数据挖掘在药物警戒中的应用

信号的产生与分析是药品不良反应监测的关键环节。随着自发呈报系统数据库中数据的逐步增多,现有的常规分析方法已暴露出其局限性。因此,引入新的数据挖掘技术对药品不良反应信号的生成十分必要。目前各国普遍使用的数据挖掘方法主要有频数法与贝叶斯法。通过介绍两种方法的原理、产生信号的标准、在各种组织中的应用现状及其相互比较,以期为药品不良反应监测提供技术支持。

Application of Data Mining Algorithms in Pharmacovigilance

The timely detection and analysis of signals is the key point in adverse drug reactions(ADR) monitoring. Conventional methods have their disadvantages for large scale database,so it is necessary to optimize signal detection. Data mining is a technique which may be used to detect signals more ef-ficiently. Frequency method and Bayesian method are the two main data mining algorithms used in pharmacovigilance nowadays. By explaining how the currently used algorithms work,discussing about the threshold for generating a signal and offering a comparison of the different methodologies,provide technique support for ADR monitoring.