|
|
| |||||||||||||||||||||||
Our guidelines regarding redundant publication are published each week on the Information for Authors page. In a practice followed by many journals, we ask authors to send us copies of any manuscripts closely related to the manuscript they want us to consider for publication. This allows us to decide whether there is excessive overlap between two manuscripts or whether the results of a single study are inappropriately divided into two or more papers. In the trade, the latter practice is sometimes referred to as "salami slicing."
The reasons for preventing redundant publication are not arbitrary. As earlier editorials have pointed out, multiple reports of the same observations can overemphasize the importance of the findings, overburden busy reviewers, fill the medical literature with inconsequential material, and distort the academic reward system.1,2
The results of huge clinical trials or epidemiologic studies with multiple and unrelated end points, such as the GUSTO (Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries) trial of thrombolytic therapy after acute myocardial infarction, the Framingham Heart Study, or the Physicians' Health Study, could not be reported as a single study. It often takes years to collect and analyze such data, and it is legitimate to describe important outcomes of such studies separately. On the other hand, reports of studies involving several dozen patients should not be split into overlapping manuscripts. Because the line between appropriate and inappropriate practice is not always clear, it might be helpful to provide several concrete examples.
First, some examples of overlapping publications: Two years ago we accepted a paper on bone lesions in patients with chronic renal failure. We asked a distinguished nephrologist to write an editorial to accompany the paper. While preparing the editorial, the nephrologist came across a study published in a specialty journal several months earlier. It was written by the same authors, described the same patients, and reported virtually the same end points. The authors had not told us they had published similar data elsewhere. Although we were well along in the production process, we pulled the paper. This example of fragmenting the results of a single study and reporting them in several papers is not unique. Several months ago, for example, we received a manuscript describing a controlled intervention in a birthing center. The authors sent the results on the mothers to us, and the results on the infants to another journal. The two outcomes would have more appropriately been reported together. We also received a manuscript on a molecular marker as a prognostic tool for a type of cancer; another journal was sent the results of a second marker from the same pathological specimens. Combining the two sets of data clearly would have added meaning to the findings.
After we published a recent study describing diagnostic tests on 101 consecutive patients with suspected traumatic rupture of the thoracic aorta,3 we learned that the report was remarkably similar to two papers that had been published in the surgical literature.4,5 One, published only two months earlier, described 160 patients.5 The other, published two years earlier, described 69 patients.4 All three papers were from the same institution; three persons were listed as authors on all three papers, and two others were listed as authors on two of the papers. Before we accepted the paper for publication, we were not informed about the report on 160 patients, although it had already been accepted for publication elsewhere. The paper published two years earlier was also not brought to our attention by the authors, although we were aware of it because it was listed in the bibliography. No information was given in any of the papers about which patients were being reported on two or more times. In fact, as the letter from Drs. Smith and Kearney in the Correspondence section of this issue of the Journal indicates,6 some patients in fact were reported on in all three papers, providing a misleading impression of the number of patients studied and the value of the tests.
It is surprising that these practices still occur, despite growing attention and nearly universal disapproval. Most of the time the redundant publication is quickly exposed by readers. In addition, an investigator's peers often recognize a succession of "least publishable units." The motivation for publishing two or more papers when one would do is not always clear. In some instances authors have argued that they were interested in getting the information to different audiences. In others, they have claimed that they perceived the overlap to be far less substantial than did the editors. Finally, there is reason to suspect that the academic incentive system fosters a desire by authors to lengthen their bibliographies. Ways of counteracting this distorted incentive have been proposed7 but have not been universally implemented.
We are not eager to act as prior-publication police, and we do not regularly search the literature to determine whether an author has committed one of the several forms of redundant publication. But we have rewritten the relevant portion of our Information for Authors as follows: "Authors should submit to the Editor copies of any published papers or other manuscripts in preparation or submitted elsewhere that are related to the manuscript to be considered by the Journal" (we formerly asked for "copies of any related manuscripts"). We will continue to rely on the honesty and judgment of authors in informing us of any work of theirs that is related to a manuscript they are submitting to the Journal.
When preparing a manuscript, authors might heed the advice offered previously.2 In deciding whether reports are redundant, authors should ask themselves whether a single paper would be more cohesive and more informative than two. When there is any doubt, authors should submit with their manuscripts any other papers possibly representing duplication or fragmentation of results, whether published, submitted for publication, or already accepted for publication.
Jerome P. Kassirer, M.D.
Marcia Angell, M.D.
References
| |||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
