一种以电子病历为核心的多数据库数据集成方法

A Design Approach for Data Integration Based on Electronic Medical Records with Multi-database

介绍了一种以电子病历(EMR)为数据核心,并融合了个性化治疗资源库的数据集成方法。该方法整合了多数据格式的临床数据集,如电子病历、抗生素知识库、随访数据库、肿瘤患者标本库和基因治疗数据库。数据集成目的是用于归纳与分类疾病的诊断,分析与对比治疗前后的临床效果,并预测和挖掘疾病治疗的路径。分析了两种常用数据格式的数据抽取与集成性能,提出了以下要求:数据集成方法要满足对多数据格式抽取的要求,以适应不同医疗数据资源的整合;数据集成是在多数据源的大数据环境下工作,所以集成方法须对数据的抽取速度性能作压力测试;数据集成是数据读入并写入目标数据库的过程,因此集成方法中要包含能灵活定义、易于调整、隐私数据查询安全的抽取规则组件,以及清晰友好的集成查询界面。最后通过3组实验,说明了基于EMR的集成架构和方法能解决多种数据格式的临床数据集成问题;利用MyBatis组件完成了源数据库表与抽取规则的映射工作,过滤了集成过程中的隐私数据;使用SilverLight组件的WEB呈现技术,给用户提供了友好便捷的数据查询平台。目前该数据集成方式已应用在国家重大专项—肝肿瘤样本库专题下7个分中心的临床数据集成工作。

This paper presents a design approach for data integration based on electronic medical records and personalized treatment database. The data integrating approach combines multi-application clinical data sets, such as electronic medical records, knowledge base of antibiotics, follow-up database, cancer specimens library and gene therapy database. Data integration is intended for induction and classification of disease diagnosis, analysis and comparison of the clinical effects before and after treatlnent, and to predict and mine the clinical paths. This paper analyzes the performance of data extraction with two kinds of data formats, and proposes：data integration means extraction with multiple data formats in order to adapt to different medical data resource integration. The process of data integration is in a multi-data source environlnent commonly, so we need test extraction speed performance with big data. Data integration is the process of reading in and writing to the target database, a fle：dble integrated architecture is necessary. In this architecture some components are available, such as data extracting definition, privacy rules adjustment and friendly query interface. Finally, the paper shows that the approach solves the multi-application of the clinical data integration issues through three set of experiments. MyBatis cotnponents can complete the source data table mapping with extraction rules, and filter the privacy data in the process. The Silverlight component based on web display technology provides users with a friendly and convenient query platform. At present, the data integration approach has been applied at seven sub-center data integration which is the major projects for national science and technology foundation of China.