AI-Based Medicine Shortage Prediction Using Machine Learning

Medicine shortages can create serious problems in healthcare facilities because patients may not receive medicines when they are needed. Predicting medicine shortages in advance can help healthcare facilities take timely action and improve medicine availability. This research presents an AI-based approach for predicting future medicine shortages using machine learning techniques and healthcare inventory data from Sierra Leone. The dataset contains information related to medicines, healthcare facilities, districts, dates, received quantity, consumption, opening balance, and previous stockout status. After data preprocessing and medicine selection, a future shortage target was created to predict whether a medicine shortage would occur in the following month. Three machine learning models, namely Random Forest, Gradient Boosting, and XGBoost, were developed and evaluated. The models were evaluated using Accuracy, Precision, Recall, F1-score, and ROC-AUC. To examine the ability of the models to work with healthcare data from different geographical areas, a cross-district evaluation was also performed using districts that were not included during model training. The results show that machine learning can identify patterns in healthcare inventory data that are useful for predicting future medicine shortages. The proposed approach can support healthcare facilities in identifying potential shortages earlier and can contribute to better medicine inventory planning and availability.