Clinicians can ask questions about what has changed since a previous visit, which recent test results need attention, whether medications have changed, and what referrals or unresolved issues remain. ChatGPT then pulls together relevant information from the authorized record and points users back to the supporting chart data.
“As a pilot partner, we’re exploring how the new EHR integration with ChatGPT for Healthcare can help clinical teams understand what has changed and what matters most across a complex patient record,” said Suresh Gunasekaran, president and CEO of UCSF Health. “By bringing relevant information together more quickly and comprehensively, the technology has the potential to reduce time spent synthesizing data and give clinicians more time with patients.”
OpenAI is also adding the Healthcare Public Data plugin, which connects ChatGPT with nine official healthcare information sources. Those include ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed and PubMed. The plugin gives healthcare teams direct access to structured records, fields, identifiers and source versions. OpenAI says that can make it easier to verify details such as medication labeling, clinical trial requirements, coverage policies and provider information without searching each database separately. A research team, for example, could compare eligibility requirements across recruiting studies in ClinicalTrials.gov, while pharmacy teams could use DailyMed to review current medication labels and warnings. Other teams could combine research, trial information and Medicare coverage data in the same workflow.
OpenAI said it has worked with hundreds of physicians across 60 countries, 49 languages and 26 medical specialties to evaluate and improve healthcare responses. Those physicians have reviewed more than 700,000 model outputs, according to the company.
For the new EHR capabilities, physicians evaluated ChatGPT across 27 clinical scenarios, including appointment preparation, medication review, clinical timelines and handoff summaries. Across 4,363 ratings, 99.1% of responses were judged safe, OpenAI said.
The company separately tested the public healthcare data connections with physicians reviewing responses to clinical questions built around large U.S. datasets. More than 93% of responses received a rating of “good” or better for accuracy across each of the five connectors tested. DailyMed produced the highest reported result at 98.6%, followed by PubMed at 97.3%. ClinicalTrials.gov and RxNorm each scored 95.1%, while CMS Coverage received 93.2%.
The new integrations sit inside ChatGPT for Healthcare’s existing governed workspace. OpenAI says the product includes controls such as role-based access, single sign-on and audit logs, while administrators determine which integrations are enabled and existing permissions continue to control what users can see.
Customers with an applicable Business Associate Agreement can also use ChatGPT Work, Codex, apps and plugins within the same workspace for HIPAA-compliant workflows.
ChatGPT for Healthcare customers can ask workspace administrators to enable the Epic integration and Healthcare Public Data plugin. Eligible U.S. ChatGPT for Clinicians users can install the public data plugin, but the EHR integration is not available to individual accounts.
This analysis is based on reporting from OpenAI.
Image courtesy of OpenAI.
This article was generated with AI assistance and reviewed for accuracy and quality.
