Pattern Recognition or Medical Knowledge? The Problem with Multiple-Choice Questions in Medicine
This highlights a critical flaw in evaluating LLMs for medicine, potentially misleading assessments of clinical reasoning, and calls for more robust evaluation methods.
The study investigated whether multiple-choice question (MCQ) benchmarks overestimate clinical understanding in large language models (LLMs) by creating a fictional medical benchmark with an imaginary organ, finding that models achieved 64% accuracy while physicians scored only 27%, indicating reliance on pattern recognition rather than reasoning.
Large Language Models (LLMs) such as ChatGPT demonstrate significant potential in the medical domain and are often evaluated using multiple-choice questions (MCQs) modeled on exams like the USMLE. However, such benchmarks may overestimate true clinical understanding by rewarding pattern recognition and test-taking heuristics. To investigate this, we created a fictional medical benchmark centered on an imaginary organ, the Glianorex, allowing us to separate memorized knowledge from reasoning ability. We generated textbooks and MCQs in English and French using leading LLMs, then evaluated proprietary, open-source, and domain-specific models in a zero-shot setting. Despite the fictional content, models achieved an average score of 64%, while physicians scored only 27%. Fine-tuned medical models outperformed base models in English but not in French. Ablation and interpretability analyses revealed that models frequently relied on shallow cues, test-taking strategies, and hallucinated reasoning to identify the correct choice. These results suggest that standard MCQ-based evaluations may not effectively measure clinical reasoning and highlight the need for more robust, clinically meaningful assessment methods for LLMs.