Med-PRM: Medical Reasoning Models with Stepwise, Guideline-verified Process Rewards
This addresses the critical need for accurate diagnosis and patient care in medicine by enabling fine-grained error correction in reasoning processes, representing a strong specific gain rather than a broad paradigm shift.
The paper tackled the problem of large language models struggling to localize and correct errors in clinical reasoning steps by introducing Med-PRM, a process reward modeling framework that verifies each step against medical knowledge bases, resulting in state-of-the-art performance with up to 13.50% improvement on base models and over 80% accuracy on MedQA using small-scale models.
Large language models have shown promise in clinical decision making, but current approaches struggle to localize and correct errors at specific steps of the reasoning process. This limitation is critical in medicine, where identifying and addressing reasoning errors is essential for accurate diagnosis and effective patient care. We introduce Med-PRM, a process reward modeling framework that leverages retrieval-augmented generation to verify each reasoning step against established medical knowledge bases. By verifying intermediate reasoning steps with evidence retrieved from clinical guidelines and literature, our model can precisely assess the reasoning quality in a fine-grained manner. Evaluations on five medical QA benchmarks and two open-ended diagnostic tasks demonstrate that Med-PRM achieves state-of-the-art performance, with improving the performance of base models by up to 13.50% using Med-PRM. Moreover, we demonstrate the generality of Med-PRM by integrating it in a plug-and-play fashion with strong policy models such as Meerkat, achieving over 80\% accuracy on MedQA for the first time using small-scale models of 8 billion parameters. Our code and data are available at: https://med-prm.github.io/