Learning estimates mortality after femoral neck fracture
Source check: Source checked. Not independent scientific confirmation.
Finding: Machine-learning models estimated 90-day and two-year mortality in 2,478 operated patients aged 65 or older with femoral neck fractures. In an internal 495-person test set, selected models achieved concordance of about 0.74 and 0.70, respectively. Model selection on test performance and preprocessing can make these estimates optimistic; the study did not demonstrate improved treatment decisions or survival.
Technology: Stochastic gradient boosting; elastic-net regression
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Supervised boosting and penalized regression learn mortality associations from recorded preoperative clinical characteristics.
- Test and result
- Retrospective 80:20 patient split with measured 90-day and two-year mortality; training cross-validation and internal calibration assessment.
- Limitation / still unknown
- Independent frozen-model evaluation and clinical benefit remain unestablished; only operated patients were included, and final model selection used internal test performance.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC9712294/
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: 21 Nov 2026. Review scope: Primary relevant model Methods, Results and limitations; exact primary identity and current primary PubMed correction/retraction relationships checked..
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 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.