ChatGPT Health adds Epic integration for clinicians to import patient data

OpenAI said contiguous that it is integrating ChatGPT Health with Epic’s physics wellness grounds (EHR) system, which holds information for implicit 325 cardinal patients, to fto clinicians import diligent information and usage AI to inquire questions. The institution said that successful definite systems, it volition besides integrate ChatGPT straight wrong EHR workflows.

With ChatGPT, clinicians tin present entree accusation similar assignment notes, laboratory results, medications, and specializer documentation and summarize them easily. They tin besides look astatine diligent history, place changes, and hole for aboriginal appointments. OpenAI said that successful immoderate deployments, clinicians tin straight entree ChatGPT for getting pre-visit reappraisal and gathering objective timelines without leaving a diligent chart. The institution specified that the integration allows lone read-only entree to wellness records, and AI doesn’t constitute thing back.

OpenAI is besides adding a caller Healthcare Public Data plugin, which tin fetch accusation from sources including ClinicalTrials.gov, CMS Coverage, RxNorm, DailyMed, and PubMed. The institution said that the caller plugin tin assistance healthcare workers synthesize information for proceedings eligibility criteria, medicine identifiers, sum argumentation versions, and supplier records.

The institution said that isolated from enabling these caller features, it is allowing organizations with a Business Associate Agreement to usage ChatGPT Work, Codex, apps, and connectors successful their workspace for compliant workflows.

Last month, OpenAI rolled retired ChatGPT for wellness to each U.S. consumers. The institution said astatine that clip that radical are asking 300 cardinal health-related queries to ChatGPT each week. Despite galore heavy integrations with aesculapian systems, the institution has maintained that AI is not suitable for diagnosis oregon treatment.

OpenAI said that it collected implicit 4,300 responses from physicians crossed 27 objective usage cases, including pre-visit review, objective timelines, medicine review, and handoff summaries, and recovered that 99.1% of responses were safe. However, adjacent a fewer unsafe answers tin pb to harmful results for humans.

A fewer days earlier this rollout, a Florida-based pastor sued the company, alleging that ChatGPT gave him a near-fatal recommendation. The institution was besides sued successful May by household members of a idiosyncratic who blamed ChatGPT for giving wrongful proposal related to dosage.