Machine Learning-Driven Query Optimization for Secure and Efficient Blockchain Electronic Medical Record Management

Abstract

The increasing volume of electronic medical records (EMRs) has created a growing need for healthcare information systems that can provide secure, reliable, and efficient access to patient data. Although blockchain offers strong security, immutability, transparency, and controlled data sharing, the increasing size and complexity of blockchain records can affect query performance and data retrieval. This study proposes a Machine Learning-Driven Query Optimization Framework for secure and efficient blockchain-based EMR management. The framework integrates machine learning, permissioned blockchain, smart contracts, and security mechanisms to optimize query execution while protecting sensitive medical information. Machine learning is employed to analyse query patterns and support the selection of efficient execution strategies, thereby reducing processing overhead and improving system responsiveness. The proposed framework was evaluated by experts using functional, security, privacy, reliability, scalability, usability, and query-efficiency criteria. The results showed positive expert assessments across all criteria, with Query Efficiency recording the highest mean score (4.28), followed by Functionality (4.23) and Security (4.16). These findings suggest that integrating machine learning with blockchain can improve the efficiency of EMR data retrieval without undermining essential security and access-control requirements. The framework provides a foundation for intelligent and scalable healthcare information management and can support secure information exchange among authorized healthcare stakeholders. Further empirical testing with real-world healthcare workloads is recommended to validate its performance and scalability.