A new nomogram for the individualized prediction of children's mortality risk in pediatric intensive care unit

Area of Science:

• Pediatric critical care medicine
• Biostatistics
• Clinical informatics

Background:

• Pediatric intensive care units (PICUs) manage critically ill children.
• Accurate mortality risk prediction is crucial for patient management and resource allocation.
• Existing models may not fully capture the complexity of pediatric critical illness.

Purpose of the Study:

• To develop and validate a novel nomogram for predicting mortality risk in children admitted to PICUs.
• To create an easily applicable tool for individualized risk assessment.
• To improve clinical decision-making in pediatric critical care.

Main Methods:

• Retrospective analysis of 10,538 children from the PICU Public Database.
• Development of a multivariate logistic regression model incorporating age and physiological indicators.
• Presentation of the risk model as a nomogram and internal validation of its performance.

Main Results:

• The nomogram includes predictors such as neutrophils, platelets, albumin, lactate, and oxygen saturation.
• The model demonstrated effective discriminatory power with an area under the receiver operating characteristic (ROC) curve of 0.7638.
• Internal validation showed sustained discriminative ability with an ROC curve area of 0.7404.

Conclusions:

• The developed nomogram provides an effective tool for individualized mortality risk prediction in PICU patients.
• The model is user-friendly and can assist clinicians in assessing risk for critically ill children.
• This nomogram contributes to enhanced risk stratification and management strategies in pediatric critical care.