Can Triage Nurses Accurately Predict Patient Dispositions in the Emergency Department?

Contemporary emergency departments experience crowded conditions with poor patient outcomes. If triage nurses could accurately predict admission, one theoretical intervention to reduce crowding is to place patients in the admission cue on arrival to the emergency department. The purpose of this study was to determine if triage nurses could accurately predict patient dispositions.

This prospective study was conducted in a tertiary academic hospital's emergency department using a data collection tool embedded in the ED electronic information system. Study variables included the predicted and actual disposition, as well as level of care, gender, age, and Emergency Severity Index level. Data were collected for 28 consecutive days from September 17 through October 9, 2013. Sensitivity and specificity, positive and negative predictive values, and accuracy of prediction, as well as the associations between patient characteristics and nurse prediction, were calculated.

A total of 5,135 cases were included in the analysis. The triage nurses predicted admissions with a sensitivity of 71.5% and discharges with a specificity of 88.0%. Accuracy was significantly higher for younger patients and for patients at very low or very high severity levels.

Although the ability to predict admissions at triage by nurses was not adequate to support a change in the bed procurement process, a specificity of 88.0% could have implications for rapid ED discharges or other low-acuity processes designed within the emergency department. Further studies in additional settings and on alternative interventions are needed.