Lumpy Demand Forecasting for Slow-moving Medicines: A Case Study of Community Hospital Thailand

Abstract

Introduction: The small community hospitals, a case study in Thailand always face inventory management problems such as shortages and overstocks of medicines because of unpredictable demand, particularly irregular demand. The irregular demand which zero demand were appeared many periods and demand values can vary greatly. It was known as lumpy demand. It obviously appears in case of slow moving vital medicines that they are critically needed for the patients. Objective aimed to examine the effectiveness of forecasting method for the slow-moving medicines in small community hospitals. Methods: The study compared the effectiveness of two forecasting methods; Croston’s method (CR) and the Teunter, Syntetos, and Babai’s (TSB) method. The simulation was analyzed using historical data, obtained from a community hospital in Thailand.  The data were collected from January 2015 to December 2016. Indicators used in this study are the mean square error (MSE) and shortage. Results: Result pointed out that the pattern of irregular demand and interval to adjust the smoothing constant values affect to the effectiveness of forecasting method.TSB method outperformed the Croston’s method, indicated by smaller MSEs, when applied to V1’s pattern whereas Croston’s method outperformed in case of shortage indicator. Both Croston’s and TSB method obtained smaller MSEs when the smoothing constant values were remained the same to 8 and 12 weeks. However, the numbers of shortage for both methods were high. Conclusion: When the forecasting method were applied to forecast irregular demand, demand pattern, the variation of demand and demand interval should be concerned. Because those issues affect to forecasting value that it could lead to shortage or over inventory.