Integration with Epic Systems
OpenAI today announced a direct integration between ChatGPT Health and Epic, the electronic health record system storing data for 325 million patients. This move allows clinicians to pull records directly into their AI workflows. Doctors can now access appointment notes, lab results, medication lists, and specialist documentation through the interface. The tool summarizes this information to help staff prepare for upcoming patient visits.
Technicians designed this feature with specific constraints. It operates as a read-only bridge. ChatGPT does not write information back into the patient records at this stage. Some clinical environments will feature the AI directly inside existing EHR windows to prevent software switching. It acts as an assistant for pre-visit preparation rather than an automated scribe or diagnostic machine.
Data Access and Research Tools
Beyond record imports, OpenAI launched a Healthcare Public Data plug-in. This tool connects to public databases such as ClinicalTrials.gov, PubMed, and DailyMed. Clinicians use it to find medication identifiers or verify coverage policies through CMS data. It helps medical staff verify eligibility for clinical trials against specific patient data sets.
Organizations holding a Business Associate Agreement can now use ChatGPT Work, Codex, and various connectors within their digital environment. This setup allows for compliance-focused workflows. It marks a transition from consumer-facing health features toward institutional medical adoption. The platform architecture allows hospitals to handle these requests while adhering to security standards for sensitive health data.
Clinical Safety and Legal Scrutiny
OpenAI reports that it analyzed 4,300 responses from doctors across 27 different use cases. These include handoff summaries and medication reviews. The company claims 99.1% of these outputs met safety standards. Still, medical professionals acknowledge that even a single error in a clinical setting carries significant weight.
Public confidence faces challenges from ongoing litigation. A Florida pastor recently sued the company, claiming an AI recommendation nearly resulted in his death. Another suit from May involves a family who claims the system provided incorrect medication dosage advice. Despite these cases, the company continues to maintain that its product is not a tool for diagnosis or treatment.
Industry Context and Future Outlook
This expansion follows the July rollout of ChatGPT for health to all U.S. consumers. OpenAI states that individuals currently perform 300 million health-related queries on the platform every week. The shift toward professional medical settings represents a pivot from broad information retrieval to specific clinical support. Industry observers watch to see how hospitals balance the efficiency of AI-assisted data summary with the risks of hallucinations. Regulatory bodies will likely continue to monitor how these tools integrate into high-stakes care environments where precision remains the standard.