Researchers at the University of Pittsburgh Medical Center’s Institute for Precision Medicine used OpenAI’s GPT-4 to interpret pharmacogenomic testing results and provide answers.
The chatbot sourced from Clinical Pharmacogenomics Implementation Consortium’s prescribing guidelines and relevant literature, according to the American Society of Health-System Pharmacists.
Researchers noted strong performance after adjusting the tool for boundaries and emphasizing the necessity for evidence-based answers, but they said more data is needed before clinical application.
The Journal of the American Medical Informatics Association published the results.
