Cost–Utility Analysis of Treatment Sequences for Moderate-to-Severe Crohn’s Disease

Objectives The clinical pathway for patients with moderate-to-severe Crohn’s disease (CD) typically includes sequential pharmacologic treatment as well as surgery, but positioning of different therapies within these sequences remains challenging. Cost–utility analysis rarely captures these sequences and does not incorporate registry data on long-term effectiveness. In this study, we aim to overcome these limitations. Methods We developed an individual state transition model with four health states (active disease, remission, and remission due to surgery and death), five sequential treatment lines, and surgery. Efficacy data from network meta-analyses (NMA) for biologic naive and biologic exposed patients were combined with Dutch registry data to forecast long-term benefit, calculate costs, and estimate utilities. Analyses had a Dutch societal perspective with a lifetime time horizon. Costs were reported in 2023 euros and discounted with 3%. Effects were reported in quality-adjusted life years (QALYs) and discounted with 1.5%. The cost-per-QALY threshold was €20,000. Deterministic analyses for the base case, three scenarios (including recently published trials or price declines for ustekinumab), and one-way sensitivity analysis were run with 30,000 patients. The probabilistic sensitivity analysis was conducted by sampling 1000 patients in 1000 model runs. Results When opting for step-up sequences, the most cost-effective sequence (out of 156 sequences) starts with either azathiopurine/6-mp or methotrexate and is followed by combination therapy (infliximab + azathioprine) when patients discontinue their first line owing to disease activity or discontinuation. The most cost-effective top-down sequence (out of 72) starts with combination therapy (infliximab + azathioprine). After two lines of treatment, differences in cost-effectiveness between biologics become smaller. To be equally cost-effective as anti-tumor necrosis factor (TNF) combination therapy, a price decline for ustekinumab (biosimilars) of 81% is required or 50% to become the preferred option after combination therapy. Validation against external data suggested good predictive capabilities of the model. Conclusions Integrating NMA and registry data improves the quality of cost-effectiveness models for treatment sequences in CD. This open-source model can be easily updated for future therapies and holds the potential to become a standard model for use in clinical guideline development and the economic evaluation of new drugs.