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.
A paper published in one of our journals (paper A) provoked the submission of a correspondence article claiming that a minor conclusion of the paper was a misinterpretation and erroneous. The point in contention was a question of zoomorphology and our paper’s conclusions were based on analysis using a non-invasive technique while the rebuttal relied on more traditional techniques. We are bringing this case to COPE because although it appears to be in the process of being amicably resolved, with a clear resolution of the scientific issues, it has highlighted an area of confusion about the use of privileged information.
The authors’ of the correspondence article (rebutting authors) originally expressed anger and surprise that the paper A contained this error, because they thought they had clearly laid the issue to rest in an earlier rebuttal of a previously published paper making similar errors (paper B). Although this first rebuttal had not yet appeared in print, it had been considered and accepted for publication by the one of the authors of paper A, in his/her capacity as the editor of another journal. Furthermore, this first rebuttal not only challenged the findings of paper B, it also specifically called into question the interpretation of some website data which was included (unmodified) in paper A.
We sent the correspondence article for peer review, and the reviewers supported the soundness of the rebuttal data presented and the alternative morphological interpretation. The reviewers appeared inclined towards the view that the perpetuation of the “wrong” interpretation in paper A was surprising and did not reflect well on the authors of paper A. However, they also indicated that given the close chronology of the various publications, this was a grey area, and not germane to the scientific case for publishing the second rebuttal. We therefore asked the correspondence authors to revise their text to keep the focus on resolving the scientific questions.
Having decided we should, in principle, accept and publish the correspondence article, the authors of paper A were invited to submit a signed response. In this they have clearly acknowledged that the data presented by the authors in both their rebuttals fully support the conclusions reached in these rebuttals and that some of their own data had been misinterpreted in paper A. They also explained that they were already convinced by the first rebuttal which one author had seen in his/her capacity as an editor, and the other had reviewed. However, they had felt it would not be ethical to make use of this privileged information to modify their own paper (paper A) shortly before final acceptance.
We are inclined to accept this as the personal view of the authors of paper A but question whether they adopted the best ethical course.
Questions for COPE • What is COPE’s view? • How should editors and reviewers proceed when they have access to privileged information which suggests that their own work should be modified or corrected? • Is there an ethical responsibility to avoid letting known errors into the scientific literature which was transgressed in this case?
The Forum questioned the authors’ use of the term “privileged information”. The Forum agreed that the authors had acted wrongly. They could have delayed publication of their paper until after the information was in the public domain. The authors should have contacted the publisher and asked them to hold back on publication, explaining the reasons why. Although there was no major misconduct, a correction should appear in the journal (in addition to the correspondence) so that the article will be permanently linked to it.
The case was successfully and amicably resolved. All parties found the advice from COPE very helpful.
We submit to COPE a case regarding the suspected multiple publication of research on four separate occasions in four different journals.
Close inspection of the articles in question revealed that the author had directly copied and reused extensive sections of text, including tables in all four articles. After this matter was bought to the attention of the Editor of Journal A, the chronology detailed below was pieced together. It demonstrates that there is significant overlap in the dates the articles were submitted and copyright forms signed. This, coupled with the extent of the similarities between the articles, lead us to believe that the author would have been aware they were submitting near identical articles to multiple journals over a short space of time.
Upon this realisation the authors were contacted according to the COPE guidelines and the lead author cooperated with us in putting together the information below. Because of the complexity of the case, the number of journals and the time frames involved, we would be grateful if COPE could advise us on the correct course of action.
• In January 2002, three UK authors submitted a paper to Journal A based on the lead author’s PhD thesis. This paper was accepted by Journal A in October 2003 and a copyright assignment form was signed in November 2003. This article appeared in Journal A in January 2005.
• In July 2002, the lead author presented a conference paper based on the same PhD research and submitted it to Journal B. This too was accepted and a copyright agreement was signed in June 2003. This article went on to appear in Journal B in January 2004. This paper acknowledges the conference it was given at.
• In June 2003, the lead author gave another conference paper based on their PhD research which was subsequently published in Journal C in December 2003. The copyright assignment form for this article was signed in September 2003.
