A group of authors from the same specialty unit published a study in Journal A on all prehospital X procedures. They then sent another paper on X procedure in a subgroup of patients to Journal B. Paper B references paper A, but does not make it apparent that there is any overlap in these studies. On questioning by editor B, they stated that no patients in paper B were included in the previous study. Paper A studied all prehospital X procedures between February 1998 and February 2001 and states that all patients requiring X procedure were included. Paper B includes a subgroup of patients requiring X procedure between March 1998 and March 2002 and states that all patients were included. The two papers have similar methodology, use the same equipment and analysis; large sections of the text are identical; half of the references are the same; and the patients come from the same geographical area. The two papers cover overlapping periods and are undertaken by the same organisation. Both state all patients are included; no mention is made of any exceptions. However, the authors state that no patients were included in both studies. Therefore either patients from the subgroup in the period March 1998 to February 2001 were included in both studies, or the subgroup were somehow allocated to only one study. Any such allocation is not described in the text. The authors clearly state that there is no overlap of patients between the two studies but the editors feel that the article should not be published. Is this course of action correct? Should this be explored further to determine if the patient groups are completely different. Should a request to see the original database be made?
_ There appeared to be some evidence of misbehaviour on the authors’ part. _ The editors need to go back to the authors and explicitly challenge them on their assertion that there were two non-overlapping patient sets. _ In some situations it was permissible to publish studies of subgroups, but there had to be full disclosure of that fact and very good reasons for doing so. _ If the editors request the raw data, the journal rather than the institution should analyse them first. _ It is preferable to request the raw data in electronic format. Investigating raw data can incur substantial costs and it is the institution’s responsibility to investigate its own staff. _ But the editors should make it clear that if there are still unhappy with the explanation they will contact the authors’ institution to request it also reviews the researchers’ raw data. _ How can the editors pursue this course of action if they do not wish to publish the paper? If the editors told the authors of their intention not to publish, their position would be weakened and the authors might not bother to reply. _ Editors are privileged whistleblowers as they are harder to attack than a colleague expressing concerns over someone else’s work. Also, authors believe that editors are powerful and so it is still possible to get a reply from authors even if the paper is no longer in consideration. _ Ultimately, editors have the ability to publish an account of any misconduct in their journals. _ Editors should make it clear to authors when they did not wish to publish a paper. _ Editors should also make it clear to authors that there are concerns about a paper and that they would still pursue the issue even after rejection.