As is well known,it is impossible to allocate the drug users in a randomized way under the post-marketing surveillance.Therefore,it is hard to control the confounding factor in signal detection of adverse drug reaction(ADR) databases.It is important to control the potential confounding effects during the analysis of ADR databases and detect signals in a more accurate way.This article briefly used the method of stratification analysis in controlling confounding factors.We analyzed the spontaneous reporting database of shanghai in 2009 with some methods.We compared the difference between stratification analyze and non-stratification analysis.We expect that this article can provide some effective information for the post-marketing surveillance.