ChatGPT for Healthcare gained broader data access on September 1. Health systems can now attach their Epic electronic health record environments to the workspace, OpenAI said. A new Healthcare Public Data plugin also ties the product to nine official medical datasets.
The Epic hookup gives clinicians a read-only window into authorized patient context. Instead of digging through appointment notes, lab results, medications and specialist documents, users can pose questions grounded in the chart. OpenAI pointed to two supported modes. Record data can be pulled into ChatGPT, or ChatGPT can be embedded inside the EHR layout in supported deployments so clinicians never leave the patient chart.
UCSF Health signed on as a pilot partner. Its president and chief executive, Suresh Gunasekaran, said the tool can help teams quickly grasp what changed across a complex record, potentially cutting the time spent synthesizing data and giving clinicians more of it back for patients.
The plugin consolidates connectors for ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed and PubMed, among others. Teams can pull specific records, fields, identifiers and versions, then compare trial eligibility criteria, medication labels and coverage policy versions without opening each source one at a time. OpenAI described the connection as one-way, with chart data feeding answers but no model output flowing back into the record. The features are meant to complement existing EHR workflows, not replace them.
