PHDMF: A Flexible and Scalable Personal Health Data Management Framework Based on Blockchain Technology

【Author】 Ma, Liangxiao; Liao, Yongxiang; Fan, Haiwei; Zheng, Xianfeng; Zhao, Jintao; Xiao, Ziyi; Zheng, Guangyong; Xiong, Yun

【Source】FRONTIERS IN GENETICS

【Abstract】Currently, most of the personal health data (PHD) are managed and stored separately by individual medical institutions. When these data need to be shared, they must be transferred to a trusted management center and approved by data owners through the third-party endorsement technology. Therefore, it is difficult for personal health data to be shared and circulated over multiple medical institutions. On the other hand, the use of directly exchanging and sharing the original data has become inconsistent with the data rapid growth of medical institutions because of the need of massive data transferring across agencies. In order to secure sharing and managing the mass personal health data generated by various medical institutions, a federal personal health data management framework (PHDMF, ) has been developed, which had the following advantages: 1) the blockchain technology was used to establish a data consortium over multiple medical institutions, which could provide a flexible and scalable technical solution for member extension and solve the problem of third-party endorsement during data sharing; 2) using data distributed storage technology, personal health data could be majorly stored in their original medical institutions, and the massive data transferring process was of no further use, which could match up with the data rapid growth of these institutions; 3) the distributed ledger technology was utilized to record the hash value of data, given the anti-tampering feature of the technology, malicious modification of data could be identified by comparing the hash value; 4) the smart contract technology was introduced to manage users' access and operation of data, which made the data transaction process traceable and solved the problem of data provenance; and 5) a trusted computing environment was provided for meta-analysis with statistic information instead of original data, the trusted computing environment could be further applied to more health data, such as genome sequencing data, protein expression data, and metabolic profile data through combining the federated learning and blockchain technology. In summary, the framework provides a convenient, secure, and trusted environment for health data supervision and circulation, which facilitate the consortium establish over medical institutions and help achieve the value of data sharing and mining.

【Keywords】personal health data; blockchain; smart contract; data provenance; data sharing

【标题】PHDMF：基于区块链技术的灵活可扩展的个人健康数据管理框架

【摘要】目前，大部分个人健康数据（PHD）由个体医疗机构单独管理和存储。当这些数据需要共享时，必须通过第三方背书技术传输到可信的管理中心，并由数据所有者批准。因此，个人健康数据难以在多个医疗机构之间共享和流通。另一方面，由于需要跨机构进行海量数据传输，直接交换和共享原始数据的方式与医疗机构数据快速增长的趋势不符。为了安全共享和管理各医疗机构产生的海量个人健康数据，开发了联邦个人健康数据管理框架（PHDMF，），具有以下优点：1）利用区块链技术建立数据多个医疗机构的联合体，可以为会员扩展提供灵活、可扩展的技术方案，解决数据共享过程中的第三方背书问题； 2）采用数据分布式存储技术，个人健康数据可以主要存储在原医疗机构中，海量数据传输过程无用武之地，与这些机构数据快速增长相匹配； 3）利用分布式账本技术记录数据的哈希值，鉴于该技术的防篡改特性，可以通过比较哈希值来识别对数据的恶意修改； 4）引入智能合约技术管理用户对数据的访问和操作，使数据交易过程可追溯，解决了数据来源问题； 5) 以统计信息代替原始数据，为元分析提供可信计算环境，通过结合，可信计算环境可以进一步应用于更多的健康数据，如基因组测序数据、蛋白质表达数据、代谢谱数据等通过结合联邦学习和区块链技术。综上所述，该框架为健康数据的监管和流通提供了一个便捷、安全、可信的环境，有利于联合体组建医疗机构，实现数据共享和挖掘的价值。

【关键词】个人健康数据；区块链；智能合约；数据来源；数据共享

【发表时间】2022

【收录时间】2022-07-06

【文献类型】Article

【论文大主题】区块链联邦学习

【论文小主题】两者结合

【影响因子】4.772

【翻译者】石东瑛