A Short Survey of AI-driven Load Distribution for Electronic Health Record Management with Recent Deployed Techniques, Research Gaps and Challenges

With the arrival of Artificial Intelligence (AI), the Internet of Things (IoT), machine learning, and deep learning methodologies, data-driven medical applications have appeared as a promising solution for implementing scalable and robust prognostic and diagnostic approaches with the utilization of medical data. These technologies attained a lot of interest from academia and industry and led to insightful advancements in healthcare systems. The AI-driven techniques related to medical applications still face several issues including privacy, security, and Quality-of-Service (QoS) standards. The complex machine learning models are effectively trained via Federated Learning (FL) in a distributed manner, this advanced technique has become an attractive research area, especially in processing healthcare data at the network edge in a decentralized way to address security concerns and preserve privacy of the information. The availability of FL in the medical domain has several benefits, but it has not been explored to its extent. Existing review works related to FL mainly concentrated on the importance of FL in diverse applications, but no brief or comprehensive survey exists on the distribution of Electronic Health Records (EHR). Thus, this review work focuses on various techniques of AI-driven load distribution for federated networks on EHR. For this purpose, a detailed literature survey is conducted on AI-driven load distribution framework for federated networks on EHR from the previous decades. The AI techniques employed for disease detection are analyzed and categorized. The dataset details, performance metrics, and simulation tools are gathered and listed for a better understanding of the process. Subsequently, the research gaps on the conventional AI-driven load distribution framework for federated networks on EHR are provided to facilitate future work.