Healthy Skepticism Library item: 17031

Warning: This library includes all items relevant to health product marketing that we are aware of regardless of quality. Often we do not agree with all or part of the contents.

Publication type: news

A dubious practice, any way you slice it
The Toronto Star 2010 Jan 15
http://www.thestar.com/sciencetech/science/article/751006--a-dubious-practice-any-way-you-slice-it?bn=1

Abstract:

Study says data on drug was cut up and recycled several times to create extra articles in journals

Full text:

Glen Spielmans was watching TV one night when a commercial for the antidepressant Cymbalta came on.

The frequently seen ad talked about how “depression hurts” and offered Cymbalta, a blockbuster drug by pharmaceutical giant Eli Lilly, as a way to alleviate the pain.

Spielmans, an associate professor of psychology at Metropolitan State University in Saint Paul, Minn., decided to investigate the claim and ended up stumbling onto a growing area of concern in academic circles.

Salami slicing.

Simply put, salami slicing is the practice of carving up the data from clinical trials into smaller bite-sized pieces and then using those slices to write several papers for publication in competing academic journals.

As he went through every study he could find on Cymbalta – which is also known by its generic name, duloxetine – Spielmans began feeling as if he was reading the same study over and over again, sort of an academic déjà vu.

“There just aren’t that many data to do all these secondary analyses on,” Spielmans said in an interview. “That raised a flag for me.”

So Spielmans dug deeper and, last week, co-published a study in the journal Psychotherapy and Psychosomatics looking at 43 Cymbalta articles containing recycled data from only a handful of clinical trials. Two of those trials were recycled some 33 times each.

Most of the studies used slices from more than one clinical trial. One grouping of six original clinical trials was sliced up for use in some 20 separate, but very similar, journal articles.

The most extreme example was a group of authors using one database to write two separate articles for competing journals: the Annals of Clinical Psychiatry and the Journal of Clinical Psychopharmacology. One study was based on eight weeks’ observation of patients taking the drug, while the other was based on 12 weeks’ observation.

Not surprisingly, the two papers came to identical conclusions – that patients benefit by switching to Cymbalta from other drugs.

Spielmans says such practices distort the medical literature, making a drug appear to have more scientific support than it actually deserves.

Not all academics are convinced the salami slicing situation is dire, however.

“I would be careful of anything that’s given a folksy name,” says Dr. Ralph Meyer, director of the National Cancer Institute of Canada’s Clinical Trials Group at Queen’s University.

There are extreme examples in which data are overly sliced up to generate extra journal articles for the researchers, says Meyer, who has run several clinical cancer trials in his career.

But there can often be good reasons for dividing data into smaller chunks. Clinical trials are so costly to run, they are often designed with the intention of dividing the data up later.

“Clinical trials are a lot of work. They’re a lot of expense,” he says.

Trials that Meyer has run, for instance, have looked at such standard questions as the safety and effectiveness of a new treatment but have also used the opportunity to examine the cost-effectiveness of the procedure or the treatment’s impact on a patient’s quality of life.

“These are issues that can warrant a separate manuscript,” he says.

Jocalyn Clark, a researcher at the University of Toronto and an editor at the journal PLoS Medicine, says top journals like hers require all authors to declare if any part of the underlying research has been – or soon will be – published elsewhere.

Universities and hospitals need to make it clear to their researchers the practice is not acceptable, she says.

Clark says academic journal editors are increasingly taking a dim view of researchers who take part in salami slicing.

“If something fishy is found out about an author, editors and journals get very, very upset,” she says. “Your reputation can be damaged.”