Using rules with patient data improves clinical guideline decisions
More Reasoning Is Not Always Better: Executing Clinical Guidelines with LLMs and Rules
Computation and Language
Summary
Large language models can help doctors by suggesting treatments, but if they try to think through each step too much, they can be slow and make mistakes. The researchers created a system called MedDEx that uses AI to understand patient information and then follows clear rules to decide treatments based on medical guidelines. This method is more accurate and faster than letting AI handle all the reasoning itself. Their results show that adding more AI thinking steps does not always lead to better medical decisions.
What this means in practice
- •For hospital data teams: Implement systems that derive patient information with AI then apply fixed clinical rules to improve accuracy and reduce computation.
- •For medical software developers: Build clinical decision support tools separating patient data extraction from guideline logic to improve treatment recommendations.
Authors
Xiaoqian Lu, Guangfu Guo, Long Cheng, Kunpeng Liu
Abstract
Large language models (LLMs) have shown considerable promise in clinical decision support. However, existing approaches often rely on multi-step reasoning or multi-agent collaboration to execute clinical guidelines, introducing substantial computational overhead without reliably preventing decision errors. To address these limitations, we propose a framework for Medical Decision Execution (MedDEx) that integrates structured patient states with deterministic rule execution to support guideline-based treatment planning. MedDEx separates clinical record interpretation from guideline execution, leveraging LLMs to extract patient information and generate treatment plans while assigning guideline-defined decision logic to a rule executor. Experiments across multiple clinical guideline benchmarks demonstrate that MedDEx improves decision accuracy while reducing inference costs. Further analyses highlight the critical roles of structured patient representations and deterministic rule execution, suggesting that additional LLM reasoning does not necessarily translate into better guideline adherence.