PICU Infection Risk Stratification Using Admission and Reassessment Features

The PhysioNet Paediatric Intensive Care database provided 684 infection-related PICU admissions for this analysis. The dataset was created by Yuhang Wang to develop and validate an early risk prediction model for short-term major adverse events. It was last updated on June 4, 2026.

Use Cases

• Develop early warning models for major adverse events based on admission (M0) features.
• Validate reassessment-oriented risk stratification using features from the 12–36 hour (M1) window.
• Compare machine learning model performance (LASSO, random forest, XGBoost, ensemble) for clinical prediction.
• Perform feature importance analysis (e.g., SHAP) to understand contributions of M0 and M1 variables.

Strengths

• Includes 684 infection-related PICU admissions from the PhysioNet database.
• Model performance was validated in a de-identified timestamp-ordered cohort.
• The random forest model achieved an AUC of 0.724 in the internal test set.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is very small at 16.9 KB, indicating limited scope.

Provenance

Source

PhysioNet Paediatric Intensive Care database

Collection Method

Features were extracted from admission (0–6 h) and reassessment (12–36 h) windows.

Freshness

Last updated 2026-06-04 05:44:15