A Markov model for optimal intervals in prostate cancer screening programs in the era of individual baseline PSA values

Abstract

4618 Background: Once a prostate cancer screening program using prostate-specific antigen (PSA) has been established and is operational, the issue of the re-screening interval for patients with prior negative PSA results comes into question. We approached this issue using a decision-analytic model in which participants were stratified by baseline PSA values. Methods: Based on an actual contemporary screening program, we established a Markov decision analytic model of prostate cancer screening with different re-screening interval strategies determined by individual baseline PSA levels. Cost-effectiveness analysis was performed to determine the incremental cost-effectiveness ratio (ICER; costs per quality-adjusted life years) for each strategy. Results: Re-screening strategies with biennial PSA testing in men with PSA levels ≤3.0 and ≤4.0 ng/ml are dominated by biennial screening in men with PSA ≤2.0 ng/ml, which is the most cost-effective strategy. ICERs for annual screening and biennial screening strategies in men with PSA ≤1.0 were US$25,220 and US$10,557, respectively, with respect to strategies with biennial screening in men with PSA ≤2.0 ng/ml. On sensitivity analyses, ICER for annual screening with respect to biennial screening in men with PSA ≤2.0 ng/ml will be larger under conditions with a lower biopsy participation rate in PSA-positive participants, or a lower cancer detection rate on biopsy. It was also observed that ICER for an annual screening strategy is larger in an initial age cohort of men in their 50’s and 70’s than in their 60’s. Conclusions: Our model shows that cut off of baseline PSA for annual PSA testing in prostate cancer mass screening should be determined as 0.0 to 2.0 ng/ml based on clinical and economical variables in view of the budget of each society. No significant financial relationships to disclose.