As editor of journal A I am handling a manuscript by an author and it is likely to be accepted, although this is not yet decided. As a reviewer for journal B, I have since been asked to review a manuscript by the same author that uses similar material and comes to a similar conclusion, but pushes the presentation of the results a little further. My gut feeling is that there is insufficient novelty for journal B. However, my problem is how do I tell journal B without compromising my role as editor of journal A?
Following advice from a COPE council member, I decided to review the manuscript for the second journal and in my comments to the authors I suggested that in order to sharpen their conclusions this study might be better embedded in a larger study. In my covering note to the editor, I wrote “Unfortunately, there may be an ethical issue to do with this manuscript. I have consulted COPE, without of course mentioning the journal, the authors or the subject of the research, but they cannot bring it to their Forum until March, which is long after the review deadline”.
The “larger study” that I mentioned is the work that is under review in the first journal. The covering note to the editor does not breach confidentiality with respect to the first journal but it does alert the editor of the second journal that there “may” be a problem and that he or she can decide to wait for a decision by the Forum.
I have since found out that Journal 2 has now rejected the manuscript, with another reviewer’s report identifying the same general scientific problems as I did in my comments to the authors. So, my cautionary covering note to the editor about a potential conflict of interest seems not to have been needed.
Does the Forum think that I handled the situation correctly or is there anything else I could have done?
The Forum agreed that the editor acted correctly by raising his conflict of interest although some suggested it might have been easier to tell the other editor that he could not review the paper as he was reviewing a similar paper submitted to his own journal by the same author. The Forum noted that it is fine for an editor to raise a problem with another editor. The Forum did raise the ethical issue of the author failing to declare a similar paper submitted to another journal. Most journals would require a copy of the other article. Had the author signed a form to say that the work was not under consideration at any other journal? It may be the case that the editor should review his guidelines to authors to ensure that such guidance is clear.
We have recently developed and begun to put into practice a policy on collection and declaration of conflicts of interest statements from any individual involved in contributing to or reviewing our pathways. This policy includes members of our editorial team, and contributor and reviewer members from our stakeholder groups. We have devised a standard form to collect these statements in a uniform way, and publish the declarations of interest from all contributors (both inhouse and external), but collect and do not make public those of peer reviewers.
Stakeholder group A, who provides clinical contributors and peer reviewers for some of our pathways, has a pre-existing declaration of interest policy for any review committee member performing work on behalf of the group which does not align with our own. Stakeholder A’s review committees undertake a number of functions for the group themselves; review of our pathways is an additional role they have taken on over the past 18 months. They collect the statements on a voluntary basis annually (or as new committee members join) and it is their current policy not to share those declarations with any outside group. The stakeholder group has suggested that it would be willing to sign an annual declaration to confirm that their committee members do not have any conflicts of interest.
However, we have requested that if an allegation of conflict of interest is raised with us, we require a process to ensure that the group’s committee make the individual members’ names and declarations available to us. We await the stakeholder’s views on this proposal, but would welcome COPE’s views:
Is it sufficient to collect a statement from the group to say that they have reviewed all of the individual statements for the committee members and determined that there is no troubling statement?
Is it reasonable to request a change in the stakeholder’s policy to ensure our access to the detail of the committee members’ individual statements in case of any allegation of conflict of interest?
The Forum questioned the logic of why you would obtain conflict of interest statements but not be able to see them? The rationale for competing interest statements is that they should be transparent. The advice was to obtain the names of the people and approach the individuals if possible. Also, the Forum reasoned that it is not sufficient to have annual statements—competing interest statements need to be up to date. Some suggested changing the wording and asking, for example, for any relevant interests (as opposed to conflicts) which might elicit more useful information.
We have agreed an exception to our conflict of interest for stakeholder group A. We will collect, hold but will not display the conflict of interest statements for anyone peer reviewing our pathways for accreditation purposes. We have agreed with stakeholder group A that they will provide an annual statement of assurance which describes the issue, outlines what they do to collect declarations from members and agrees a process on what happens should we need to investigate any declarations.
During review of a manuscript submitted to our journal, a dispute arose over some of the data used in the database that was described in the submitted paper.
The authors listed several preferred reviewers and also one non-preferred reviewer (without giving reasons). The journal’s submission site states that the editors will consider the authors’ preferred suggestions but are under no obligation to use any or all of them and that the editors reserve the right to approach non-preferred referees. Authors are asked to outline in their comments to the editor any particular reasons for requesting exclusion.
