Using Standardized Cases to Investigate Emergency Department Triage Disparities by Race and Ethnicity

Observational studies have found racial/ethnic differences in ED triage, with Black and Hispanic patients assigned less acute Emergency Severity Index (ESI) scores than White patients with the same chief complaints. However, observational studies are confounded by the effects of structural racism on healthcare access, and it is uncertain if such differences reflect appropriate triage assessments or are due to bias in triage. Therefore, we used video-based simulated cases to investigate if there are differences in triage assessments by patient race/ethnicity.

We developed an electronic survey with 4 video-based cases depicting ED triage patients, with correct triage scores drawn from the ESI implementation handbook. We filmed each case with 3 standardized patients (SPs) from different racial/ethnic groups (White, Black, Hispanic). We randomized participants to view only one version of each case, then assign the SP an ESI score and answer other questions. We recruited ED nurses from 2 academic and 3 community EDs. We used descriptive statistics to compare assessments by SP race/ethnicity. We fit a generalized linear mixed effect model to estimate odds of under-triage by SP race/ethnicity, adjusting for SP sex as a fixed effect and accounting for correlation among observations with random intercepts for case, nurse, and nurse's ED site.

Among 805 invited nurses, 191 (24%) responded to the survey, completing 721 total cases. 74% of those completing the survey were current or former triage nurses. Across all cases, under-triage was more common for Hispanic (24%) and Black (22%) versus White SPs (18%; p=0.48). There was no difference in triage destination (trauma bay, room, hallway, waiting room) by race/ethnicity (p=0.60). A lower proportion of Hispanic SPs were perceived as very likely to be admitted (37%) compared to Black (48%) and White SPs (49%; p=0.02). In adjusted analyses, Hispanic SPs had significantly greater odds of under-triage compared to White SPs (aOR 1.68, 95% CI 1.002-2.82). Odds of under-triage was not significantly different for Black, versus White, SPs (aOR 1.53, 95% CI 0.91-2.56).

Our results suggest bias in triage may contribute to racial/ethnic triage disparities. Given the importance of triage in patients' care and outcomes, this study reinforces the critical need to investigate approaches to ensure all ED patients receive unbiased triage assessments.