PrecepTron enables large scale evaluation of medical AI reasoning
Scaling Clinical Judgment to Evaluate Medical AI
Artificial Intelligence
Summary
Evaluating how well AI models think like doctors is hard because it usually needs many doctors reviewing answers, which is expensive and slow. The authors created PrecepTron, an AI trained to judge medical AI answers like a real doctor, using only a small set of examples from physicians. They also made a big set of doctor-scored questions called GRAND-ROUNDS. With PrecepTron, they could check how medical AI models reason without needing doctors to grade everything, and even explore new ways to test AI on medical cases bit by bit.
What this means in practice
- •For hospital data teams: Automate scoring of medical AI responses to enable rapid quality checks in clinical AI tools without requiring physician time.
- •For clinical ai developers: Use PrecepTron to benchmark and improve AI models’ clinical reasoning consistently across multiple medical tasks and datasets.
Authors
Thomas A. Buckley, Zahir Kanjee, Peter G. Brodeur, Byron Crowe, Anthony M. Pettinato, Aashna P. Shah, Adrian D. Haimovich, Liam G. McCoy, Daniel Restrepo, Ethan Goh, Jonathan H. Chen, Laura Zwaan, Katherine E. Goodman, Daniel J. Morgan, Raja-Elie E. Abdulnour, Adam Rodman, Arjun K. Manrai
Abstract
Blinded physician evaluation has been considered by many to be the gold standard for assessing clinical reasoning in large language models (LLMs). This is difficult to scale; thus, prior studies typically rely on small physician panels, often from a single institution or specialty, which both limits the scientific questions investigated and makes it unclear whether findings would be reproduced with a different set of evaluators. To more rigorously and scalably study clinical reasoning in AI models, here we introduce PrecepTron, an LLM fine-tuned for physician-level evaluation of open-ended responses. PrecepTron was trained using low-rank adaptation (LoRA) of a 32-billion-parameter model on a small number of physician examples. We also release GRAND-ROUNDS, a new large-scale physician-annotated benchmark of 9,217 scores by 11 physicians across seven studies. We show that frontier LLMs in typical "LLM-as-a-judge" approaches often disagree with physicians and with each other, but fine-tuning PrecepTron on a small number of cases enables physician-level consistent scoring across tasks. We use PrecepTron to reproduce headline findings from five influential studies assessing LLMs for clinical care in JAMA, Science, and Nature Medicine without new human grading. Using PrecepTron, we then pose new questions about how LLMs reason in medicine that would have been infeasible with human grading alone, including measuring the diagnostic accuracy of frontier LLMs when clinical cases are provided piecemeal, even token by token. Together, PrecepTron and GRAND-ROUNDS provide a foundation for reproducible, large-scale study of how LLMs reason in medicine. All code, data, and labels are made freely available for researchers.