Predicting transfers to intensive care in children using CEWT and other early warning systems

Area of Science:

• Pediatric critical care medicine
• Clinical informatics
• Patient safety research

Background:

• The Children's Early Warning Tool (CEWT) is a widely adopted system for monitoring hospitalized children's risk of deterioration.
• Existing tools like Bedside Pediatric Early Warning Score (Bedside PEWS) and Between the Flags (BTF) are used alongside CEWT.
• The pediatric Calculated Assessment of Risk and Triage (pCART) is a newer, machine learning-based tool for risk assessment.

Purpose of the Study:

• To compare the predictive performance of CEWT against Bedside PEWS, BTF, and pCART.
• To evaluate the effectiveness of these early warning scores in identifying children at high risk for ICU transfer.
• To assess the diagnostic accuracy of different pediatric early warning systems.

Main Methods:

• A retrospective observational study of over 50,000 pediatric admissions (2009-2019) at a major children's hospital.
• Comparison of prediction performance using the area under the receiver operating characteristic curve (AUC).
• Analysis of alert rates at various score thresholds for each tool.

Main Results:

• CEWT showed good discrimination for predicting ICU transfers (AUC 0.74-0.79).
• In initial cohorts, CEWT performed comparably to Bedside PEWS (AUC 0.76) but better than BTF (AUC 0.66).
• The pCART tool demonstrated the highest predictive performance, particularly in later years (AUC 0.84).

Conclusions:

• CEWT is a valuable tool for identifying children at risk of ICU transfer.
• The pCART system exhibited superior predictive accuracy compared to CEWT, Bedside PEWS, and BTF.
• Further research may explore the integration of advanced tools like pCART into clinical practice.