Development and Validation of a Prediction Model for the Number of Patients Visiting Emergency Departments

ABSTRACT

PURPOSE: We aimed to develop and validate a prediction model for the number of patients visiting emergency departments. METHODS: Enrolled patients were from eleven regional emergency departments (EDs) (level-1) that inputted information on emergency patients into the National Emergency Department Information System since 2004. We developed the automated regressive integrated moving average (ARIMA)-based prediction model using a dataset covering 2005 to 2007. To validate the prediction model, we performed Bland-Altman plot analysis for a new dataset, that of 2008, calculating the agreement rate. RESULTS: The total number of enrolled patients was 1,532,294. Of these, 844,802 (55.1%) were male and mean age was 36.5. The ARIMA (1, 1, 1) (1, 1, 1) 7 was selected as the best-fit prediction model. When we tested the validity using Bland-Altman plots, the agreement rate was 96.4% (95% CI, 94.0%~98.1%). Non-agreement dates were national holidays (n=9), and the other weekdays (n=4), respectively. CONCLUSION: We developed the ARIMA-based prediction model for emergency patients at regional EDs. The model showed a very high validity.