A groundbreaking study published in JAMA demonstrates that artificial intelligence can dramatically accelerate patient screening and enrollment in clinical trials, potentially transforming the traditional clinical research process. Mass General Brigham researchers found that their AI-powered screening tool significantly outperformed conventional manual screening methods in a heart failure clinical trial.
The study, which involved 4,476 randomized patients, compared the performance of a novel AI tool called RECTIFIER (RAG-Enabled Clinical Trial Infrastructure for Inclusion Exclusion Review) against traditional manual screening methods. The AI system was tasked with evaluating patient eligibility for the Co-Operative Program for Implementation of Optimal Therapy in Heart Failure (COPILOT-HF) trial.
Superior Screening Efficiency
The AI-assisted screening process demonstrated remarkable efficiency, successfully screening 458 eligible patients compared to 284 patients identified through manual screening by study staff. More importantly, this enhanced screening capability translated into superior enrollment numbers, with 35 patients ultimately enrolling in the AI-screened group versus 19 in the manually screened group.
"The rate of enrollment in the AI-enabled arm was almost double the rate of enrollment in the manual arm. This means that AI could almost halve the time it takes to complete enrollment in a trial," explained Dr. Ozan Unlu, Clinical Informatics fellow at Mass General Brigham and Cardiovascular Medicine fellow at Brigham and Women's Hospital.
Advanced Technical Capabilities
RECTIFIER's sophisticated analysis capabilities include reviewing clinical notes and electronic health records to assess various eligibility criteria, including:
- Patient symptoms
- Chronic disease status
- Current medication regimens
- Past medication history
- Other relevant clinical parameters
To maintain quality control, study staff conducted brief reviews of AI-identified eligible patients to verify the assessment's accuracy.
Addressing Bias Concerns
Given historical concerns about AI bias in healthcare, the researchers conducted thorough analyses of enrollment patterns across race, gender, and ethnicity. Comparing manual and AI-assisted screening methods, they found no significant demographic differences, suggesting the tool maintains fair and equitable patient selection.
Future Implications
Samuel Aronson, Executive Director of IT and AI Solutions for Mass General Brigham Personalized Medicine, expressed enthusiasm about the results: "Seeing this AI capability accelerate screening and trial enrollment this substantially in the context of a real-world randomized prospective trial is exciting. We look forward to using this capability to assist as many trials as we can."
This study builds upon previous research published in NEJM AI, which initially validated RECTIFIER's accuracy in retrospective health record reviews. The current findings confirm the tool's effectiveness in active clinical settings, suggesting a promising path forward for accelerating clinical research timelines and potentially reducing costs associated with trial recruitment.
