Integrating EHR, Pharmacy and Claims Data to Advance Real-World Evaluation of GLP1 Receptor Agonists on Cardiometabolic Outcomes - ABSTRACT Glucagon-like peptide-1 receptor agonists (GLP1-RAs) are widely used for diabetes, obesity and cardiometabolic risk reduction, with demonstrated benefits for weight, glycemia, blood pressure, and cardiovascular outcomes. However, rapid clinical uptake has outpaced insurance coverage, leading many patients to obtain GLP-1RAs through out-of-pocket (cash) purchases, manufacturer coupons, or nontraditional dispensing channels. As a result, studies relying solely on insurance claims often incompletely capture GLP1- RA exposure, creating uncertainty around adherence, persistence, and real-world effectiveness. This project integrates electronic health records (EHR), pharmacy dispensing data, and Medicare claims to comprehensively evaluate GLP-1RA exposure measurement approaches and generate robust evidence on medication use and blood pressure effects in routine care. Aim 1 will quantify misclassification of GLP-1RA exposure by comparing Medicare claims and EHR prescribing data against a pragmatic reference standard of pharmacy dispensing, and will characterize how misclassification varies across indications, patient subgroups, products, and time. Aim 2 will use linked EHR-dispensing data (UF EHR+) to estimate real-world adherence, persistence and discontinuation patterns for GLP1-RAs across clinically relevant subgroups. Aim 3 will emulate a target trial using UF EHR+ data to estimate the causal effects of GLP1-RA initiation on blood pressure change among adults with hypertension, leveraging clone-censor-weighting methods to address treatment timing and avoid immortal time bias. By integrating multi-source data, this study overcomes key limitations of prior observational research by more accurately capturing medication exposure, adherence, and real-world blood pressure effects. Findings will clarify cardiovascular implications of GLP1-RA therapy in hypertensive patients, identify persistence patterns across patient groups, and provide methodologic insights for future real world evaluations of cardiometabolic therapies. The project also provides structured, mentored training in exposure validation, longitudinal medication analysis, causal inference, and project leadership, directly supporting my progression toward research independence. Beyond the research aims, the fellowship delivers coordinated training in advanced pharmacoepidemiology, real-world data methods, and leadership development through targeted coursework, directed readings, and short courses guided by a multidisciplinary mentorship team. Additional activities, including seminars, professional development workshops, and regular mentor meetings, will further strengthen my methodological and professional skills. Training and research activities will occur primarily within the UF College of Pharmacy Department of Pharmaceutical Outcomes and Policy, which provides a robust data and computational environment. Collectively, this integrated research and training plan will provide the analytic, methodological, and leadership competencies required for my transition to an independent investigator.
