Authors and publication practices.

Although authors are usually considered to be the main perpetrators of research and publication misconduct, any person involved in the process has the potential to offend. Editors may breach ethical standards particularly with respect to conflicts of interest. In the same way that authors are now required to declare competing interests, notably commercial affiliations, financial interests and personal connections, so must editors. Editors can influence the chances of acceptance or rejection of a paper by reviewer selection. Reviewers should also be ready to disclose conflicts of interest. They must ensure that their reviews are evidence based and free from destructive criticism driven by self interest. It seems likely that ultimately we will progressively move towards 'open' peer review in which both the authors and the reviewers are known to each other. There is an urgent need for increased transparency of the relationship between editors and owners. The events of the last few years indicate that unless this interface is fully understood by all parties, conflicts may arise. There is also a need for a radical overhaul in the relationship between journals, journal editors and the biomedical industry. It is now increasingly accepted that all clinical trials should be registered in a centrally held database and that protocols should include the primary and secondary outcome measures and the intended approach to data analysis thereby avoiding opportunistic post hoc analyses. However, the even more radical proposal that journals should cease to publish clinical trials sponsored by industry deserves wider debate.