A 14-year-old Hispanic girl presents to the Emergency Department with her mother for suicidal ideation after a conflict at home. The girl is quiet and cooperative. Her mother, who speaks primarily Spanish, is trying to explain the situation. The nurse assigns an ESI level 2, the same score given to nearly every child who walks through the door with a behavioral health complaint. But does that score accurately capture this patient’s needs?
A new multicenter PECARN study published this week in JAMA Network Open takes a close look at triage accuracy for pediatric behavioral health ED visits. The findings: mistriaging errors are common, and they are not equally distributed [1].
Hoffmann et al. analyzed 78,411 ED visits by children aged 5 to 17 with behavioral health chief concerns across 15 PECARN Registry EDs from 2021 to 2023 [1]. They classified each visit as appropriately triaged, overtriaged, or undertriaged using vital signs, Glasgow Coma Scale, pain scores, emergency medication use, resource utilization, and disposition. Of the 74,564 visits with complete data:
- 57% were overtriaged
- 34% were appropriately triaged
- 8.5% were undertriaged
How ESI Handles Behavioral Health
The Emergency Severity Index (ESI) is used in over 90% of US EDs [2]. It sorts patients into 5 acuity levels. Level 1 is for patients needing lifesaving interventions. Level 2 is for high-risk situations, confused patients, or those in severe pain. Levels 3 through 5 are based on anticipated resource needs. In this study, 83.5% of all behavioral health visits were triaged as ESI level 2.
To assess triage accuracy, the authors compared each child’s assigned ESI level against what actually happened during their visit.
Overtriage
Overtriage means a child was assigned a higher acuity score than their clinical course supported. For a child assigned ESI level 2, overtriage was defined as meeting ALL of the following [1]:
- Stable vital signs within 2 hours of arrival (heart rate and respiratory rate not high risk for age, SpO2 ≥93% or not recorded)
- Pain score <7 (or not recorded)
- GCS of 15 (or not recorded)
- No emergency medications during the visit
In other words, the triage nurse predicted high acuity, but the visit didn’t bear that out.
Undertriage
Undertriage means the opposite: a child was assigned a lower acuity score than their clinical course warranted. For example, a child triaged as ESI level 4 (expected to need 1 resource) who ended up being admitted, needing emergency medications, or using multiple resources. The triage nurse underestimated how sick the child was or how much care they would need.
Undertriage Disproportionately Affects Minority Children and Spanish-Speaking Families
The most concerning equity finding was in undertriage.
After adjusting for clinical and visit characteristics, undertriage was significantly more likely for Hispanic children (AOR 1.46), non-Hispanic Black children (AOR 1.28), and children whose families preferred Spanish (AOR 1.31), all compared to non-Hispanic White and English-speaking patients [1]. The authors point to implicit clinician bias, systemic racism, and underutilization of professional interpreters as likely contributors.
The safety implications are real. Children whose acuity is underestimated may face longer waits, miss time-sensitive interventions, or leave the ED without being seen despite elevated risk.
Overtriage Was Common
More than half of all visits (57%) were overtriaged [1]. These children received a higher acuity triage score than their clinical course supported.
The strongest predictor was age. Children aged 5-9 had over 4-fold higher adjusted odds of overtriage compared to those aged 10-14 (AOR 4.43), possibly because younger children have a limited ability to communicate their symptoms and needs.
To a lesser degree, non-Hispanic Black children also had higher adjusted odds of overtriage compared to non-Hispanic White children (AOR 1.17). The authors cite research on adultification, the tendency to perceive Black youth as older or more threatening than they are, as a potential contributor. This means Black children in this study were more likely to be both undertriaged and overtriaged compared to White children. The errors are not unidirectional. They likely reflect different biases operating at different points in care.
Take Home Points
- The ESI has limited ability to differentiate pediatric behavioral health presentations. In this study, 83.5% of behavioral health visits were triaged as ESI level 2.
- UNDERTRIAGE was more likely for Hispanic children (AOR 1.46), non-Hispanic Black children (AOR 1.28), and Spanish-speaking families (AOR 1.31), raising concerns about missed acuity in these groups.
- OVERTRIAGE occurred in 57% of visits, driven most strongly by younger age (AOR 4.43 for ages 5 to 9) and to a lesser degree by non-Hispanic Black race (AOR 1.17).
References
- Hoffmann JA, Foster AA, Rojas CR, et al. Overtriage and undertriage of children presenting to the emergency department for behavioral health. JAMA Netw Open. 2026;9(3):e263042. Full text
- McHugh M, Tanabe P, McClelland M, Khare RK. More patients are triaged using the Emergency Severity Index than any other triage acuity system in the United States. Acad Emerg Med. 2012;19(1):106-109. doi:10.1111/j.1553-2712.2011.01240.x
Professor and Digital Innovation Lab Director
Department of Emergency Medicine
University of California, San Francisco
@M_Lin
