The application of machine learning for identifying frailty in older patients during hospital admission.
Source check: Source checked. Not independent scientific confirmation.
Finding: Among 3,367 older inpatients, 2,843 were labeled frail; SVM achieved internal test AUROC 0.811, accuracy 73.15%, precision 94.53% and F1 82.10%. Age and routine clinical features ranked among influential predictors. Randomly held-out test set from the same single hospital and six-month data window; no external/prospective validation. The high frailty prevalence and single-center retrospective sample limit transfer; model performance is not proof of improved patient outcomes.
Technology: Compared logistic regression, random forest, XGBoost and SVM on routine admission-period clinical data; SHAP/PDP interpreted frailty predictions.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Compared logistic regression, random forest, XGBoost and SVM on routine admission-period clinical data; SHAP/PDP interpreted frailty predictions.
- Test and result
- Among 3,367 older inpatients, 2,843 were labeled frail; SVM achieved internal test AUROC 0.811, accuracy 73.15%, precision 94.53% and F1 82.10%. Age and routine clinical features ranked among influential predictors.
- Limitation / still unknown
- Randomly held-out test set from the same single hospital and six-month data window; no external/prospective validation. The high frailty prevalence and single-center retrospective sample limit transfer; model performance is not proof of improved patient outcomes.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://link.springer.com/article/10.1186/s12911-024-02684-z
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: The BMC primary full text reports a machine-learning frailty screen trained on 3,367 hospitalized older adults (2,843 frail); the internal test set gave SVM AUROC 0.811, accuracy 73.15%, precision 94.53% and F1 82.10%, with age and routine clinical features important to predictions. This is a specific clinical screening result for acute-care inpatients. The first located result-bearing version is the publisher article dated 2024-09-27; exact-title/DOI searches found no earlier matching preprint, which does not establish global priority. Retrospective single-center design and internal testing limit generalization; no clinical utility or intervention benefit is claimed..
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 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.