New NIH tool uses genAI to connect volunteers to clinical trials
Researchers at the National Institutes of Health are using large language models to develop an artificial intelligence framework to streamline the clinical trial matching process and more quickly match potential volunteers to relevant trials listed on ClinicalTrials.gov.
By comparing its accuracy with three human doctors, researchers found that the tool, TrialGPT, achieved nearly the same level of accuracy, according to an NIH announcement this month.
WHY IT’S IMPORTANT
Because finding the right clinical trial for a patient is both time and resource intensive, researchers at the National Library of Medicine and the National Cancer Institute developed the TrialGPT framework to streamline this.
The new clinical trial matching algorithm analyzes patient summaries for relevant medical and demographic information, then identifies clinical trials for which a patient is eligible and excludes trials for which they are not eligible.
TrialGPT produces an annotated list of clinical trials – ranked by relevance and eligibility – that physicians can use to discuss clinical trial options with their patients. The AI tool also explains how someone meets the enrollment criteria for the study, which is critical to its effectiveness.
To assess how well TrialGPT predicted whether a patient met a specific clinical trial requirement, the researchers compared the tool’s results with those of three human physicians who assessed more than 1,000 patient-criterion pairs, according to NIH.
“Machine learning and AI technology show promise in matching patients to clinical trials, but their practical application in diverse populations remains to be explored,” Stephen Sherry, acting director of NLM, said in a statement.
The researchers also conducted a pilot user study and found that when physicians used TrialGPT, they spent 40% less time screening patients but maintained the same level of accuracy.
Note that TrialGPT relies on OpenAI’s GPT-series LLMs, such as GPT-3.5 and GPT-4, and the latter is closed-source and only accessible via commercial applications or API, the researchers said in their report. report.
For their research, published in Nature Communications and co-authored by collaborators from the Albert Einstein College of Medicine, the University of Pittsburgh, the University of Illinois Urbana-Champaign and the University of Maryland, College Park, the research team received an innovation award and the study will further assess the model’s performance and fairness in real-world clinical settings, NIH said.
THE BIG TREND
The use of AI to improve patient recruitment, retention, and clinical trial outcomes began before OpenAI launched its generative AI model ChatGPT. During the COVID-19 pandemic, oncology organizations used healthcare data to find ways to find patients across the country who would qualify for trials, even if they weren’t physically there.
In driving decentralized clinical trials, increased adoption of AI has helped promote healthcare equity and trial diversity, said Jeff Elton, CEO of ConcertAI, a provider of data and AI SaaS platforms for clinical trial optimization. to research.
“With integrated digital exams, clinical trials are an integral part of the care process itself, rather than imposed on it,” Elton said Healthcare IT news.
“Trials should not impose higher burdens on healthcare providers and patients than the standard of care.”
Reducing friction during the clinical trial lifecycle is critical to giving patients access to trial therapies, said Seth Howard, vice president of research and development at Epic.
The electronic health record provider implemented data-driven clinical trials matchmaking two years ago. Using the anonymized Cosmos dataset, Epic offers providers who register for the service the ability to match clinical trial opportunities from sponsors with a count of their organization’s eligible patients.
Many healthcare systems have also been testing using analytics applications that can surface clinical trial opportunities for patients using their organization’s EHR data. In October, Microsoft announced new AI tools that will allow healthcare systems to build their own custom AI tools for many administrative needs, including clinical trial matching.
However, AI bias is still a concern for clinical outcomes.
It could emerge in any algorithm development pipeline and worsen health care inequalities, researchers from the Yale School of Medicine said in new research published earlier this month.
ON THE RECORD
“Our research shows that TrialGPT can help physicians more efficiently connect their patients to clinical trial opportunities and save valuable time that could be better spent on more demanding tasks that require human expertise,” said Zhiyong Lu, senior investigator at NLM and the corresponding author of the study. a statement.
“This study shows that we can use AI technology responsibly so that physicians can connect their patients even faster and more efficiently to a relevant clinical trial that may be of interest to them,” Sherry added.
Andrea Fox is editor-in-chief of Healthcare IT News.
Email: afox@himss.org
Healthcare IT News is a HIMSS Media publication.