Conflicts of interest can severely affect scientific objectivity, or at least appear to do so. But there are no uniform rules for disclosure and management of conflicts, so authors need to rely on journal guidelines and advice from their institutions and other organizations.
A year or so ago, 62 scientists and physicians wrote a letter to the National Library of Medicine urging that PubMed – the biggest, publicly funded database of academic abstracts in the field of medicine and life science – provide conflict of interest (CoI) disclosures. Citing published work, the correspondents argued that the results of industry-funded research are likely to favour industry interests. Indeed, critics have long suspected CoI, especially in the area of nutrition and health. For instance, science journalist Leonid Schneider highlighted apparent CoI in research about the healthiness of chocolate in a blog post last year. The experts’ letter seems to have had its intended effect, for PubMed has now started to include CoI statements along with the abstracts.
What is conflict of interest?
CoI is defined and understood somewhat differently by different organizations and publishers. The International Committee of Medical Journal Editors (ICMJE), for example, defines CoI as follows: “A conflict of interest exists when professional judgment concerning a primary interest (such as […] the validity of research) may be influenced by a secondary interest (such as financial gain)”. Nature’s definition includes not only potentially real CoIs that can influence the behaviour of scientists or the content of the publication but also the mere perception of there being a CoI. The ICMJE stresses that CoIs can involve authors, reviewers, editors and members of the editorial board of the journal.
Financial CoIs, such as those arising from industry-sponsored research, are probably the most obvious type. Authors (and editors) can encounter financial CoIs when they are or have been employed by an organization that is likely to benefit financially from the research or because of personal financial interests, such as when they hold patents or stocks of a company that can experience financial gain or loss upon publication of the paper.
But there are many other types of CoI. Such non-financial CoIs can include a membership on an advisory board or consulting assignments for an educational company, as exemplified in the guidelines of PLOS journals. Peer reviewers might come across a paper that is poorly written and not appropriate for publication, but use the ideas for their own research. A reviewer might also be friends with the authors of the paper, be a direct competitor (often unavoidable in highly specialised fields), or have political, ideological or religious convictions related to the paper’s topic that hinder an unbiased review. Editors too may encounter instances when they handle papers submitted by contacts from their previous academic work.
How should we deal with conflicts of interest?
The examples discussed above show that the very nature of academic and editorial work sometimes makes CoI unavoidable, as PLOS Medicine editors point out too. Indeed, the editors of this journal consider expertise itself as one kind of CoI.
"As it is hard to judge one’s own bias, authors should leave it to the editors to make informed judgements."
So, what should authors do when they are aware of CoIs in the publication process? An essential step that many editors and leaders of research organizations recommend is disclosure. As it is hard to judge one’s own bias, authors should leave it to the editors to make informed judgements. This recommendation of disclosure seems prudent because CoIs that were undisclosed to begin with but later revealed can destroy a researcher’s reputation and credibility. This is also the case when the CoI has not actually lead to any documented bias or loss of scientific objectivity. Note that many top journals, such as Nature and Science, do not forward CoI disclosures to peer-reviewers, leaving the judgements to better trained editors.
Disclosure rules differ from journal to journal (Table 1). Whereas some journals from smaller publishers sometimes do not have CoI policies, journals from larger publishers usually require disclosure. For example, the publisher Elsevier demands disclosure of all potential CoIs, either with the specification of the conflicts or by declaring that no CoIs exist. Nature adheres to a similar policy but only requires financial CoI statements. By contrast, PLOS journals require explicit disclosure of non-financial CoIs too, but a declaration is not needed if no CoI exists. Undisclosed CoIs can generally result in either a correction or, in severe cases, a retraction and ban from publishing in the respective journal for a period of time.
To assist authors who are unsure about how to disclose CoIs, the ICMJE has developed a CoI disclosure form for authors that can be used for submission to any journal and is encouraged especially for ICMJE member journals. The form includes statements about CoIs, the origin of the study’s funding and the extent to which the authors have access to the study data. Jeffrey Perkel, editor at Nature, advises researchers to check with their institutions when they are unsure about potential CoI. He writes that most research institutions have policies, offer training and have offices that can guide the identification and management of CoIs.
Peer reviewers for Nature are required to exclude themselves when there is significant CoI of any kind. In any case, the peer reviewer must inform the editor who can then take potential CoIs into account when considering the reviewer’s comments on the article.
To sum up, authors should disclose potential CoIs in accordance with journal guidelines, consult their institutions if needed and familiarize themselves with the guidance provided by publishers and other organizations. Transparency is key for peer-reviewers too, and all CoIs should be reported to the editor.