药品不良反应多维关联规则挖掘及预警模型构建  

Excavation of Multidimensional Association Rules in Adverse Drug Reaction and Construction of Warning Model

作　　者：冯变玲[1] 杨世民[1] 贺小红[2] 雷祎[1] 

机构地区：[1]西安交通大学医学院药学系,西安710061 [2]北京工业大学经济与管理学院

出　　处：《中国药事》2012年第10期1076-1082,共7页Chinese Pharmaceutical Affairs

基　　金：国家食品药品监督管理局政策研究课题资助项目(编号2011-ZXKT-04)

摘　　要：目的找寻不良反应、药品、用药人群间的关联关系,构建关联模型,为临床安全合理用药,有效降低由药源性疾病带来的额外经济负担,提高临床治疗效果提供必要的保障。方法应用SQL构建多维数据仓库,运用数据挖掘技术中的关联规则挖掘数据中潜在的有价值的信息。结果与结论得到江苏省2004～2009年6年多维关联结果以及6年纵向多维关联规则比较结果,共获得7组具有强的关联性的三维关联组合和5组具有一定关联性的三维关联组合。Objective To find the association relationship among drug,drug users and adverse drug reaction（ADR）,so as to build the ADR warning model,provide necessary guarantee for the drug safety and reduce the additional financial burden arising from drug-induced diseases and improve therapeutic effect.Methods ADR multidimensional data-base was built by SQL Server 2005 Analysis Service（SSAS）,the data mining techniques of association rules was adopted to find the potential relationship among drug,ADR and drug users.Results and Conclusion Multidimensional association results and the comparison results of the longitudinal multidimensional association of Jiangsu Province between 2004 and 2009 were achieved,seven combinations of three-dimensional association were found with a strong correlation and five with a certain relevance.

关 键 词：多维关联规则 药品不良反应 预警模型 

分 类 号：R994.11[医药卫生—毒理学] TP311[医药卫生—药学]