Skip to main content
Back to News Hub
🟢TechCrunch AI
September 1, 2026
ChatGPT

ChatGPT Health adds Epic integration for clinicians to import patient data

Overview

OpenAI has announced that ChatGPT Health will integrate with Epic's electronic health record system to provide clinicians with read-only access to patient data. This system allows healthcare providers to review medical histories, generate patient timelines, and compile notes without altering records. While safety tests showed a high accuracy rate, OpenAI continues to emphasize that the AI is not designed for direct medical diagnosis or treatment.

Key Takeaways

  • OpenAI is connecting ChatGPT Health with Epic, an electronic health record system containing records for over 325 million patients.

    Through this setup, medical professionals can review lab results, appointment notes, and medication histories directly within their workflows.

  • The company clarified that the software operates with read-only permission, meaning the tool will not write back into health records or make direct modifications to patient files.

    To further support medical workflows, OpenAI released a Healthcare Public Data plug-in capable of extracting information from public repositories like PubMed, RxNorm, and ClinicalTrials.gov.

  • Additionally, organizations holding a Business Associate Agreement can deploy tools such as ChatGPT Work and Codex for compliant administrative tasks.

    Despite physician evaluations indicating that 99.1% of tested clinical responses were safe, OpenAI stresses that the technology is not meant to diagnose or treat patients.

  • The new integration connects ChatGPT Health to Epic's electronic health record platform, which stores information for over 325 million patients.

    Clinicians receive read-only access to records, ensuring the artificial intelligence cannot write or alter data directly within patient charts.

  • An internal evaluation of over 4,300 physician responses across 27 clinical use cases showed that 99.1% were considered safe.

Stats & Key Facts

  • #OpenAI is connecting ChatGPT Health with Epic, an electronic health record system containing records for over 325 million patients.
  • #Despite physician evaluations indicating that 99.1% of tested clinical responses were safe, OpenAI stresses that the technology is not meant to diagnose or treat patients.
  • #The new integration connects ChatGPT Health to Epic's electronic health record platform, which stores information for over 325 million patients.
  • #An internal evaluation of over 4,300 physician responses across 27 clinical use cases showed that 99.1% were considered safe.

OpenAI is connecting ChatGPT Health with Epic, an electronic health record system containing records for over 325 million patients. Through this setup, medical professionals can review lab results, appointment notes, and medication histories directly within their workflows. The company clarified that the software operates with read-only permission, meaning the tool will not write back into health records or make direct modifications to patient files.

To further support medical workflows, OpenAI released a Healthcare Public Data plug-in capable of extracting information from public repositories like PubMed, RxNorm, and ClinicalTrials.gov. Additionally, organizations holding a Business Associate Agreement can deploy tools such as ChatGPT Work and Codex for compliant administrative tasks. Despite physician evaluations indicating that 99.1% of tested clinical responses were safe, OpenAI stresses that the technology is not meant to diagnose or treat patients.

The new integration connects ChatGPT Health to Epic's electronic health record platform, which stores information for over 325 million patients. Clinicians receive read-only access to records, ensuring the artificial intelligence cannot write or alter data directly within patient charts. OpenAI introduced a Healthcare Public Data plug-in that gathers information from sources like PubMed, RxNorm, and ClinicalTrials.gov.

For more details please read the original article at TechCrunch AI.

Continue Learning

Comments

Comments appear only after moderation. Your email identifies your submission to the moderator and is never displayed here.

No approved comments yet.

Originally published by TechCrunch AI
Read the original