A BERT-based recommender system for secure blockchain-based cyber physical drug supply chain management (Jun, 10.1007/s10586-023-04088-6, 2023)

Drug Supply Chain Management (DSCM) can be one of the most affected streams in healthcare due to pandemics. The delivery of medicine to patients through DSCM is a complex process. Also, DSCM has several challenges, including counterfeiting, fraud, and the availability of medicine. Therefore, there is a need for security and intelligence strategies to remove pharmaceutical fraud, which remains a significant challenge since ensuring fair and secure access to medicine, services, and assistance is essential in Cyber-Physical Systems (CPS)-based DSCM. The existing CPS-based DSCM systems do, however, have some limitations in security, intelligence, planning, scheduling, quality, and logistics. This paper proposes a secure drug supply chain management framework that can acheive more security and intelligence via machine learning models. The proposed framework utilizes Bidirectional Encoder Representations from Transformers (BERT)-based and machine learning-based attack detection modules to provide more intelligence and security in blockchain-based DSCM. Evaluation results show that BERT-based recommender systems ideally suggest appropriate alternative drugs that are close to 99 % similar to the prescribed medication based on public datasets. Moreover, attack detection in the proposed framework provides significant accuracy, precision, recall, and F-measure results in threat detection (phishing, scamming, and abnormal transactions) in the blockchain layer . © 2023, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.