Background: To evaluate the various treatments for patients with hormone-refractory prostate cancer with bone metastases using a Markov model.
Methods and materials: The base case to be evaluated was of a man with hormone-refractory prostate cancer. The evaluated palliative treatments were pain medication only, chemotherapy consisting of mitoxantrone and prednisone, and single- and multifraction radiotherapy (RT). A literature search was used to generate the transition probabilities and patient utilities. Modeling was used to generate the cost estimates. Expert opinion was used to generate utilities and cost estimates in the absence of literature data. Second-order Monte Carlo simulation produced incremental cost-effectiveness scatterplots and 95% confidence ellipses.
Results: Pain medication had the least expected mean cost of all the treatment options at 11,700 US dollars but also the second lowest quality-adjusted survival at 5.75 quality-adjusted life months. Chemotherapy had the highest expected mean cost, 15,300 US dollars, but the lowest quality-adjusted life months (4.93). Incremental cost-effectiveness analysis revealed that single-fraction RT was the most cost-effective treatment, with a cost of 6,857 US dollars/quality-adjusted life year; multifraction RT had an incremental cost-effectiveness ratio of 36,000 US dollars/quality-adjusted life year. Chemotherapy was dominated by pain medication.
Conclusion: Within the limits of the established model, single-fraction RT was the most cost-effective palliative treatment compared with pain medication, chemotherapy, and multifraction RT. The use of this model allowed comparison of different treatment regimens that could never be evaluated together in a randomized clinical trial.