Using UNSILO tools to detect salami slicing

Salami slicing (also known as “redundant publication” is the colourful term to describe publishing multiple articles from a single research project, when a single article would be more sensible. Why would researchers do this? One argument is that pharmaceutical companies sometimes try to improve the credibility of a new drug by publishing not one, but 20 or 30 scholarly articles about it. All the articles come from the same team of authors, and all relate to the same underlying trial, but medical practitioners, when approached by pharmaceutical company sales people, tend to be more convinced if many papers appear to demonstrate the efficacy of the new drug. And, of course, for any researcher, to have multiple publications to your credit does not do your academic reputation any harm.

Identifying examples of salami slicing is a challenge. A 2013 article (Smolčić 2013) stated “There is no software application or algorithm for detection of salami publication. Identifying this type of publication misconduct is complex because salami publications do not often include text plagiarism so that manuscripts can easily evade strict software checking.”

However, the UNSILO Technical Checks provide, I believe just such a (very necessary) check.It would be all but impossible for any human editor to identify potential examples of salami slicing without a close knowledge of the subject and an encyclopedia awareness of what has recently been published. Nor could a simple string matching algorithm be successful, because such articles are unlikely to have any identical phrasing. Nonetheless, there is a real problem to be identified, and we believe the UNSILO Technical Checks can help address it.

As an example, I attach an article about a drug “Duloxetine” that was identified in a 2010 paper as an example of salami slicing. The paper that identified the problem is “A case study of salami slicing: pooled analyses of duloxetine for depression.” (Spielmans 2010). Using one of the papers identified in that article we fed it into the UNSILO Technical Checks and ran the diplomatically named check “related papers by manuscript authors”. This is how it appears:  

As always with the UNSILO checks, we make no claim that the articles identified are actually evidence of salami slicing: we leave that for the publisher to judge. But we feel that the articles revealed here largely appear to correlate the Spielmans article, which identified 43 “pooled analysis” articles that, thesy claim, frequently reused material from a much smaller number of clinical trials. As Spielman writes: “For example, one study compared the safety and efficacy of duloxetine in the treatment of African-Americans to Caucasians. Using data from the same underlying clinical trials, another publication compared duloxetine’s efficacy and safety between Hispanics and Caucasians.”

We would welcome feedback from publishers and users of the tool with more examples of suspected salami slicing.

References

Smolčić, V.Š., 2013. Salami publication: definitions and examples. Biochem Med (Zagreb) 23, 237–241. https://doi.org/10.11613/BM.2013.030

Spielmans, G.I., Biehn, T.L., Sawrey, D.L., 2010. A case study of salami slicing: pooled analyses of duloxetine for depression. Psychother Psychosom 79, 97–106. https://doi.org/10.1159/000270917


Receive an email every time we publish a new blog post