Although frowned upon, “salami slicing” — the practice of cutting up of a single set of research data into many different papers for publication — is well known in academia. Researchers, particularly ones starting out in their career, do it mostly because of the intense pressure to “publish or perish.”
But salami slicing is used for more than boosting the careers of young academics hoping for tenure. Pharmaceutical companies carve their research into these “least publishable units” for purely financial reasons: to promote and market their products.
And they do it rather brazenly, as Glen Spielmans, PhD, an associate professor of psychology at Metropolitan State University in St. Paul, reports in a study published last month in the journal Psychotherapy and Psychosomatics. [Full disclosure: I teach in the writing department at Metro State.]
In the study, Spielmans examined 43 pooled analyses (studies that combine data from several clinical trials into a single, larger database) of the anti-depressant duloxetine, marketed by Eli Lilly. He found that at least 20 of those separately published analyses used data from the same six clinical trials, and several failed to answer any unique research questions.
Amazingly, six of those pooled analyses included data from only two clinical trials. That raises the question, as Spielmans and his co-authors note in their study, “of why pooled analyses based on only 2 trials warrant publication in the absence of novel scientific data.”
Also revealing (but perhaps not all that surprising) was the finding that 88 percent of the analyses had at least one author who was employed by Eli Lilly. That suggests that the company was quite aware of how the data was being repackaged and repurposed.
“Salami publications may be more representative of propaganda than of actual contributions to science,” Spielmans and his co-authors concluded.
To impress physicians
“There’s a lot of redundancy in the analyses,” Spielmans told me in a recent phone interview. “They’re recycling the same studies’ data.”
Here are some examples of such recycling, as described by Spielmans and his co-authors in their study:
[T]he potential interaction of gender and race with treatment response is certainly a topic of interest. However, given that Latinos, Blacks, and Whites all tended to respond similarly to duloxetine and that gender also did not moderate outcome, it seems that such data could easily have been presented in a single publication rather than in 3 published pooled analysis. Likewise, data on safety outcomes were sliced into several pieces. It is unclear how an individual publication is needed to report data separately for safety outcomes such as nausea, sexual side effects, urinary side effects, and several others.
Why would drug companies dice the data into so many different pieces?
“The industry in their trade publications makes it really clear that the reason they do studies is to sell product, and that’s it,” said Spielmans. “The more publications [your studies appear in], the more you’re going to be read by physicians who can prescribe your drug.”
But numbers matter for another reason. “Doctors get visited by drug reps who hand them a list and say, ‘We’ve done dozens of studies.’ Doctors, of course, are going to be impressed with that,” Spielmans explained.
What the doctors are unlikely to know, unless they read the studies very, very carefully, is that the data for many of those studies come from the same small set of clinical trials.
A widespread practice
Eli Lilly is not alone in slicing up its data for different publications. “I think if people really wanted to dig through the studies, they would find this practice across a wide variety of drugs,” Spielmans said.
In fact, Spielmans decided to do this study after researching the effectiveness of another anti-depressant. “I kept running into what seemed to be redundant studies, and I thought, ‘I should look into this later,’ ” he said.
What does his findings mean for health consumers who want to make sure they’re taking appropriate medications?
“It’s very important for consumers, if possible, to learn a little about research and about how to understand the findings of research,” Spielmans said. “Theoretically, their doctors are doing that for them. But, in the end, it makes sense for a consumer to be his or her best advocate when finding out which treatments are safe and effective.”