That Sounds Familiar

What’s behind paper retractions? (10)
by Adam Marcus and Ivan Oransky, Labtimes 04/2012

Photo: zettberlin / Source: PHOTOCASE

When is using some of the same data acceptable?

A research group conducts a study of the effects of treatment Y on disease X in 18 patients and publishes their findings in the

Journal of Medical Plausibility

. A year later, they submit a second paper looking at treatment Y and disease X, this time in 27 patients – the initial set plus nine more. In both cases, the scientists report similar rates of efficacy and side effects.

Question: Have they committed publishing misconduct?

True duplication of a paper is a straightforward issue. As a 2004 editorial in the Journal of Antimicrobial Chemotherapy noted, “Duplicate publication is unacceptable because it has the potential to skew the evidence base. If publication of the same data is not openly acknowledged, readers will reasonably assume that two different studies are being described and are likely to place greater reliance on evidence that appears to have been corroborated in more than one study. It can also affect the outcomes of meta-analyses used to establish best practice.” In a recent issue of The Oncologist, editor Bruce Chabner echoed these comments.

We certainly have seen this first-hand. In one example we covered on our blog, two groups of researchers in Korea published a study in related journals on the same group of cadavers. The result was a retraction. We also wrote about the retraction of a case study, again from Korea, by researchers, whose colleagues in a separate department at the same institution had published the report earlier. This time, the papers appeared in the journals two disparate medical specialties – dermatology and plastic surgery.

But the question becomes a bit murkier when it involves marginal but real additions to an original study. The problem has become particularly acute with the arrival of software programmes and online services that screen for signs of plagiarism. Spend a little time on sites like Deja vu – a paper-screening system that bills itself as a “database of highly similar citations” – and it’s easy to become convinced that the scientific literature is overrun with kudzu.

As the operators of the site (which is based at Virginia Tech University in the United States) note, however, “There are many definitions of the terms ‘duplicate publications’ and ‘plagiarism’, and the characterisation of a citation or document is subjective. All entries in Deja vu are based on a measure of similarity and are presented in such a way as to enable the user to quickly compare citations. The user must make the ultimate determination as to how they wish to use this information.”

So, where does that leave researchers and journals? We’re afraid we don’t know. Consider the hypothetical we raised earlier: Admittedly, the follow-up study added little to what the researchers already knew about the intervention. But some physicians who read the second report might be reassured by the confirmation that the drug seems safe and effective in a larger population of patients.

Does the publication pad the literature? Perhaps. It’s not typical “salami slicing”, since the later data wasn’t available when the first paper was written.

Hardly irrelevant

But is it mere fluff? Not necessarily. After all, studies that look at subsets of previously investigated populations might not carry as much weight as the entire group but they’re hardly irrelevant.

We doubt this issue is going to resolve itself anytime soon. So, in the meantime, researchers should be as transparent as possible about the provenance of their manuscripts. If they are trying to publish a follow-up analysis, let the journal editors know that up-front. It might help avoid a retraction later. As the International Committee of Medical Journal Editors puts it:

“When submitting a paper, the author must always make a complete statement to the editor about all submissions and previous reports (including meeting presentations and posting of results in registries) that might be regarded as redundant or duplicate publication. The author must alert the editor if the manuscript includes subjects about which the authors have published a previous report or have submitted a related report to another publication. Any such report must be referred to and referenced in the new paper. Copies of such material should be included with the submitted manuscript to help the editor decide how to handle the matter.”

We’ll duplicate – we mean, agree with – that.

(The authors run the blog Retraction Watch:

Last Changed: 13.07.2012