A case study of salami slicing: pooled analyses of duloxetine for depression

Publishing separate, yet very similar pieces of a single dataset across multiple papers is known as 'salami slicing'. This practice may be motivated by researchers wishing to increase their publication counts and by the desire to increase exposure of their findings. 'Salami slicing' may also be used by the drug industry to help widely disseminate positive findings regarding its products. Journal editors across many scientific disciplines have bemoaned such duplicative publications on several occasions. However, little research has been conducted on the frequency of such publication practices, and findings have been inconsistent. No research has investigated whether 'salami slicing' may also occur in publications presenting results from pooled analyses of clinical trials.

We examined the scientific literature on duloxetine as a treatment for depression, examining how data from clinical trials were reported across 43 pooled analyses.

The vast majority of pooled analyses (88%) had at least one author who was employed by the manufacturer of duloxetine. Several pooled analyses based on highly overlapping clinical trials presented efficacy and safety data that did not answer unique research questions, and thus appeared to qualify as salami publications. Six clinical trials had their data utilized as part of 20 or more separately published pooled analyses.

Such redundant publications add little to scientific understanding and represent a poor use of peer reviewer and editorial resources.