Background: A redundant publication is one which duplicates previous, simultaneous, or future publications by the same author or group or, alternatively, could have been combined with the latter into one paper. As there is no information about the extent of this problem in the surgical literature, we set out to assess the incidence, spectrum, and salient characteristics of redundant publications in 3 leading surgical journals.
Methods: Original articles (excluding reviews, editorials, abstracts, and letters) published during 1998 in the journals Surgery, The British Journal of Surgery, and Archives of Surgery were searched by using the on-line search engine PUBMED. Each original article was scrutinized to identify redundancy by combining the names of the first, second, and last authors with a few key words from the title. Papers were defined as "suspected" redundant publications if they were found to address the same topic as the "index" article and shared some or most of the elements of methodology, results, or conclusions. The full versions of all suspected papers were retrieved and compared with the index articles. A grading system was developed to define several types of redundant publications: A. "dual"; B. "potentially dual"; C. "salami-slicing."
Results: A total of 660 articles were screened. There were 92 index articles (14%) leading to 147 suspected papers found in other journals, representing some potential form of a redundant publication. The vast majority of suspected papers were published within approximately a year of the index paper and were not cited by the latter. Most (69%) of the suspected papers were also published in surgical journals. Only 12 (8.1%) appeared in, or originated from, a "local-foreign" journal. Twenty (13.6%) of the suspected papers met the criteria for dual publications, 50 (34%) for potentially dual publications, and 77 (52.4%) were considered products of salami-slicing.
Conclusions: Almost 1 in every 6 original articles published in leading surgical journals represents some form of redundancy. Current on-line search technology provides an effective tool for identifying and tracing such publications, but it is not used routinely as part of the peer review process. Redundancies occur in several well-defined patterns; the phenomenon is widespread, and it cuts across the entire spectrum of surgeons in the United States and abroad. Redundant publications must be recognized not as a mere nuisance but as a real threat to the quality and intellectual impact of surgical publishing.