On the assessment of adverse drug reactions from spontaneousreporting systems: the influence of under-reporting on oddsratios

A well-known problem in spontaneous reporting systems (SRSs) for adverse drug reactions (ADRs)is under-reporting, that is, the problem that not all occurrences of ADRs are reported to the SRS. Welook at the question of how to draw statistical conclusions from analyses of SRS data using reportingodds ratios. We will show that certain under-reporting problems play no role in assessing ADRs fromSRSs: the results from the analyses turn out to be biased by some specic under-reporting problems,but not by others. SRS data can be particularly useful for the assessment of drug–drug interactions. Ifthe assumption holds that there is an under-reporting problem for a rst drug, and an under-reportingproblem for a second drug, but that these two under-reporting problems do not inuence each other,then reporting odds ratios estimated from SRSs are useful for signalling drug–drug interactions in theADR-experiencing population. Similar results hold for covariate–drug interactions. We illustrate ourresults using two examples. Copyright?2002 John Wiley & Sons, Ltd.