Development and verification of a prediction model for delirium in critically ill children

Ting-Ting Xu; Yan Li; Cong-Hui Fu; Min-Jie Ju; Ji Liu; Xiao-Ya Yang; Wen-Juan Tang; Wei-Ying Zhang

doi:10.1016/j.anr.2025.05.008

Development and verification of a prediction model for delirium in critically ill children

Asian Nurs Res (Korean Soc Nurs Sci). 2025 Jun 19:S1976-1317(25)00066-0. doi: 10.1016/j.anr.2025.05.008. Online ahead of print.

Authors

Ting-Ting Xu¹, Yan Li², Cong-Hui Fu², Min-Jie Ju², Ji Liu², Xiao-Ya Yang², Wen-Juan Tang³, Wei-Ying Zhang⁴

Affiliations

¹ Department of Nursing, Shanghai Children's Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China.
² Department of Critical Care Medicine, Shanghai Children's Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China.
³ Department of Nursing, Shanghai Children's Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China. Electronic address: wjtang2014@163.com.
⁴ Department of Nursing, Shanghai East Hospital, School of Medicine, Tongji University, Shanghai, China. Electronic address: zhangwy_cn@126.com.

PMID: 40543881
DOI: 10.1016/j.anr.2025.05.008

Abstract

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 multivariate 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 85.61% and a specificity of 76.07%, with a Youden index of 0.62. The validation analysis demonstrated an AUC of 0.88 (95% CI: 0.86-0.90). The Hosmer-Lemeshow goodness-of-fit test yielded a 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.

Keywords: child; delirium; risk factors.