Cure Models for Health Technology Assessment: Can They Be Trusted for Decision-Making?

ISSUE:

New treatments for cancer and other diseases increasingly appear to offer the potential for cure. But enforcing an assumption of cure when fitting models to data from clinical trials is associated with substantial uncertainty due to small sample sizes and limited follow-up periods. Therefore, health technology assessment (HTA) decision-makers may be reluctant to utilise cure models. This Issue Panel will debate whether we should be more open to using cure models to inform HTA decision-making, and whether new methodological developments mean we can place greater trust in these models.

OVERVIEW:

Parametric survival models are used to extrapolate beyond clinical trial follow-up periods to estimate the long-term survival benefits associated with new health technologies. ‘Standard’ parametric models are likely to extrapolate poorly if some patients are cured. For new treatments that are potentially curative, such as immunotherapies and chimeric antigen receptor (CAR) T-cell therapies, using cure models to extrapolate survival is logical. However, cure models may estimate cure fractions inaccurately, leading to biased estimates of treatment effects and cost-effectiveness, especially when fitted to data with limited follow-up and small sample sizes.

Dr John will introduce the session (2 minutes). Professor Latimer will present flexible parametric non-mixture cure models, arguing that these offer advantages over more commonly used models (11 minutes). Dr Felizzi will focus on novel ‘informed’ mixture cure models (11 minutes). Dr Whittington will draw on published examples of cure models used to support HTA decision-making, discussing their practical and technical strengths and limitations (11 minutes). Dr John will summarise the panel presentations, offering her perspective as a NICE Appraisal Committee Chair, and a general practitioner (8 minutes), and will lead the debate and discussion with the audience (17 minutes).

Stakeholders with an interest in the assessment of potentially curative treatments will find this session interesting and stimulating.