Economic Model of Alzheimer's Disease That Incorporates the Uncertainty Associated With Measuring Efficacy in Clinical Trials
Author(s)
Javier Mar, PhD, MD1, Arantzazu Arrospide, Ph.D.2, Ron Handels, PhD3, Myriam Soto-Gordoa, PhD4.
1Economic evaluation of chronic diseases, Biogipuzkoa Research Institute, Donostia-San Sebastián, Spain, 2Basque Government Health Department, Vitoria-Gasteiz, Spain, 3Maastricht University, maastricht, Netherlands, 4Mondragon University, Mondragón, Spain.
1Economic evaluation of chronic diseases, Biogipuzkoa Research Institute, Donostia-San Sebastián, Spain, 2Basque Government Health Department, Vitoria-Gasteiz, Spain, 3Maastricht University, maastricht, Netherlands, 4Mondragon University, Mondragón, Spain.
OBJECTIVES: Attention has recently focused on proposals for new approaches to reproduce changes in Alzheimer's Disease (AD) patient trajectories in the medium term based on the results of ongoing clinical trials and taking into account uncertainty about their efficacy. We adapted a previous AD economic model, incorporating the uncertainty associated with measuring efficacy in clinical trials, and validate the model to evaluate a disease-modifying treatment intervention for patients in the mild cognitive impairment stage.
METHODS: A discrete event simulation model was built using data from a synthetic clinical trial dataset (IPECAD collaboration) to model typical patient-level natural history trajectories of CDR-SB scores from mild cognitive impairment to severe dementia and death using mixed regression models for repeated measures (MMRM). The model, implemented using Python's SimPy, compares two cloned cohorts of 100,000 individuals (treated and non-treated) that differed in their progression to dementia. Probabilistic sensitivity analysis was performed through 1000 simulations and bootstrapping the synthetic data. As the MMRM coefficients are correlated, this variability was incorporated into the model using Cholesky decomposition. Uncertainty about treatment effect waning was addressed by scenario analysis (optimistic and pessimistic).
RESULTS: The incremental cost-effectiveness ratio of a hypothetical treatment was above €30,000/quality-adjusted life year (QALY) in the pessimistic scenario and around that threshold in the optimistic scenario. The cost-effectiveness plane showed more variability in the incremental cost than in the incremental utility in both scenarios. Treatment dominated for thresholds above $40,000/QALY in the optimistic scenario and above $60,000/QALY in the pessimistic scenario.
CONCLUSIONS: We describe an innovative approach by applying probabilistic sensitivity analysis to two scenarios and shaping individual cognitive trajectories on a continuous scale. Incorporating long-term effectiveness and multi-sectoral costs of dementia, along with advanced methodologies, will provide a new framework for decision-making in the market access process for new preventive interventions and disease-modifying treatments.
METHODS: A discrete event simulation model was built using data from a synthetic clinical trial dataset (IPECAD collaboration) to model typical patient-level natural history trajectories of CDR-SB scores from mild cognitive impairment to severe dementia and death using mixed regression models for repeated measures (MMRM). The model, implemented using Python's SimPy, compares two cloned cohorts of 100,000 individuals (treated and non-treated) that differed in their progression to dementia. Probabilistic sensitivity analysis was performed through 1000 simulations and bootstrapping the synthetic data. As the MMRM coefficients are correlated, this variability was incorporated into the model using Cholesky decomposition. Uncertainty about treatment effect waning was addressed by scenario analysis (optimistic and pessimistic).
RESULTS: The incremental cost-effectiveness ratio of a hypothetical treatment was above €30,000/quality-adjusted life year (QALY) in the pessimistic scenario and around that threshold in the optimistic scenario. The cost-effectiveness plane showed more variability in the incremental cost than in the incremental utility in both scenarios. Treatment dominated for thresholds above $40,000/QALY in the optimistic scenario and above $60,000/QALY in the pessimistic scenario.
CONCLUSIONS: We describe an innovative approach by applying probabilistic sensitivity analysis to two scenarios and shaping individual cognitive trajectories on a continuous scale. Incorporating long-term effectiveness and multi-sectoral costs of dementia, along with advanced methodologies, will provide a new framework for decision-making in the market access process for new preventive interventions and disease-modifying treatments.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
EE402
Topic
Economic Evaluation, Health Technology Assessment, Methodological & Statistical Research
Topic Subcategory
Trial-Based Economic Evaluation
Disease
Mental Health (including addition), Neurological Disorders, No Additional Disease & Conditions/Specialized Treatment Areas