树状扫描统计量在药品安全性监测中的应用  

The Application of Tree-based Scan Statistic in Drug Safety Surveillance

作　　者：李海龙 赵厚宇 詹思延[1] 

机构地区：[1]北京大学医学部公共卫生学院,北京100191

出　　处：《药物流行病学杂志》2018年第3期209-213,共5页Chinese Journal of Pharmacoepidemiology

基　　金：国家自然科学基金重大研究计划培育项目(编号:91646107)

摘　　要：本文介绍药品上市后安全性监测中一种新的不良反应信号挖掘方法——树状扫描统计量方法(treebased scan statistic,TreeScan)。TreeScan方法是基于纵向的电子医疗记录数据库来实现的,通过将诊断映射到树状结构中,计算树中各个部分的药物-事件对的相对风险,并计算对数似然比检验统计量,利用蒙特卡洛模拟得出P值。TreeScan方法是药物安全监测中有效的数据挖掘工具,其发现的统计显著信号可以作为药物流行病学研究的候选药物-事件对,进一步细致地分析验证其因果关系。本文首先简要概括TreeScan方法的基本原理,然后结合实例对其应用过程进行阐述。To introduce a novel signal detection method for adverse drug reaction,tree-based scan statistics（ TreeScan）,in post-marketing drug safety surveillance. First,we give a brief overview of the fundamental principle of the TreeScan method. Then we describe the detail of the method through its application examples,especially in drug safety data mining. The method is based on longitudinal data,to map a tree according to the basis of the hierarchical structure of events（ coding diagnosis）,calculate the relative risk of each part of the tree. Monte-Carlo method is used to generate the adjusted P values for multiple testing problem. The TreeScan method can be successfully applied as a data mining tool to generate candidate drug-event pairs in drug safety surveillance using observational data.

关 键 词：扫描统计量 药物流行病学 药品安全 主动监测 

分 类 号：R181.35[医药卫生—流行病学]