidence in a population. We found that previously rising biopsy rates decreased significantly in 2008 and 2012 in conjunction with changes to the perceived utility of prostate cancer screening. More years of follow up are required to determine if these changes were transient or the start of broad practice changes.|