Clinical Informatics | Artificial Intelligence | Enterprise Implementation
As Associate Chief Medical Informatics Officer, I served as a primary physician leader for the implementation of AI-generated patient summaries and hospital discharge summaries across a six-hospital health system, with sites ranging from academic medical centers to community hospitals and involving clinicians across multiple specialties.
The Project
The initiative introduced AI-generated summaries designed to synthesize information from across the electronic health record and make relevant clinical information more accessible to physicians. A related AI workflow supported generation of hospital discharge summaries, with a particularly significant impact on hospitalists responsible for the majority of inpatient discharges.
My Role
My work focused on ensuring that the technology functioned appropriately within real-world clinical practice. During beta testing, I evaluated AI-generated patient summaries for clinical accuracy, usability, source selection, and quality of the resulting narrative.
Testing also demonstrated how the information selected from the medical record could influence more than factual accuracy. For example, incorporating nursing documentation could alter the context and tone of the generated summary. This feedback helped inform decisions about which portions of the chart should be reviewed and summarized and how information should be presented to clinicians.
I served as a physician voice throughout implementation, meeting weekly with Epic and internal clinical and technical teams to review updates, provide feedback, identify workflow concerns, and support continued optimization.
Go to link