[Skip to Content]
[Skip to Content Landing]
September 14, 1994

Screening for Prostate Cancer: A Decision Analytic View

Author Affiliations

From the Departments of Medicine (Drs Krahn and Detsky), Clinical Biochemistry (Dr Krahn), Health Administration (Dr Detsky), and Urology (Drs Mahoney and Trachtenberg), University of Toronto (Ontario); the Clinical Epidemiology Unit (Drs Krahn and Detsky) and the Prostate Centre (Drs Krahn, Mahoney, and Trachtenberg), The Toronto (Ontario) Hospital; Wellesley Hospital, Toronto, Ontario (Dr Mahoney); and the Division of Clinical Decision Making, Department of Medicine, New England Medical Center and Tufts University School of Medicine, Boston, Mass (Drs Eckman and Pauker).

JAMA. 1994;272(10):773-780. doi:10.1001/jama.1994.03520100035030

Objective.  —To determine the clinical and economic effects of screening for prostate cancer with prostate-specific antigen (PSA), transrectal ultrasound (TRUS), and digital rectal examination (DRE).

Design.  —Decision analytic cost-utility analysis comparing four screening strategies with a strategy of not screening. We assumed that the cancer detection rate and stage distribution were predicted by each combination of tests and that localized cancer was treated with radical prostatectomy. For each strategy, we calculated life expectancy, quality-adjusted life expectancy (QALE), and cost-utility ratios for unselected and high-prevalence populations.

Data.  —Probabilities and rates for clinical events were gathered from published data. We assessed utilities by the time—trade-off method using urologists, radiation oncologists, and internists as subjects. The Clinical Cost Manager at the New England Medical Center provided cost data.

Results.  —In unselected men between the ages of 50 and 70 years, screening with PSA or TRUS prolonged unadjusted life expectancy but diminished QALE. Screening with DRE alone yielded no reduction in mortality at any age. All programs increased costs. Results were sensitive only to assumptions about the efficacy of treatment. In high-prevalence populations, screening produced a similar pattern: gains in unadjusted life expectancy, losses in QALE, and increased costs.

Conclusions.  —Our analysis does not support using PSA, TRUS, or DRE to screen asymptomatic men for prostatic cancer. Screening may result in poorer health outcomes and will increase costs dramatically. Assessment of comorbidity, risk attitude, and valuation of sexual function may identify individuals who will benefit from screening, but selecting high-prevalence populations will not improve the benefit of screening.(JAMA. 1994;272:773-780)