• In January 2006, another paper was published in Journal D, again based on the same PhD research. We do not know when this was submitted or when the copyright form was signed.
The advice from COPE was to consult the flowchart for “redundant publication in a published paper”. The flowchart advises that you check the degree of overlap. If it is substantial, contact the authors. If an unsatisfactory response is received, the editor should consider publishing a notice of redundant publication or retraction. Sometimes the author makes a genuine mistake or the instructions to authors are not clear enough (does your journal say that submitted work should be original and not submitted elsewhere) or the author is very junior. In such cases writing to the author explaining the situation and outlining the expected behaviour is sufficient. However, if the editor is satisfied that this is not a simple error and an unsatisfactory response has been received, he should consider contacting the author’s institution and asking them to investigate. The Forum would also advise contacting the other journals.
We have agreed with the publisher that the paper will be retracted.
A paper was accepted and published in journal A which dealt with a cohort of patients with an unusual respiratory pathogen. A similar paper had been published in a US journal B a few months before. It dealt with more or less the same patients (a few more had been added) and provided some extra secondary outcome data but with the same conclusions.
The editor of journal A considered this to be duplication but the authors deny this on the grounds that there are further data.
This is a difficult issue as the editor considers it a case of duplicate publication but the authors disagree. Some journals ask authors to send related papers when submitting their articles for publication. The Forum agreed that journals should have some form of declaration in the instructions to authors or the submission information must be very clear on the rules of duplicate publication. Some argued that a rule of thumb is that if the “extra” data do not stand alone, then it is probably duplicate publication.
In this difficult area, the decision really needs to be left to the judgement of the editor and a correction published in both journals if duplicate publication is believed to have occurred. Software can sometimes help as it can give (in percentages) the amount of overlap between two papers and then editors can judge what amount is acceptable. However, if undetected, all agreed that this is a serious problem as the data may be counted twice in meta-analyses.
The editor indicated to the author that this was a case of duplicate publication and the paper was withdrawn from the website. A notice of duplicate publication was published in the editor’s journal.
Two authors submitted a case report which was interesting but not written in the style of the journal. The editor therefore invited the authors to rewrite the case report, and resubmit it. They did so within a week. The case report was sent out for peer review, accepted and published.
The head of department of one of the authors then wrote to the journal, stating that the case report had previously been published in a non-English language journal. Moreover, in that version of the case report, but not in the more recent version, the head of department had been listed as an author. The head of department had not been asked or informed about the submission to the British journal. The head of department gave the reference of the previous publication. The editor obtained the paper: it did indeed appear to be the same case, since it had essentially the same story and identical figures. The head of department was listed as an author, together with the two authors who had submitted the case report to the editor.
The authors had written a cover letter for each submission. Both cover letters stated that the case report had not been submitted to any other journal. They had been sent after the case report was published in the non-English language journal. The authors were contacted separately by email, and asked to explain the discrepancy. They were warned that if no satisfactory explanation was provided, the journal would have to retract the case report. At the time of writing, they have both written to say that they will provide an explanation, but have not yet done so.
The majority of the members of the Forum agreed that there were grounds for retraction of the paper. However, others argued that as the papers were in different languages, a “notice of duplicate publication” would be more appropriate. All agreed that this was a definite case of misconduct. There are two issues here: (1) the author issue in relation to the names on the two papers and (2) the duplicate publication issue. There may also be copyright issues with the first journal. The advice was to write to the author’s institution, but not in this case to the head of department as he may be unable to remain unbiased. It was suggested to write to the Dean of the university and ask him to investigate.
The authors wrote to us to say that the previous publication was not a scientific publication (ie, in a magazine rather than a journal). They said that the magazine had no editor, no peer reviewers and no signed author agreement. They sent us a photocopy of the instructions to authors, and highlighted a sentence saying “bear in mind that you are not writing a scientific publication”, from which they inferred that the magazine was not a scientific publication.
However, the publication is quite a well known journal in their country of origin. It does have an editor. It is indexed on PUBMED, as indeed was their original publication. The statement about “not writing a scientific publication” appeared to refer to the section, rather than to the journal as a whole: moreover, it was in the section “guide to style”, and hence appeared to be intended to encourage lucid writing.
