Table 4_Short-term risk stratification using parallel admission and reassessment features

684 infection-related pediatric intensive care unit (PICU) admissions from the PhysioNet Paediatric Intensive Care database were used to develop and validate an early risk prediction model for short-term major adverse events. The model, created by author Yuhang Wang, uses features from admission (0-6 hours) and early reassessment (12-36 hours) windows. The primary outcome was a 72-hour major adverse event, defined as new vasoactive drug use, invasive mechanical ventilation, or death.

Use Cases

• Training machine learning models for early risk prediction based on admission and reassessment features.
• Comparing model performance (LASSO, random forest, XGBoost, stacked ensemble) for clinical outcome forecasting.
• Conducting SHAP analysis to understand the complementary contributions of admission and reassessment features.

Strengths

• Model performance was validated in two cohorts: an internal test set and a de-identified timestamp-ordered validation cohort.
• The random forest model achieved an AUC of 0.724 and a PR-AUC of 0.741 in the internal test set.
• The dataset is focused on a specific, high-stakes clinical scenario: infection-related PICU admissions.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small (16.4 KB), indicating limited scope.

Provenance

Source

PhysioNet Paediatric Intensive Care database

Collection Method

Retrospective analysis of 684 infection-related PICU admissions.

Freshness

Last updated 2026-06-04 05:44:14; freshness should be verified.