Statistics from Altmetric.com
Editor,—The editorial on research misconduct (Gut 1997;41:1–2) is timely. Duplicate reports, redundant publications and “salami slicing” are products of today’s environment where academics are often judged by the length of their curriculum vitae and number of publications,1-3rather than the quality of work and whether it has any impact on current medical practice. This necessity to publish may be due to institutional pressures, personal ambition, vanity, direct financial gain, or even psychiatric illness.1 The prevalence of fraud is estimated to be around 0.1–0.4% of research studies and over 70 cases have been documented.4 ,5 About 5% of drug trials are thought to involve misconduct of some sort.6Peer review offers little or no protection against fraud.7
A few years ago, as assistant editor of a psychiatry journal, a case of duplicate publication was brought to my notice, several months after the original article had been published. The authors had sufficiently disguised the second article and it was published in a journal from another specialty. There was some debate about what constituted acceptable overlap; the paper concerned fell in the grey area and it was felt that the extent of duplication was not sufficient to prompt a retraction. Apart from informing the editor of the second journal, no further action was taken to my knowledge. However, I believe that if we were made aware of the duplicate version earlier, the paper in question would not have been accepted. Combining both manuscripts would certainly have led to a more comprehensive paper. The threshold for retracting a published article is probably much higher than the threshold for rejection before publication.
In response to a similar situation, the editors of Annals of the Rheumatic Diseases have introduced the following statement in their information for authors: “Authors must declare, and submit copies of, any manuscripts in preparation or submitted elsewhere that are closely related to the manuscript to be considered.” I believe that this may not be sufficient as it provides a loophole as to what constitutes closely related, an area which authors have exploited. Often what editors consider “closely related” is disputed by the authors.9
Therefore, in addition to copies of closely related manuscripts, it may be useful to ask authors to submit a list of all their publications (related and otherwise), including those currently in the pipeline. Random audits can be done to ensure that there are no other papers which editors consider to be similar, which authors may not otherwise declare. These audits may be restricted to those authors who have a large number of publications (say, over 25 or 50) as they are more likely to have made a career out of such practice over several years. I emphasise that the vast majority of researchers with numerous publications would not have indulged in unethical research practice, but by definition, it is in this group that one would have a higher chance of finding expert salami slicers and those with overworked photocopiers. This method is certainly not fool proof and a clever cheat is difficult to expose7 —but this at least increases the level of deception necessary for duplicate publications and may deter all except those with antisocial personality traits, nurtured in an environment that looks for quantity, not quality.
At least one journal black lists authors involved in dual publication for a period of five years.8 However, this does not prevent recalcitrant authors from perpetrating the same fraud on another journal, considering that retraction notices are not always obvious or easily accessible through a computerised database such as Medline.9 Editors have already agreed on uniform requirements for manuscripts. It may now be time for editors to maintain a common register of authors who have been black listed by one journal (similar to the Food and Drug Administration’s black list of clinical investigators),6 so that others are aware of the same. As members of the scientific community, the only way to avoid the threat of a “science police” is to encourage science to police itself.10
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.