Abstract

Objectives. To document the prevalence of ethical requirements in the instructions for authors of journals that published randomized clinical trials in 2005. Design. Using a validated computerized search strategy for retrieving abstracts of RCTs, we retrieved 13 184 references from 2056 journals. These journals were divided into journals that had published > 30 RCTs in 2005, and those that had published fewer. We included all the former and a random sample of the latter journals in the analysis. Measurements. Coders retrieved the instructions for authors for each journal. The coders identified whether 15 ethical requirements were present in each of the instructions, calculated the age of the journal, and determined where it was published. Results. Only 51% of journals required that authors document review by an ethics committee, only 55% required declaration of conflicts of interest, and only 33% required documentation of informed consent. Overall, 48% of instructions required three or fewer of 15 ethics items. Journals publishing many RCTs each year had more ethical requirements, older journals had more requirements, as did journals published in the US. Journals published in Europe had fewer. However, in multiple regression analyses involving the journal's age, location of publication, and number of RCTs published, only publishing many RCTs was associated with having more ethical requirements. Conclusions. Many journals do not require important elements of research ethics in their instructions for authors. Ethical requirements are less common in the many journals that publish few clinical trials each year, journals that have often been overlooked in prior studies of journal ethics. Keywords: Randomized clinical trials, ethical instructions, publication ethics, journal requirements