(VP084) POPULATION HEALTH VALUE OF BEING IN TARGET: RESULTS FROM THE CANADIAN MULTI-MORBIDITY MODEL FOR TYPE 2 DIABETES

Background: Diabetes affects nearly 10% of Canadians, and can lead to many diabetes-related complications. Novel therapies purport to both improve control of diabetes and reduce the risk of these complications. Microsimulation modelling allows for the synthesis of multiple data sources for comparative effectiveness analysis related to a variety of correlated outcomes. However, existing type 2 diabetes models often use data that are decades old, are derived using non-representative clinical trial populations, or do not fully model vascular risk.

METHODS AND RESULTS: We developed a novel microsimulation model using health state features including, sex, age, BMI, HbA1c, eGFR, blood pressure, lipids, smoking behaviours, diabetes duration, and history of major acute events. Each month, risk factors update, and simulants may experience acute events. Major acute events included are myocardial infarction, stroke, congestive heart failure, amputation, cancer, and death. Most risk factors and biomarkers update using correlated prediction equations from the U.S. National Health and Nutrition Examination Survey (NHANES) 2005-2017. The trajectories of other risk factors relied on prediction models and observational cohort analysis in the peer reviewed literature. Incidence of major acute events and acute event mortality rates relied on prediction equations derived from Ontario’s population-based administrative healthcare datasets (2012-2017). We calibrated long-term survival after acute events to population-based estimates of overall survival and longitudinal cohorts in the literature. We validated model predictions of event rates and overall survival to existing models of diabetes progression and risk of acute events, and to recent diabetes clinical trials. We then used the model to calculate the number of acute events that could be averted and the extension to life expectancy that could be obtained if a population of people with type 2 diabetes achieved treatment targets for HbA1c, blood pressure and LDL-cholesterol.

Conclusion: We developed a microsimulation model including detailed progression and outcomes associated with multiple comorbidities and acute events in individuals with type 2 diabetes. Prediction of acute events and trajectories of biomarkers and other risk factors relies on recent clinical cohorts and population-based data in North America. This model is capable of comparative effectiveness and cost-effectiveness data of novel therapies in diabetes.