Estimating the healthcare costs of treating prostate cancer in Australia: A Markov modelling analysis

Abstract

Purpose: To estimate the health system costs of prostate cancer by disease risk category and treatment type over 2016 to 2025 and to identify potential strategies to contain the cost increase.

Methods: A Markov cohort model was developed using clinical pathways from US prostate cancer guidelines and clinical expertise. Estimates of the probabilities of various treatments and outcomes and their unit costs were sourced from systematic reviews, meta-analyses, epidemiological publications and national cost reports. Estimated costs by stage of disease, by major treatments and by age at diagnosis were reported in 2016 US dollars. One-way and probabilistic sensitivity analyses assessed potential variation in the modeled costs.

Results: Australia-wide costs of prostate cancer were estimated at US$270.9 million in 2016 rising to US$384.3 million in 2025, an expected increase of 42%. Of this total increase, newly diagnosed low risk cases will contribute US$32.9 million, intermediate-risk US$56.8 million, high-risk US$53.3 million and advanced US$12.6 million. For men diagnosed at age 65 with low-risk disease, lifetime costs per patient were US$14,497 for surgery, US$19,665 for radiation therapies to the primary lesion, and US$9,234 for active surveillance. For intermediate- or high-risk disease, mean costs per patient were US$34,941 for surgery plus radiation and US$31,790 for androgen deprivation therapy plus radiation while advanced cancer therapies were at US$31,574 per patient. Additional costs for managing iatrogenic disease secondary to these treatments were excluded.

Conclusion: Strategies for identifying patients early before cancers have spread are critical to contain the estimated 42% increase in costs over the next decade. Increased uptake of active surveillance would also lead to substantial cost-savings in the management of low-risk prostate cancer.

Keywords: Active surveillance; Cost analysis; Markov modeling; Prostate cancer.