The Sydney triage to admission risk tool (START) to improve patient flow in an emergency department: a model of care implementation pilot study.

BACKGROUND: The Sydney Triage to Admission Risk Tool (START) is a validated clinical analytics tool designed to estimate the probability of in-patient admission based on Emergency Department triage characteristics. METHODS: This was a single centre pilot implementation study using a matched case control sample of patients assessed at ED triage. Patients in the intervention group were identified at triage by the START tool as likely requiring in-patient admission and briefly assessed by an ED Consultant. Bed management were notified of these patients and their likely admitting team based on senior early assessment. Matched controls were identified on the same day of presentation if they were admitted to the same in-patient teams as patients in the intervention group and same START score category. Outcomes were ED length of stay and proportion of patients correctly classified as an in-patient admission by the START tool. RESULTS: One hundred and thirteen patients were assessed using the START-based model of care. When compared with matched control patients, this intervention model of care was associated with a significant reduction in ED length of stay [301 min (IQR 225-397) versus 423 min (IQR 297-587) p < 0.001] and proportion of patients meeting 4 h length of stay thresholds increased from 24 to 45% (p < 0.001). CONCLUSION: In this small pilot implementation study, the START tool, when used in conjunction with senior early assessment was associated with a reduction in ED length of stay. Further controlled studies are now underway to further examine its utility across other ED settings.

Extending the Sydney Triage to Admission Risk Tool (START+) to predict discharges and short stay admissions.

OBJECTIVE: This study aims to validate previously reported triage tool titled Sydney Triage to Admission Risk Tool (START+) and investigate whether an extended version of the tool could be used to identify and stream appropriate short stay admissions to ED observation units or specialised short stay inpatient wards. METHODS: This was a prospective study at two metropolitan EDs in Sydney, Australia. Consecutive triage encounters were observed by a trained researcher and START scores calculated. The primary outcome was length of stay <48 hours. Multivariable logistic regression was used to estimate area under curve of receiver operator characteristic (AUROC) for START scores. The original START tool was then extended to include frailty and multiple or major comorbidities as additional variables to assess for further predictive accuracy. RESULTS: There were 894 patients analysed during the study period. Of the 894 patients, there were 732 patients who were either discharged from ED or admitted for <2 days. The AUROC for the original START+ tool was 0.80 (95% CI 0.77 to 0.83). The presence of frailty was found to add a further five points and multiple comorbidities added another four points on top of the START score, and the AUROC for the extended START score 0.84 (95% CI 0.81 to 0.88). CONCLUSION: The overall performance of the extended ED disposition prediction tool that included frailty and multiple medical comorbidities significantly improved the ability of the START tool to identify patients likely to be discharged from ED or require short stay admission <2 days. TRIAL REGISTRATION NUMBER: ACTRN12618000426280.

Evaluating Stroke Code Activation Pathway in Emergency Departments study.

OBJECTIVES: To describe the clinical characteristics and outcomes of Code Stroke activations in an ED and determine predictors of a final diagnosis of stroke or transient ischemic attack (TIA) diagnosis. METHODS: This was a retrospective analysis of Code Stroke activations through an ED over 2 years at a quaternary stroke referral centre. Stroke Registry data was used to identify cases with clinical information abstracted from electronic medical records. The primary outcome was a final diagnosis of acute stroke or TIA and the secondary outcome was access to reperfusion therapies (thrombolysis and or endovascular clot retrieval). RESULTS: The study analysed data from 1354 Code Stroke patients in ED. Of all Code Strokes, 51% had a stroke or TIA diagnosis on discharge. Patient characteristics independently associated with increased risk of stroke were increasing age, pre-arrival notification by ambulance, elevated BP or presence of weakness or speech impairment as the initial presenting symptoms. Dizziness/vertigo/vestibular neuritis were the most common alternative diagnoses. One hundred and thirty-five patients (10%) underwent reperfusion therapy. Pre-arrival notification by ambulance was associated with higher proportion of eventual stroke/TIA diagnosis (68% vs 46%, P < 0.001) and significantly lower door to CT and door to needle times for patients undergoing thrombolysis. CONCLUSIONS: In a cohort of patients requiring Code Stroke activation in an ED, increased age, systolic blood pressure and weakness and speech impairment increased the risk of stroke. Prehospital notification was associated with lower door to needle times for patients undergoing thrombolysis.

The Sydney Triage to Admission Risk Tool (START2) using machine learning techniques to support disposition decision-making.

OBJECTIVE: To further develop and refine an Emergency Department (ED) in-patient admission prediction model using machine learning techniques. METHODS: This was a retrospective analysis of state-wide ED data from New South Wales, Australia. Six classification algorithms (Bayesian networks, decision trees, logistic regression, naïve Bayes, neural networks and nearest neighbour) and five feature selection techniques (none, manual, correlation-based, information gain and wrapper) were examined. Presenting problem was categorised using broad (n = 20) and specific (n = 100) representations. Models were evaluated based on Area Under the Curve (AUC) and accuracy. The results were compared with the Sydney Triage to Admission Risk Tool (START), which uses logistic regression and six manually selected features. RESULTS: Sixty admission prediction models were trained and validated using data from 1 721 294 patients. Under the broad representation of presenting problem, the nearest neighbour algorithm with manual feature selection had the best AUC of 0.8206 (95% CI ±0.0006), while the decision tree with no feature selection had the best accuracy of 74.83% (95% CI ±0.065). Under the specific representation, almost all models improved; the nearest neighbour with information gain feature selection had the best AUC of 0.8267 (95% CI ±0.0006), while the decision tree with wrapper or no feature selection had the best accuracy of 75.24% (95% CI ±0.064). Eleven of the machine learning models had slightly better AUC than the START model. CONCLUSION: Machine learning methods demonstrate similar performance to logistic regression for ED disposition prediction models using basic triage information. This should be investigated further, especially for larger data sets with more complex clinical information.

The Sydney Triage to Admission Risk Tool (START): A prospective validation study.

OBJECTIVE: The present study aims to prospectively validate the Sydney Triage to Admission Risk Tool (START) to predict ED disposition. METHODS: This was a prospective validation study at two metropolitan EDs in Sydney, Australia. Consecutive triage encounters were observed by a trained researcher and START scores calculated. The primary outcome was patient disposition (discharge or inpatient admission) from the ED. Multivariable logistic regression was used to estimate area under curve of receiver operator characteristic (AUC ROC) for START scores as well as START score in combination with other variables such as frailty, general practitioner referral, overcrowding and major medical comorbidities. RESULTS: There were 894 patients analysed during the study period. The START score when applied to the data had AUC ROC of 0.80 (95% CI 0.77-0.83). The inclusion of other clinical variables identified at triage did not improve the overall performance of the model with an AUC ROC of 0.81 (95% CI 0.78-0.84) in the present study. CONCLUSION: The overall performance of the START tool with respect to model discrimination and accuracy has been prospectively validated. Further clinical trials are required to test the clinical effectiveness of the tool in improving patient flow and overall ED performance.