New Policy on Disclosures at Health Services Research

Abstract

The appropriate role of scientific journals is to disseminate the best science available to the public and the research community. Two principal mechanisms to fulfill this role are peer review and open disclosure of potential biases in conducting and reporting results. Peer review primarily focuses on what is presented in the submitted manuscript—reviewers cannot easily assess what has been omitted or whether the authors made choices in the design or execution of the study that may have biased their results. For example, it appears some key adverse data were omitted from the Vioxx Gastrointestinal Outcomes Research (VIGOR) trial reported in the New England Journal of Medicine (Curfman, Morrissey, and Drazen 2005a, b). Some authors claimed to have been unaware of the problems, raising questions about what are the roles and responsibilities of all authors in ensuring “the best science available” is published in scientific journals. Increasingly, journals have adopted more explicit policies about what to disclose and who should be informed during the process of review and publication. There are also studies showing that a project's sponsorship may be related to the findings reported. Most of these have focused on trials reporting use of drugs and devices, comparing studies sponsored by the for-profit manufacturers with those sponsored by not-for-profit and government entities (Lexchin et al. 2003; Ridker and Torres 2006), or on studies relating to tobacco use (Barnes and Bero 1996). Publication of biased studies is an insidious problem because the submitted findings may not be incorrect per se and thus are unlikely to be caught by peer reviewers. Absent or incomplete disclosure, however, prevents the consumer of such published information from applying the appropriate skepticism.