Statistical assessment of the association between vaccination and rare adverse events post-licensure

Once a new vaccine has been licensed and is in widespread use, it is important to be able to detect and investigate adverse events which may be associated with the vaccine. This is particularly important for rare events which may have been missed in pre-licensure studies. Recent studies investigating adverse events following vaccination have used either a population or a sample of the population (cohort studies), cases matched to controls (case-control studies), or just information on the cases themselves (self-controlled case-series studies). In this paper, the three methods are reviewed and compared and examples are given of recent applications. The effects of potential sources of bias are demonstrated using a data set on MMR vaccine and convulsions.