Mandatory disclosure of financial interests of journals and editors

Transparency of editor’s financial interests

The ICMJE—whose current 12 member journals include influential general and internal medicine journals such as New England Journal of Medicine, Lancet, JAMA, The BMJ, and Annals of Internal Medicine—produces guidance on the conduct, reporting, editing, and publication of scholarly work.3 Its recommendations on author declarations are followed by thousands of medical journals and those in other research fields.3 However, medical journals do not perform well in following the ICMJE standards for disclosing editors’ COIs.

In a cross-sectional analysis of a random sample of 350 medical journals listed as following ICMJE recommendations in 2018, less than 1% of the journals disclosed editors’ individual COIs, whereas 82% required the disclosure of authors’ COIs.6 Another study of websites of 130 top specialty medical journals in 2018 showed that only 45% endorsed the ICMJE recommendations and 12% disclosed individual editors’ COIs whereas 99% required authors to disclose their COIs (table 1).7 In this study, 69% of the journals were linked to a professional organisation (society, association, or college) or a public institution7; the professional community and the public expect such institutions to be transparent about their work, including journal publication.

Among the ICMJE member journals, only The BMJ discloses individual editors’ COIs on the journal website, though the Annals of Internal Medicine discloses editorial COIs for certain types of online articles.8 Most do not seem to follow their own recommendations for declaring editors’ COIs (table 1).8 This lack of transparency from the journals that created and regularly update the ICMJE recommendations is worrying, especially in view of the recent statements from the New England Journal of Medicine that their editors should have no “financial relationships with any biomedical companies” and from the Lancet that the journal complies “with additional company [Elsevier] COI policies and procedures.”9 It is not clear why these large and influential medical journals do not publicly disclose whether their editors should have no financial COIs or, if this is not a condition for the editorial team, why they do not publish the details of any such COIs.

The World Health Organization has a critical role in clinical trials transparency through the International Clinical Trials Platform and by sponsoring trial registries in developing countries, yet its flagship journal (Bulletin of the World Health Organization) does not disclose its editors’ COIs. This is despite it being an ICMJE member.8 Similarly, prominent journals linked to public health organisations like Eurosurveillance from the European Centre for Disease Prevention and Control (ECDC), and Emerging Infectious Diseases and Morbidity Mortality Weekly Report from the US Centers for Disease Control and Prevention, do not disclose their editors’ COIs.7 Whereas payments to some journal editors have been reported in the US because of the existence of a national database on industry payments to physicians,1011 there are few reports on industry payments to healthcare professionals in other countries12131415 and none regarding payments to journal editors. The lack of transparency for editors’ COIs does not seem to diminish journals’ prestige, as authors keep submitting manuscripts to journals that do not follow disclosure standards. This is troubling.

Others involved in editorial process

ICMJE recommendations also require peer reviewers to disclose COIs to editors and recuse themselves from the review process if needed.3 However, only three of the 12 ICMJE member journals had a publicly available policy on reviewers’ disclosure of COIs.8 Since reviewers’ identities are often kept confidential, the ICMJE recommendations do not require public disclosure of reviewers’ COIs.3 However, a small but increasing number of journals are publishing both the names of the reviewers and their comments together with the article. In this open peer review model16 journals should disclose reviewers’ COIs. Some journals already do this—for example, all the BMC medical journals17—showing that disclosure is feasible.18

Transparency of journals’ financial interests

Scientific journals, as business enterprises, also have financial interests such as revenues from article reprints sales, advertisements, and commercial sponsorships. Again, these are rarely disclosed. There are no policies for this type of declaration, as judged by the publishing ethics polices of major medical publishers.192021

In 2016, more than $630m (£470m; €530m) was spent on print advertisements in medical journals.22 Drug advertisements in medical journals have been a concern for some time, since they may appear to influence editorial decisions.23 Income from reprints, particularly of industry funded trials, is considered to be a financial conflict of interest.242526 In two studies of journal revenues, three journals owned by US professional organisations (New England Journal of Medicine, JAMA, and Annals of Internal Medicine) refused to provide figures of their reprints orders, in contrast to The BMJ (also owned by a professional organisation) and the commercially owned Lancet.2425 These studies were conducted almost 10 years ago, but we were still unable to find this information on these journals’ websites, even though societal pressure for transparency has significantly increased.

