Method of Forecasting Medicine Requirements Based on Consumption Data and Disease Patterns in Hospitals

Abstract

Effective drug inventory management is crucial for maintaining service quality and cost efficiency in hospitals. Inaccurate procurement planning can lead to stock shortages or surpluses, which disrupt healthcare operations. This study presents several forecasting methods to improve the accuracy of demand predictions. Various forecasting methods have been developed based on two types of forecasting approaches, namely qualitative and quantitative. Qualitative methods produce forecasts based on assessments or opinions, while quantitative techniques largely rely on data to generate prediction models. In the pharmaceutical industry, time series models are most commonly used. Forecasting demand for pharmaceutical products by exploring artificial intelligence technologies improves the forecasting process and the accuracy of demand forecasts. However, classical univariate time series methods, such as exponential smoothing and autoregressive integrated moving average (ARIMA), generally provide good results. The ARIMA method is the best method to use compared to the Exponential Smoothing method because the MAPE and MSE calculation results show that the ARIMA method has the smallest error value. Field studies and survey results show that the use of human judgement adds value to the estimates produced by quantitative/statistical models. Implementing better forecasting models and refining inventory control mechanisms based on consumption data and disease patterns in hospitals can significantly reduce stock imbalances, optimise procurement decisions, and improve hospital financial performance.