Harnessing AI and EHR data for rapid insights on GLP-1s

Traditionally, EHR data has required extensive cleaning and normalization—delaying insight generation and making large-scale, timely studies difficult. This session will explore how AI is now being used to rapidly structure and standardize clinical data from EHRs—converting unstructured inputs, such as clinical notes, into research-ready formats.

We’ll introduce Truveta Data, which includes normalized, de-identified EHR data on more than 120 million US patients. These data are updated daily and linked to closed claims for over 200 million patients, enabling longitudinal views of care and outcomes across sites and payers. With these data, researchers can evaluate prescribing patterns, clinical outcomes, and access barriers at scale—enabling studies that were previously too slow, costly, or complex to conduct.

We’ll then highlight published research on GLP-1s as a case study of what is now possible with AI-enabled real-world data. These findings reveal important trends in comparative effectiveness, discontinuation and reinitiation, and disparities in access. They also demonstrate how data extracted from clinical notes can add rich clinical context—surfacing patterns in patient behavior and treatment decisions that may be missed in structured data alone.