O1017: Determining a consensus-based measure for critical illness and injury for children cared for by emergency medical services

Background: Accurate triage of patients from the field or community EDs to higher levels of care is essential for emergency, disaster, and surge response. EMS clinicians frequently report discomfort with pediatric care and have difficulty identifying children at risk of deterioration. We aimed to use a consensus-based approach to define indicators of critical illness and/or injury in hospital settings to support the development of statistical models using EMS data to predict hospital outcomes.

Methods: We convened a 22-member eDelphi panel of clinicians across subspecialties including hospital pediatrics, pediatric emergency medicine, emergency medicine, pediatric surgery, trauma surgery, and emergency medical services. We presented diagnostic, intervention, and clinical outcome criteria for patients receiving hospital care to determine consensus-based indicators of critical illness and/or injury. Diagnostic and assessment criteria included gradations to identify a threshold for severe illness or injury. Drawing from the literature, we proposed 63 initial criteria and solicited feedback from panelists to identify additional candidate criteria. We engaged in a modified eDelphi process with four sequential rounds of voting to determine consensus priorities. Consensus was defined as ≥70% agreement.

Results: Participation ranged between 95-100% on each of the four rounds. The 22 participants were from 13 states or the District of Columbia. The sample included 13 males, with pediatric emergency medicine (n=7) and emergency medicine (n=6) as the most common specialties. The final criteria including 8 medical diagnoses (including active seizure, anaphylaxis, overdose with mental status changes, sepsis with organ dysfunction, pre-arrest states, and altered mental status), 14 trauma diagnoses (including neurotrauma with important clinical findings, major thoracic or intra-abdominal injury, major burns, and hemorrhagic shock), one trauma mechanism (gunshot wound with intra-abdominal injury), 6 assessments (including specific vital signs or mental status changes), 13 interventions (transfusions, operative care, angiography, central line, chest tube, advanced airway management, tracheostomy tube placement, pacing, defibrillation, cardiopulmonary resuscitation, cardioversion, neurointervention, and surgical airway), 6 medications (vasoactive infusions, intramuscular epinephrine, atropine, adenosine, paralytics, and hypertonic therapy), intensive care unit admission, and morgue. On the fourth round, we achieved consensus of 82% on the entire set of indicators. Areas of concern raised on the final set of indicators pertained to vital signs (including criteria oxygen saturation, bradycardia, and hypotension) and transfer from the ED to the operating room.

Conclusion: Using a modified eDelphi process, we reached consensus on 50 indicators that will be used to operationalize a hospital-based outcome of critical illness and/or injury to be predicted by prehospital data. Subsequent steps include determining the prevalence of these indicators within different hospital-based settings and the development of statistical models using prehospital factors to predict these outcomes.