NIH Uses AI, LLM to Identify Clinical Trial Participants

The National Institutes of Health’s National Library of Medicine is using artificial intelligence and large language models to identify suitable candidates for clinical trials.

To help researchers and clinicians streamline clinical trial matching, the NLM has developed a prototype called TrialGPT, a tool designed to predict three elements of patient eligibility for clinical trials. The NIH said that early results from using the tool showed its potential in patient-criterion relevance and effective ranking but demonstrated its lack of intrinsic medical knowledge, FedScoop reported. 

Dianne Babski, director of NLM’s user services and collection division, said the institution is working to address the limitations of the AI-enabled tool through effective collaboration and interdisciplinary teamwork.

The NIH has been researching how to best use AI for better medical research and health care delivery. In 2022, it announced the Bridge to Artificial Intelligence program, a $130 million commitment over four years to invest in AI use in biomedical and behavioral research.

In February, the public biomedical and health research agency sought a vendor to help improve the management of health care data through an AI-powered electronic health record platform.