If N is the size of the exposed population, i.e. the number of subjects treated or having been treated with the considered drug during the surveillance period, and p is the reference risk of a given event, i.e. the risk in this population if not exposed, then the number of fortuitous (i.e. non-causal) occurrences of this event expected during the period is N • p.
If RR is the relative risk associated with drug exposure and U the under-reporting coefficient, then the expected number of reports is (Tubert-Bitter et al., 1992):
Was this article helpful?