593 - Impact of Team Triage Implementation on Triage Discordance in a Pediatric Emergency Department

Publication Number: 1570.593

Kaileen Jafari, University of Washington School of Medicine, Seattle, WA, United States; Ashley E. Keilman, Seattle Children's Hospital/University of Washington, Seattle, WA, United States; Brian Burns, Seattle Children's, Seattle, WA, United States; Kimberly Stone, Seattle Children’s and University of Washington School of Medicine, Seattle, WA, United States; Dwight Barry, Seattle Children's, Seattle, WA, United States; Kathleen Hennings, Seattle, Seattle, WA, United States; Hiromi Yoshida, University of Washington School of Medicine, Seattle, WA, United States; Russell Migita, Seattle Children's Hospital, Seattle, WA, United States; Emily A. Hartford, University of Washington School of Medicine, Seattle, WA, United States

Background: Accurate triage is essential to safe and efficient pediatric emergency care, yet Emergency Severity Index (ESI) triage scores are often discordant with resource use and clinical severity. Team triage (TT), in which a nurse and physician or advanced practice provider jointly assess patients at Emergency Department (ED) arrival is designed to improve ED efficiency, but its impact on triage discordance is unknown.

Objective: To measure changes in discordant triage for low and moderate acuity pediatric ED visits (ESI 3-5) following TT implementation.

Design/Methods: Retrospective analysis of ED visits with ESI 3-5 for children aged 0-18 years seen in a single academic pediatric ED. Interrupted time series using generalized least squares regressions compared monthly proportion of under-triage (ESI 4,5) and over-triage (ESI 3,4) before (10/2021-09/2023) and after (10/2023-10/2024) TT implementation. Visits outside TT hours were used as a control group. Discordant triage was defined as a mismatch between ESI-predicted and observed resource use or illness severity (Table 1). Descriptive statistics were used to compare patient and visit characteristics across groups.

Results: Among 124,539 included visits, 79,938 occurred pre-TT and 44,601 occurred post-TT implementation. Pre- and post-TT groups were similar across age, race/ethnicity, language, medical complexity, ED census, and disposition (Table 2). Triage discordance was frequent pre-implementation and variable by ESI (overall 51%, range 37-61%). After TT implementation, there was a sustained reduction of 9.1% (95%CI -9.9%,-8.4%) in ESI 3 over-triage in the intervention group, not observed in the control group. Trends in ESI 4 over-triage stabilized after TT, shifting from a baseline of 46% with a slope of -0.4% (95% CI -0.5%,-0.3%) to 33% (95% CI 29%,36%). Similarly upward trend in ESI 4 under-triage leveled after implementation from a baseline of 18% with a slope of 0.3% (95% CI 0.3%, 0.4%) pre-TT to 26% (95% CI 23%, 29%) post-TT (Figure 1). No significant changes in ESI 5 under-triage occurred in the intervention group, but slope increased in control group from 0.2% pre-TT (95% CI 0.1%,0.3%) to 0.8% (95% CI 0.2%, 1.5%) post-TT.

Conclusion(s): Team triage implementation was associated with a significant reduction in triage discordance, driven primarily by a decrease in over-triage of ESI 3 visits. These findings suggest that nurse-provider collaboration at triage may improve alignment between assigned ESI levels and ED care received. Future studies should assess whether improved triage concordance translates to better patient outcomes and quality of care.

Table 1. Definition of Under- and Over-triage.
Expected resource used is based upon ESI Version 4 Manual. The definitions for Level 1-3 events were adapted from Sax et al. Emergency Severity Index Version 4 and Triage of Pediatric Emergency Department Patients. JAMA Pediatrics 2024. Level 1 events: death in the ED, high risk sepsis defined as: an ED diagnosis of sepsis and triage temperature ≥ 100.4 or ≤95, or any of these critical events or interventions within 1 hour of ED arrival, tier 1 medication administration (see below), or transfer to another facility, operating room admission, or pediatric intensive care unit admission.
Level 2 events: most level 1 events beyond the first hour, except admission to the operating room or hospital transfer. Level 2 events also included age <28 days and triage temp≥100.4, chief concern of suicidal ideation, >2 albuterol doses administered.
Level 3 events: Hospital admission/transfer beyond first hour, sedating medication, age 1-3 months with temp ≥100.4
Tier 1 medications grouped by typical indication (parenteral unless otherwise indicated): Shock (Epinephrine, Norepinephrine, Dopamine, Calcium chloride, Calcium gluconate, Bicarbonate); Cardiogenic Shock( Dobutmaine, Amiodarone, Atropine, Adenosine, Prostaglandin E, Furosemide; Anaphylaxis (Epinephrine); Hypovolemic Shock (≥2 IV weight based fluid bolus); Status Epilepticus (Ativan, Valproate, Diazepam, Phenytoin, Fosphenytonin, Phenobarbital, Levetiracetam, intranasal Midazolam); Altered mental status/hypoglycemia/hyperglycemia (Insulin Infusion, Mannitol, Hypertonic saline D10, D25, or D50 bolus, Naloxone); Status Asthmaticus (Magnesium sulfate, Methylprednisolone. Terbutaline, Epinephrine, Aminophylline)


Table 2. Patient and Visit Characteristics Pre- and Post-implementation of Team Triage(TT).


Figure 1. Interrupted Time Series Analysis of Triage Discordance Before and After Team Triage Implementation in October 2023 (dashed line).
The intervention group was defined as visits which occurred during team triage hours of operation (10am -1am Fall, Winter,Spring; 3pm-10pm in Summer). The control group was defined as visits which occurred outside the hours of team triage operation.