Development of heart and respiratory rate percentile curves for hospitalized children

OBJECTIVE: To develop and validate heart and respiratory rate per-centile curves for hospitalized children and compare their vital sign distributions to textbook reference ranges and pediatric early warning score (EWS) parameters. METHODS: For this cross-sectional study, we used 6 months of nurse-documented heart and respiratory rates from the electronic records of 14 014 children on general medical and surgical wards at 2 tertiary-care children's hospitals. We developed percentile curves using generalized additive models for location, scale, and shape with 67% of the patients and validated the curves with the remaining 33%. We then determined the proportion of observations that deviated from textbook reference ranges and EWS parameters. RESULTS: We used 116 383 heart rate and 116 383 respiratory rate values to develop and validate the percentile curves. Up to 54% of heart rate observations and up to 40% of respiratory rate observations in our sample were outside textbook reference ranges. Up to 38% of heart rate observations and up to 30% of respiratory rate observations in our sample would have resulted in increased EWSs. CONCLUSIONS: A high proportion of vital signs among hospitalized children would be considered out of range according to existing reference ranges and pediatric EWSs. The percentiles we derived may serve as useful references for clinicians and could be used to inform the development of evidence-based vital sign parameters for physiologic monitor alarms, inpatient electronic health record vital sign alerts, medical emergency team calling criteria, and EWSs.