PDF《基于 openEHR 的原型关系映射方法》

33 卷4期2014 年8月中国生物医学工程学报Chinese Journal of Biomedical Engineering Vol.

33 No.

4 August

2014 doi: 10.

3969 /j. issn. 0258- 8021. 2014. 04.

07 收稿日期: 2014- 03- 25, 录用日期: 2014- 05-

27 基金项目: 国家高技术研究发展计划(

863 计划) ( 2012AA02A601) ;

国家科技重大专项( 2013ZX03005012) * 通信作者( Corresponding author) , E- mail: lxdlxd7366@ gmail. com 基于 openEHR 的原型关系映射方法 王利闵令通 吕旭东 * 段会龙 ( 浙江大学生物医学工程与仪器科学学院,杭州 310027) 摘要: 医疗信息的复杂性和动态性给医疗信息系统带来

1 巨大挑战.openEHR 规范的两层建模思想可以提高医 疗信息系统的灵活性, 适应医疗信息需求的变化.目前虽然已有采用 openEHR 官方提出的属性 + 路径( node + path) 和openEHR 参考模型对象关系映射的 openEHR 数据存储实现方案, 但其性能无法达到临床实际应用的需要. 本研究提出一种基于 openEHR 原型和关系数据库的自动映射方法, 将原型映射为关系数据表、 原型属性映射为关 系数据表字段, 在此基础上设计了原型驱动的数据持久化和数据访问平台.基于原型关系映射方法, 对原型查询 语言进行扩展, 添加 INSERT、 UPDATE、 DELETE 等语法, 使原型查询语言支持完整的数据操作功能.结果表明, 原 型关系映射方法与属性 + 路径和 openEHR 参考模型对象关系映射方法相比, 在1231 条数据中查询

3 条数据时快

69 倍和

8 倍, 在128

141 条数据中查询

1 593 条数据时快

2 倍和

17 倍.原型关系映射和原型查询语言等技术不但 可以发挥两层建模方法在医疗信息表达、 存储和访问等方面的灵活性优势, 而且能大幅提高数据访问的性能. 关键词: 原型关系映射;

原型查询语言;

原型;

关系数据库;

openEHR 中图分类号 R318 文献标志码 A 文章编号 0258- 8021( 2014) 04- 0432-

06 Archetype Relational Mapping Based on openEHR WANG Li MIN Ling- Tong LU Xu- Dong* DUAN Hui- Long ( College of Biomedical Engineering and Instrument Science,Zhejiang University,Hangzhou, 310027,China) Abstract: Complexity and continuous evolution of healthcare information is the biggest challenge for healthcare information systems. Two- level information model from openEHR can facilitate healthcare information systems to adapt to the change of healthcare information requirements. Persistence methods such as Node + Path promoted by openEHR and openEHR reference model object relational mapping ( RMORM) cannot meet the performance requirements of clinical daily use. In this paper,we proposed a novel persistence method based on direct mapping between openEHR archetype and relational database by mapping archetype to data table and mapping archetype attribute to data table column,and designed an archetype- driven data persistence and data access platform. Based on the archetype relational mapping ( ARM) method,we extended the archetype query language ( AQL) by adding INSERT,UPDATE,and DETELE clauses to support full data manipulation. Results showed that,ARM is

69 times faster than Node + Path and

8 times faster than RMORM when querying

3 results in

1 231 records,and ARM is

2 times faster than Node + Path and

17 times faster than RMORM when querying

1 593 results in

128 141 records. With technics such as ARM and AQL, the data platform can achieve high performance while keeping the advantage of openEHR two- level modeling in healthcare information expression,storage,and access. Key words: archetype relational mapping;