Introduction: Recently, there has been a proliferation of new approved treatments for the management of prostate cancer (PCa), including numerous second-generation anti-androgens (SGAAs). Despite improving patient outcomes, SGAA use typically leads to resistance and prevents effective rechallenges. Consequently, the optimal timing of SGAA use is an open question. Although there are health economic analyses of novel PCa treatments for specific patient states, there is a lack of analogous dynamic analyses. Our paper aims to fill this gap.

Methods: We developed a Monte Carlo Markov model to simulate the management of advanced PCa to end-of-life. We modeled patients who begin in metastatic and non-metastatic hormone-sensitive PCa (mHSPC and nmHSPC), with risk stratification for mHSPC, progressing to metastatic and non-metastatic castration-resistant PCa (mCRPC and nmCRPC). Using current Canadian guidelines and reimbursement restrictions, we simulated all admissible treatment sequences over these states over a 15-year horizon and compared outcomes for each sequence.

Results: We find and report the best treatment sequences over a 15-year horizon for a variety of health outcomes as well as net health benefits (NHB), expressed as quality-adjusted life years (QALYs) minus costs for a range of willingness to pay (WTP) values. We find that early SGAA use delivers the best health outcomes for all patients as well as the best NHB for mHSPC-starting patients at moderate to high WTP (NHB gains: 0.93–1.5 QALYs at C$100k–150k WTP). However, early SGAA use is comparable to late SGAA use and even no SGAA use for nmHSPC patients at all WTP (NHB gains: -0.27–0.18 QALYs at C$50–150k WTP) and mHSPC patients at low WTP (NHB gain: 0.09–0.32 QALYs at C$100–150k).

Conclusions: We conclude that both from a health and health-economic perspective, there a wide range of treatment strategies which deliver near-best average patient outcomes. Broadly, SGAAs are more effectively used during early stages of PCa, but not using SGAAs at all is near-optimal for nmHSPC patients and/or cost-constrained payers.