OBJECTIVES: Rapid advancements in transformative therapies have shown that the prospect of cure is becoming a reality in disease areas where treatments had previously focused on delaying progression. Mixture cure models (MCMs) have been used increasingly to capture curative, long-term outcomes in health economic models for these therapies. We examined MCM use in published NICE technology appraisals (TA), and considered their respective submissions to other HTA agencies.

METHODS: A systematic review was conducted for NICE TAs published between June 2021–May 2024. All appraisals that considered MCMs were included. A targeted snowball approach was used to identify relevant TAs prior to June 2021. The following items were extracted: model structure, rationale and endpoints for MCM analysis, and methodological critique. A targeted review was subsequently performed on Canadian Drug Agency (CDA) submissions for a comparison of findings.

RESULTS: 275 TAs were identified during the search period, of which 7 (2.6%) reported use of MCMs. Prior to 2021, 8 further TAs were identified and subsequently included. Of the 15 TAs included in the review, 10 (66.7%) were in hematological cancers, 2 (13.3%) were in gynecological cancers, and 1 (6.7%) was in urolithel cancer, non-small-cell lung cancer, and melanoma respectively. All TAs used a partitioned survival model framework. There was variation in MCM application (cure point and survival endpoints). The main rationale for using MCM was the curative intent of the intervention, which corroborated strongly with observed survival data. Majority of the submissions received critique on the use of MCMs due to short-term follow-up, where the resulting uncertainty may have considerable impact on decision making. In their respective submissions to CDA, MCMs were not always submitted nor utilized.

CONCLUSIONS: Use of MCMs within HTA submissions may be applicable for treatments with curative intent, but methodological conduct and acceptance remains heterogenous among developers and HTA bodies.