Connelly Donald P, Kijsanayotin Boonchai, Speedie Stuart M
Health Informatics,University of Minnesota, Minnesota, USA.
AMIA Annu Symp Proc. 2007 Oct 11:917.
Record linkage algorithm development is a challenging task requiring large databases representative of a target population. As part of an evaluation of a health information exchange project we had the need for a linking algorithm that would not pass protected health information among organizations. We synthesized a test bed for this purpose employing a randomization process and incorporating conditions that represent the study population and possible error generating conditions in real world electronic record systems.
记录链接算法的开发是一项具有挑战性的任务,需要代表目标人群的大型数据库。作为对健康信息交换项目评估的一部分,我们需要一种不会在组织之间传递受保护健康信息的链接算法。为此,我们通过随机化过程合成了一个测试平台,并纳入了代表研究人群的条件以及现实世界电子记录系统中可能产生错误的条件。