Evidence of the validity of the Emergency Severity Index for triage in a general hospital emergency department

Objective. To determine whether the Emergency Severity Index (ESI) is valid for triage according to evidence based on classifying real patients in a general referral hospital's emergency department. Methods. Observational, cross-sectional descriptive study carried out in the emergency department of Hospital Clinico Universitario Virgen de la Arrixaca in Murcia. Thirty-two nurses used the ESI algorithm to triage 410 patients as they arrived seeking care. The results were compared to a gold standard (a triage expert's opinion, which was later confirmed by an expert committee after discussion, if necessary, of cases for which opinions were not unanimous). We calculated sensitivity, specificity, under- and over-triage rates, as well as descriptive statistics about resource assignment, exitus, patients who left without being seen, destination on discharge, and times. Results. ESI was highly correlated with resources (p = -0.717, P <.01) and moderately correlated with destination on discharge (p = 0.437, P <.01). Regarding time spent in the department, we found that patients assigned ESI levels 1 and 2 had significantly longer stays, and those assigned ESI levels 4 and 5 had significantly shorter stays (p < 0,001). Interobserver agreement was good or very good, indicating that this triage tool is reliable. Conclusions. This pilot of the ESI triage algorithm in the emergency department of a referral hospital found evidence supporting the system's validity.