Emergency severity index version 4 during the first year of implementation at an academic institution

Abstract

Background: The Emergency Severity Index (ESI) version 4 (v4) is a triage system based on vital signs, potential limb or organthreat, as well as expected resources needed in the emergency department (ED).
Objective: The purpose of this study was to examine accuracy and misclassification rate of ESI triage over one year following implementation.
Methods: This was a retrospective analysis of ED encounters from January 2011 to 2012. Charts were selected in one-week intervals every 12 weeks for one year (months 1, 3, 6, 9, and 12). Each encounter was reviewed to determine post hoc ESI level based on care in the ED. Descriptive statistics was used to compare the agreement between initial triage and post hoc ESI levels. Sensitivity and specificity for each level was determined. Kruskal Wallis test (KW) and Mann Whitney U (MWU) was used toassess differences in initial versus post hoc ESI levels by month to explore change in accuracy over time.
Results: Five hundred and sixty separate ED encounters were included. Agreement was observed in 301 triage encounters (53.8%). Overestimation of the triage level occurred in 131 (23.4%) encounters, while the triage level was underestimated in 128 (22.9%) encounters. There was a significant decline in accuracy during the year (KW = 10.2; p = .037); with the greatest dropbetween month 1 and 9 months (MWU 4,859; p = .035). Sensitivity ranged from 24% to 76% and specificity ranged from 61% to 99%, based on ESI level.
Conclusions: Enhanced education and quality improvement process is necessary to improve overall accuracy rates at this site.