We therefore retracted the publication.
We also wrote to the professor who had let us know about the duplicate publication, to thank him.
We received a case report from one of the authors of this paper, and the professor who had told us about the duplicate publication. We rejected it on the grounds that it was good, but not exceptional.
Author X recently published a paper in Journal Y and has asked for the paper to be retracted. The reason given is that part of the data presented in the paper was published without the permission of a colleague, who is not listed as an author of the paper (and probably does not qualify for full authorship). This colleague is now seeking to publish the data in another journal and it is implied that Author X is also a co-author on the second paper (which has been submitted, not yet accepted). During correspondence with Journal Y, Author X has confirmed that the data presented in the published paper are 'accurate and reproducible in all respects' and the conclusions of the paper are not affected. Author X acknowledges that this is a dispute regarding the use of data without permission, and understands that retraction is a serious matter. However rather than publish a correction, Author X prefers to retract the published paper in order to maintain a good relationship with colleagues. The decision to retract was reached through discussion with the researchers involved, and has not been requested by the authors' institution. As far as Journal Y knows, the institution has not been involved. Can and should Journal X refuse to publish a retraction on these grounds?
As the author has clearly stated that the data are correct, and the only dispute is a small section of the paper that was published without permission, the committee felt that a retraction is not necessary. As the degree of overlap is so small, this is unlikely to constitute duplicate submission. The second paper could cite the first paper and make a note that the data were published previously in error. The journal could publish a correction or an acknowledgement but the committee felt that the editor should hold firm and not agree to a retraction. The committee felt that contacting the author’s institution was not necessary in this instance.
The editor wrote to Author X and explained that COPE committee members had agreed that there were insufficient grounds to retract the paper published by Journal Y. Instead the editor recommended that Journal Y publish a correction to the paper acknowledging the input of the colleague who generated the data in question. The editor also noted that it should be made very clear to the editors of the other journal that the data had been previously published in Journal Y. Author X passed this recommendation onto his colleagues who then agreed to publication of a correction rather than retraction of the paper. Author X expressed appreciation to us and to COPE for helping to resolve this matter.
After peer-review, a general medical journal published a household survey of violence following a coup against the country’s elected President. The survey revealed high levels of violence and human rights abuses, only a small minority of which were attributed to supporters of the deposed regime. The manuscript stated that none of the interviewers had political affiliations and the authors declared that they had no conflict of interest.
Within days of publication the Editor was contacted by an expatriate from the country and by a local aid worker who expressed incredulity over the fact that the findings attributed so little of the violence to supporters of the deposed President’s political party, It was also pointed out that one of the authors was acquainted with the deposed President and had previously published pieces under a different name which were supportive of him. Some of these pieces were cited in the manuscript.
The author admitted that she had done this and the co-author, her thesis supervisor, stated that he was aware of these facts and did not consider them a conflict of interest.
Not satisfied by the responses from the authors, the Editor asked the Dean of the authors’ institution to undertake an internal investigation to verify that the data had been coded accurately. Results are expected by the end of 2006.
This interesting case prompted much discussion. The committee felt that the conflict of interest should have been identified in the peer-review process and were surprised by the reviewers’ responses and their failure to pick up on the political bias. The committee agreed that the editor has a duty to his readers to inform them that an investigation is ongoing. He should tell his readers that there have been allegations made about the paper but that it is not possible to establish the truth as yet. Hence the advice was to issue a statement of concern in the journal or possibly write an editorial highlighting all sides of the issue.
Following discussion of the case, a statement was published in the journal, a summary of which is given below.
In response to credible allegations that one author’s former activities might constitute an undisclosed conflict of interest, the journal began an inquiry. The authors’ institution was asked to investigate the matter, and the issue was referred to the Committee on Publication Ethics (COPE).
