Creative Commons source reading · Prediction and forecasting
The prediction of essential medicine demand using machine learning and traditional methods on EPSS Gondar hub 2018–2022 data
Biadigilign TM; Tagele MJ; Wondimu A; Ayenew W
Originally published in Explor Res Clin Soc Pharm · 2026-05-12
From the article
Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.
Abstract
AbstractIntroductionNo prior research has utilized predictive models for forecasting essential medicine demand in Ethiopia. Therefore, this study aimed to identify the most accurate predictive model for predicting essential medicine demand at public health facilities via traditional and machine learning techniques in Northwest Ethiopia.MethodsA cross-sectional study was conducted from September to December 2023 using five years of consumption data from the Ethiopian pharmaceutical supply service. The top ten essential medicines were identified based on defined daily dose utilization. Demand prediction was performed in Python using nine models, naïve, ARIMA, linear regression, support vector machine, random forest, artificial neural network, k-nearest neighbor, gradient boosting, and extreme gradient boosting. Model accuracy was…
Discussion
The evaluated models demonstrated notable differences in forecasting performance. Among all approaches, XGB achieved the highest predictive accuracy, with the lowest RMSE (0.18) and a high-test R2 (0.87). Compared with the naïve baseline model, XGB reduced prediction error by approximately 67% and improved explanatory performance by more than 39 percentage points. In practical terms, this corresponds to an estimated reduction of about 887 DDDs in forecast error per medicine per facility, highlighting the model's potential to reduce overstocking, medicine expiry, and stock-out risks in health facilities.41 SVM and GB also showed strong predictive performance, indicating that ML approaches are better able to capture the nonlinear and heterogeneous consumption patterns commonly observed in essential medicine demand. XGB's superior…
Conclusion
This study found that XGB achieved the best performance for forecasting EM demand among the evaluated models in public health facilities of Northwest Ethiopia. The improved predictive accuracy of the model may help strengthen medicine availability by supporting better procurement planning, reducing stock-outs, and minimizing medicine wastage. The findings also indicate that ML-based forecasting could be integrated into the EPSS system to support data-driven decision-making. Linking the model with LMIS dashboards may improve real-time monitoring of medicine consumption, stock status, and procurement planning. Effective implementation, however, will require adequate digital infrastructure and practical training for supply chain personnel. Future studies should validate the model across multiple EPSS hubs using more detailed and…
Editorial reading guide
- What data created and tested the model, and was evaluation independent?
- What is the cost of a false positive, false negative or out-of-distribution decision?
