DDO: Dual-Decision Optimization for LLM-Based Medical Consultation via Multi-Agent Collaboration

2025/05/24 by Jia, Zhihao, Jia, Mingyi, Duan, Junwen +1
#Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #FOS: Computer and information sciences #Multiagent Systems (cs.MA)

paper · doi:10.48550/arxiv.2505.18630

Abstract

Large Language Models (LLMs) demonstrate strong generalization and reasoning abilities, making them well-suited for complex decision-making tasks such as medical consultation (MC). However, existing LLM-based methods often fail to capture the dual nature of MC, which entails two distinct sub-tasks: symptom inquiry, a sequential decision-making process, and disease diagnosis, a classification problem. This mismatch often results in ineffective symptom inquiry and unreliable disease diagnosis. To address this, we propose DDO, a novel LLM-based framework that performs Dual-Decision Optimization by decoupling the two sub-tasks and optimizing them with distinct objectives through a collaborative multi-agent workflow. Experiments on three real-world MC datasets show that DDO consistently outperforms existing LLM-based approaches and achieves competitive performance with state-of-the-art generation-based methods, demonstrating its effectiveness in the MC task. The code is available at https://github.com/zh-jia/DDO.

Citations

• MEDDxAgent: A Unified Modular Agent Framework for Explainable Automatic Differential Diagnosis
• Citrus: Leveraging Expert Cognitive Pathways in a Medical Language Model for Advanced Medical Decision Support
• Memorize and Rank: Elevating Large Language Models for Clinical Diagnosis Prediction
• Inferring from Logits: Exploring Best Practices for Decoding-Free Generative Candidate Selection
• Feedback-Aware Monte Carlo Tree Search for Efficient Information Seeking in Goal-Oriented Conversations
• MAGDA: Multi-agent guideline-driven diagnostic assistance
• Enhancing Healthcare LLM Trust with Atypical Presentations Recalibration
• Large Language Models for Disease Diagnosis: A Scoping Review
• CoD, Towards an Interpretable Medical Agent using Chain of Diagnosis
• Qwen2 Technical Report
• medIKAL: Integrating Knowledge Graphs as Assistants of LLMs for Enhanced Clinical Diagnosis on EMRs
• MediQ: Question-Asking LLMs and a Benchmark for Reliable Interactive Clinical Reasoning
• Group Robust Preference Optimization in Reward-free RLHF
• MedAdapter: Efficient Test-Time Adaptation of Large Language Models towards Medical Reasoning
• Agent Hospital: A Simulacrum of Hospital with Evolvable Medical Agents
• Multicalibration for Confidence Scoring in LLMs
• ChiMed-GPT: A Chinese Medical Large Language Model with Full Training Regime and Better Alignment to Human Preferences
• GPT-4 Technical Report
• Diaformer: Automatic Diagnosis via Symptoms Sequence Generation
• End-to-End Knowledge-Routed Relational Dialogue System for Automatic Diagnosis
• Proximal Policy Optimization Algorithms

Related

• MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making
• MDTeamGPT: A Self-Evolving LLM-based Multi-Agent Framework for Multi-Disciplinary Team Medical Consultation
• A Demonstration of Adaptive Collaboration of Large Language Models for Medical Decision-Making
• DoctorAgent-RL: A Multi-Agent Collaborative Reinforcement Learning System for Multi-Turn Clinical Dialogue
• Beyond Direct Diagnosis: LLM-based Multi-Specialist Agent Consultation for Automatic Diagnosis