How to Share Medical Data Belonging to Multiple Owners in a Secure Manner

Changsheng Zhao, Wei Song, Zhiyong Peng

Abstract: Over the past few years, we have witnessed that the big data analysis technologies based on machine learning and deep learning have brought great change to medical research. The medical data is a special kind of sensitive data, which is contributed by multiple users. For example, the prescriptions are generated by doctors but also contain the patients’ symptoms and history of diseases. However, the existing data sharing solutions focus on protecting the privacy of the single data owner while ignoring the privacy belonging to multiple owners. In this paper, we first identify this problem and propose a novel secure medical data sharing scheme based on the Chinese remainder theorem to protect the privacy of multiple owners in the medical data. To better support the scenario of medical data sharing, we implement the proposed scheme over a medical blockchain system to build a medical data sharing prototype, and dynamically manage two kinds of subscribers based on logical key hierarchy. The performance analysis validates the security and efficiency of the proposed method for the data sharing scenario of medical data from multiple owners.