Predicting In-Hospital Mortality in Pediatric Sepsis: Machine Learning Development and Multicenter Validation.
Source check: Source checked. Not independent scientific confirmation.
Finding: Eight algorithms were evaluated for in-hospital mortality using predictors from the first 24 hours of pediatric ICU admission. XGBoost had AUC 0.821 (95% CI 0.751–0.892) and was externally assessed on the PIC pediatric and MIMIC-IV adult databases; the latter tests biological generalizability, not pediatric clinical deployment.
Technology: Eight machine-learning algorithms use first-24-hour ICU variables; SHAP interpretation and spline calibration were applied to the selected XGBoost model.
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Eight machine-learning algorithms use first-24-hour ICU variables; SHAP interpretation and spline calibration were applied to the selected XGBoost model.
- Test and result
- The retrospective development cohort was from Chongqing Children’s Hospital. XGBoost achieved AUC 0.821 (95% CI 0.751–0.892); external assessment used the PIC pediatric database and adult MIMIC-IV data.
- Limitation / still unknown
- Development was retrospective at one hospital; the external MIMIC-IV cohort was adult and supports biological generalizability only. No prospective pediatric workflow or mortality benefit was tested.
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/PMC13544363/ · https://europepmc.org/article/MED/42699765
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.