Time Series Modeling and Forecasting of the Patients’Inflow and Admission in the Hospitals: A cases study of LUMHS Hospital Jamshoro Pakistan

Abstract

The patients’ crowding in the hospitals is an international phenomenon that demands much attention to avoid harm to the lives of patients. The quantitative based models have been successfully investigated to predict the crowding of patients. Thus, the main objective of this study is to probe a statistically feasible forecasting model capable of estimating the crowding of patients (patients’ inflow and patients’ admission specifically). As a case study, the Liaquat University of Medical and Health Sciences (LUMHS) Hospital Jamshoro was chosen. The patients’ secondary data was collected form hospital and commercial computational software MATLAB was used to carry out all the calculations and manipulations by writing a concise user defined program (code). The Autoregressive Integrated Moving Average (ARIMA) modeling approach is adopted to investigate the best forecasting model. It is found that among the various six combinations of ARIMA (p,d,q) the ARIMA (1,0,1) are the best fit models for the patients’ inflow and the patients’ admission respectively; having the lowest AIC, BIC and p-values. Since the forecast accuracy contains minimal contains minimal errors thus forecast trends show very good results. The presented procedure can be helpful to manage the patients’ volume in the hospitals and can also predict the future trend of patients’ inflow and patients’ admission with good accuracy.