Decoding Complications: Large Language Model-Based Classification of Free-Text Entries in Clinical Stroke Registries
Source check: Source checked. Not independent scientific confirmation.
Finding / tested result: The primary source Methods/Results confirm the represented result: An LLM-guided taxonomy classifies 9,146 free-text stroke-registry adverse-event entries, with frozen GPT-5.4 labels matching clinician consensus on 300 held-out entries (weighted F1 96.5%) and recovering 19.6% more adverse-event instances than structured fields.
Technology: GPT-5.4 with an LLM-guided taxonomy
Family GPT-5: profile · models page
Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.
- AI contribution
- GPT-5.4 with an LLM-guided taxonomy is used in the reported analysis or evaluation.
- Limitation / still unknown
- The source reports one registry and a 300-entry clinician-consensus holdout; this does not establish performance at other registries or patient benefit.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://doi.org/10.1007/s10916-026-02486-y
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 3/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.