Related Experiment Video
Updated: Sep 18, 2025
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Development and verification of a prediction model for delirium in critically ill children
Ting-Ting Xu1, Yan Li2, Cong-Hui Fu2
1Department of Nursing, Shanghai Children's Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China.
Insights
A new prediction model can identify delirium in critically ill children. Key risk factors include mechanical ventilation, benzodiazepines, young age, multiple catheters, and physical restraints, aiding early detection and care.
Area of Science:
- Pediatric Intensive Care
- Critical Care Medicine
- Clinical Prediction Models
Background:
- Delirium is a frequent complication in pediatric intensive care units (PICUs).
- Early identification of delirium is crucial for improving outcomes in critically ill children.
- Existing prediction tools for pediatric delirium are limited.
Purpose of the Study:
- To develop and validate a prediction model for delirium in critically ill children.
- To identify independent predictors of delirium in this population.
- To enhance early detection and management strategies for pediatric delirium.
Main Methods:
- Prospective cohort study of 1,047 critically ill children in a tertiary PICU.
- Multivariate logistic regression analysis to derive a risk prediction model.
- Nomogram construction and validation using ROC curve analysis and calibration curves.
Main Results:
- Delirium occurred in 26.6% of the study population.
- Independent predictors included mechanical ventilation, benzodiazepines, age ≤ 2 years, ≥ 3 catheters, and physical restraints.
- The model showed high sensitivity (85.61%), specificity (76.07%), and an AUC of 0.88, indicating good predictive performance.
Conclusions:
- A validated prediction model for delirium in critically ill children was developed.
- This model enables accurate risk assessment by nurses.
- It has the potential to improve nursing care quality for critically ill children.
Purpose:
Delirium is a common syndrome in the intensive care unit (ICU), with a high incidence in critically ill children. This study aims to develop and validate a prediction model for delirium in critically ill children, which could potentially enhance early identification and management strategies.
Methods:
In this prospective cohort study, we collected 1,047 critically ill children admitted to the pediatric intensive care unit (PICU) of a tertiary children's hospital from November 2021 to November 2023. Based on the risk prediction model derived from multiple logistic regression analysis performed with SPSS software, a nomogram was constructed using R software. The model's predictive performance was evaluated through analysis of the area under the curve (AUC) of the receiver operating characteristic curve (ROC) and the calibration curve for discriminatory ability and accuracy.
Results:
Among the 1,047 critically ill children, delirium occurred in 26.6% of cases. Mechanical ventilation, benzodiazepines, age ≤2 years, number of catheters ≥3, and physical restraints were independent predictors of delirium in critically ill children. The predictive model demonstrated a sensitivity of 85.6% and a specificity of 76.1%, with a Youden index of .62. The validation analysis demonstrated an AUC of .88 (95% CI: 0.86-0.90). The Hosmer-Lemeshow goodnessoffit test yielded a x2 value of 15.23 (P > .05), demonstrating satisfactory discriminatory performance and good calibration of the predictive model.
Conclusions:
This study provided a predictive model for the occurrence of delirium in critically ill children, enabling nurses to accurately assess delirium risk and enhance the quality of nursing care for this vulnerable patient population.
