Supervised machine learning models for predicting sepsis-associated acute kidney injury in children: a real-world evaluation
Machine learning identified early warning signs of kidney injury in children with sepsis, a major cause of death in pediatric intensive care.
This retrospective study of 2424 pediatric sepsis patients (Phoenix criteria) developed supervised ML models to predict sepsis-associated AKI, identifying key clinical predictors under the new Phoenix definition. The study provides actionable predictive tools for early AKI identification in a high-mortality pediatric population, though specific AUC values were not available in the retrieved abstract portion.
What the study was
- Study design
- Retrospective cohort study
- Population
- Pediatric patients with sepsis (Phoenix criteria) at Children's Hospital, Nanjing (n=2424; 484 with AKI)
- Sample size
- 2424
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- World Journal of Pediatrics
Why it surfaced
Relevant real-world ML application for pediatric critical care. Abstract truncated — specific AUC and full validation details not retrievable; classification_confidence set to medium.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.