Relative Survival Modeling for Appraising the Cost-Effectiveness of Life-Extending Treatments: An Application to Tafamidis for the Treatment of Transthyretin Amyloidosis with Cardiomyopathy

Robert Young, Jack Said and Sam Large
Additional contact information
Robert Young: Health Economics and Outcomes Research Ltd, Cardiff, UK
Jack Said: Pfizer Ltd, Tadworth, UK
Sam Large: Pfizer Ltd, Tadworth, UK

Medical Decision Making, 2025, vol. 45, issue 6, 726-739

Abstract: Background Economic evaluations for life-extending treatments frequently require clinical trial data to be extrapolated beyond the trial duration to estimate changes in life expectancy. Conventional survival models often display hazard profiles that do not rise as expected in an aging population and require the incorporation of external data to ensure plausibility. Relative survival (RS) models can enable the incorporation of external data at model fitting. A comparison was performed between RS and “standard†all-cause survival (ACS) in modeling outcomes from the tafamidis for the treatment of transthyretin amyloid cardiomyopathy (ATTR-ACT) trial. Methods Patient-level data from the 30-mo ATTR-ACT trial were used to develop survival models based on parametric ACS and RS models. The latter was composed of an expected hazard and an independent excess hazard. Models were selected according to statistical goodness of fit and clinical plausibility, with extrapolation up to 72 mo validated against ATTR-ACT long-term extension (LTE) data. Results Information criteria were too similar to discriminate between RS or ACS models. Several ACS models were affected by capping with general population mortality rates and considered implausible. Selected RS models matched the empirical hazard function, could not fall below general population hazards, and predicted well compared with the LTE data. The preferred RS model predicted the restricted mean survival (RMST) to 72 mo of 51.0 mo (95% confidence interval [CI]: 46.1, 55.3); this compared favorably to the LTE RMST of 50.9 mo (95% CI: 47.7, 53.9). Discussion RS models can improve the accuracy for modeling populations with high background mortality rates (e.g., the ATTR-CM trial). RS modeling enforces a plausible long-term hazard profile, enables flexibility in medium-term hazard profiles, and increases the robustness of medical decision making. Highlights To inform survival extrapolations for health technology assessment, a relative survival model incorporating external data per the recommendations of the National Institute for Health and Care Excellence (NICE) Decision Support Unit was used in support of the NICE evaluation of tafamidis for treatment of transthyretin amyloid cardiomyopathy (ATTR-CM). Relative survival modeling allowed selection of a broader range of hazard profiles compared with all-cause survival modeling by ensuring plausible long-term predictions. Predictions from plausible relative survival models of overall survival in patients with ATTR-CM, extrapolated from the ATTR-ACT trial, validated very well to outcomes after a doubling of follow-up and demonstrated improved precision and accuracy versus parametric all-cause survival models.

Keywords: Transthyretin amyloid cardiomyopathy; relative survival modeling; survival extrapolation; NICE technology appraisal (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://journals.sagepub.com/doi/10.1177/0272989X251342459 (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:sae:medema:v:45:y:2025:i:6:p:726-739

DOI: 10.1177/0272989X251342459

Access Statistics for this article

More articles in Medical Decision Making
Bibliographic data for series maintained by SAGE Publications ().