LO62: Variability in triage performance for chest pain patients in two Canadian cities

Abstract

Introduction: CTAS triage acuity determinations are used to prioritize patients, describe illness acuity, and compare casemix across institutions. The latter functions assume reliable application in diverse settings, but no studies have evaluated this using actual triage data. Methods: This administrative database study included all patients with a triage complaint of chest pain (CP) in Vancouver (2012-16) and Calgary (2016). We stratified patients into high vs. non-high severity groups based on discharge diagnoses. High severity diagnoses included all patients with aortic pathology, ACS, shock or arrest states, as well as patients requiring admission because of pulmonary embolism, dysrhythmias, CHF, neurologic or respiratory conditions. We dichotomized patient triage assignments to high (CTAS 1,2) vs. low (3,4,5) acuity, then constructed 2x2 tables correlating CTAS acuity with disease severity. Main outcomes included the proportion of CP patients triaged to high acuity categories and CTAS sensitivity for high severity conditions. Results: We studied 97277 Vancouver and 18622 Calgary patients. Age (mean, 54.8 years), sex (53.5% male) and casemix distributions were similar between cities, although Calgary had more high severity conditions (15.0% v. 10.5%) and a higher admission rate (22.5% v. 21.4%). Calgary triage nurses placed more patients in high acuity triage categories (85.1% vs. 45.2%) and achieved higher sensitivity for severe illness (96.2% vs. 76.2%); however, they were less accurate (28.7% vs. 60.3%) and less specific (16.8% vs. 58.4%). The proportion of CP patients triaged into high acuity categories ranged from 79% to 87% across four Calgary hospitals and from 28% to 62% at five Vancouver hospitals. Conclusion: This study shows profoundly different triage categorization at different sites seeing similar patient populations. Triage nurses are taught to strive for high sensitivity, but there may be operational consequences if specificity drops too low and large numbers of non-severe patients are triaged into high acuity categories. It is not clear which approach is better but these data suggest CTAS should not be used to compare patient acuity or complexity across different hospitals or regions.