Aligning Clinical Needs and AI Capabilities: A Survey on LLMs for Medical Reasoning

2026-07-08

AI Summary

This study addresses the critical challenge of systematically aligning large language models’ reasoning capabilities with real-world clinical demands to enhance their reliability and applicability in healthcare settings. The authors propose the first analytical framework that integrates Miller’s pyramid of clinical competence with deductive, inductive, and abductive reasoning paradigms, enabling the construction of a benchmark dataset spanning five levels of medical reasoning. Through multidimensional evaluation of 18 state-of-the-art models, the study reveals that specialized models excel in diagnostic tasks, whereas general-purpose models demonstrate superior performance in clinical decision support and patient–physician communication. The findings also highlight persistent challenges in hallucination control, data scarcity, and practical deployment in clinical workflows.

Abstract

Large language models (LLMs) have emerged as important tools in healthcare, showing growing potential for clinical reasoning and patient care. This survey examines recent progress in medical LLMs, focusing on reasoning applications and requirements. We present a dual-view approach that connects clinical practice with computational methods. On the clinical side, we establish a five-level competency scheme following Miller's Pyramid, progressing from knowledge recall to dynamic case management. On the computational side, we link deductive, inductive, and abductive reasoning patterns to common medical goals and tasks. We also introduce a benchmark dataset spanning five levels of medical reasoning capability and report results on 18 state-of-the-art models, revealing that medical specialist models excel in diagnosis-centric tasks while general models lead in decision support and dialogue. We conclude by discussing current progress and open challenges, including data limitations, hallucination, and grounding issues, and outline directions toward safer, more reliable, and workflow-ready systems.

Problem

• clinical reasoning
• large language models
• medical AI
• competency alignment
• healthcare applications

Innovation

• medical reasoning
• large language models
• Miller's Pyramid
• reasoning patterns
• benchmark dataset