Journals that generate considerable revenues from reprint sales are usually influential journals.2425 Furthermore, higher numbers of reprints may result in more citations, which in turn boosts the journal’s impact factor and prompts trial sponsors to request more reprints of newly published articles. Industry funded trials published in highly influential journals are more often cited than trials with non-industry funding.2527 The use of reprints as a marketing tool likely affected this finding, although the fact that industry trials more often report positive outcomes28 could also explain why they are more cited.2930

It is very unusual for journals to report their annual revenues, but BMJ started to report this information for its journals in 2017.31 Even if it is understandable that commercial journal owners might want to keep their revenues confidential, it is difficult to justify the lack of transparency about revenues by journals belonging to professional or public organisations.

Opening up

Medical journals should start disclosing COIs for individual editors, and commercial revenues from commercial sources should be disclosed each year. Journal editors should also be responsible for ensuring that everyone involved in the editorial decisions publicly discloses their COIs.

Although there are proposals to establish nationally funded open access registries to allow complete disclosure of COIs,32 we do not believe that registries alone can provide transparency in research publishing. Disclosure of individual editors’ COIs belongs on journals’ websites or, in mega-journals that have many editors, in the articles, as these are where most readers would most likely look for this type of information.

We believe that authors’ and readers’ trust in journals will increase if journals show themselves to be as transparent about editorial and journal COIs as they require their authors to be. We call for all medical journals to disclose their editors’ COIs on the journals’ website and keep this information updated. Another possibility is to publish the information on COIs regularly in their journals so that such declarations become a part of the published record and indexed in bibliographical databases, just as it is for the journal authors when they publish their articles.

The approach may need to differ for large editorial teams, such as in mega-journals (eg, BMJ Open, Plos One, Scientific Reports) that have hundreds of editors. The name of the editor and their COIs could be reported at the place in the article where authors’ COIs are disclosed. Another option is to link the editor’s name in the published article to the journal site where all editors’ COIs are reported and updated.

We suggest that journals belonging to the same organisation—whether a professional organisation (eg, American Heart Association, British Medical Association) or a group of related journals (eg, Lancet, Nature Reviews)—should have a uniform approach to disclosure of editors’ COIs. Currently, the few publishers that disclose editors’ COIs, implement the policy inconsistently. For example, within BMJ, the flagship journal (The BMJ) discloses the COIs of all editorial team members, but other BMJ journals disclose conflicts for one or a few members of their editorial teams.7 Consistency across journals in the same group would increase the trust of the professional and general public in the transparency and consistency of publishers’ policies and work.

Transparency is not enough

Transparency about potential COIs is a solution to a rather simple problem compared with judging their relevance and how to act on them. Firstly, there is no consensus on when a financial conflict of interest is serious enough to be disclosed: how high should the financial stakes be, how close to the topic of the article, how distant in the past, or how near in terms of the relationship (spouse, family member, close friend, institutional colleague, collaborator in earlier projects, etc). Secondly, it is also difficult to define and measure non-financial conflicts. Thirdly, decisions need to be made about when a conflict of interest is so large that being transparent is insufficient and any involvement should be avoided. For example, if a behavioural therapy is designed by a psychotherapist who also wrote a handbook on it, trains new therapists to apply it in practice, and has a substantial income from these, we believe that this person should not take part in efficacy studies of that behavioural therapy, or reviewing such studies, or making editorial decisions on them. Finally, for conflicts where transparency is enough there remains the almost unanswerable question of how to weigh them: should we downgrade only the level of evidence or also the strength of the reported effects?