The institution audited 100 questionnaires selected by computerised randomisation. Outcome details on the original handwritten records corresponded with the project’s computerised database. The overall distribution of rapes and murders were re-analysed according to alleged perpetrators, and the results agreed with the published findings. Outcomes were then compared by political affiliation of the interviewer and for the author’s own data (as an interviewer). Again, there was no evidence of systematic bias. On the basis of this investigation, the journal has confidence in the authors’ findings as published.
COPE recommended that readers should be made aware that the author had published as a reporter under another name, and that failure to disclose a separate name, under which relevant material had been published and cited in her paper, constitutes an undeclared conflict of interest. The journal’s position on transparent disclosure of potential conflicts of interest is in accordance with guidelines established by the International Committee of Medical Journal Editors. The journal has made this position prominently available to readers and to authors, and stated clearly that incomplete disclosures will be amended in a published statement in the Department of Error section, which will also be linked electronically to the publication in electronic databases. Such a correction for this study appears in today’s journal.
To realise their full potential to benefit populations, research findings must influence practice. Intelligent debate is part of that process. The journal encourages genuine debate, and will always consider seriously allegations of scientific misconduct. It is unfortunate, however, that in this case much of the debate was aimed at exploiting historical divisions in the country in question. That process has obscured the message of the authors’research and detracted from the real issue—the welfare of civilians in that country—to whom attention should now turn.
A reader contacted the editorial staff of Journal A after noticing that Journal B, which is primarily non-English, had published a paper that was remarkably similar. The editor of journal A contacted the editor of Journal B. Both editors reviewed the two papers and agreed that the paper from Journal B contained methods, results, and conclusions that formed a part of the paper from Journal A.
The authors were contacted by the editor of Journal A and asked for an explanation. They have replied that the two papers were intended for different audiences, and since the paper in Journal A was a more comprehensive study, this did not constitute duplicate publication. The papers were submitted at approximately the same time, and neither paper referenced the other.
The committee felt that the authors’ excuse that the papers were intended for different audiences did not stand up. If the two papers were basically the same, with most of the data in the tables being duplicated (overlapping data) but with an extension of the data in one of the papers, then this was clearly a case of duplicate, or “redundant”, publication. It would have been acceptable for the author to publish both papers, provided permission to do so had been obtained from the editors of both journals, and that the matter was acknowledged in both papers. However, neither paper referenced the other.
Therefore, the committee advised that Journal A should issue a notice of redundant publication. Ideally, the authors should issue the notice, but probably the editor would have to do it. When writing to the authors, the editor does not need to go into great detail, just point out that it has been discovered that this particular study has been published in full or part by another journal, and that consequently a notice of redundant publication will be published in the journal.
This case was resolved by publishing a notice of redundant publication. The author was not willing to write the correction initially, but did approve the final draft that went on to be published.
An English language journal received a study describing a randomised controlled trial. The paper was accepted and published several months later. Five months after its publication the editors were informed that a similar study had been published in a German language journal two years earlier. Three of the four authors were involved. It had been carried out over the same time period, using the same methods, and had arrived at the same conclusion. The only difference was that the earlier publication involved 168 cases and the later one 200. The second article did not reference the earlier publication nor had the editors been made aware of its existence. Had the editors known about the earlier publication the chances of publication of the second paper would have been significantly weakened. The lead author was contacted to explain. He apologised for the oversight, acknowledging that the second article should have referenced the first, but arguing that the first article would not have been generally accessible to most of the journal’s readership. He continued that the second publication expanded on the earlier pilot study of 32 patients, while the discussion considered other countries. The tables and text were original and not copied from any previous article, he said. He added that the authors would be pleased to acknowledge in an erratum that this was an unintentional oversight. - Does this constitute dual publication? - Should the authors’ apology be accepted or should the matter be taken further?
- This case should really be called redundant publication. - The explanation that this was an oversight is simply not credible. - The question is how will other people know that this is an expansion of the first study? - The editors should write to the authors’ institutions and post a notice of redundant publication.
