| Pharmacovigilance attempts to detect potential correlations between certain drugs and adverse reactions to these drugs through the analysis of spontaneous reporting data. A variety of mathematical methods are used in a signal detection style to search out these greater than expected pairs of drugs and adverse reactions. First, a theoretical examination of these methods as well as their properties and their place in the current pharmacovigilance literature is conducted. Second, the introduction of a method commonly used in categorical data analysis that has not been applied to pharmacovigilance is considered. Then, a data example considering the effects of Vioxx on hypertension will be looked at from the perspective of the method's abilities to flag this known interaction. Finally, a simulation study is conducted to examine the estimators in action. This is followed by a discussion on power and type one error of the different methods. |
