Analysis of Secure EHR Storage Methods on Blockchain and Integrating ML to Predict Chronic Kidney Disease

An Electronic Health Record(EHR) is the digital form of the patient’s health data, including vital signs, demographic details, clinical notes, X-rays, medical images, etc. When discussing EHR processing and management, storing this vast information is the main challenge. Although cloud computing’s centralized storage model is one of the best ways to store large amounts of data, it is not secure. Blockchain technology has the potential to revolutionize EHR storage systems and provide data security. Directly keeping the vast EHR data on blockchain may be costly and inefficient. The best method to store the data is off-chain in a Blockchain-based EHR system. Implementing machine learning (ML) in the healthcare industry aims to anticipate diseases earlier so patients can receive higher-quality medical care. Integrating these two disruptive data-driven technologies can highly improve the quality of healthcare. First, this paper analyzes three different off-chain EHR storage methods: Storj, InterPlanetary File System(IPFS), and CosmosDB, based on their Storage and Access Time. From the experimental results, IPFS has the fastest Storage and Access Time. Second, we integrate blockchain and ML technology to predict Chronic Kidney Disease(CKD) from the CKD dataset stored on the IPFS. Finally, using IPFS storage, a record is accessed, and the prediction time is 0.4 sec for detecting CKD or NOCKD is measured.