广东省药品不良反应监测数据库的信号检测

Signal Detection of Adverse Drug Reaction Monitoring Database in Guangdong Province

对广东省自发呈报的药品不良反应系统数据库进行信号检测。方法应用贝叶斯判别可信区间递进神经网络模型比值失衡测量法,对广东省2002年4月28日到2007年7月10日上报的所有数据,进行信号检测。结果信号检测方法能发现9 144条药品/药类-不良事件/不良事件类组合,其中2 529条有信号,56条是强信号。结论信号检测有助于发现新的药品不良反应信号,及突出的药品安全性问题,可以极大地提高药品不良反应监测工作的效率,是将来药物警戒的研究重点之一。

To use the method of signal detection to evaluate the spontaneous reporting database of adverse drug reactions （ADRs） held by the Guangdong Monitoring Cemre. Bayesian Confidence Propagation Neural Network, an internationally recognized measure of disproportionality for quantitative signal detection was introduced to test the adverse drug reaction reports collected from April 28, 2002 to July 10, 2007 in Guangdong. The method of signal detection was capable of finding out 9144 drug-ADR combinations from the database. 2529 of them were suspected of drug-ADR signals and 56 were potentially serious signals. The signal detection can greatly enhance the efficiency of adverse drug reaction monitoring and mine new drug-ADR signals and outstanding safety. It will play a greater role in pharmacovigilance in the future.