OpenAI Launches Epic EHR Integration and Official Data Plugins for ChatGPT for Healthcare

OpenAI has introduced new Epic Electronic Health Record (EHR) integration capabilities and the Healthcare Public Data plugin to ChatGPT for Healthcare. This enables healthcare professionals to review and utilize approved patient context and structured information from official sources in one place.

Epic EHR Integration

Healthcare organizations can now connect their Epic environments to ChatGPT, allowing them to retrieve patient information via natural language queries instead of searching through appointment notes, test results, medication information, and specialist documents. For example, users can ask "changes since the last visit" or "recent test results to review before today's appointment," and ChatGPT will summarize relevant information and provide links to the original charts.

This integration supports two experiences:

• EHR Context in ChatGPT: Bring approved patient information into ChatGPT to review medical history, identify changes, and prepare for care.
• ChatGPT in EHR: Directly invoke ChatGPT within supported EHR environments to perform AI-powered tasks without leaving the chart.

Suresh Gunasekaran, CEO of UCSF Health, evaluated that this technology has the potential to reduce time spent synthesizing data and allow clinicians to focus more on patient care.

Healthcare Public Data Plugin

Beyond patient information, healthcare professionals require official data on research, drugs, and coverage. The new Healthcare Public Data plugin connects directly to nine official medical data sources, including PubMed, ClinicalTrials.gov, CMS Coverage, RxNorm, and DailyMed.

Through this plugin, teams can perform tasks based on specific records, fields, identifiers, and versions. For example, research teams can compare clinical trial enrollment status and eligibility criteria, while pharmacy teams can verify the latest drug labels and warning information. This helps compare and verify accurate information without having to search each source individually.