AI-Selected Trial Candidates are More Diverse
Could artificial intelligence-enabled medical chart review improve the speed, accuracy, and equity of study enrollment?
-
1
An AI system analyzed electronic medical records to identify eligible patients for a clinical trial for a rare heart condition.
-
2
The AI platform processed 1,476 patient records in six days, assessing numerous eligibility criteria efficiently.
-
3
AI-identified candidates for the trial were more diverse, with 37% being Black compared to only 7% from manual identification.
-
4
The AI system demonstrated 96.2% accuracy in answering trial-relevant questions and 99% reliability in excluding ineligible patients.
-
5
AI-assisted screening may broaden recruitment and reduce bias in clinical trials, improving efficiency in patient identification.
