A Differential-Privacy-Based Blockchain Architecture to Secure and Store Electronic Health Records

Abstract

As humanity enters the information age, the amount of digitized personal information grows daily. With growing connections between the digital and real-world, privacy information becomes more and more at risk, most especially information pertaining to one's electronic health records, or EHRs. The consequences of improper EHR security are shown in [12] and [13]. Innovations in security, namely Blockchain and differential privacy, provide data centers a powerful tool to combat would-be belligerents and secure patient data. We propose a novel blockchain architecture that utilizes the discrete M-band wavelet transform with Laplace-Sigmoid noise that allows connected centers to perform relevant research while also securing sensitive EHRs and protecting patient identities. We then simulate training machine learning models using our system and show that they perform with high accuracy.