OpenAI connects Epic records and public health data to ChatGPT

OpenAI says healthcare organizations can now connect supported Epic environments to ChatGPT for Healthcare. The announcement, published September 1, 2026, introduces two additions: a way to review authorized patient-record context and a Healthcare Public Data plugin for structured access to official sources. OpenAI News reports that the tools are designed for governed, compliance-conscious workflows rather than disconnected chatbot searches.

Clinicians can ask ChatGPT to summarize what changed since a patient’s last visit, review recent lab results, identify medication updates, or surface unresolved referrals. In supported deployments, ChatGPT can also appear inside the EHR layout, allowing staff to use AI without leaving the chart. The system points back to relevant chart information so users can review the underlying context.

The new plugin connects nine public healthcare sources, including PubMed, ClinicalTrials.gov, DailyMed, RxNorm, and CMS Coverage. That lets teams compare trial eligibility, verify medication labels and identifiers, examine coverage policies, and combine evidence from multiple sources in a focused, source-backed workflow.

The announcement also includes evaluation results. Physicians assessed responses across 27 clinical use cases and rated 99.1% of 4,363 responses as safe. In a separate test of five public-data connectors, at least 93.2% of responses received a “good” or better accuracy rating; results ranged from 93.2% for CMS Coverage to 98.6% for DailyMed.

For AI builders, the development highlights the value of connecting models to permissioned enterprise systems and authoritative datasets. It also underscores that healthcare deployments need workflow integration, traceability, and domain-specific testing—not just a capable general-purpose model.

Source: OpenAI News