VeriSim: A Configurable Framework for Evaluating Medical AI Under Realistic Patient Noise

2026-04-11

AI Summary

This study addresses a critical gap in current medical AI evaluation benchmarks, which overlook communication noise arising in real clinical settings—such as patient memory lapses, limited health literacy, or anxiety—leading to inflated performance estimates that diverge from real-world effectiveness. To bridge this gap, the authors propose VeriSim, a configurable patient simulation framework that systematically models six empirically grounded noise dimensions derived from medical communication literature. VeriSim employs a hybrid verification mechanism combining UMLS and large language models to ensure injected noise remains medically plausible. Experiments reveal that realistic noise reduces diagnostic accuracy by 15–25% and increases dialogue length by 34–55% across seven open-source medical large language models, with standard fine-tuning proving insufficient for robustness. Clinician evaluations confirm high simulation fidelity (κ > 0.80), exposing a substantial sim-to-real performance gap.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

Medical large language models (LLMs) achieve impressive performance on standardized benchmarks, yet these evaluations fail to capture the complexity of real clinical encounters where patients exhibit memory gaps, limited health literacy, anxiety, and other communication barriers. We introduce VeriSim, a truth-preserving patient simulation framework that injects controllable, clinically evidence-grounded noise into patient responses while maintaining strict adherence to medical ground truth through a hybrid UMLS-LLM verification mechanism. Our framework operationalizes six noise dimensions derived from peer-reviewed medical communication literature, capturing authentic clinical phenomena such as patient recall limitations, health literacy barriers, and stigma-driven non-disclosure. Experiments across seven open-weight LLMs reveal that all models degrade significantly under realistic patient noise, with diagnostic accuracy dropping 15-25% and conversation length increasing 34-55%. Notably, smaller models (7B) show 40% greater degradation than larger models (70B+), while medical fine-tuning on standard corpora provides limited robustness benefits against patient communication noise. Evaluation by board-certified clinicians demonstrates high-quality simulation with strong inter-annotator agreement (kappa > 0.80), while LLM-as-a-Judge serves as a validated auxiliary evaluator achieving comparable reliability for scalable assessment. Our results highlight a critical Sim-to-Real gap in current medical AI. We release VeriSim as an open-source noise-injection framework, establishing a rigorous testbed for evaluating clinical robustness.

Problem

• medical AI
• patient noise
• clinical robustness
• communication barriers
• Sim-to-Real gap

Innovation

• patient simulation
• realistic noise injection
• medical LLM robustness
• UMLS-LLM verification
• Sim-to-Real gap