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.
The authors of a paper are in disagreement over whether the paper should be retracted. One group of authors (group 1) wishes to publish a correction, and another group (group 2) feel that is inadequate, and the paper should be retracted. Group 2 is concerned that one of the authors, author X, in group 1 is guilty of scientific misconduct. The remaining group 1 authors do not support this claim. The institution under whose auspices the research was conducted has carried out an investigation, and has apparently found evidence to believe that scientific misconduct has occurred. The institution has contacted the group 1 authors and has demanded that the authors retract the paper. The group 1 authors do not feel that the investigation has been properly conducted, and have declined to retract the paper. The institution now plans to contact the journal to request that the paper be retracted.
If the journal does not retract the paper, what other options are available to highlight the dispute between the authors?
Should this dispute be brought to the attention of readers, given that that case for retraction is inconclusive?
With the two institutions involved being in two different countries, this has not helped to improve communications between the two groups of researchers. It was suggested that the editor should contact the institution where the group 1 investigators are located, informing them of the dispute between the two groups of authors, and pointing out that there has been an investigation at the group 2 authors’ institution into the matter. The editor should also try to obtain the results of the investigation carried out by the group 2 authors’ institution before making any decision regarding whether to retract the paper or publish a correction. Publication of a statement of concern might be a consideration at this time, but such statements should only be published when there is very strong supporting evidence to take such action. Ideally, the two institutions need to contact one another and carry out a full investigation. Only when all the facts are available and agreed upon by both parties can the editor make a final decision.
The main institution has had two people review the case, and they agree that there are corrections that need to be published, but disagree about whether there is any misconduct. The institution is now proposing that we publish a correction, possibly with a statement from a couple of the authors to the effect that they are removing their names from the paper because they no longer feel that the conclusions are justified. We are now waiting for the institution to finalise this with the authors and then we will publish the correction.
The authors of a paper published in another journal wrote to the editor of Journal A, complaining of apparent blatant plagiarism of their work by N et al. , whose paper had been published in the journal earlier in the year. Further investigation revealed that the text of the two papers was almost identical. S et al. had used one drug and N et al. had used a different one of the same class. The published results in the second paper closely matched those of the first. The paper also seemed to have been copied entirely from the first paper, including the ethics committee approval. The editors wrote to N et al. asking for an explanation, evidence of the raw data, and copy of the ethics committee approval. The time line of ethical approval, submission, and publication meant that it would have been difficult to have recruited for, and completed, an eight week treatment study. N responded, stating that ethical approval was not required even though it was a double blind, placebo controlled study in children, and so had not been sought. The author also claimed that the lack of response from the ethics committee was synonymous with approval. The author then claimed to have sought ethical approval retrospectively, and a letter from the ethics committee was sent to the editors. When the editors attempted to contact this committee they were passed onto another ethics committee in a different area. The letter was sent to the author’s institution head. Unsigned letters and emails, purporting to be from the co-authors’ head of institution, were sent to the journal, and the author supplied an Excel spreadsheet detailing data from just 15 patients. A review of the author’s publication history revealed that s/he had changed “routes” over the past 5-6 years, publishing only fairly brief reports.
- The letter from the ethics committee chairman might have been fabricated. - A spreadsheet on 15 patients is unacceptable; the original data should definitely be available for such recent research. - The editors should write directly to both the current and previous institution heads. - The editors should also consider contacting the author’s regulatory body. - The journal should also write to the co-authors’ head of institution as they seemed to have taken complimentary authorship. - The journal should retract the article if it felt that there was sufficient evidence to suggest fraud; but if not, it should certainly publish a notice of concern.