MEDICHAIN : BLOCKCHAIN & MACHINE LEARNING HYBRID FRAMEWORK FOR SECURE MEDICAL RECORD SHARING

Abstract

The digitization of Electronic Medical Records (EMRs) has dramatically improved operational efficiency within healthcare institutions but has introduced significant vulnerabilities regarding data privacy, structural integrity, and unauthorized systemic access. Traditional centralized architectures present a singular point of failure, leaving sensitive records prone to silent tampering and large-scale data breaches. MediChain is a proposed multi-layer security paradigm intended to mediate these issues without the high overhead of public ledgers. By integrating a custom Python-based permissioned blockchain to ensure an immutable audit trail, coupling it with AES-256-CBC local encryption for robust off-chain data confidentiality, and deploying a real-time Isolation Forest machine learning model, MediChain synthesizes a comprehensive defensive ecosystem. The platform functions as a unified framework that enforces patient-governed smart-contract consent while mathematically isolating anomalous behavioral access patterns. This paper presents the architecture and efficacy of this hybrid methodology, demonstrating an effective, decentralized-adjacent solution designed specifically for restricted-access clinical networks.