Predictive Analytics on Drug Shortages and Future Challenges: A Review

Drug shortages are a major global public health threat that disrupts patient care; increases costs, medication errors, and mortality; and forces treatment interruptions.Several countries have adopted mandatory reporting systems and national monitoring platforms, thus highlighting the need for predictive analytics.Prediction research generally includes two approaches: forecasting demand and assessing supply disruption risks.Demand forecasting applies models such as the autoregressive integrated moving average to sales, search, and inventory data, whereas supply risk assessment uses machine learning and artificial intelligence to detect vulnerable drugs or suppliers, as shown in studies conducted in the U.S., Canada, China, and France.Combining these approaches could enable proactive shortage management.However, certain limitations remain, and limited data quality and accessibility hinder predictive accuracy.Centralizing shortage data from all stakeholders across the pharmaceutical supply chain, including manufacturers, distributors, healthcare providers, and patients, can improve data availability.Despite these advances, the predictive accuracy of current analytical tools is limited, thereby highlighting the need to explore more advanced tools.Overall, embedding predictive analytics tools into clinical practice is essential; however, their effectiveness remains a policy challenge.