DEVELOPMENT AND VALIDATION OF A DECISION-ANALYTIC MODEL USING REAL-WORLD DATA FOR EARLY HEALTH TECHNOLOGY ASSESSMENT OF CARDIOVASCULAR DISEASE RISK SCREENING WITH A POLYGENIC RISK SCORE
Author(s)
Cindy Nguyen, MSc1, Markus Haug, MSc2, Hedwig M. Blommestein, PhD1, Raivo Kolde, PhD2, Maureen Rutten-van Mölken, PhD1.
1Erasmus School of Health Policy and Management, Erasmus University Rotterdam, Rotterdam, Netherlands, 2University of Tartu, Tartu, Estonia.
1Erasmus School of Health Policy and Management, Erasmus University Rotterdam, Rotterdam, Netherlands, 2University of Tartu, Tartu, Estonia.
OBJECTIVES: Although statins reduce cardiovascular disease (CVD) risk, most European countries lack screening programs to identify high-risk individuals. The European Society of Cardiology recognizes that a polygenic risk score (PRS) for coronary artery disease could improve primary prevention targeting, but does not yet recommend its routine use, citing the need for evidence on long-term clinical outcomes and cost-effectiveness. This study describes the development of a decision-analytic model for CVD risk screening and presents initial validation against real-world outcomes.
METHODS: Model development was informed by clinical guidelines, literature, and expert opinion. Real-world data (RWD) from the Estonian biobank were used to determine baseline cohort characteristics and transition probabilities. Individuals were stratified by clinical risk using SCORE2, genetic risk using PRS (top 20% classified as high PRS), and age. The model’s standard-of-care arm was validated using five observed RWD outcomes: CVD risk-group distribution, CVD state distribution, overall survival, cumulative incidence of major adverse cardiovascular events (MACE), and survival following first-MACE.
RESULTS: The final model is a microsimulation consisting of five health states: Pre-first MACE, Preventive treatment due to Screening, Post-first MACE, Death from first MACE, and Death Other. After imputation, the mean cohort age was 57 years (SD: 9); 39.1%, 40.8%, and 20.1% were classified as low-to-moderate, high, and very high SCORE2 risk, respectively, with consistently 20% classified as high PRS within each SCORE2 risk group. Modelled outcomes closely matched RWD outcomes. Notably, survival outcomes differed by less than 2% at all timepoints, while incidence of MACE was within 3% at all timepoints after the third year.
CONCLUSIONS: We developed and validated a novel decision-analytic model to evaluate the cost-effectiveness of CVD screening programs. The model can simulate multiple strategies incorporating PRS with conventional risk factors to identify the optimal screening program. These findings can inform policies on integrating genetic risk assessment into standard clinical practice.
METHODS: Model development was informed by clinical guidelines, literature, and expert opinion. Real-world data (RWD) from the Estonian biobank were used to determine baseline cohort characteristics and transition probabilities. Individuals were stratified by clinical risk using SCORE2, genetic risk using PRS (top 20% classified as high PRS), and age. The model’s standard-of-care arm was validated using five observed RWD outcomes: CVD risk-group distribution, CVD state distribution, overall survival, cumulative incidence of major adverse cardiovascular events (MACE), and survival following first-MACE.
RESULTS: The final model is a microsimulation consisting of five health states: Pre-first MACE, Preventive treatment due to Screening, Post-first MACE, Death from first MACE, and Death Other. After imputation, the mean cohort age was 57 years (SD: 9); 39.1%, 40.8%, and 20.1% were classified as low-to-moderate, high, and very high SCORE2 risk, respectively, with consistently 20% classified as high PRS within each SCORE2 risk group. Modelled outcomes closely matched RWD outcomes. Notably, survival outcomes differed by less than 2% at all timepoints, while incidence of MACE was within 3% at all timepoints after the third year.
CONCLUSIONS: We developed and validated a novel decision-analytic model to evaluate the cost-effectiveness of CVD screening programs. The model can simulate multiple strategies incorporating PRS with conventional risk factors to identify the optimal screening program. These findings can inform policies on integrating genetic risk assessment into standard clinical practice.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
HTA245
Topic
Health Technology Assessment, Real World Data & Information Systems, Study Approaches
Topic Subcategory
Decision & Deliberative Processes
Disease
Cardiovascular Disorders (including MI, Stroke, Circulatory), Genetic, Regenerative & Curative Therapies, Personalized & Precision Medicine