Comment from PubMed Commons 31 May 2017 Doug Berger | May 29 2017 23:54 ESTMethodologic problems in this meta-analysis:(1) Trials with subjective endpoints (like depression) and the inability to maintain double-blinding should be…