The paper was initially positively reviewed by two referees, one of them a preferred referee, and minor revision was requested. Neither of the initial referees responded to the invitation to review the revision (they neither declined, nor agreed, just did not respond) and as the senior author had already published with virtually all important scientists in this small field, the associate editor decided to invite the non-preferred reviewer. The non-preferred reviewer and one new reviewer agreed to review.
Shortly after accepting, the non-preferred reviewer emailed the editorial office asking whether it was journal policy to publish a reference to a database without any scientific study based on these data (the journal answered yes, citing a previously published database description) and stating “I discovered by examining this database that the senior author has used my data without permission. This applies to hundreds of measurements in country 1 and in the mountains, but also to data from country 2”.
The authors state on the database website that the database contains published data and in the manuscript that the “measurements in the database have been collected over the last 20 years from various sites around the world (references given) and are included with permission from the collectors”. The journal informed the non-preferred reviewer that, from what the authors were stating, all data seemed to be in the public domain, but advised him to put his concerns into his review or contact the authors directly, if the reviewer preferred.
By the time the reviewer received this reply he had already submitted his review and informed the journal that he also sent his full review to the senior author. The reviewer also pointed out in an email to the editorial office that the measurements published in one of the cited studies (of which the referee was the senior author) were not published in the form as they appear now in the database, and then accused the senior author of having used information which he collected as a member of the reviewer’s research group a long time ago, and publishing it without the reviewer’s permission. The reviewer thought that the reference to the paper was not sufficient and also stated that the conditions with the mountain data were more serious.
In the review, the referee pointed out three problems: (1) unethical behaviour on the part of the senior author. The reviewer stated that he could not remember having given permission to use the data, but had specifically asked the senior author not to use the mountain data for anything until the reviewer had finished his analysis. He included the original email in the review. (2) potential dual publication. The reviewer stated that because no analysis of the data was presented in the paper, a simple report on the existence of the database was not suitable for publication in a journal and went on to point out that a first version of this database was already published (as cited on the database website, but not in the paper) as a preprint article. As a comparison of the two manuscripts did not reveal any obvious differences in content, the reviewer questioned whether this constituted dual publication. (3) questions over the use of a particular method. The reviewer called the use of this method “scientifically unacceptable”, stating that the problems with this method have been published and concluding that it was unacceptable for the senior author to ignore these arguments (this was not mentioned by any of the other reviewers).
Checking the date of the above mentioned email and the publication date of the preprint article, the journal realised that the reviewer emailed the senior author shortly after the document was posted online. The dispute had thus been going on for over a year by the time the referee was invited to review, but the reviewer had not declared a conflict of interest.
The associate editor, having read both reviews (the second being positive), recommended rejecting the manuscript and inviting a resubmission once the authors either were able to present the permit obtained from the non-preferred reviewer or removed all unpublished data for which no data were available.
In the meantime (and before any action was taken), the senior author, prompted by the reviewer’s email to him, emailed the managing editor and the associate editor, informing them that the non-preferred reviewer used to be the senior author’s PhD supervisor. The senior author rejected as incorrect any claims over the inappropriateness of the method and the statement that the reviewer had not given permission to use what the referee claimed were his data. As the review implies several forms of unethical behaviour on the senior author’s part, he felt the need to clarify.
(1) the criticised method has been used in top tier journals such as Science, Nature and PNAS and has therefore been through rigorous quality control. Since the reviewer’s evaluation of the method did not contain any concrete arguments, the senior author assumed that the reviewer was referring to a polemic about the method by another author, published in the same journal in which the senior author countered these arguments. The method continues to be the most widely used, cited over 90 times (the associate editor points out that it was cited mainly by the senior author’s main group, but that none of the other reviewers have criticised the method). The senior author further points out that this method is not the only one used in the database.
(2) re the inappropriate use of data from country 1, the senior author assures the journal that the referee had been asked and had given permission (an email was attached), but that if the referee wanted to retract the permission, the authors would remove the data in question. Re the data from the mountains and the email sent by the reviewer, the senior author claimed that the email had been taken out of context. The senior author had indeed asked the reviewer about the possibility to perform a separate analysis on the mountain data. According to the senior author it was this request about the separate analysis that the reviewer declined and the request had not referred to making publicly available the data that had already been published in one of the references.
(3) re dual publication, the senior author stated that a beta version of the manuscript, with a completely different code and user interface and only a very reduced set of data, was posted on a preprint server, not a regular journal publication (the associate editor saw no major changes in the number of data entries and that both discuss the database in a similar way, but thought that the submitted manuscript was more detailed than the earlier (preprint) article. Figures are not identical, but basically follow the same scheme. The associate editor left it up to the editor whether he considered a full-length paper that builds upon something that has been published on a preprint server as sufficiently novel).
