Editorial independence

A meta-analysis was published in a journal ahead of print, and then subsequently in print. Several months later, the journal was contacted by a faculty member at a university not connected with the study. The reader outlined three general concerns with the meta-analysis. The concerns were discussed by the editorial team, including the statistical editor, and it was decided that the overall results of the meta-analysis were not affected. The complainant persisted in the critique, which was relayed to the authors of the original article. The first author provided a detailed response to the issues raised; the journal did not request an erratum at that time.

A discussion followed between the complainant and the editorial office about the extent to which the issues raised were errors or ‘a matter of opinion’. One error was clear and was corrected by an erratum. In the journal’s view, the other issues raised were open to interpretation. The complainant was invited to write a letter to the editor, but they declined and persisted that an erratum should be published or that the journal should consider retracting the article.

Because some of the points of critique were of general interest to the field, and also in an effort to resolve the issue of the continued critique by the complainant, the journal decided to publish an editorial comment authored by the editor in chief and the statistical editor. The editorial comment paraphrased the complainant’s concerns and added a few additional considerations relevant to the interpretation of the meta-analyses. The complainant was given the opportunity to review the editorial comment and the journal also asked his permission to be acknowledged for bringing the issue to the journal’s attention. The complainant gave his feedback, which was incorporated, and also his permission to be acknowledged. The journal also gave the authors of the original article the opportunity to respond to the editorial comment. The authors wrote a scholarly response and also provided re-analyses of the data when excluding the contested studies. The editorial comment was published, together with the acknowledgment, the author’s response and the erratum.

Before the publication of the editorial comment, the journal held a conference call with the complainant. The complainant had published a paper on a related issue in the preceding year, but the study published in the journal was different from the previous meta-analyses from the complainant’s own group. This had not been disclosed to the journal previously. It became also clear that the complainant had not contacted the authors of the article to share his concerns, and the journal encouraged him to do so.

Subsequently, the complainant raised new issues with the original meta-analyses, which they discussed with the authors. The complainant insisted that an additional erratum was needed; the authors of the meta-analysis are now preparing a second erratum. The complainant was again invited to write a letter to the editor, and again declined.

The journal is concerned that the second erratum will not satisfy the complainant and that this issue will not end unless the journal agrees to retract the original article. The assessment of the journal is that there are insufficient grounds to retract this article. One of the complainant’s recent emails to the authors says that they will contact the authors’ university and colleagues and the funding agencies that support the work of this research group about the purported errors made by this team. The complainant has also requested that the journal involves the publisher’s "ethical committee dealing with the publication process".

Question(s) for the COPE Forum

• What is a journal’s responsibility to minimise potential damage that readers can do to the reputation of the author where the reader disagrees with the editorial team and the authors of the original article on whether an error has been made versus a difference in opinion?
• What are the options for a journal to respond to (unreasonable) requests from readers regarding the content published in the journal and/or request to retract a paper if the editorial team considers the concern not sufficiently problematic to result in retraction?
• What could the editorial team have done to better respond to the reader’s concerns?

In 2016, group A published manuscript X in our journal. In early 2017, group B submitted a comment critical of the published manuscript. Following peer review, in accordance with the journal’s then active policy, the comment was rejected from further consideration. The policy allowed for the author of the original article to be one of the peer reviewers of the comment.

The lead author of group A, acting as one of the three referees for the comment, indicated in their confidential comments to the editor that group A would be submitting a correction to address issues arising from the comment. Group A duly submitted and published a correction to manuscript X. Soon after, the journal was contacted by a legal representative of group B to express their concern over the publication of the correction by group A. The representative indicated a concern that the unpublished comment submitted to the journal contributed in part to the submission and publication of the correction. Group B researchers considered that the submission of their comment under the journal’s then active comment/reply policy had allowed the authors of the correction to prepare their manuscript using material that they had been privy to only via their involvement in the peer review of the comment, and that this fact had not been acknowledged in the correction.

Group B requested the journal withdraw the correction and re-open the peer review of the comment. As the journal’s management team considered that the first request would leave an error in the scientific record uncorrected and the second request was unlikely to result in a change of outcome, the journal instead investigated the matter raised by the representative Group B, with the goal of preparing a new correction for publication to take into account the facts of the matter following the investigation.

The investigation identified an error on the part of the administrative team that contributed to this situation; namely, failing to ensure the authors of the correction provided due acknowledgement of the provenance of the correction. As part of the investigation, the journal contacted group A for their input. The authors agreed they should have included an acknowledgement, but not having seen similar acknowledgements on other corrections published by the journal declined to include one in their correction. However, group A also noted that they had exchanged email correspondence with group B, prior to the submission of the comment, about some of the matters subsequently included in the comment. Group A have been at pains to stress that their correction was not primarily prompted by the comment.

The journal has engaged with both parties to find a mutually agreed statement on the chain of events that contributed to the publication of the correction, with a view to republishing the correction to clarify both the scientific record and the sequence of events. This has resulted in a great deal of time and effort being expended on several draft statements prepared by the journal over the previous 14 months.

As the matter remains unresolved between the two groups, the journal’s team has elected to publish nothing at all. The groups have been informed of this, and that the journal remains amenable to publishing a statement if the two parties are able to agree a form of words between themselves.

Nevertheless, the publisher regularly reviews its working practices and editorial policies, and this case has contributed to a change of the policies enacted by the publisher to reduce the likelihood of similar sequences of events and outcomes in future. Taking our experience in this case into account and aiming to address potential future conflicts of interest in submitted comments, a new comment/reply policy has been adopted. In hindsight, the previous comment/reply policy was problematic for a number of reasons, including the potential conflict of interest in having the author of the original paper being involved in the peer review of the submitted comment.

Question(s) for the COPE Forum
• Given the apparent impossibility of the two parties agreeing a form of words, and the threats of legal action and publishing their own version elsewhere, is the journal justified in choosing NOT to publish anything? Would it be better to publish the journal’s view anyway and accept the potential risks?
• While recognising the publisher’s original comment/reply policy contributed to this matter, does the Forum have any advice on how the publisher/journal could or should handle similar disputes in future? The policy has been amended to reduce the possibility of conflicts of interest.
• How far should the publisher go in trying to resolve disputes between groups (especially where, as in this case, only one party has actually published in the journal)?