Analysis of medication demand forecasting using single exponential smoothing method based on consumption patterns at Bali Mandara Hospital, Indonesia 10.55131/jphd/2026/240108

Abstract

Drug planning requires accurate forecasting to determine the right type and amount of drugs. The Single Exponential Smoothing (SES) method is used in this study due to its ability to produce more accurate forecasts than other methods. Drugs are classified based on consumption patterns into Smooth, Intermittent, Erratic, and Lumpy to evaluate the accuracy of the SES method on each pattern. This descriptive and non-experimental study uses retrospective data from 469 drugs at Bali Mandara Hospital from January 2023 to December 2023. Data were classified based on consumption patterns with CV² and ADI parameters, and forecasting was performed using the Eviews-12 application and evaluated with the MAPE accuracy indicator using Microsoft Excel. The clustering results showed 335 drugs with smooth consumption patterns, three with intermittent consumption patterns, 50 with lumpy consumption patterns, and 81 with erratic consumption patterns. SES gives the best results in a smooth pattern, with 71.43% of forecasts falling into the accurate, good, or reasonable category. The smooth pattern also shows highly accurate (MAPE <10%) and good (MAPE 10%-20%) forecasting rates that are higher than other consumption patterns. This study concludes that drugs categorized as smooth consumption patterns produce more accurate forecasts than other consumption patterns (intermittent, lumpy, and erratic).