Sensitivity and specificity of the Manchester Triage System for patients with acute coronary syndrome.

INTRODUCTION: Monitoring the performance of triage systems is essential to optimize their usefulness in each hospital. The objective of this study was to assess the sensitivity of the Manchester Triage System in assigning high priority to patients with acute coronary syndrome, and the sensitivity and specificity of the combinations of flowcharts and discriminators routinely used for triage in the Emergency Department of Hospital de S. João, Porto.
METHODS: We analyzed a systematic sample of one third of emergency department episodes in 2007, corresponding to 53,039 episodes, of which 307 were cases of acute coronary syndrome. For each episode, we considered the final diagnosis by the Emergency Department attending physician, which was confirmed in 95% of cases. The combinations of flowchart and discriminators suggestive of acute coronary syndrome were defined, based on typical symptoms, including precordial chest pain with or without radiation and/or syncope.
RESULTS: The sensitivity of the Manchester Triage System in assigning high priority (immediate--red, or very urgent--orange) to patients with acute coronary syndrome was 87.3% (95% confidence interval [CI]: 83.1-90.6). The proportion of false negatives was higher in extreme age-groups. The sensitivity of combinations of flowchart and discriminators suggestive of acute coronary syndrome was 74.3% (95% CI: 69.1-78.8), lower in women (67.3% vs. 77.7%), and the specificity was 97.4% (95% CI: 97.2-97.5). The likelihood ratio of combinations of flowchart and discriminators suggestive of acute coronary syndrome was 28.3 (95% CI: 26.0-30.8) and that of other combinations was 0.26 (95% CI: 0.22-0.32), less informative in ruling out acute coronary syndrome in women (0.33 vs. 0.23). DISCUSSION: The main limitations of this study derive from the impossibility of including only patients who came directly to this Emergency Department and of testing the validity of the final diagnosis when acute coronary syndrome was not identified in the Emergency Department. However, using electronic data records completed as a matter of routine enabled analysis of a large database, which reflects the practical value of the Manchester Triage System in this context, and most cases identified as acute coronary syndrome were confirmed.
CONCLUSIONS: The Manchester Triage System has a high sensitivity in assigning high priority (immediate/very urgent) to patients with acute coronary syndrome. The combinations of flowchart and discriminators defined as suggestive of acute coronary syndrome have very high specificity and moderate sensitivity, but the system has the potential to become more sensitive with no loss in specificity.