AHA Scientific Sessions 2026 Abstract · Nov 2, 2026
From Blanket Denial to Targeted Coverage: A "PREVENT"-Based Tiered Framework for GLP-1 Reimbursement
By Atreja A, Hanks P, Kay L et al.
Background: Cardiovascular Disease (CVD) affects approximately 1 in 20 U.S. adults, accounts for roughly 2,500 deaths daily, and costs an estimated $200 billion annually to the healthcare system. Glucagon-Like Peptide-1 (GLP-1) receptor agonists such as semaglutide have demonstrated efficacy in addressing multiple CVD risk factors by suppressing appetite, delaying gastric emptying, and affecting blood pressure, lipid profiles, and glycemic control. However broader insurance coverage poses a substantial financial burden to payers. Hypothesis: There are specific, definable clinical conditions under which GLP-1 coverage represents a justifiable financial investment for insurers. Methods: We integrated three data sources: 1. Predicting Risk of Cardiovascular Disease Events (PREVENT) score calculator developed by the American Heart Association quantifies 10-year Atherosclerotic Cardiovascular Disease (ASCVD) event risk using inputs such as Body Mass Index and Systolic Blood Pressure. 2. The National Health and Nutrition Examination Survey (NHANES) dataset provides individual-level health and cardiovascular data for a nationally representative adult cohort. 3. The Medical Expenditure Panel Survey (MEPS) dataset supplied inpatient cost data, from which we derived a log-normal expenditure distribution fitted to 835 ASCVD hospitalizations. We then conducted a three-stage Monte Carlo simulation with, in stage 1, published semaglutide trial data was used to build normal distributions reflecting the drug's effect on each PREVENT input variable and this was applied to NHANES patients. In stage 2, baseline and modified PREVENT scores were used to simulate ASCVD event frequency over a 10-year period across 1,000 trials per patient. In stage three, each simulated event was assigned a cost sampled from the MEPS-derived log-normal distribution, allowing direct comparison of expected expenditures with and without semaglutide. Results: ASCVD hospitalization savings alone (up to $500 per patient over 10 years) were insufficient to offset the cost of semaglutide at any risk level. However, incorporating a scale factor to account for broader GLP-1 benefits, two clinically meaningful subgroups were indentified (high-risk cohort where PREVENT ≥20%, and a moderate-risk cohort PREVENT 10-20%) where total savings plausibly justify coverage. We recommend a tiered coverage framework: covering semaglutide unconditionally for the high-risk cohort, require step therapy with lifestyle intervention prior to coverage for moderate-risk, and deny coverage for the low-risk cohort. Conclusions: Insurers may operate at a net loss if drug expenditures exceed downstream savings on cardiovascular hospitalizations. Our analysis has shown that expanding coverage for GLP1s can prove to be a cost-effective strategy for payers for high-risk and moderate-risk cohorts based on PREVENT scores.