Duplicate and salami publications.P Abraham
Keywords: Duplicate Publication, Ethics, Human, Periodicals,
Though there is a high degree of trust and veracity in medical publications, an occasional report of duplicate and salami publications bothers the medical fraternity. In recent years, there have been more than 100 cases of notices of duplicate or redundant publications in various journals. Byron Bailey, Editor, Archives of Otolaryngology - Head and Neck Surgery, found that of 1000 authors chosen at random from his journal, 201 published 644 articles which could be labelled as duplicate. During their work on impact of duplicate publication on meta-analysis, Tramer et al found that 17% of the manuscripts were covertly duplicated. Similarly, during a meta-analysis on exogenous surfactant, Raju found 6 of 35 articles to contain overlapping data.
Duplicate or redundant publication has been defined as publication of a paper that overlaps substantially with one already published. To avoid confusion and the possibility of authors finding flaws in what is ‘substantial’, a few journals have provided more specific definitions. For practical purposes, any article that has similar or near similar hypothesis, sample characteristics, methodology, results and conclusions to a published article is a duplicate article, and if it is republished without the knowledge of the editors, it is called duplicate publication. The authors of the duplicate manuscript may be the same, as in most cases, but it may also be a different author(s) publishing the same article without the knowledge of the initial author(s).
Publication of a manuscript in any form, electronic or print, in a journal or scientific booklet of a conference, and in any language, constitutes a prior publication. Prior publication of only the abstract of a manuscript may not be considered as duplicate publication; however, the editor of the journal should be informed about it. Publication of the full article in a different language and for a different reader base can be permitted provided the editors of both the journals agree for the same and consider it beneficial for the readers.
On the other hand, slicing of data from a single research process or gathered during a single study period, into different pieces, creating individual manuscripts from each piece, and publishing these to different journals or even the same journal is called ‘salami publication’ or ‘salami slicing’. Publishing the radiological findings of cases of cerebral hydatid disease to a radiological journal and the neurosurgical aspects of it to a neurological journal is salami slicing. Exceptions are large clinical or epidemiological studies with results that cannot be published simultaneously or are such that they address different and distinctly important questions. For other studies focused to a small population size, with overlapping results, it is best to publish the results together to form a single more cohesive and informative paper than two, without being excessively long.
Apart from the desire to see voluminous curriculum vitae, number of publications is seen as the fast-track route for promotions and academic advancement. Overinflated bibliography is also utilised for grant sanctioning at places where length rather than the quality of one’s bibliography is a selection criteria. An occasional individual may claim that the duplicated article is not ‘substantially’ similar, he was not aware of the rules, the previous publication was in a public domain and not protected by copyright rules, or was in a journal not indexed with Index Medicus or other databases.
Waste of finite resources: Journals have limited number of pages and resources; one duplicate article means that another deserving article may not be published. In peer-reviewed journals repetitive publications eat away the time and effort of reviewers. Duplication of articles doubles the efforts indexing authorities have put in, for initially indexing the article and then retracting it. In fact, MEDLINE has duplicate publication as one of the publication types.
Overloading medical information: With unnecessary and repetitive data, there is inflation of the already overloaded medical information base, thereby increasing the time to search for relevant information. As of today, there are more than 10 million documents indexed with MEDLINE; if about 17% of these are duplicate, imagine the saving in resources by deleting them.
Overemphasis of findings: Tramer et al showed that duplication led to an overestimation of ondansetron’s antiemetic efficacy by 23%. Researchers should recognise the overlap of data during a meta-analysis; the validity of meta-analyses done without looking into this problem is questionable.
Legal question: Duplicate publications are serious violations of copyright laws.
Last but not the least, they are unethical.
In case of salami publications, readers have to read more than one article to get full information and a clear idea.
At institutes where promotions, rewards and grants are given based on the number of publications, a non-meritorious person may get ahead by virtue of an over-inflated bibliography.
