Retraction or expression of concern?
Case text (Anonymised)
Shortly after publishing a short report, another group involved in similar work accused one of the authors (A) of the short report of fabricating and/or stealing data from their lab. The other group also stated that author A’s conclusions about an image published in the short report were wrong.
We asked to see author A’s original data and talked to his co-authors and the institution where his studies were reportedly carried out. We were satisfied that the data presented in the short report were real and the author’s own, and there was no evidence of data fabrication or theft.
We asked for expert opinion on author A’s interpretation of the image. Three experts thought that the author could not draw the conclusions he had based on the scan he presented in the publication and that reference images produced from the original data were needed to support his conclusions. We went back to author A, told him we were satisfied that there was no evidence of data fabrication or theft, but that we did think he needed to provide more data to support his claims. He did provide some additional images. However, our experts’ view was that the data provided did not verify the author’s claims. He had used images from published articles as reference images, and not reference images from his original data.
In the meantime, the other group submitted a correspondence article explaining how their own studies conflict with author A’s claims. This was sent for review. The reviewer felt that author A’s article should not have been published, but that the other group needed to provide some more experimental detail and data. At this point we approached author A telling him that we did not think the data he provided supported his claims and that he might want to consider retracting the article. Author A responded by sending us several opinions from ‘experts’ he had found arguing that the article should not be retracted. None of these ‘experts’ is an expert in the imaging technique used. He also said he would now be able to provide reference images from his original data, although we have not seen them and do not know, without checking with our experts, whether or not they would be enough.
The short report describes an invasive clinical intervention and makes claims about its efficacy. This is a controversial area of research, and our concern is about leaving something that may not be accurate in the public domain, but we also do not feel that the author has intentionally misled us or the public. We feel at this point that the best course of action is to publish the critic’s correspondence article, along with the authors’ response, and let the public judge for itself. However, given the clinical nature of the short report and the doubts raised about the veracity of author A’s claims, we would like the committee’s opinion on whether the publication of the correspondence piece and the authors’ response is enough, whether we would be justified in publishing an expression of concern about author A’s article or whether there are grounds to retract.
The Forum agreed that the editor had done the right thing by allowing the authors to comment. Exchange of correspondence is ideal in such cases as this will be linked permanently to the original article and so the debate will be in the public domain. The Forum did not think the article should be retracted. Also, the Forum advised against publishing an expression of concern as this indicates that there is something wrong with the data. All agreed that the editor had done all he could except perhaps to write an editorial on peer review and post publication comment.
We did not publish an expression of concern but encouraged the ‘other group’ to submit their revised correspondence. They decided they did not want to, so we have not taken any further action.