O1023: Over- and Under-Triage of Children Presenting to the Emergency Department with Behavioral Health Symptoms: A Multicenter Retrospective Study

Background: Accurate triage upon emergency department (ED) arrival enables anticipation of resource needs and identification of patients who can safely wait for care. The Emergency Severity Index (ESI), a 5-level acuity scale, is the most widely used triage system in US EDs. Rates of mistriage (i.e., resource use discordant with ESI prediction) are high among pediatric ED visits for acute injuries and medical illnesses. However, the performance of ESI in triaging children presenting with behavioral health symptoms is not well studied. We assessed visit characteristics associated with over- and under-triage of children presenting to the ED with behavioral health symptoms.

Methods: We conducted a retrospective cross-sectional study of ED visits by children 5-17 years old presenting with a behavioral health chief complaint to any of 15 US EDs participating in the Pediatric Emergency Care Applied Research Network (PECARN) Registry from 2021-2023. Aligned with ESI Handbook criteria, we defined appropriate triage, over-triage, and under-triage using combinations of first-obtained vital signs, Glasgow Coma Score, and pain score; receipt of emergency medications; number of distinct resource types used (e.g., laboratory tests, imaging studies); and disposition. We used multivariable logistic regression to assess ED visit characteristics associated with (1) over-triage and (2) under-triage, adjusting for year and site effects.

Results: We included 78,650 ED visits by children with a behavioral health chief complaint (47.5% 10-14 years old, 60.5% female, 58.6% publicly insured). There were 33,595 (42.7%) visits classified as appropriately triaged, 34,827 (44.3%) visits classified as over-triaged, and 6,380 (8.1%) visits classified as under-triaged. The adjusted odds of over-triage were higher for age 5-9 years (aOR 2.71, 95% CI 2.56, 2.87) compared to 15-17 years and for visits by Black non-Hispanic patients (aOR 1.09, 95% CI 1.05, 1.14) compared to White non-Hispanic patients. The adjusted odds of under-triage were higher for visits by Hispanic (aOR 1.42, 95% CI 1.29, 1.57), Black non-Hispanic (aOR 1.16, 95% CI 1.09, 1.24), and patients of other/unknown race and ethnicity (aOR 1.15, 95% CI 1.04, 1.26) compared to White non-Hispanic patients; for Spanish language preference (aOR 1.36, 95% CI 1.17, 1.57) compared to English language preference; for presence of a complex chronic condition (aOR 1.33, 95% CI 1.21, 1.46) relative to absence; and for arrival during a non-school month (aOR 1.31, 95% CI 1.23, 1.39) compared to a school month.

Conclusion: Over-triage rates among children presenting to the ED with behavioral health symptoms are high. The likelihood of over- and under-triage differs by sociodemographic and clinical characteristics. Prospective studies are needed to assess the performance of ESI among children presenting to the ED with behavioral health symptoms, and to examine how triage systems can be designed to inform accurate resource allocation for all children.