The senior author finally adds that because of previous similar experiences, the authors had listed the reviewer as non-preferred.
The corresponding author emailed the editorial office offering to remove from the database any data that have been collected with the reviewer’s participation but emphasised that by doing so, the authors do not acknowledge any form of wrongdoing on their part, but seek to make it easier for the journal to make a decision.
After discussions between the editors, the editorial office and the publisher, the decision was to reject the manuscript but to invite resubmission. The letter pointed out (a) the disputed use of data in the database; (b) the question of dual publication; and (c) the scientific criticisms expressed by both referees. It of course included both reviews in full.
The editor confirmed that the journal did not have a policy on whether they will accept material that has already been posted on a preprint server. Preprint servers are not peer-reviewed, so some journals will accept material that has already been posted, but all such prior presentations should be disclosed to the editor in the covering letter. All agreed that the editor should not get involved in an author dispute. The editor could suggest that the authors find an independent adjudicator, that both sides find acceptable, if the authors cannot come to an agreement among themselves. It was suggested that the editor could go to the author’s institution and ask them to adjudicate. The editor should also clarify their policy on preprint servers in the journal’s instructions to authors.
We have received threats of legal action from the authors of a manuscript rejected by our journal, henceforth referred to as journal A. These “aggrieved” authors claim that their manuscript was unfairly reviewed by a close competitor, who then used some of their findings in a paper subsequently published in journal B, without either attribution or citation.
The “accused” scientist had indeed reviewed the paper for journal A, and the date on which he/she had first been sent the paper preceded that of his/her own submission to journal B. The steps of our investigation were as follows:
The aggrieved author was asked to provide additional details on which aspect of his work he/she suspected of being unethically used, and he/she identified a particular paragraph in journal B’s paper, which named two genes which our authors claimed to have identified for the first time in the particular bacterial genus studied in both papers.
Meanwhile, the “accused” scientist was asked to respond to the accusation. His response identified the paragraph in question as being a small area of overlap between the two papers, however, he categorically denied that the content of this paragraph drew in any way on the information presented in the manuscript which he had reviewed for journal A. The accused backed up this denial by sending us a copy of an earlier version of the paper, which had been submitted to and rejected by a previous journal (journal C) months before he had first reviewed the complainant’s paper in journal A.
We confirmed this by contacting the editors of journal C who, after obtaining permission, provided us with an independent copy of the manuscript that had been submitted to his journal. On examination we found that the paragraph in question had remained unchanged, and that the description of the two genes was indeed present before any submission to journal A took place.
We agreed with the accused that this data analysis was a very minor part of the paper published in journal B.
Questions for COPE • At this point, we feel that our investigation has exonerated the accused reviewer of one allegation (unethically using information obtained during the peer review process in his/her own publication). Does COPE agree?
• If the manuscript submitted to journal C (providing independent confirmation of the accused’s defence) had not been available, how would such a case be investigated?
• The other allegation (of the reviewer causing the authors’ manuscript to be unfairly rejected) remains unresolved. The reviewer denies misconduct, but there is at least the appearance of misconduct on the basis of conflict of interest. However, we do not think that any further investigation can resolve this issue. Does COPE agree?
• The aggrieved authors have asked for a correction to acknowledge their work (which was published in yet another journal one month before journal B published its article). While the reviewer did not “steal” any data or ideas, he may have unfairly “squashed” the authors’ publication. However, the data analysis in question is a very minor point in the article published in journal B, and the authors’ work may simply be independent corroboration. At this time we do not feel a correction is warranted because we have no evidence of wrongdoing. Does COPE agree?
• Are there other options that might be used in place of a correction?
The Forum agreed that the accused reviewer had been exonerated. The advice from the Forum was that the editor should obtain permission from the reviewer to contact the authors, tell them that their allegations were unfounded and explain the situation to them. The authors should be informed that the reviewer has given his permission for this disclosure as the editor is not obliged to reveal the names of reviewers if the journal operates a closed peer review system. The editor might also suggest that the authors may like to consider an apology to the reviewer. However, some Forum members voiced concerns that the reviewer did have a conflict of issue and that he should have declared this initially.
The reviewer was notified (as were the editors of the other journals) that he had been exonerated, with thanks for his patience and for his cooperation throughout the investigation. The author was also contacted and told that the reviewer had been exonerated. The authors did not formally apologise to the reviewer.
