中国药物警戒 ›› 2014, Vol. 11 ›› Issue (8): 492-494.

• 综述 • 上一篇    下一篇

错误率控制的药品不良反应信号检测方法

郭晓晶, 叶小飞, 张筱, 许金芳, 贺佳*   

  1. 第二军医大学卫生统计学教研室,上海200433
  • 收稿日期:2016-02-03 修回日期:2016-02-03 出版日期:2014-08-08 发布日期:2016-03-02
  • 通讯作者: 贺佳,女,教授·博导,药物流行病学及新药评价的统计分析。E-mail：hejia63@yeah.net
  • 作者简介:郭晓晶,女,讲师,药物流行病学。
  • 基金资助:
    基金项目：国家自然科学基金资助项目（No; 81373105, No.81202285, No.81072388）,2010年上海市领军人才计划（022）,上海市自然; 科学基金项目（12ZR1453700）,上海市循证公共卫生重点学科（12GWZX0602）

Signal Detection Methods Based on False Discovery Rate of Adverse Drug Reaction

GUO Xiao-jing, YE Xiao-fei ,ZHANG Xiao ,XU Jin-fang, HE Jia   

  1. Department of Health Statistics, Second Military Medical University, Shanghai 200433, China
  • Received:2016-02-03 Revised:2016-02-03 Online:2014-08-08 Published:2016-03-02

PDF (PC)

266

摘要： 常用的药品不良反应信号检测方法均是基于不相称测定理论,这些方法在信号检测过程中发挥了重要的作用。然而目前的信号检测过程存在一些缺陷,如多重性问题等,导致信号检测的准确性降低。通过介绍多重性问题对信号检测准确性的影响,和目前解决该问题的常用方法,如阳性结果错误率（false discovery rate, FDR）方法等。

关键词: 多重假设检验, 信号检测, 药品不良反应, 阳性结果错误率

Abstract: Conventional signal detection methods based on disproportionality analysis are playing an important role in the process of signal detection of adverse drug reaction. However, these methods have their own drawbacks and the detection results are influenced greatly by the multiple hypothesis testing. These problems should be solved through new analysis strategy. In this article, we present the impact of multiple hypothesis testing on the signal detection and introduce the false discovery rate (FDR) methods that should solve the problem.

Key words: multiple hypothesis testing, adverse drug reaction, signal detection, false discovery rate

中图分类号: