Should methods of correction for multiple comparisons be applied in pharmacovigilance?