Beyond Idealized Patients: Evaluating LLMs under Challenging Patient Behaviors in Medical Consultations

TL;DR

This paper introduces CPB-Bench, a bilingual benchmark evaluating medical LLMs' responses to challenging patient behaviors, revealing specific failure patterns and the limited effectiveness of intervention strategies.

Contribution

It presents a new benchmark dataset for testing LLMs on realistic, challenging patient behaviors in medical dialogues and analyzes model performance and failure modes.

Findings

Models struggle with contradictory and medically implausible patient inputs.

Intervention strategies show inconsistent improvements and may cause unnecessary corrections.

The dataset and code are publicly released for further research.

Abstract

Large language models (LLMs) are increasingly used for medical consultation and health information support. In this high-stakes setting, safety depends not only on medical knowledge, but also on how models respond when patient inputs are unclear, inconsistent, or misleading. However, most existing medical LLM evaluations assume idealized and well-posed patient questions, which limits their realism. In this paper, we study challenging patient behaviors that commonly arise in real medical consultations and complicate safe clinical reasoning. We define four clinically grounded categories of such behaviors: information contradiction, factual inaccuracy, self-diagnosis, and care resistance. For each behavior, we specify concrete failure criteria that capture unsafe responses. Building on four existing medical dialogue datasets, we introduce CPB-Bench (Challenging Patient Behaviors Benchmark), a bilingual (English and Chinese) benchmark of 692 multi-turn dialogues annotated with these behaviors. We evaluate a range of open- and closed-source LLMs on their responses to challenging patient utterances. While models perform well overall, we identify consistent, behavior-specific failure patterns, with particular difficulty in handling contradictory or medically implausible patient information. We also study four intervention strategies and find that they yield inconsistent improvements and can introduce unnecessary corrections. We release the dataset and code.