Although it's not typical, in a scholarly publication such as the BMJ, to add a personal perspective to a commentary, I believe that in this situation my background is relevant to the discussion. My doctoral dissertation, written in 1988, dealt with the topic of publication bias, which was well described in the Ottawa statement appearing in this issue.1 Since then, I've contributed to several studies of factors affecting publication, including an early empirical demonstration of publication bias.2

About six months ago, I moved from a university, where I had spent 15 years, to a position in a large pharmaceutical research and development group. Registration and disclosure of the results of clinical studies have, not surprisingly, been topics of numerous conversations where I work.

At …