Recent Advances in Blockchain-Based Hybrid Contextual-ATNet Approach for Daily Diabetes Management: Predicting Insulin Dosage for Improved Control: A Systematic Review

Abstract

Diabetes mellitus is a rapidly growing global health concern requiring continuous and precise insulin management to prevent severe complications. Traditional approaches often fail to achieve optimal glycemic control due to the dynamic and individualized nature of glucose-insulin interactions. The increasing availability of continuous glucose monitoring and patient data has created opportunities for intelligent, personalized decision-support systems.

This paper presents a systematic review of blockchain-integrated Contextual-ATNet architectures for insulin dosage prediction. The model leverages attention-based temporal networks and contextual embeddings to capture complex physiological patterns from multivariate data, including glucose levels, diet, and activity. Blockchain technology enhances this framework by enabling secure, decentralized data management, supporting federated learning, and ensuring data integrity through smart contracts and consensus mechanisms.

Applications include real-time glucose forecasting, personalized insulin recommendations, and remote patient monitoring. Empirical results demonstrate improved prediction accuracy, extended forecasting horizons, and enhanced privacy preservation compared to traditional models. Despite these advancements, challenges related to scalability, interoperability, and clinical deployment remain. This review highlights the potential of combining deep learning and blockchain to develop secure, efficient, and personalized diabetes management systems.