With the availability of computerised medical databases, including databases of dissertations and scientific meetings, and various Internet-based search engines, it will be difficult for authors to hide their crime. On finding a duplicate publication, journals can publish a notice about it. The notice is cited in MEDLINE, and this can have serious impact on the prestige of the author; the effort of inflating the curriculum vitae goes in vain. The journal editor can send the information to other journals, and to the head of the institute, who can impose a ban for a specified period from publishing. Like-minded editors have come together to condemn this unethical practice and formed groups to help each other in the investigation of such cases and to ban such authors in their respective journals. At the most grievous end, authors can face a civil suit for infringement of international copyright laws.
A majority of biomedical journals now follow the Uniform Requirements for Manuscripts Submitted to Biomedical Journals and supplemental statements from the International Committee of Medical Journal Editors. These clearly state that ‘manuscripts will be reviewed for possible publication with the understanding that they are being submitted to one journal at a time and have not been published simultaneously or already accepted for publication elsewhere’.
Journals can ask authors to submit ‘copies of any and all published articles or other manuscripts in preparation or submitted elsewhere that are related to the manuscript to be considered by the Journal’, as is being done by the New England Journal of Medicine, instead of just asking for ‘copies of any related manuscript’.
Journals should insist that all authors sign the copyright agreement transferring the entire rights to the Journal.
In the event of discovering a duplication of publication, the editor of the other journal should be promptly informed and both the manuscripts should be sent for review to one or more independent referees to opine on the possibility and degree of overlap. The corresponding author(s) of both the manuscripts are informed about the incident and their explanation should be sought. If the conclusion of this procedure is that the mishap has occurred, a notice of the same with the cross-reference should be published in both the journals at prominent places. With the occasional incident of surprise-authorships coming up, it may be worth informing about the affair to all the authors separately. The head of the institutes of the authors should also be intimated. Depending on the policy of the journal a ban on publication of articles by the concerned authors can be imposed.
Though not everybody agrees, the journals can have an active role in prevention. During the editorial review process, a few journals provide reviewers with MEDLINE abstracts. At that time, an added vigil can be kept for the possibility of duplicate publication. The ethics committees of the journals can review for the possibility of duplicate publication. This should be very easy with the use of the ‘Related Articles’ feature of PubMed.
Organisations and institutes
One of the prime factors leading to duplication and slicing of publications is the role publications play in obtaining grants and promotions. If our system can be modified to look for quality rather than quantity, the need to have large number of publications will be obviated. A researcher should be asked to list five of his publications according to his choice and the decision to sanction grant or promotion should be based on these five.
Some credit should be given to how many times a person was acknowledged for his help by his colleagues in their publications. This should help in reducing the number of ghost and gift authorships as well as duplicate publications.
Institutional and departmental heads and local ethics committees should have a supervisory role over the publications from the institutes. The institutes and/or departments should keep a record of publications from their researchers and academicians. Prompt investigation and follow up by the institutes are a must in the event of a report from a journal editor about duplicate publication.
Indexing authorities should publish complete details of duplicate publications in their database, including the titles of both journals and names of all the authors.
Readers and reviewers
Readers, reviewers, and anyone who is in touch with recent medical literature should promptly inform both journals, if they find or suspect such an incident.
Authors are key persons in this entire issue. They should understand their moral responsibilities as well as respect the rights of journals to receive original manuscripts. The authors’ act is disservice to the readers and is a breach in the agreement with the journal to provide authentic data.
Journal editors rely on authors’ sincerity and judgement in informing editors about the possibility of similarity of their earlier publications to manuscripts under consideration with the journal. In case authors have doubts about these matters they should clarify the situation with the journal editor, must submit copies of any closely related manuscript, and must inform about prior publication of abstracts and publication in any other language or format.
‘If you have but one great study, submit it in its best, most complete form to the most appropriate journal. Do not publish in instalments, do not publish it in variants, do not publish it more than once’ .
There is a lack of notices of duplicate publications in Indian medical literature, giving a false impression that the Indian medical community is very faithful. It is hard to believe that this practice is not followed by Indian academicians. Considering this to be a global problem, there is an urgent need for Indian biomedical journal editors to come forward to form a group to fight this crime. There is a need for a consensus among Indian journal editors on steps to be taken in case such incidents occur and for their prevention.