A journal published an animal study on the use of drug X for the treatment of clinical condition A. The authors did not declare any competing interests. A few months after publication, a journalist contacted the editors to say that the corresponding author had several patents on drug X, was listed as an inventor of the drug, and that the public charity of which he is the director recently announced that they were seeking approval for clinical trials of drug X in condition B. He also said the corresponding author co-owned a commercial company with whom the charity does business. The authors were asked to clarify any competing interests, and were directed to the journal’s policy, which is posted online as part of the instructions for authors. The corresponding author replied saying that the commercial company was a subsidiary of the charity and had no ownership rights to the drug. He did not specifically say whether there were any competing interests to declare. Instead, he wanted to know how this matter had come to light. During a telephone conversation he confirmed that the charity had applied for a patent for the use of the drug for the treatment of clinical condition A, but it was yet to be approved. He also said that no company had been licensed to develop/manufacture the drug for either condition. However, he also said that if the drug was ever licensed, the charity might choose to pass on some of the royalties to the inventors. It was suggested that the following competing interest statement should be published: “Authors 1, 2 and 3 are employees of the not for profit institute, which has applied for a patent for the use of drug X in the treatment of clinical condition A. The institute is a public charity that currently holds patents for the use of drug X in other clinical conditions, with Author 1 listed as an inventor.” - Should the authors have included a competing interest declaration on the manuscript? - If so, should it be any different to what was suggested? - In general, what should be done about studies on drugs that are potentially lucrative? Should the authors declare if they “might” make money? - In this case, what does “not-for-profit” really mean? - And what does all this mean for researchers at universities that make money through spin-off companies?
- In the interests of transparency the authors should have declared a conflict of interest. - The holder of a patent should declare that they hold such a patent even if any profits made go elsewhere, because they stand to derive an intangible benefit from being associated with that patent. - There is nothing wrong with having a conflict of interest, but it must be declared. This allows readers to decide for themselves the relevance of the conflict of interest to the paper’s conclusions. - It is good practice for journals to ensure that they have a clear policy on conflict of interest. - Some journals send the paper back to the original reviewers and ask them to comment on whether the conflict of interest would have changed their opinion on the paper. - Retraction of a paper for an undeclared conflict of interest would only be considered in very serious cases where this seemed to undermine the validity of the data. - The publication of a correction about the undeclared conflict of interest is usually the most appropriate course of action. - The reference to the not-for-profit institution was a red herring as there was a potential for the authors to still receive both financial and non-financial benefits from the product. - The editor should publish a correction stating the nature of the undeclared conflict of interest. - The journal should also consider whether they need to state their conflict of interest policy more clearly.
A new correction was drafted, which excluded the authors’ affiliation to a non-profit organisation. The editors contacted the corresponding author to let him know of the change, why it had been done, and to ask for his approval of the reworded correction. The new correction read as follows: “In our recent article, we failed to declare the following competing interests: A, B, and C are employees of X Institute, which has applied for a patent for the use of drug N in the treatment of haemorrhagic fever. The Institute currently holds patents for the use of N in other clinical conditions, with D listed as an inventor.” The corresponding author said that he would not agree with a declaration that did not mention that his institution is a public charity. He also stated that he felt the guidelines for declaring competing interests posted in the instructions for authors were not explicit enough. The editors responded that the “non-profit” status of the institution was not relevant, potential competing interests might still exist, and that it should be declared. The author refused to consent to a declaration that did not mention “non-profit” and wanted to contact COPE to discuss the matter in more detail. The author also threatened legal action if the editors proceeded to publish the correction without his consent. He suggested a new draft for the text of the correction. In our recent article , we did not declare the patent rights related to N technology. Since they may potentially be viewed as non-financial competing interests, despite the fact our employer is a non-profit public charity, we are now providing the following supplemental information: A, B and C are employees of X Institute, which has applied for a patent for the use of N in the treatment of haemorrhagic fever. X Institute currently holds patents for the use of N in other clinical conditions, with D listed as an inventor.” To clarify the journal’s position, the editors decided to post the following comment on the correction, which the author also agreed to: “[The publisher] thanks the authors for clarifying their competing interests and wishes to make clear its view that even employees of non profit public charities may have competing interests (financial or otherwise) and that it is always best to err on the side of declaring these. However, at the time of publication of the original article, our competing interests policy was not sufficiently explicit on this point. We will, therefore, be updating it in due course.”