Forecasting the Amount of Drug Procurement: A Case Study of a Community Hospital

Abstract

Objective: To study the accuracy of forecasting methods for the amount of drug procurement using the simple moving average method and the exponential smoothing method. Methods: This study examined only 61 drugs classified as group A according to the VEN and SDE systems in a F2 community hospital. The study used data on actual drug use from the past 3 fiscal years (2016-2018) for forecasting the values of drug procurement. The forecasted values were then compared with actual drug use data in fiscal year 2019. The comparison of forecasting accuracy was expressed by mean absolute percentage error (MAPE). Results: Historical data on the use of drugs in group A suggested the formless of data. Forecasts of the single exponential smoothing method with a smoothing coefficient of 0.9 were of the highest accuracy, with the lowest MAPE value of 19. When classified by the MAPE accuracy level, single exponential smoothing where the smoothing coefficient was 0.9 had the highest percentage of accuracy (MAPE less than 10) of 24.6, which was significantly higher than those of other methods (P<0.001). The essential drug group had a statistically significantly higher percentage of accuracy (11.7% with MAPE less than 10) than that in the vital group (6.3% with MAPE less than 10) (P=0.010). Conclusion: The results of this research are useful to pharmacists in hospitals with the same context. The single exponential smoothed forecasting method should be considered for applying to plan the procurement of group A drug with formless movement in order to reduce the inaccuracies in forecasting drug procurement quantities and prevent drug shortages.