Machine Learning for Donor and Resource Management: A Systematic Review Applied to Thalassemia Care

Abstract

The management of Thalassemia depends on the timely and continuous availability of compatible blood transfusions, which places ongoing coordination demands on patients, donors, and blood banks. Machine learning (ML) has increasingly been applied to healthcare logistics in recent years, particularly for studying donor behavior, forecasting blood inventory, and supporting operational decision-making. However, these efforts are largely oriented toward system-level efficiency, with limited focus on transfusion-dependent patient care. This paper reviews existing studies that apply ML techniques to blood donor prediction, inventory management, and donor retention. The surveyed work shows that ML-based approaches can improve population scale planning and resource utilization. At the same time, most solutions are designed for aggregate optimization and do not directly address the continuity required in chronic transfusion settings. For individuals with Thalassemia, care depends less on short-term efficiency and more on the sustained availability of suitable donors over long periods. Based on the reviewed literature, there is a need for ML frameworks that make use of longitudinal data and explicitly support patient-linked donor coordination in long-term transfusion management.