2019/01/05 by Avishek Choudhury, Choudhury, Avishek
Medicine · Decision Sciences · #Emergency and Acute Care Studies #Forecasting Techniques and Applications
paper · pdf · doi:10.48550/arxiv.1901.02714
Background: The stochastic behavior of patient arrival at an emergency\ndepartment (ED) complicates the management of an ED. More than 50% of hospitals\nED capacity tends to operate beyond its normal capacity and eventually fails to\ndeliver high-quality care. To address the concern of stochastics ED arrivals,\nmany types of research has been done using yearly, monthly and weekly time\nseries forecasting. Aim: Our research team believes that hourly time-series\nforecasting of the load can improve ED management by predicting the arrivals of\nfuture patients, and thus, can support strategic decisions in terms of quality\nenhancement. Methods: Our research does not involve any human subject, only ED\nadmission data from January 2014 to August 2017 retrieved from the UnityPoint\nHealth database. Autoregressive integrated moving average (ARIMA), Holt\nWinters, TBATS, and neural network methods were implemented to forecast hourly\nED patient arrival. Findings: ARIMA (3,0,0) (2,1,0) was selected as the best\nfit model with minimum Akaike information criterion and Schwartz Bayesian\ncriterion. The model was stationary and qualified the Box Ljung correlation\ntest and the Jarque Bera test for normality. The mean error (ME) and root mean\nsquare error (RMSE) were selected as performance measures. An ME of 1.001 and\nan RMSE of 1.55 was obtained. Conclusions: ARIMA can be used to provide hourly\nforecasts for ED arrivals and can be utilized as a decision support system in\nthe healthcare industry. Application: This technique can be implemented in\nhospitals worldwide to predict ED patient arrival.\n