联合用药不良反应信号检测的两种基线模型及应用

Application of Two Baseline Models in Detecting Signals of Drug-Drug Interactions

目的构建两种基线模型——乘法模型和加法模型,结合药物不良反应自发呈报系统实际数据,快速有效地检测联合用药交互作用信号。方法利用上海市药品不良反应监测中心自发呈报的2007～2008年数据,分别采用两种基线模型进行药物交互作用的信号筛选,并经过统计学检验,确定有统计学关联的信号。结果加法模型初步筛选出可疑交互作用信号210例,经过统计学检验,产生统计学关联信号30例;乘法模型初步筛选信号151例,有统计学关联的信号81例。结论自发呈报系统是监测联合用药不良反应极其重要的数据来源。加法模型初步筛选交互作用信号具有较高的敏感度,但有统计学关联的可疑信号相对较少;乘法模型进一步验证了加法模型检测信号的强度。

Objective Two baseline models were developed to detect signals in drug-drug interactions（DDI） effectively with the spontaneous reporting database. Methods Adverse drug reaction reports submitted to Adverse Drug Reaction Monitoring Centre of Shanghai from January 1,2007 to December 31,2008 were used as a material in our study. The multiplieative model and the additive model were employed to generate signals of DDI. Furthermore, we documented the statistical correlation signals according to statistical test. Results The additive model generated 210 suspected signals, including 30 statistical correlation signals. While only 151 signals were detected by the multiplicative model, which included 81 statistical correlation signals. Conclusion The spontaneous reporting database is a foremost resource for detecting signals of DDI. The additive model is more sensitive than the multiplieative model in initial screening, but less statistical correlation signals. The additive model may further validate the strength of the signal detected by the multiplicative model.