Effect of age adjustment on two triage methods

Most emergency departments rely on acuity assessment, triage, to recognize critically ill patients that need urgent treatment, and to allocate resources according to need. The accuracy of commonly used triage instruments such as the Emergency Severity Index (ESI) is lower for older adults compared to young patients. We aim to examine, whether adjusting the triage category by age leads to improvement in sensitivity without excessive increase in patient numbers in the higher triage categories. The primary outcome measure was 3-day mortality and secondary outcomes were 30-day mortality, hospital admission, and HDU/ICU admissions.

We gathered data of all adult patients who had an unscheduled visit to any of our three emergency departments within one month. The data was analysed for 3-day mortality, 30-day mortality, hospital admission, and high dependency unit or intensive care unit (HDU/ICU) admission. The analysis was run for both the standard ESI triage method and a local 3-level Helsinki University Hospital (HUH) method. A further analysis was run for both triage methods with age adjustment. Net reclassification improvement values were calculated to demonstrate the effect of age adjustment.

Thirteen thousand seven hundred fifty-nine patients met the study criteria, median age was 57. 3-day mortality AUCs for unadjusted HUH and ESI triage were 0.77 (0.65-0.88) and 0.72 (0.57-0.87); 30-day mortality AUCs were 0.64 (0.59-0.69) and 0.69 (0.64-0.73); hospital admission AUCs were 0.60 (0.68-0.71) and 0.66 (0.65-0.68) and HDU/ICU admission AUCs were 0.67 (0.64-0.70) and 0.82 (0.79-0.86), respectively. Age adjustment improved accuracy for 30-day mortality and hospital admission. With the threshold age of 80, AUCs for 30-day mortality were 0.73 (0.68-0.77) and 0.77 (0.73-0.81) and for hospital admission, 0.66 (0.65-0.67) and 0.72 (0.71-0.73) for the HUH and ESI triage. The effect was similar with all cut off ages.

Moving older adults into a more urgent triage category based on age, improved the triage instruments' performance slightly in predicting 30-day mortality and hospital admission without excessive increase in patient numbers in the higher triage categories. Age adjustment did not improve HDU/ICU admission or 3-day mortality prediction.