基于随机增量抽样的ADR信号监测与预警研究  

Signal Detection and Early Warning Research of Random Incremental Sampling Adverse Drug Reactions

作　　者：孙越泓[1] 魏建香[2] 支云龙 孙骏[3] 徐厚明[3] 李明[3] 

机构地区：[1]南京师范大学数学科学学院,南京210023 [2]南京邮电大学 [3]江苏省药品不良反应监测中心

出　　处：《药物流行病学杂志》2014年第5期298-302,共5页Chinese Journal of Pharmacoepidemiology

基　　金：教育部人文社会科学研究青年基金(编号:12YJCZH179);江苏省"六大人才高峰"2009年度D类资助项目

摘　　要：随着药品不良反应监测数据的日益增多,利用总体样本数据来进行信号检测的工作量大且及时性差。基于江苏省药品不良反应中心提供的两年的10万余条监测数据,以增量抽样的样本作为对时间序列的模拟,采用IC(Information Component)方法研究抽样样本量对信号检出率的影响。仿真实验结果表明:利用随机增量抽样法能有效地减少信号检出时所需的样本量,提高了信号检出的效率和及时性,并具有一定的药物风险预警能力。With the increasing of adverse drug reaction monitoring data, it was tedious and timely to use the total data for signal detection. Based on more than 100000 monitoring data provided by the adverse drug reaction monitor center of Jiangsu province in two years, an increment sampling method was used to achieve the simulation of time series data. The IC （ Information Component） method was adopted to analyze the detection rate of signal varying with sampling size. The simulation experiment results showed that the method proposed could reduce the required sample size, improve the efficiency and timeliness of the signal detection, and had a certain drug risk early warning capability.

关 键 词：药品不良反应 IC方法 增量抽样 信号检测 风险预警 

分 类 号：R181.35[医药卫生—流行病学]