the full context of the quote you used from my comment is that mbct "failed to show a significant benefit vs the control". Your shortened quotation misses that

the full context of the quote you used from my comment is that mbct "failed to show a significant benefit vs the control". Your shortened quotation misses that very important context. The control exists to account for all the other confounding factors, besides mindfulness, involved in mbct. When you control for those, and mindfulness is the only variable that differs between active arm and control arm, there is no significant benefit to mindfulness. So one can reasonably conclude that other things like being in a therapy group, learning about depression, etc were driving the benefit, not mindfulness Regarding the subgroup finding: as far as I know, the study was not designed or powered to rigorously measure the impact of the interventions on this subgroup. If they did another good study looking just at this subgroup and the results were positive, that's a different thing. The first thing investigators do when a study fails the primary endpoint is mine for subsets of the data that show a positive result. That doesn't make the result meaningless, but should invite skepticism Reproducibility is a huge problem in scientific research. These are not determinate systems, and a huge number of things can bias a result, intentionally or otherwise. Scientific publications, even in good journals, need to be viewed with an appropriate amount of skepticism According to NICE, one of the leading agencies on cost effectiveness and clinical guidelines, MBCT is only recommended for prevention of relapse in patients who have had three or more episodes of depression. CBT and augmenting antidepressants work just as well. Funny thing about depression studies is most interventions, pill or psychotherapy (or even placebo sometimes) tend to have similar levels of effect