【Author】
Liu, Wei; Feng, Wenlong; Yu, Benguo; Peng, Tao
【Source】UBIQUITOUS SECURITY
【Abstract】The sharing of Electronic Medical records (EMRs) has great positive significance for research of disease and doctors' diagnosis. However, patients' EMRs are usually distributed in the databases of multiple medical institutions. Due to the insecurity of the network environment and distrust of other parties, EMR owners worry about data insecurity and privacy leakage, which makes sharing with other parties difficult. Patients worry about the loose control of their health data as well. To solve this problem, we present a solution for the EMRs data sharing based on blockchain and federated learning, which will provide data security and patients' privacy. Firstly, we propose a method for EMRs data retrieval records and sharing records as transaction records adding to the blockchain, and design the two algorithm processes, respectively. Secondly, federated learning is used to help EMRs data owners to build a model based on the original data. The data owner only shares the model instead of the original data. Finally, by security and privacy analytics, we analyzed the advantages and influence of the proposed model. Overall, the evaluation shows that the proposed solution is significantly superior to the previous models and achieves reasonable efficiency for sharing EMRs data.
【Keywords】Electronic medical records; Security sharing; Privacy; Blockchain; Federated learning
【标题】基于区块链和联邦学习的电子病历共享安全和隐私
【摘要】电子病历（EMR）的共享对于疾病研究和医生诊断具有重要的积极意义。然而，患者的电子病历通常分布在多个医疗机构的数据库中。由于网络环境的不安全和对其他方的不信任，EMR 拥有者担心数据不安全和隐私泄露，这使得与其他方共享变得困难。患者也担心对其健康数据的控制松懈。为了解决这个问题，我们提出了一种基于区块链和联邦学习的 EMR 数据共享解决方案，这将提供数据安全和患者隐私。首先，我们提出了一种将 EMR 数据检索记录和共享记录作为交易记录添加到区块链的方法，并分别设计了这两种算法流程。其次，联邦学习用于帮助 EMR 数据所有者基于原始数据构建模型。数据所有者只共享模型而不是原始数据。最后，通过安全和隐私分析，我们分析了所提出模型的优势和影响。总体而言，评估表明，所提出的解决方案明显优于以前的模型，并且在共享 EMR 数据方面达到了合理的效率。
【关键词】电子病历；安全共享；隐私;区块链；联邦学习
评论