Effects of a triage process conversion on the triage of high-risk presentations

Acad Emerg Med. 2008 Oct;15(10):916-22. doi: 10.1111/j.1553-2712.2008.00236.x. Epub 2008 Sep 10.


Objectives: The objective was to determine effects of a modification in triage process on triage acuity distribution in general and among patients with conditions requiring time-sensitive therapy.

Methods: The authors retrospectively reviewed triage acuity distributions before and after modification of their triage process that entailed conversion from the Canadian Triage and Acuity Scale (CTAS) to the Emergency Severity Index (ESI). The authors calculated the ratio of the odds of being triaged to a nonemergent level (3, 4, or 5) under ESI to the odds of being triaged as nonemergent under CTAS. The authors calculated sensitivity and specificity of triage to an emergent acuity level (1 or 2) for identifying patients with common presentations who required time-sensitive care.

Results: There were shifts from higher to lower acuity levels for all subsets, with odds ratios ranging from 2.80 (95% confidence interval [CI] = 2.75 to 2.86) for all patients to 21.39 (95% CI = 14.66 to 31.21) for patients over 55 years of age with a chief complaint of chest pain. The sensitivity of triage for identifying abdominal pain patients requiring admission to an intensive care unit (ICU) or operating room (OR) or emergency department (ED) death was 80.7% (95% CI = 73.2 to 86.5) before versus 50.8% (95% CI = 43.5 to 58.1) following the transition to ESI. Specificity under CTAS, 55.2% (95% CI = 54.0 to 56.4), was significantly lower than under ESI, 83.6% (95% CI = 82.7 to 84.4). The authors found similar effects for patients presenting with chest pain.

Conclusions: Monitoring for changes in the sensitivity of the triage process for detecting patients with potentially time-sensitive conditions should be considered when modifying triage processes. Further work should be done to determine if the decreased sensitivity seen in this study occurs in other institutions converting to ESI, and potential causative factors should be explored.

MeSH terms

  • Adult
  • Emergency Service, Hospital*
  • Female
  • Humans
  • Male
  • Middle Aged
  • Outcome and Process Assessment, Health Care
  • Patient Admission / statistics & numerical data
  • Retrospective Studies
  • Sensitivity and Specificity
  • Severity of Illness Index
  • Triage*