OpenAI adds Epic EHR integration to ChatGPT for Healthcare
OpenAI has introduced an Epic electronic health record integration for ChatGPT for Healthcare, alongside a Healthcare Public Data plugin that connects users to nine official public healthcare sources. The additions are designed to let healthcare organizations review authorized patient context and structured external information within a governed ChatGPT workspace.
The Epic integration supports two deployment models. Organizations can bring authorized patient information from a supported EHR into ChatGPT, or, in supported deployments, integrate ChatGPT into the EHR layout so staff can use AI-assisted workflows without leaving the patient chart. OpenAI says the system can surface relevant chart information and point users back to supporting records.
Epic context for clinical review
Clinicians can use the integration to review changes since a patient’s last visit, examine recent laboratory results, identify medication changes and specialist recommendations, and prepare for outstanding follow-ups or referrals. The intended inputs include appointment notes, lab results, medications and specialist documentation.
UCSF Health is participating as a pilot partner. Its president and CEO, Suresh Gunasekaran, said the organization is exploring whether the integration can help clinical teams identify what has changed and what matters most in complex records, while frontline teams validate the capabilities in practice.
For organizations considering clinician-facing access, ChatGPT tools for U.S. medical practice provides context on ChatGPT tools for U.S. medical practice and the evolving availability of healthcare-focused capabilities.
Plugin connects nine official data sources
The Healthcare Public Data plugin provides dedicated connectors to nine official sources. OpenAI named ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed and PubMed among them. Users can work with specific records, fields, identifiers and versions, rather than conducting separate searches across each website.
OpenAI describes uses such as comparing trial eligibility criteria through ClinicalTrials.gov, confirming current medication labels and warnings through DailyMed, and combining research, active trials and Medicare coverage information for population-health planning. The company says ChatGPT for Healthcare already supports clinical questions and medical-research synthesis through trusted clinical search across thousands of medical sources.
Evaluation and enterprise controls
OpenAI says it works with hundreds of physicians across 60 countries, 49 languages and 26 medical specialties. Those physicians have reviewed more than 700,000 model responses. In an evaluation of connected EHR context across 27 clinical use cases, physicians issued 4,363 ratings and judged 99.1% of responses safe. The use cases included pre-visit review, clinical timelines, medication review and handoff summaries.
In a separate two-round assessment of nuanced questions using large U.S. healthcare datasets, more than 93% of responses for each of five connected data sources were rated “good” or better for accuracy. OpenAI frames these evaluations as measures of how the system surfaces relevant information, cites supporting evidence and prepares material for clinician or staff review.
ChatGPT for Healthcare includes role-based access, single sign-on and audit logs. OpenAI says customers with an applicable Business Associate Agreement can use ChatGPT Work, Codex, apps and plugins for HIPAA-compliant workflows. Healthcare leaders should treat the release as a configuration and governance decision: confirm eligibility, define authorized data access, and retain clinical review of generated outputs before using the tools in care or operations.