As discussed above, it is highly probable that the number of reports involving Drug 1 (k1) and Drug 2 (k2), respectively, are only a part of the numbers a1 and a2 of cases which have occurred. The issue is that the magnitude of this under-reporting may differ across the two drugs compared (Begaud et al., 1991). If the respective values of underreporting coefficients, U1 and U2, were assessable, one could write:
Therefore, the relative risk for Drug 1 compared with Drug 2 would be:
Was this article helpful?