DEVELOPMENT AND STRUCTURAL BENCHMARKING OF THE ASSURE-SSP MODEL: ASUNDEXIAN REFERENCE ECONOMIC MODEL IN SECONDARY STROKE PREVENTION
Author(s)
Hubert Polek, MSc1, Michal Pochopien, MSc, PhD1, Luke Bamber, MSc2.
1Clever-Access, Krakow, Poland, 2Bayer AG, Wuppertal, Germany.
1Clever-Access, Krakow, Poland, 2Bayer AG, Wuppertal, Germany.
OBJECTIVES: To develop a de novo cost-effectiveness model of asundexian in combination with antiplatelet therapy in patients with non-cardioembolic ischaemic stroke or high-risk transient ischaemic attack, and to assess its structural validity through benchmarking against published models identified via a systematic literature review.
METHODS: A cohort-based Markov state-transition model was developed to reflect the OCEANIC-STROKE population, with health states defined by modified Rankin Scale (mRS 0-6). Patients entered the model according to 90-day post-event functional outcomes, with subsequent transitions driven by secondary stroke. Stroke risk was modelled as time-dependent, using 3-month cycles in the acute phase and extrapolated over a lifetime horizon. Major bleeding, myocardial infarction and transient ischaemic attack were included as clinically relevant events alongside cause-specific mortality. Treatment effects were parameterised using hazard ratios from OCEANIC-STROKE. A systematic literature review was conducted to characterise existing models and benchmark structural assumptions related to health states, time horizons and secondary stroke risk modelling.
RESULTS: The review identified 14 intervention models across 12 publications, six genotype-testing models and nine additional methodological frameworks. Structural approaches were heterogeneous, including Markov (n=6), hybrid decision tree-Markov (n=4), decision tree (n=2) and simulation-based models. Lifetime horizons predominated, while cycle lengths ranged from monthly to annual. Secondary stroke was modelled using constant, time-stratified, or age- and treatment-dependent risks. These findings supported key structural choices, including granular mRS 0-6 health states, explicit secondary stroke pathways, inclusion of bleeding events and alignment with the anticipated clinical profile of Factor XIa inhibition.
CONCLUSIONS: This literature-informed modelling framework aligns with trial endpoints while reflecting the diversity of published approaches. It provides a transparent and flexible platform for cost-effectiveness analyses and supports adaptation across settings, in which structural assumptions around secondary stroke and disability are key drivers of outcomes.
METHODS: A cohort-based Markov state-transition model was developed to reflect the OCEANIC-STROKE population, with health states defined by modified Rankin Scale (mRS 0-6). Patients entered the model according to 90-day post-event functional outcomes, with subsequent transitions driven by secondary stroke. Stroke risk was modelled as time-dependent, using 3-month cycles in the acute phase and extrapolated over a lifetime horizon. Major bleeding, myocardial infarction and transient ischaemic attack were included as clinically relevant events alongside cause-specific mortality. Treatment effects were parameterised using hazard ratios from OCEANIC-STROKE. A systematic literature review was conducted to characterise existing models and benchmark structural assumptions related to health states, time horizons and secondary stroke risk modelling.
RESULTS: The review identified 14 intervention models across 12 publications, six genotype-testing models and nine additional methodological frameworks. Structural approaches were heterogeneous, including Markov (n=6), hybrid decision tree-Markov (n=4), decision tree (n=2) and simulation-based models. Lifetime horizons predominated, while cycle lengths ranged from monthly to annual. Secondary stroke was modelled using constant, time-stratified, or age- and treatment-dependent risks. These findings supported key structural choices, including granular mRS 0-6 health states, explicit secondary stroke pathways, inclusion of bleeding events and alignment with the anticipated clinical profile of Factor XIa inhibition.
CONCLUSIONS: This literature-informed modelling framework aligns with trial endpoints while reflecting the diversity of published approaches. It provides a transparent and flexible platform for cost-effectiveness analyses and supports adaptation across settings, in which structural assumptions around secondary stroke and disability are key drivers of outcomes.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EE317
Topic
Economic Evaluation
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory)