2317-P: Real-World Trajectories of Type 2 Diabetes Pharmacotherapy in the United States: A Nationwide Clustering Analysis of Electronic Health Records

Introduction and Objective: Clinical guidelines for type 2 diabetes recommend timely use of newer glucose-lowering therapies, yet real-world treatment pathways remain highly variable. Longitudinal medication trajectories at national scale are poorly characterized. Methods: We conducted a retrospective cohort study using de-identified electronic health record data from a nationwide research network. Adults newly diagnosed with type 2 diabetes in early 2019 without prior glucose-lowering therapy were followed through 2024. Medication prescriptions across nine drug classes were encoded in twelve semiannual intervals. Unsupervised clustering identified distinct pharmacotherapy trajectories. Changes in body mass index (BMI) and hbA1c were summarized over time. Logistic regression evaluated associations between cluster membership and patient characteristics. Results: Among 9,327 individuals, we identified thirty medication trajectory clusters spanning monotherapy, dual therapy, complex therapy, variant therapy, and GLP-1 transitions. GLP-1 trajectories were associated with the greatest reductions in BMI and hbA1c but older adults and Hispanic individuals were underrepresented in early GLP-1 transition clusters (Figure 1). Conclusion: Real-world pharmacotherapy trajectories for type 2 diabetes are highly heterogeneous, with delayed and uneven adoption of newer therapies. Disclosure P. Kukhareva: None. Funding National Institutes of Health (1K01DK142045)