MEASURING THE BENEFIT OF PSA SCREENING FROM THE PATIENT AND SOCIETAL PERSPECTIVES

Purpose: To measure the potential benefit of prostate specific antigen (PSA) screening from the patient and societal perspectives.

Method: We use a POMDP model that represents the progression of patients through health states of no prostate cancer (PCa), PCa not detected, PCa detected and treated, and death. Decisions to conduct a PSA test, biopsy, and treat the patient are made annually. The probability that the patient has PCa is estimated using Bayesian updating based on all available PSA test results as the patient ages. The model is solved to determine the optimal combined PSA testing schedule and biopsy referral policy from the patient and societal perspectives. The patient perspective seeks to maximize expected quality adjusted life years (QALYs). The societal perspective maximizes rewards for QALYs, based on a societal willingness to pay, minus costs of PSA tests, biopsies and treatment. Transition probabilities are estimated using a dataset of 11,872 men residing in Olmstead County, MN from 1988 to 2006. Other parameters are drawn from secondary sources in the medical literature. We measure the incremental benefit of the optimal policy over the traditional guideline of regular annual PSA screening with biopsy referral for a PSA value exceeding 4.0 ng/mL.

Result: The optimal policy from the patient perspective suggests annual PSA screening of the population until age 76. The optimal policy from the societal perspective suggests selective PSA screening based on the patients age and probability of having PCa. The societal perspective suggests that all PSA screening should be stopped after age 71 independent of the probability of PCa. Based on 2006 U.S. population estimates, the optimal policy from the patient perspective yields an annual incremental benefit, over the traditional guideline, of 134,000 QALYs for the U.S. population. From the societal perspective the annual incremental benefit is $U.S. 865 million based on a willingness to pay of $U.S. 50,000/QALY.

Conclusion: Optimal PSA screening of the population from the societal perspective suggests screening should be stopped 5 earlier than from the patient perspective. Regular annual PSA screening is optimal from the patient perspective; selective screening based on the probability of PCa is optimal from the societal perspective. Evaluation of PSA-based screening based on the traditional guideline underestimates the potential benefits of PSA-based screening.