OpenAI's ChatGPT Health is integrating with Epic's electronic health record system, allowing clinicians to import patient data for summarization and analysis. The integration provides read-only access and includes a new Healthcare Public Data plugin.
This integration aims to streamline clinical workflows and improve access to patient information for healthcare providers by leveraging AI, potentially enhancing efficiency and patient care. It also marks a significant step in the adoption of AI tools within the sensitive healthcare sector.
OpenAI has announced a significant integration between its ChatGPT Health service and Epic's widely-used electronic health record (EHR) system. This collaboration will enable clinicians to import patient data directly from Epic, which serves over 325 million patients, allowing them to use AI to analyze and summarize information such as appointment notes, laboratory results, and medication histories.
The integration is designed for read-only access, ensuring that the AI does not alter patient records. Clinicians can leverage ChatGPT to review patient history, identify changes, and prepare for appointments. In some implementations, ChatGPT will be accessible directly within EHR workflows, enabling tasks like pre-visit reviews and the creation of clinical timelines without leaving the patient chart.
In addition to the Epic integration, OpenAI is introducing a new Healthcare Public Data plugin. This plugin can access information from various health-related databases, including ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed, and PubMed, assisting healthcare professionals in synthesizing data for trial eligibility, medication details, and coverage policies.
Furthermore, OpenAI is extending its compliance offerings by allowing organizations with a Business Associate Agreement (BAA) to utilize ChatGPT Work, Codex, apps, and connectors for compliant healthcare workflows. This move follows OpenAI's recent rollout of ChatGPT for health to all U.S. consumers, with a significant volume of health-related queries already being processed weekly.
OpenAI has emphasized the safety of its AI in clinical settings, citing feedback from over 4,300 physicians across 27 use cases, where 99.1% of responses were deemed safe. However, the company acknowledges the potential for harm from even a small number of unsafe answers, referencing recent lawsuits alleging harmful advice from ChatGPT.