Machine learning prediction of long-term postoperative pneumonia risk: a retrospective cohort study.
Source check: Source checked. Not independent scientific confirmation.
Finding: A retrospective 365-day postoperative-pneumonia risk model was tested on a separate internal set; the low 7.4% PPV narrows its potential to screening/rule-out use.
Technology: 11,655 surgical encounters, 238 365-day pneumonia events; RF with oversampling AUC .886, sensitivity 85.4%, specificity 77.4%, PPV 7.4%, NPV 99.6%; internal separate test and 5-fold CV.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- 11,655 surgical encounters, 238 365-day pneumonia events; RF with oversampling AUC .886, sensitivity 85.4%, specificity 77.4%, PPV 7.4%, NPV 99.6%; internal separate test and 5-fold CV.
- Test and result
- 11,655 surgical encounters, 238 365-day pneumonia events; RF with oversampling AUC .886, sensitivity 85.4%, specificity 77.4%, PPV 7.4%, NPV 99.6%; internal separate test and 5-fold CV.
- Limitation / still unknown
- The result is limited to the reported study population, data and evaluation; broader generalization or real-world benefit was not established.
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/s12911-026-03604-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: 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 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.