Pediatric Vital Sign Distribution Derived From a Multi-Centered Emergency Department Database

Area of Science:

• Pediatric Emergency Medicine
• Clinical Informatics
• Biostatistics

Background:

• Existing vital sign thresholds in pediatric emergency department (ED) screening tools may not align with observed vital signs.
• There is a need for updated vital sign data that reflects current pediatric populations.

Purpose of the Study:

• To develop accurate heart rate (HR) and respiratory rate (RR) centile rankings and z-scores for pediatric ED screening tools.
• To compare derived centiles with existing guidelines and previous studies.

Main Methods:

• Analysis of ~1.2 million non-admitted pediatric records from 169 EDs (2009-2013).
• Development of centile curves and z-scores using generalized additive models.
• Split-sample validation and comparison with prior data and Pediatric Advanced Life Support (PALS) guidelines.

Main Results:

• Empirical 95th centiles for HR and RR were determined from the large dataset.
• Derived centiles were higher than previously published results.
• Both HR and RR centiles deviated from PALS guideline recommendations.

Conclusions:

• Centiles derived from real-world pediatric ED data can improve screening tools by stratifying severity of vital sign deviation.
• These centiles can inform modifications for electronic and paper-based screening tools.
• Potential applications include improved bedside alarm limits beyond the ED.