MTDiag: A Multi-Turn Diagnostic Dataset Towards Clinically Meaningful LLM Evaluation

Abstract:Clinical diagnosis is fundamentally interactive and incremental, yet the dominant paradigm for evaluating Large Language Models (LLMs) in medicine remains static QA benchmarks or template-based dialogues. These benchmarks say little about whether a model can serve as a diagnostic agent in a dynamic clinical encounter, with LLMs showing significant accuracy and reliability degradation in multi-turn settings. To address this issue, we present MTDiag, a large multi-turn diagnostic dialogue dataset constructed from three heterogeneous sources: DDXPlus, MIMIC-IV, and published case reports (AJCR), covering common ED presentations as well as long-tail rare and atypical conditions. All cases are normalized into a canonical schema anchored in the most comprehensive and widely-adopted medical knowledge bases (UMLS concept identifiers, with ICD-10 diagnosis codes). We release the schema, a UserLM-8B-based utterance-generation pipeline, and the physician-validated dataset that converts structured clinical evidence into natural-language utterances. Importantly, we introduce and motivate clinical knowledge-grounded metrics for evaluating LLMs as diagnostic agents, beyond diagnostic accuracy, for the task of multi-turn differential diagnosis.

Computation and Language (cs.CL)

arXiv:2608.25085 [cs.CL]

(or arXiv:2608.25085v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2608.25085 arXiv-issued DOI via DataCite (pending registration)

Chouayfati, Pia, et al. "MTDiag: A Multi-Turn Diagnostic Dataset Towards Clinically Meaningful LLM Evaluation." Proceedings of the 27th Annual Meeting of the Special Interest Group on Discourse and Dialogue. 2026

From: Alexander Fichtl M.Sc. [view email] [v1] Tue, 25 Aug 2026 19:27:10 UTC (1,719 KB)