Triage nurse assessment of the need for admission from an ED based on clinical judgement outperformed several routinely used scoring systems
Triage nurse assessment based on clinical judgement alone, of whether a patient should be admitted after visiting an emergency department (ED), has been shown to be far better than several scoring systems. This was the finding of a study by a team from UOC Pronto Soccorso e Medicina d’Urgenza, Milan, Italy.
Overcrowding in an ED leads to an increased waiting time and some evidence shows that reducing overcrowding is linked with better clinical outcomes. Consequently, ED staff require some form of rapid assessment of patients to ensure appropriate disposition. Although several scoring tools such as the Ambulatory (AMB score), the Glasgow Admission Prediction (GAP) and the Sydney Triage to Admission Risk Tool (START) have good predictive accuracy, such tools have not yet proven their worth. However, an alternative to the use of assessment tools would be for nurses to use their clinical judgement but a recent systematic review concluded that ‘triage nurse prediction of disposition is not accurate enough to expedite admission for ED patients.‘
Nevertheless, there is currently a lack of data comparing individual scoring tools with triage nurse assessment and for the present study, the Italian team decided to compare these existing tools with the clinical judgement of nurses in predicting hospital admission.
They conducted a prospective, single-centre, observational study at a tertiary teaching hospital which has around 70,000 adult ED visits each year. For the study, the triage nurse calculated a patient’s AMB, GAP and START scores and estimated the probability of admission according to their clinical judgement using a 0 to 100 scale. Though the nurses collected the data to calculate each score, this was determined by the investigator team so that the nurses were blinded to the final score. Their own assessments were dichotomised for the purposes of analysis, with a greater than 50% estimated probability, used to define a prediction of admission. The primary outcome of the study was hospital admission and receiver operating characteristic curves were generated for the accuracy of predictions and the area under the curve (AUC) for each tool compared.
A total of 1710 patients with a median age of 54 years (49.3% male) visited the hospital ED and were included in the analysis and among whom, 310 (18%) were subsequently admitted from the ED.
The AUCs were 0.77, 0.72, 0.61 and 0.86 for the AMB, GAP, START and triage nurse clinical assessment respectively and the nurses’ clinical judgement was significantly higher than the AUC of all the other tools and for all comparisons (p < 0.0001).
In a separate analysis, age, years of experience as a nurse, years of experience as an ED nurse and years of performing triage were found not be related to the nurse’s ability to predict triage and hence regression analysis of this data was not undertaken.
Commenting on their findings, the authors noted how this was the first study which directly compared the currently available tools and they concluded that while clinical judgement was subjective, it still provided good predictive accuracy and was superior to any of the other tools available.