To the Editor Wallach et al1 must be commended for their demanding Original Investigation, a meta-epidemiological survey, examining data sets of trials with subgroup claims. Their conclusions that “…subgroup analyses are often spurious and should be considered hypothesis generating”1(p559) and “…when subgroup corroborations are attempted, the initially observed subgroup differences are not demonstrated again”1(p559) deserve the strongest support and also repentance.