We publish “mini-reviews” of published articles. Our faculty of eminent researchers and clinicians write these evaluations. One of the conditions we insist on from our faculty is that they may not evaluate work on which they are an author. We received a review of a paper, the authorship of which was listed as:
Name A, Name B, Name C; study group X
As the reviewer was a member of “study group X”, we rejected the submission. The reviewer wrote back claiming that:
“I am not an author on the “Name A” paper. The latter is a substudy, which uses the “study group X” database. While we entered some patients into the main trial, in no way do I fulfil authorship criteria given that I never even saw a draft of the paper nor knew it was being submitted.”
We published the review with the declaration above listed as a “competing interest”. We then informed the journal in question.
Did we act correctly?
The Forum thought that this was a grey area. Even though the reviewer is not directly involved in the study, is he sufficiently removed? The case also highlights the fact that the definition of an author can be vague and different journals have different criteria. Some suggested that the journal should consider disqualifying a reviewer with any involvement in a study from publishing reviews concerning that study. This would give the reader more confidence in the system. All agreed that the best course of action is to have a policy on this issue for future such incidences.
The advised policies which members of the Forum suggested have not yet been implemented due to a massive ongoing redesign project for our website which has put many things on hold. It is unlikely this will be resolved until the new year; however, we have taken the Forum’s comments on board and intend, when the time comes, to implement the suggestions in the most thorough and transparent way possible.
An online post-publication literature evaluation service, aiming to highlight the best articles in medicine, received an evaluation of an article whose authors were based at the same institution as the evaluator. The editor asked the contributor if he/she had any involvement in the study and received the following response: “I am based at the university but did not participate in the design of the study. I occasionally cross-covered some of the patients when a co-author was out of town”. During an in-house editorial meeting, it was decided that the editor should ask the evaluator for a more detailed declaration and asked whether he/she had advised or contributed in any other way, other than treating the patients according to the protocol? The editor is awaiting feedback; the editor presented the case in an editorial meeting, raising the potential point of concern of whether we can accept this evaluation, considering the evaluator has covered some of the patients in the study and, therefore, may have a certain level of involvement.
(1) Would you have taken a different course of action?
(2) With regards to involvement and contributing to studies, where does contributorship begin?
The committee advised that the evaluation service should have a policy of not accepting evaluations if the evaluator is from the same institution as the authors. The committee felt that a clear code of practice on this issue would prevent any future such cases from arising.
The editorial guidelines have now been updated and the editor considers this case now closed.
An online post-publication literature evaluation service, aiming to highlight the best articles in medicine, received an evaluation of an article on which the evaluator was listed as an author on PubMed. The editor queried the evaluation and the evaluator replied explaining s/he had no involvement with the study but had commented on it. When the editor looked at the full text HTML version on the journal website, the evaluator was not listed as an author on the article but featured as an author on the side-bar search tool. When the editor looked for the commentary in the HTML version, it could not be found. Only when the editor looked at the full text PDF version could the commentary then be read, where it was included on the last page of the article. Essentially, the article and commentary were published as one. The editor discussed the issue within an in-house editorial meeting, to ascertain whether the evaluation should be accepted, given that although the evaluator commented on the page, s/he was listed as an author. It was also discussed whether the evaluator might be too involved with the original article and it might be construed as self-promotion.
Questions we would like to ask COPE:
(1) Would you have taken a different course of action?
(2) Would you feel differently if the competing interest was that he/she was the peer reviewer?
(3) This is a form of consensus paper; therefore, it would be hard to find any expert/evaluator who was not involved in some way (it is very difficult to find anyone who hasn't been involved or consulted in the writing process). Could you offer any advice on how we could get round this?
(4) Since the evaluator appears as an author on PubMed, but not on the journal’s website, should it be our responsibility to raise this issue with the journal?
The committee thought that this might simply be an error. In the first instance, the advice from committee was to contact PubMed and the journal in question, to see if an error has been made, as the paper and commentary, although linked, should be two distinct entities. Further advice was to re-check the contributorship statements from the evaluator.
An online post-publication literature evaluation service aiming to highlight the best papers in medicine, received an evaluation of a basic science study funded by an NIHM grant. The evaluator declared in his/her competing interests that he/she is the director of a project that included the evaluated study as one of its components. The overall project was funded by an NIHM grant and paid the salary of the study’s first author. The evaluator did not supervise the study or write the paper. As competing interests were declared, the editor decided to publish the evaluation.
The editor presented the case in an editorial meeting, raising the following concerns:
Can we accept this evaluation considering the evaluator is the director of the overall project and may gain from promoting this study’s results?
The meeting made the following decision:
As competing interests were declared, the meeting felt we could go ahead with publishing the evaluation.
Competing interests declaration: I am the director of the grant that paid some of the expenses for the conduct of this study, including the first author’s salary. I did not supervise the work or write the paper, as you can see I am not a co-author. If you look at the paper, it indicates the financial support of a NIMH grant. I am the director of the overall project that included a component directed by the first author that conducted this study.
