We reviewed our original submission at the proof stage and noted two errors. These errors impact on the discussion about the original review as both Hengartner and Davies and Read mention them in their responses to our commentary:
In the section on study quality, originally we stated: “Thus, in the trial by Montgomery et al. (Montgomery et al., 2005), DESS score was higher during placebo treatment than active treatment”, when pointing out the possibility of a nocebo effect.
This should be corrected to “Thus, in the trial by Montgomery et al. (2004), DESS score was numerically higher during active treatment than placebo” as we realise we cited the wrong Montgomery study, and the sentence did not make logical sense because of this citation error. The correct citation is: Montgomery SA, Kennedy SH, Burrows GD, Lejoyeux M, Hindmarch I. Absence of discontinuation symptoms with agomelatine and occurrence of discontinuation symptoms with paroxetine: a randomized, double-blind, placebo-controlled discontinuation study. International clinical psychopharmacology. 2004;19(5):271–80. We fully acknowledge this error, and the (correct) inference made by Davies and Read and Hengartner when commenting on this specific point of our commentary. We apologise for any confusion caused by this genuine error.
We would also like to clarify the section in Selective reporting where we originally stated “As noted above, the authors do not mention withdrawal symptoms occurring in people receiving placebo, as opposed to active drug (Oehrberg et al, 1995), which invalidates statements about the frequency of withdrawal due specifically to anti-depressants. ” This should read: “As noted above, the authors do not mention withdrawal symptoms occurring in people receiving placebo, as opposed to those ceasing active drug (Oehrberg et al 1995), which invalidates statements about the frequency of withdrawal due specifically to antidepressants.” We believe this clarifies our argument, but does not alter the meaning. In the study by Oehrberg and colleagues, people received placebo then drug or placebo for 12 weeks then placebo for three weeks. They then reported the placebo-placebo group to have withdrawal symptoms. We added this point to make it clear to the reader that those people continuing to receive placebo developed withdrawal symptoms.