OBJECTIVES: Demand forecasting is a challenge which requires relevant data and advanced statistical procedures to address opportunities. On the pharmacy perspective, high inventory may increase the inventory holding and storage costs of medications and increase the chance of a medication reaches its expiration date. On the other hand, low inventory may increase the chances of stockouts, when a medication is not available to patients. Thus, optimizing the demand forecasting system would financially benefit any pharmacy. This study applied analytical methods to the demand forecasting of the top-ten most-prescribed medications in a specialty pharmacy.

METHODS: The data consisted of 26 months of pre-recorded real-world dispensing data of the top-ten most prescribed medications in the pharmacy: erenumab-aooe, fremanezumab-vfrm, bictegravir, mycophenolate mofetil, galcanezumab-gnlm, etanercept, cannabidiol, rimegepant, tacrolimus and temozolomide. The variables considered for the forecasting models were weekly demand and the date of purchase of each medication. Two commonly used demand forecasting models, ARIMA (Autoregressive Integrated Moving Average) and Long Short-term Memory (LSTM), were tested to determine the accuracy of forecasting. The accuracy of the models of each medication was assessed by Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) accuracy metrics.

RESULTS: The study developed an accurate forecasting model for demand of the top-ten most-prescribed medications in a specialty pharmacy. The selected forecasting method was ARIMA. The results revealed high accuracy for two medications (bictegravir and temozolomide), reasonable accurate predictions for seven medications (erenumab-aooe, fremanezumab-vfrm, galcanezumab-gnlm, rimegepant, etanercept, cannabidiol and tacrolimus) and only one medication (mycophenolate mofetil) model was considered as not accurate.

CONCLUSIONS: The ARIMA model accurately forecasted the demand for nine medications, demonstrating either high accuracy or reasonable precision. Therefore, employing data-driven analytical methods appears to be a good option to demand forecasting compared to reliance on pharmacists’ experimental forecasting method.