Model to predict inpatient mortality from information gathered at presentation to an emergency department: The Triage Information Mortality Model (TIMM).

OBJECTIVES: To derive and validate a mortality prediction model from information available at ED triage. METHODS: Multivariable logistic regression of variables from administrative datasets to predict inpatient mortality of patients admitted through an ED. Accuracy of the model was assessed using the receiver operating characteristic area under the curve (ROC-AUC) and calibration using the Hosmer-Lemeshow goodness of fit test. The model was derived, internally validated and externally validated. Derivation and internal validation were in a tertiary referral hospital and external validation was in an urban community hospital. RESULTS: The ROC-AUC for the derivation set was 0.859 (95% CI 0.856-0.865), for the internal validation set was 0.848 (95% CI 0.840-0.856) and for the external validation set was 0.837 (95% CI 0.823-0.851). Calibration assessed by the Hosmer-Lemeshow goodness of fit test was good. CONCLUSIONS: The model successfully predicts inpatient mortality from information available at the point of triage in the ED.

Predicting admission of patients by their presentation to the emergency department.

OBJECTIVE: The present study aims to determine the importance of certain factors in predicting the need of hospital admission for a patient in the ED. METHODS: This is a retrospective observational cohort study between January 2010 and March 2012. The characteristics, including blood test results, of 100,123 patients who presented to the ED of a tertiary referral urban hospital, were incorporated into models using logistic regression in an attempt to predict the likelihood of patients' disposition on leaving the ED. These models were compared with triage nurses' prediction of patient disposition. RESULTS: Patient age, their initial presenting symptoms or diagnosis, Australasian Triage Scale category, mode of arrival, existence of any outside referral, triage time of day and day of the week were significant predictors of the patient's disposition (P < 0.001). The ordering of blood tests for any patient and the extent of abnormality of those tests increased the likelihood of admission. The accuracy of triage nurses' admission prediction was similar to that offered by a model that used the patients' presentation characteristics. The addition of blood tests to that model resulted in only 3% greater accuracy in prediction of patient disposition. CONCLUSIONS: Certain characteristics of patients as they present to hospital predict their admission. The accuracy of the triage nurses' prediction for disposition of patients is the same as that afforded by a model constructed from these characteristics. Blood test results improve disposition accuracy only slightly so admission decisions should not always wait for these results.