Bedside clinical prediction tool for mortality in critically ill children

Area of Science:

• Pediatric Critical Care Medicine
• Clinical Prediction Modeling
• Hemostasis and Thrombosis

Background:

• Pediatric intensive care unit (PICU) mortality remains a significant challenge.
• Identifying high-risk pediatric patients is crucial for resource allocation.
• Predictive tools are needed to improve outcomes for critically ill children.

Purpose of the Study:

• To develop a prediction score for 28-day PICU mortality.
• To utilize clinical parameters and hemogram data for risk stratification.
• To identify key predictors of mortality in critically ill children.

Main Methods:

• Retrospective study of children (1 month to 18 years) admitted to the PICU.
• Data collected between January 2018 and December 2022.
• Multivariable logistic regression used to develop the prediction score, assessing calibration and discrimination.

Main Results:

• The prediction score incorporates male gender, vasoactive drug use, and red blood cell distribution width (RDW) and platelet distribution width (PDW) levels.
• Scores range from 0 to 8, with a cutoff of 5 identifying high-risk patients.
• The model showed excellent discrimination (AuROC 0.86) with 82.8% sensitivity and 73.1% specificity.

Conclusions:

• A novel score combining clinical and hemogram data can predict 28-day mortality in critically ill children.
• The developed score shows potential for clinical application in risk stratification.
• External validation is recommended before widespread clinical implementation.