I am pleased to be able to respond to the letter from Drs Owens & House with which I am in partial agreement. I am happy to confirm that the cited study (Thompson et al, 2000) was designed by me, carried out by Eli Lilly in a UK primary care context, and was analysed and written up by my colleagues and I. These facts are acknowledged in the primary research publication in the American Journal of Psychiatry.

I am also happy to confirm that, along with occasional paid lectures, I currently hold a consultancy with Organon UK and have recently held a similar consultancy with Janssen and with Philips, but not with Eli Lilly. Readers of the Journal might also wish to know that these interests have been declared to the Royal College of Psychiatrists and are known to my employing University.

I completely agree that conflicts of interest should be declared. However, I wonder whether the editors of the New England Journal of Medicine would include in their censorship policy the authors of editorials in which the objective is solely to comment on the methodological adequacy of an original article. If so, that appears to me to be more a prohibition on freedom of speech than anything related to evidence-based medicine, and would have precluded my criticisms of the Barbui & Hotopf article. If that is the objective of Drs Owens & House, then I cannot agree with them on that point.

Finally, I have not been able to find any scientific points in Owens & House's letter to which I can respond and I therefore assume that they are in full agreement with my criticisms. Otherwise, I am sure that they would have presented rational arguments against my analysis instead of taking their argument ad hominem.