Disparities in Pediatric Emergency Department Triage Among Encounters Resulting in Admission

Most US emergency departments (EDs) use the Emergency Severity Index (ESI) for triage, which assigns patients scores ranging from 1 (most urgent) to 5 (least urgent).1,2 The algorithm involves a combination of objective assessments as well as those subject to bias. As such, several general ED-based studies have found evidence of racial and ethnic triage disparities,3–5 although data in pediatric settings are limited. We examined the association between several patient demographic variables and undertriage at a large pediatric ED.We conducted a retrospective cross-sectional study of ED encounters at a large, urban, quaternary pediatric ED between January 2018 through August 2023. We included all ED encounters resulting in admission, including transfers. We excluded encounters with an ESI score of 1 or missing. We used electronic medical record (EMR) data to evaluate age group, sex, race and ethnicity, primary language, and insurance type. Race and ethnicity were reported into the EMR by self-report during registration. We used race and ethnicity groups defined by the EMR and categorized American Indian, Native American, Native Hawaiian, Pacific Islander, and Non-Hispanic Other into the “Other” category because of small sample sizes. We selected non-Hispanic white as our reference category given our hypothesis that minoritized racial and ethnic groups had higher odds of undertriage. Our outcome was ESI score, which we dichotomized to high (2 or 3) or low (4 or 5), and we defined undertriage as an encounter assigned a low ESI score and resulting in admission.6 Logistic regression was performed to determine associations between demographic variables and undertriage. Data collection was performed via blind abstraction from the EMR, and analyses were determined a priori. This study was not considered human subjects research by our institutional review board.We analyzed 44 273 ED encounters that led to hospital admission. Among these, 40 941 encounters (92.5%) were triaged high acuity, and 3332 encounters (7.5%) were triaged low acuity (Table 1).Younger age groups had higher odds of undertriage, as did female patients compared with male patients (adjusted odds ratio [aOR], 1.12; 99% CI, 1.02–1.23). Odds were also elevated for Hispanic or Latinx (aOR, 2.43; 99% CI, 2.07–2.84), Non-Hispanic Black (aOR, 1.60; 99% CI 1.34–1.92), and Asian (aOR, 1.94; 99% CI, 1.49–2.51) patients compared with non-Hispanic white patients. Patients whose primary language was Spanish had increased odds of undertriage (aOR, 1.27; 99% CI, 1.12–1.44) compared with those whose primary language was English. Additionally, patients with public insurance were more likely to be undertriaged than those with private insurance (aOR, 1.43; 99% CI, 1.26–1.62) (Table 2). Variance inflation factor values were all less than 3, indicating that multicollinearity did not significantly affect the model.In this study, we identified disparities in undertriage by age group, sex, race and ethnicity, primary language, and insurance type. Recent evidence has found that mistriage with the ESI was common.7 Our findings align with previous research highlighting disparities in the pediatric ED triage process. Zook et al reported that African American, Hispanic, and American Indian patients received lower-acuity triage scores compared with white patients.8 Metzger et al found that non-white patients received lower triage scores than their white counterparts, even when adjusted for illness severity.9 Rojas et al found similar disparities for pediatric patients whose caregivers preferred languages other than English.6 Disparities among minoritized groups may reflect interpersonal bias, language and communication barriers, or structural racism, as these groups often face unreliable access to outpatient care,10 potentially leading to more complex issues or influencing providers’ disposition decisions owing to anticipated lack of follow-up. Further research should investigate explanatory mechanisms for observed triage disparities.Our study has several limitations, including a potential misclassification of undertriage, inaccuracies in EMR data, and limited generalizability. We assumed independence of observations given that our unit of analysis was the encounter. Despite these limitations, our findings are consistent with results from other health systems that use the ESI.This study contributes to the body of literature that underscores the need for systematic efforts to address implicit bias and structural racism in the pediatric triage process. Initiatives should also target reducing overall undertriage rates, which may improve disparities. Additionally, our simplified definition of undertriage may facilitate ongoing monitoring of triage processes and outcomes as part of equity-focused institutional quality-improvement initiatives.We thank Kimberly DeNicolo, Jill Deferville, Annie Quinn, Lauren Mrozek, and Rachel Schultz for their input and feedback.