Superhuman performance of a large language model on the reasoning tasks of a physician

2024-12-14

arXiv.org

AI Summary

This study systematically evaluates large language models (LLMs) on complex clinical diagnostic reasoning across five dimensions: differential diagnosis generation, diagnostic and management reasoning, triage, probabilistic reasoning, and clinical decision calibration. Using the OpenAI o1-preview model augmented with chain-of-thought prompting, outputs were rigorously adjudicated by physician experts applying psychometric standards. To our knowledge, this is the first empirically validated study demonstrating LLM superiority over human clinicians in specific clinical reasoning tasks—namely, differential diagnosis generation and diagnostic/management reasoning quality—outperforming both historical physician cohorts and prior LLMs. However, the model fell short of human performance in probabilistic reasoning and triage. The work advances AI evaluation paradigms for real-world clinical settings by establishing a novel methodology centered on expert adjudication and standardized psychometric frameworks, thereby setting a benchmark for rigorous, clinically grounded LLM assessment.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Performance of large language models (LLMs) on medical tasks has traditionally been evaluated using multiple choice question benchmarks. However, such benchmarks are highly constrained, saturated with repeated impressive performance by LLMs, and have an unclear relationship to performance in real clinical scenarios. Clinical reasoning, the process by which physicians employ critical thinking to gather and synthesize clinical data to diagnose and manage medical problems, remains an attractive benchmark for model performance. Prior LLMs have shown promise in outperforming clinicians in routine and complex diagnostic scenarios. We sought to evaluate OpenAI's o1-preview model, a model developed to increase run-time via chain of thought processes prior to generating a response. We characterize the performance of o1-preview with five experiments including differential diagnosis generation, display of diagnostic reasoning, triage differential diagnosis, probabilistic reasoning, and management reasoning, adjudicated by physician experts with validated psychometrics. Our primary outcome was comparison of the o1-preview output to identical prior experiments that have historical human controls and benchmarks of previous LLMs. Significant improvements were observed with differential diagnosis generation and quality of diagnostic and management reasoning. No improvements were observed with probabilistic reasoning or triage differential diagnosis. This study highlights o1-preview's ability to perform strongly on tasks that require complex critical thinking such as diagnosis and management while its performance on probabilistic reasoning tasks was similar to past models. New robust benchmarks and scalable evaluation of LLM capabilities compared to human physicians are needed along with trials evaluating AI in real clinical settings.

Problem

• Evaluating LLM performance on clinical diagnostic reasoning tasks
• Comparing AI and physician diagnostic accuracy in emergency settings
• Assessing LLM superhuman abilities in medical decision-making and reasoning

Innovation

• LLM outperforms physicians in clinical reasoning
• Five experiments validate AI diagnostic accuracy
• Real-world ER study confirms LLM superhuman performance