Presentation (CTI)

OBJECTIVES: Health economic models often assume continuous treatment use over long time horizons; however, in real-world clinical practice, patients frequently discontinue treatment due to adverse events, lack of efficacy, or personal preference. Ignoring treatment discontinuation may bias long-term cost-effectiveness estimates. This study evaluated the impact of alternative treatment discontinuation methods on lifetime cost-effectiveness outcomes.
METHODS: A cohort-based Markov model with annual cycles and a lifetime horizon (40 years) was developed from a healthcare payer perspective. Two treatment strategies were compared: (1) continuous treatment until death and (2) treatment with an annual discontinuation risk and three health states included in the model were: on-treatment, post-discontinuation, and death. Treatment discontinuation was modelled as a constant annual hazard, assuming an exponential process over time. Annual discontinuation rates (15%), mortality risks, utilities, and costs were taken from the published literature. Post-discontinuation was associated with higher mortality and lower health-related quality of life. Costs and QALYs were discounted at 3% annually. Scenario analyses compared no discontinuation versus discontinuation, and one-way sensitivity analyses assessed the influence of discontinuation rates and post-discontinuation outcomes.
RESULTS: Assuming continuous treatment resulted in higher lifetime QALYs (12.3 vs. 8.2), and higher total costs ($230,722 vs. $92,490) compared with models incorporating treatment discontinuation. Accounting for treatment discontinuation reduced health benefits by one third and the resulting incremental cost-effectiveness ratio was approximately $33,700 per QALY gained, which was substantially higher than estimates assuming continuous treatment. Sensitivity analyses showed that higher discontinuation rates and poorer post-discontinuation health outcomes had the greatest impact on increasing ICERs.
CONCLUSIONS: Adding treatment discontinuation in health economic models can significantly affect cost-effectiveness estimates and post-discontinuation health outcomes leads to more conservative and realistic results. Economic models developing for HTA submissions should transparently justify discontinuation assumptions and explore alternative scenarios to support robust decision making.