Patient Visit Forecasting at Emergency Department using Autoregressive Integrated Moving Average (ARIMA) and Exponential Smoothing Method in RSUD Kembangan

Abstract

The situation in the Emergency Department (ED) at RSUD Kembangan is generally overcrowded where many patient’s arrival is unpredictable. Based on the results data in 2015-2019, patient visits to the emergency department tend to increase by around 42% per year. The limited number of beds and medical personnel causes a decrease in productivity and mobility when conducting health services. Therefore, forecasting for patient visit is needed to minimize these problems. This study aims to predict patient visits at the Emergency Department in RSUD Kembangan using Autoregressive Integrated Moving Average (ARIMA) and Exponential Smoothing. Secondary data obtained from April 2015 to June 2019 retrieved from RSUD Kembangan. The results showed that the ARIMA model (1,1,2) was chosen as the best model with MSE 22600.3 and MAPE 10.6 while Exponential Smoothing from Brown showed MSE 26900.6 and MAPE 11.8. ARIMA (1,1,2) has the smallest error size parameter so that a suitable model is applied in forecasting the number of emergency patient visits at RSUD Kembangan in the future.