MediChainGuard: A Permissioned Blockchain-Based Framework for Patient-Centric Access Control and AI-Assisted Threat Detection in IoMT

Abstract

The rapid expansion of the Internet of Medical Things (IoMT) has enhanced remote patient monitoring, connected diagnosis, and continuous healthcare delivery. However, centralized healthcare systems remain vulnerable to single points of failure, unauthorized access, insider attacks, data tampering, and privacy breaches, creating significant challenges for secure Electronic Health Record (EHR) management. This paper presents MediChainGuard, a permissioned blockchain-based framework for patient-centric access control and AI-assisted threat detection in IoMT environments. The proposed framework integrates Decentralized Identity (DID) for secure authentication, blockchain-enabled patient consent management, a hybrid Role-Based Access Control (RBAC) and Attribute-Based Access Control (ABAC) model for context-aware authorization, a Break-Glass mechanism for emergency access, and SHA-256-based integrity verification for secure EHR protection. Furthermore, an AI-assisted anomaly detection module employing Isolation Forest-based analysis continuously monitors authentication logs, blockchain transactions, and user behavioural patterns to identify suspicious activities and cyber threats in real time. Experimental evaluation conducted in a simulated IoMT environment demonstrates improvements in authentication efficiency, authorization accuracy, EHR integrity protection, and threat detection performance while maintaining scalability through a permissioned Hyperledger Fabric architecture. The proposed framework provides a secure, transparent, auditable, and privacy-preserving solution for next-generation healthcare systems and establishes a foundation for integrating blockchain, artificial intelligence, and decentralized identity into secure IoMT ecosystems. Experimental evaluation was conducted using a simulated healthcare environment consisting of 15,000 patients, 4,000 healthcare professionals, and 25,000 IoMT devices over six months of operational activity.