MIMIC-SR-ICD11: A Dataset for Narrative-Based Diagnosis
This work addresses the challenge of leveraging self-reports for more accurate diagnosis in healthcare, though it appears incremental as it builds on existing datasets and methods.
The authors tackled the problem of disease diagnosis from narrative clinical reports by introducing MIMIC-SR-ICD11, a dataset aligned with WHO ICD-11 terminology, and LL-Rank, a re-ranking framework that outperformed a baseline method across seven model backbones, with gains attributed to isolating semantic compatibility from label frequency bias.
Disease diagnosis is a central pillar of modern healthcare, enabling early detection and timely intervention for acute conditions while guiding lifestyle adjustments and medication regimens to prevent or slow chronic disease. Self-reports preserve clinically salient signals that templated electronic health record (EHR) documentation often attenuates or omits, especially subtle but consequential details. To operationalize this shift, we introduce MIMIC-SR-ICD11, a large English diagnostic dataset built from EHR discharge notes and natively aligned to WHO ICD-11 terminology. We further present LL-Rank, a likelihood-based re-ranking framework that computes a length-normalized joint likelihood of each label given the clinical report context and subtracts the corresponding report-free prior likelihood for that label. Across seven model backbones, LL-Rank consistently outperforms a strong generation-plus-mapping baseline (GenMap). Ablation experiments show that LL-Rank's gains primarily stem from its PMI-based scoring, which isolates semantic compatibility from label frequency bias.