Assessment of forecasting models for patients arrival at Emergency Department

Abstract

The unpredictability of arrivals to the Emergency Department (ED) of a hospital is a great concern of the management. The existence of more complex pathologies and the increase in life expectancy originate a higher rate of hospitalization. The hospitalization of patients via ED upsets previously programmed services and some cancellations may occur. The Hospital's ability to predict turnout variations in the arrivals to the ED is fundamental to the management of the human resources and the required number of beds.Braga Hospital, in Portugal, is the subject of this work. Data for ED arrivals in 2 years (2012-2013), the test period, was studied and forecasting models based on time series were built. The models were then tested against the real data from the evaluation period (2014). These models are of ARIMA (AutoRegressive-Integrated-Moving Average) type, used software was the Forecast Pro.