Presentation (CTI)

OBJECTIVES: Recent HTAs of Alzheimer’s disease (AD) DMTs, lecanemab and donanemab, highlight key modelling challenges driven by limited data. These challenges reflect not only data limitations but also the fundamental nature of the disease, with gradual onset, overlapping domains of function, and outcomes that are often difficult to quantify with standard instruments. As more high-cost DMTs enter the pipeline, the way that uncertainty is handled in AD models will be increasingly consequential for access and affordability decisions. The objective is to examine how uncertainty has been addressed in recent HTA evaluations of lecanemab and donanemab and explore the implications for modelling and decision-making frameworks.
METHODS: A targeted review of prior HTA submissions and appraisals from NICE was conducted. Key model inputs, structural assumptions, extrapolation approaches, and uncertainty analyses (e.g. sensitivity, scenario, structural) were extracted.
RESULTS: NICE appraisals of lecanemab and donanemab identified shared uncertainties, including limited long-term data and difficulty linking short-term clinical changes to meaningful outcomes. Models varied in assumptions around disease progression, institutionalisation, and caregiver burden. Cost-effectiveness was highly sensitive to assumptions about treatment effect duration and waning, often based on limited evidence. As current standard of care does not involve intravenous therapy, demonstrating cost-effectiveness relative to usual care is especially challenging. Structural uncertainty was also underexplored.
CONCLUSIONS: Uncertainty in AD modelling stems from the diffuse, slowly progressive nature of the disease and the difficulty of capturing long-term clinical benefit in a measurable way. These features challenge traditional HTA approaches, which often rely on discrete, near-term outcomes and well-validated surrogate measures. There may be a need for evolution to better reflect the complexities of diseases like AD; for example, extrapolation methods, expanded consideration of caregiver and societal impact, or the use of adaptive evidence standards in settings of high unmet need.