NIH's TrialGPT Uses AI to Accelerate Clinical Trial Recruitment
核心洞察
The NIH has developed TrialGPT (搜索), an AI-driven large language model, to streamline patient matching for clinical trials, addressing a major bottleneck in research.
TrialGPT (搜索) demonstrated accuracy comparable to human experts in matching patients to relevant trials listed in clinicaltrials.gov, potentially improving patient enrollment rates.
A pilot study showed that clinicians using TrialGPT (搜索) spent 40% less time screening patients while maintaining the same level of accuracy in patient-trial matching.
Researchers at the National Institutes of Health (NIH) have developed an artificial intelligence (AI) tool called TrialGPT (搜索) to improve patient recruitment for clinical trials. Patient recruitment is a significant hurdle in clinical research, often leading to delays and increased costs. TrialGPT, a large language model (LLM) derived from ChatGPT, aims to solve this problem by identifying relevant clinical trials for eligible patients and providing summaries explaining their eligibility. The research was published in Nature Communications.
Addressing Recruitment Bottlenecks
It is estimated that 80% of clinical trials fail to recruit the required number of patients on time, which can be attributed to inefficiencies in traditional recruitment methods. Approximately 40% of cancer (搜索) trials fail due to insufficient patient enrollment, according to Zhiyong Lu of the NIH's National Library of Medicine (NLM). TrialGPT (搜索) was developed in collaboration with scientists at the National Cancer Institute (NCI).
"TrialGPT (搜索) could help clinicians connect their patients to clinical trial opportunities more efficiently and save precious time that can be better spent on harder tasks that require human expertise," said Lu.
TrialGPT Performance and Accuracy
TrialGPT (搜索) was tested on its ability to match patients to suitable trials listed in the clinicaltrials.gov database. The AI tool achieved an accuracy rate comparable to that of human clinical research experts. In a pilot user study, clinicians using TrialGPT spent 40% less time screening patients while maintaining the same level of accuracy.
Clinician Efficiency
The study involved two human clinicians reviewing anonymous patient summaries and matching them to clinical trials. One clinician manually reviewed patient summaries to assess eligibility, while the other used TrialGPT (搜索) for the same task. The results indicated a significant time saving for clinicians using the AI tool.
Future Directions
The team behind TrialGPT (搜索) has been invited to participate in a Director's Challenge Innovation Award to further test the LLM in real-world clinical settings. This will help assess its performance across diverse populations and refine its application in clinical practice.
Stephen Sherry, acting director of the NLM, noted, "This study shows we can responsibly leverage AI technology so physicians can connect their patients to a relevant clinical trial that may be of interest to them, with even more speed and efficiency."
AI is also being explored for increasing patient retention and identifying high-risk patients who can be encouraged to participate in trials.