260 Impact of Patient Ethnicity and Language Proficiency on Emergency Department Triage

Abstract

In order to decrease wait times and lengths of stay, many emergency departments (EDs) are implementing new triage and patient flow systems. One of these, the “split flow” model, divides the ED into high and low acuity treatment areas. Patients undergo a rapid triage assessment consisting only of demographics and chief complaint before moving to the treatment area. In the absence of vital signs and any physical examination in triage, this initial quick assessment may be highly subject to the biases of those performing the evaluation. Multiple studies have shown that ethnicity and provider intrinsic biases play roles in patient care. The split-flow model may be particularly susceptible to these biases as no objective indicators of health are used during quick triage. We determine if ethnicity or language proficiency affects triage to the high or low acuity treatment areas in the ED. We are unaware of prior studies evaluating this in the split flow model or any other ED triage system. Retrospective chart review of adult patients presenting to two large urban EDs with a combined annual census of 190,000 visits/year. The Emergency Severity Index (ESI), ethnicity, age, language proficiency and ED destination (high or low acuity) were collected over a three-month period for patients presenting with two common ED complaints, chest pain or abdominal pain. Data was analyzed using descriptive statistics and Fisher’s exact test analysis. 2983 charts were included of which 231 (8%) were white and 2653 (89%) were non-white and 87 (3%) were of unknown ethnicity. 2407 (81%) spoke English and 576 (19%) did not speak English. Amongst 1174 patients presenting with chest pain (85 white (7%), 1040 non-white (89%), and 49 unknown ethnicity (4%)) white patients were statistically significantly more likely to be triaged to the high acuity side of the ED (75 of 85 (88%)) than their non-white counterparts (795 of 1040 (76%)) (p-value 0.013). Among 1809 patients presenting with abdominal pain (146 white (8%), 1613 non-white (89%), 50 unknown ethnicity (3%)) there was no statistically significant difference in triage to high vs low acuity sides of the ED (80 of 146 (55%) for white; 845 of 1613 (52%) for non-white) (p = 0.6). No statistically significant difference in triage was found between English and Non-English speakers for either patients presenting with chest pain or abdominal pain. In this study the “split flow” model for ED triage resulted in a statistically significant proportion of non-white patients presenting to the ED with chest pain to be triaged to the low acuity side of the ED as compared with their white counterparts. While the reasons for this may be multifactorial, the inherent susceptibility to bias of a triage system that lacks objective indicators of health seems likely to be a contributing factor. It is unclear why a difference was found in patients with chest pain as opposed to abdominal pain, though it is concerning given the high morbidity diagnoses associated with a chest pain presentation. Similarly, we did not find any differences in relation to language for triage. As more hospitals move to integrate triage systems that decrease wait times and promote rapid patient flow it will be important to delineate potential biases in the system that may prove detrimental to patient care. Further research is needed to assess how biases may affect patient outcomes.