[Estimating standard performance measures of opportunistic screening for prostate cancer]

Abstract

Objective: to estimate the standard performance measures of prostate-specific antigen (PSA) screening as implemented in the existing health care system and to compare the observed results with those fom the European Randomized Screening for Prostate Cancer (ERSPC) trial.

Design: in a consecutive series of men living in two Italian health districts, aged > or = 30 years, with at least one total PSA test between 2000 and 2003, those subjects putatively tested for screening purposes were identified using record linkage with multiple local health databases. This was also used to determine the outcomes of follow-up of subjects with positive test result (PSA > or = 4 ng/ml) at 12 and 24 months of observation.

Setting: clinical chemistry laboratories, outpatient urology clinics, pathology departments, and mortality registries of the health districts of Ravenna and Forlì, and the Romagna Cancer Registry.

Participants: 52,513 total subjects, 42,398 of whom putatively tested for screening purposes.

Main outcome measures: the most common performance measures of cancer screening.

Results: the 2-year screening rate increased until 80 years of age. In the age range 55-69 years, i.e. the target age of the ERSPC trial, the screening rate was 38.1%, and the rate of positive test results (9.1%) was within the expected range. The PSA level, but not the subjects age, was strongly associated with follow-up procedures. After 24 months of observation, 24.4% PSA-positive patients received no further evaluation, 54.8% underwent repeat testing as the initial follow-up action, and 44.0% were referred for urologic assessment. The biopsy rate was 25.3%, the positive predictive value of PSA testing 10.1%, the detection rate of prostate cancer 9.4 per thousand, and the detected/expected ratio 8.4.

Conclusion: compared with methods and results of the ERSPC trial, the management of PSA-positive subjects was much more conservative, and the yield of disease much smaller.