404-P: Cardiovascular Risk Scoring and Stratification in Patients with Type 2 Diabetes Enrolled in a Medicare Advantage Plan

Current tools to identify patients at risk for cardiovascular (CV) disease or events are variable, limiting their utility and generalizability. The purpose of this study was to develop a predictive model to identify patients with type 2 diabetes (T2D) who are ≥65 years of age, enrolled in a Medicare Advantage Prescription Drug (MAPD) plan and at risk for future CV events. This retrospective cohort study used administrative claims, laboratory and enrollment data from 1/2011-12/2018. Patients with T2D were identified from 1/2012-12/2013 using diagnosis codes and/or claims for antihyperglycemic medications. The dependent variable was the first composite CV event defined as ≥1 inpatient hospitalization for myocardial infarction, ischemic stroke, unstable angina, or heart failure or any evidence of revascularization. Independent variables for the prediction survival model included baseline demographic and clinical characteristics prognostic for the dependent variable. C-statistic, accuracy, sensitivity, and specificity were used to assess model performance. Risk ranking was conducted, whereby patients were classified as low, medium or at high risk of an event based on probability distribution derived cut-points. A total of 362,791 patients with T2D were identified. The proportion of patients with ≥1 CV event was 18.0% up to 5 years after T2D identification. The final model included 42 demographic and clinical variables. The C-statistic was 0.68, and accuracy, sensitivity and specificity were 0.63. Results were consistent across the training, test and holdout datasets suggesting internal validity. Up to 5 years after identification, 11% of patients classified as low, 27% as medium, and 51% as high risk had a future CV event. A predictive model for composite CV events utilizing administrative claims to identify and risk stratify patients may be used to scale focused interventions to high-risk T2D patients in a MAPD population to prevent future CVD events.