Would you have taken a different course of action? In particular, do you think it was right to publish the evaluation and is the declaration of interests sufficient in its current form?
Seeing that approximately 1% of evaluators declare any competing interests when probably 90% (again, a rough guess) have interests with regards to the articles they evaluate, would you have any suggestions on how we could improve the situation?
The chair commented that the post-publication literature evaluation services do raise some very interesting and new issues not encountered in paper journals. The majority view of the committee seemed to be that there was a clear and unacceptable conflict of interest and some argued that the evaluation should not have been published. However, the committee advised that it might have been useful to have determined in advance the exact involvement of the author in the study and to publish a more detailed competing interests statement. The editor should consider whether the service’s published advice on competing interests is sufficiently comprehensive.
An online post-publication literature evaluation service, aiming to highlight the best papers in medicine, received an evaluation of a cost effectiveness study assessing a new therapy. The evaluator quoted estimated amounts of cost saved when using the new therapy compared with other therapies, naming the manufacturer of the new therapy. The evaluator declared in his/her competing interests that s/he had “received honoraria, financial support and […] been on several advisory boards” for the company that developed the therapy, mentioning the company name. Whether the evaluator was involved in the development of the therapy in question is unknown.
The editor presented the case in an editorial meeting, raising the following potential points of concern:
(1) Does inclusion of manufacturer information, a practice that is quite common in research articles to identify specific devices, become a form of advertising when appearing in a post-publication evaluation?
(2) Is it permissible to mention amounts of cost saved if quoting directly from the study or does that again amount to advertising?
(3) Can we accept this evaluation considering the evaluator’s involvement with the manufacturer?
The meeting made the following decisions:
In answer to (1), the meeting considered that it definitely felt awkward and clearly unnecessary to include the manufacturer’s information in the comments. It was therefore taken out.
In answer to (2), as the relevant statements merely paraphrased the study’s findings, they were considered acceptable.
In answer to (3), as competing interests were declared, the meeting felt that the evaluation should be accepted.
Questions we didn’t consider at the time but have since asked ourselves:
(4) Should we have asked if the evaluator was involved in the study itself and should this information have been included in the competing interests declaration?
(5) Does naming the manufacturer in the competing interests declaration amount to advertising? Should it therefore only have said “received honoraria from the therapy’s manufacturer”, not including the name?
Would COPE suggest a different course of action and how would COPE answer the subsequent two questions?
The committee agreed that it was best to find out the exact involvement of the author with the company and with the development of the therapy. The author’s involvement should be transparent. If there is a conflict of interest, it might be worth discovering whether it is a personal/specific or non-personal/non-specific conflict. If the latter, a conflict of interest statement should be published and the author allowed to publish the review. However, if the conflict of interest is personal/specific, the author should not be allowed to publish the review. The committee questioned the purpose of the review in the first instance as the reviewer is clearly not independent. The committee felt that it was imperative for the reviewer to publish the conflict of interest statement and so allow readers to make up their own minds.
After further investigation by the editor, it was discovered that the evaluator's conflicts of interest were too great, leading to rejection of the review on the basis of bias.
An online post-publication literature evaluation service that publishes only positive reviews, aiming to highlight the best papers in medicine, received an evaluation of a paper that had been published in a journal for which the evaluator of the paper acts as editor in chief. The evaluator did not declare any competing interests but the editor dealing with the evaluation knew about his/her role in the journal. The editor decided to reject the evaluation because he/she felt that the editor in chief was effectively using the evaluation service to press release a paper from his/her journal.
Do you agree or would you have handled the case differently?
This is a case of self selection. The evaluation service does not state in its guidelines that an evaluator cannot select papers from his or her own journal, so there is a case for tightening up its written instructions to prevent a recurrence. However, it was agreed that this behaviour is acceptable if the evaluator declares a conflict of interest. As long as the evaluation is published together with the conflict of interest noted, there can be no argument. The committee felt that editors shouldn’t be excluded from evaluating articles in their own journals, especially in the case of highly specialist fields, but the evaluator must declare a conflict of interest. One solution would be to obtain two separate evaluations, but again a conflict of interest must be declared. It was suggested that instead of the editor, it might be more appropriate for a member of the editorial board to write an evaluation.
After the discussion at the COPE meeting, and following further discussion with the editorial team, we have now changed our guidelines to say that editors may evaluate papers from their journals as long as they declare their relationship with the journal. We have also started collecting data on evaluators’ involvement with any journals so that in the long run we will be able to request a competing interest declaration if applicable.