Related Experiment Videos
Short-term risk stratification using parallel admission and reassessment features in PICU patients with infection
Yuhang Wang1, Cuie Chen1, Guocheng Jiang1
1Department of Pediatrics, Yiwu Maternity and Children Hospital, Yiwu, Zhejiang, China.
Insights
This study developed a risk model to predict major adverse events in pediatric intensive care unit patients with infection. The random forest model showed moderate success, highlighting potential for early risk stratification.
Area of Science:
- Pediatric critical care medicine
- Clinical informatics
- Machine learning in healthcare
Background:
- Infection in the pediatric intensive care unit (PICU) is associated with significant morbidity and mortality.
- Early identification of high-risk patients is crucial for timely intervention and improved outcomes.
- Existing risk prediction models may not fully capture dynamic changes in patient status during PICU stay.
Purpose of the Study:
- To develop and validate an early risk prediction model for short-term major adverse events (MAE) in pediatric intensive care unit (PICU) admissions with infection.
- To utilize features from initial admission (M0) and early reassessment (M1) for enhanced prediction accuracy.
- To compare the performance of different machine learning models for this prediction task.
Main Methods:
- Utilized the PhysioNet Paediatric Intensive Care database, including 684 infection-related PICU admissions.
- Defined primary outcome as 72-hour MAE (vasoactive drug use, mechanical ventilation, or death).
- Extracted features from M0 (0-6h) and M1 (12-36h) time windows; compared LASSO, random forest (RF), XGBoost, and stacked ensemble models using AUC, PR-AUC, and calibration metrics.
Main Results:
- The random forest (RF) model demonstrated the best performance, achieving an AUC of 0.724 and PR-AUC of 0.741 in the internal test set.
- In a validation cohort, RF achieved an AUC of 0.718 and PR-AUC of 0.766.
- SHAP analysis confirmed the complementary value of both M0 and M1 features for prediction, though calibration was attenuated in the validation cohort.
Conclusions:
- A combined M0+M1 prediction framework offers moderate discrimination for early risk stratification in infection-related PICU admissions.
- The developed model shows potential utility for reassessment-oriented clinical decision-making.
- Further external validation and recalibration are necessary before widespread clinical implementation.
Objective:
To develop and validate an early risk prediction model for short-term major adverse events (MAE) in pediatric intensive care unit (PICU) admissions with infection using admission (M0) and early reassessment (M1) features.
Methods:
Using the PhysioNet Paediatric Intensive Care database, 684 infection-related PICU admissions were included in the final landmark-defined analytic cohort. The primary outcome was 72-hour MAE, defined as new vasoactive drug use, invasive mechanical ventilation, or death within 72 h after PICU admission. For the primary M0 + M1 analysis, prediction targeted events occurring after completion of the reassessment window. Features were extracted from M0 (0-6 h) and M1 (12-36 h). LASSO, random forest (RF), XGBoost, and a stacked ensemble were compared. Performance was assessed in an internal test cohort and a de-identified timestamp ordered validation cohort using AUC, PR-AUC, calibration metrics, Brier score, and decision curve analysis.
Results:
Among the candidate models, RF showed the most favorable overall numerical performance. In the internal test set, RF achieved an AUC of 0.724 and a PR-AUC of 0.741. In the de-identified timestamp-ordered validation cohort, the corresponding values were 0.718 and 0.766. Calibration was good in the internal test set but attenuated in the later cohort, suggesting the need for recalibration. SHAP analysis indicated complementary contributions of M0 and M1 features.
Conclusions:
A parallel M0 + M1 framework showed moderate discrimination and potential utility for reassessment-oriented early risk stratification in infection-related PICU admissions. Further external validation and recalibration are needed before broader application.
Related Concept Videos
01:29Endocarditis IV: Nursing Management
01:28Acute Pyelonephritis II: Diagnostic Studies and Management
01:29Pulmonary Embolism II: Diagnostic Studies and Interprofessional Care
01:13Radiological Investigation III: Pulmonary Angiogram and PET Scan
Pulmonary Angiogram
A Pulmonary Angiogram is an invasive procedure involving injecting a contrast medium through a catheter threaded into the pulmonary artery or the right side of the heart to visualize the pulmonary vasculature. Computed Tomography (CT) scans have mainly replaced this...
01:30Pneumonia III: Complications and Assessment