Actually the reverse, we should be skeptical of subgroup analyses, because they are often fishing for results, have issues with multiple comparisons, and are typically underpowered.