Abstract:It is becoming increasingly emphasis on the importance of LLM participating in clinical diagnosis decision-making. However, the low specialization refers to that current medical LLMs can not provide specific medical advice, which are more like a medical Q\&A. And there is no suitable clinical guidance tree data set that can be used directly with LLM. To address this issue, we first propose LLM-executavle clinical guidance tree(CGT), which can be directly used by large language models, and construct medical diagnostic decision-making dataset (MedDM), from flowcharts in clinical practice guidelines. We propose an approach to screen flowcharts from medical literature, followed by their identification and conversion into standardized diagnostic decision trees. Constructed a knowledge base with 1202 decision trees, which came from 5000 medical literature and covered 12 hospital departments, including internal medicine, surgery, psychiatry, and over 500 diseases.Moreover, we propose a method for reasoning on LLM-executable CGT and a Patient-LLM multi-turn dialogue framework.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper aims to solve the problem of the lack of professionalism of large - language models (LLMs) in clinical diagnosis decision - making in the current medical field. Specifically, the existing medical LLMs are unable to obtain complete information through multi - round inquiries and provide specific medical advice, and are more like a simple medical Q&A system. In addition, there is currently no clinical guidance tree dataset suitable for direct use in combination with LLMs. To solve these problems, the author proposes the following points: 1. **Propose a clinically - executable clinical guidance tree (CGT) for LLMs**: - This new decision - tree structure can be directly used in large - language models, enabling LLMs to reason and generate more reliable medical responses. - CGT uses natural language to represent node content, enhancing the clarity of information transmission and improving compatibility with LLMs. 2. **Construct a medical diagnosis decision - making dataset (MedDM)**: - Extracted and transformed from the flowcharts in clinical practice guidelines, a dataset containing 1,202 decision trees was constructed, covering 5,000 medical literatures and 12 hospital departments, involving more than 500 diseases. - The construction process of the dataset includes steps such as image extraction, flowchart recognition, text recognition, and manual verification to ensure the accuracy and integrity of the data. 3. **Propose a generation framework based on decision - retrieval**: - This framework can handle medical decision - making knowledge and generate diagnostically - dialog - response - rich and accurate information. 4. **Develop a clinical decision - making reasoning engine (CDM Engine)**: - Based on the CGT executable by LLMs, using the powerful analytical capabilities of LLMs, combined with patient information for step - by - step reasoning, and taking different measures according to different reasoning results. - The engine can not only make multi - step decisions, but also generate follow - up questions for multi - round conversations with patients, and finally make accurate diagnoses for patients. Through these innovations, the paper hopes to improve the application effect of medical LLMs in actual medical scenarios and enhance their professionalism and accuracy.

MedDM:LLM-executable clinical guidance tree for clinical decision-making

Improving Clinical Expertise in Large Language Models Using Electronic Medical Records

Evaluating large language model workflows in clinical decision support: referral, triage, and diagnosis

MedKP: Medical Dialogue with Knowledge Enhancement and Clinical Pathway Encoding

Large language models in solving clinical dilemmas - advantages and drawbacks

LLM-Mini-CEX: Automatic Evaluation of Large Language Model for Diagnostic Conversation

Enhancing Large Language Models for Clinical Decision Support by Incorporating Clinical Practice Guidelines

From Beginner to Expert: Modeling Medical Knowledge into General LLMs

CliBench: A Multifaceted and Multigranular Evaluation of Large Language Models for Clinical Decision Making

MEDIC: Towards a Comprehensive Framework for Evaluating LLMs in Clinical Applications

medIKAL: Integrating Knowledge Graphs as Assistants of LLMs for Enhanced Clinical Diagnosis on EMRs

The Application of LLMs for Radiologic Decision-Making

Lingdan: enhancing encoding of traditional Chinese medicine knowledge for clinical reasoning tasks with large language models

AI Hospital: Interactive Evaluation and Collaboration of LLMs As Intern Doctors for Clinical Diagnosis

Generation of guideline-based clinical decision trees in oncology using large language models

DISC-MedLLM: Bridging General Large Language Models and Real-World Medical Consultation

Large Language Models and Medical Knowledge Grounding for Diagnosis Prediction

The Potential of LLMs in Medical Education: Generating Questions and Answers for Qualification Exams

MDAgents: An Adaptive Collaboration of LLMs for Medical Decision-Making

Systematic analysis of ChatGPT, Google search and Llama 2 for clinical decision support tasks