aiqbench analysis
This research proposes a retrieval-augmented framework to reconstruct temporally grounded patient trajectories by aligning unstructured clinical narratives with structured EHR data. The study demonstrates that integrating these modalities generally improves temporal performance metrics without compromising event extraction accuracy, though significant uncertainty remains due to the limited sample size of 40 cases. These findings advance the capability of clinical language models to understand disease progression through time rather than just static snapshots.
Text Knows What, Tables Know When: Clinical Timeline Reconstruction via Retrieval-Augmented Multimodal Alignment
arXiv:2605.15168v2 Announce Type: replace-cross Abstract: Clinical language models increasingly operate over electronic health records (EHRs), yet patient records are not stored as temporally grounded trajectories. Clinical notes describe symptoms, assessments, and disease progression, but often compress or narratively reorder events. Structured EHR rows provide timestamps for labs, medications, vitals, and procedures, but capture only part of the clinical story. We formulate clinical timeline reconstruction as retrieval-augmented temporal grounding: constructing a patient trajectory by using narrative text for event semantics and structured rows as partial temporal evidence. We introduce a scaffolded workflow that extracts central narrative events, builds an initial temporal scaffold, attaches non-central events, and calibrates timestamps using retrieved structured EHR rows. We evaluate on 40 discharge summaries, including 15 i2b2-derived and 25 MIMIC-IV summaries, each with manual gold-standard timelines and aligned structured EHR data. Across models, multimodal calibration left event match rates largely unchanged and generally improved temporal performance: mean paired case-level multimodal-unimodal differences were positive in 7 of 12 model-metric comparisons across concordance and AULTC, with none negative. However, uncertainty was substantial given the 40-case sample; paired case-level bootstrap intervals excluded zero only for the DeepSeek V3.2 AULTC improvement. A gap analysis shows that 35.1% of text-derived events have no structured counterpart. These findings support treating structured EHR data as partial temporal evidence for narrative-derived patient trajectories.
Abstract:Clinical language models increasingly operate over electronic health records (EHRs), yet patient records are not stored as temporally grounded trajectories. Clinical notes describe symptoms, assessments, and disease progression, but often compress or narratively reorder events. Structured EHR rows provide timestamps for labs, medications, vitals, and procedures, but capture only part of the clinical story. We formulate clinical timeline reconstruction as retrieval-augmented temporal grounding: constructing a patient trajectory by using narrative text for event semantics and structured rows as partial temporal evidence. We introduce a scaffolded workflow that extracts central narrative events, builds an initial temporal scaffold, attaches non-central events, and calibrates timestamps using retrieved structured EHR rows. We evaluate on 40 discharge summaries, including 15 i2b2-derived and 25 MIMIC-IV summaries, each with manual gold-standard timelines and aligned structured EHR data. Across models, multimodal calibration left event match rates largely unchanged and generally improved temporal performance: mean paired case-level multimodal-unimodal differences were positive in 7 of 12 model-metric comparisons across concordance and AULTC, with none negative. However, uncertainty was substantial given the 40-case sample; paired case-level bootstrap intervals excluded zero only for the DeepSeek V3.2 AULTC improvement. A gap analysis shows that 35.1% of text-derived events have no structured counterpart. These findings support treating structured EHR data as partial temporal evidence for narrative-derived patient trajectories.
Accepted for oral presentation at the Pacific Symposium on Biocomputing (PSB) 2027. Sayantan Kumar, Shahriar Noroozizadeh, Juyong Kim (authors contributed equally)
Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Machine Learning (stat.ML)
arXiv:2605.15168 [cs.CL]
(or arXiv:2605.15168v2 [cs.CL] for this version)
Full text under the publisher's open licence · 256 words
Related coverage
Microsoft’s AI chief says superintelligence is near, but won’t take your job
The Verge · Jun 8, 2026
Agentic conversational video intelligence built on AWS
AWS · Sep 23, 2026
The payload: the semantic layer, memory, and the loop that fills it
Neo4j · Aug 10, 2026
Presentation: The Future of Engineering: Mindsets That Matter When Code Isn’t Enough
InfoQ · Jul 28, 2026