Using Vital Signs to Predict Disposition of Children Transported to the Hospital by Emergency Medical Services.

Abstract

To compare prehospital and emergency department (ED)-based vital sign assessments for children with out-of-hospital emergencies and to evaluate the relationships between vital signs and mental status assessment on patient mortality or intensive care unit (ICU) admission. We performed a retrospective, multiagency and multicenter study of pediatric transports from the scene to the hospital within an integrated health system between 2014 and 2023. We compared 3 sets of vital signs (first prehospital, last prehospital, and first ED). We evaluated the multivariable association of vital signs with outcomes of ICU admission and in-hospital mortality. We included 68 489 children. ICU admission occurred in 2.4%, and in-hospital mortality in 0.1%. All vital signs demonstrated high agreement between the first prehospital, last prehospital, and ED phases of care (Gwet's AC1 > 0.80). For ICU admission, the area under the receiver operator characteristic curve of the initial prehospital model (0.79, 95% CI: 0.77-0.80) was lower than the last prehospital (0.81, 95% CI: 0.80-0.83) and first-ED model (0.83, 95% CI: 0.82-0.85). Models predicting ICU admission had moderate accuracy (sensitivities of 0.72-0.79; specificities of 0.80-0.87). The area under the receiver operator characteristic curve for models predicting mortality based on prehospital vital signs and ED vital signs reached highly accurate criteria (initial prehospital: 0.91, 95% CI: 0.85-0.97; last prehospital: 0.95, 95% CI: 0.91-0.99; first ED: 0.96, 95% CI: 0.93-0.99). Models for mortality demonstrated high sensitivity (0.90-0.98) and specificity (0.90-0.97). Across models, altered mental status showed the strongest associations with both outcomes, and predictive performance improved modestly from initial prehospital to ED assessment. Prehospital and ED vital signs are predictive of in-hospital outcomes. Following prospective evaluation, these findings may inform pediatric clinical decision-making in prehospital and initial hospital management. These results support the use of age-adjusted physiologic measures together with mental status assessment to strengthen early risk stratification of children transported by emergency medical services.