Prostate cancer screening: a decision analysis

Abstract

Background: The issue of whether to screen men for prostate cancer is controversial. No randomized clinical trials have been completed to confirm the efficacy of screening for prostate cancer. We created a mathematical model of the clinical risks and benefits of screening for prostate cancer.

Methods: A Markov decision-analytic model evaluated the outcomes of annually screening asymptomatic men for prostate cancer beginning at age 50 years. The screening and testing algorithm included the digital rectal examination, transrectal ultrasound, and prostate-specific antigen test. A sample of 10 male patients with no history of prostate disease were interviewed to assess their utilities (preferences) regarding the various adverse outcomes of prostate cancer treatment.

Results: The model indicated that no screening was preferred to screening when patients' utilities were considered (24.14 vs 23.47 quality-adjusted life years expected). The optimal decision was sensitive to the utilities of impotence and urethral stricture, the most common adverse outcomes for patients under the age of 65 years. When adverse outcomes of treatment were ignored, screening was favored (24.86 vs 24.22 years of life expectancy.

Conclusions: When quality-of-life preferences of men are considered, the annual screening of asymptomatic patients for prostate cancer is not recommended.