Industry Authors and Positive Studies: Analysis Makes the Link in Cardiovascular Device Trials

One in five published papers evaluating interventional cardiology devices list at least one coauthor employed by industry, according to a new analysis of the literature from 2010 to 2012. That proportion rises to more than one-quarter when looking at only randomized controlled trials, with such studies being more likely to report positive outcomes than those lacking industry collaboration.

Yet senior author Giora Weisz, MD (Shaare Zedek Medical Center, Jerusalem, Israel; Cardiovascular Research Foundation, New York, NY), cautioned to TCTMD in an interview against assuming these results point to a problem. “Definitely we shouldn’t look at this as something we want to discredit,” he said. “There are many very high-level scientists that work for industry, and they deserve to publish their study results like anyone else. There’s a lot of good science going on industry.”

Weisz along with coauthors Nathaniel R. Smilowitz, MD (NYU Langone Medical Center (New York, NY), and Altaf Pirmohamed, MD (NewYork-Presbyterian/Columbia University Medical Center), published the findings online earlier this week in JAMA Internal Medicine.

That does not mean, though, that the patterns should go without merit further scrutiny, the investigators say. “Industry employees, academic coauthors, and journal editors should consider the potential implications and perceptions of [industry-employee authors] of scientific manuscripts,” they write. “Specific oversight of [such employees], with detailed disclosure of their role in a publication, may be warranted to ensure the reporting of evidence without industry bias.”

Looking for Links

The researchers identified 357 studies evaluating interventional cardiology devices that were published in nine of the field’s top journals between 2010 and 2012. One-fifth (21.8%) listed one or more industry employees, who represented 4.0% of coauthors named on the papers. Furthermore, 11.2% of papers listed at least two industry coauthors and 6.2% at least three.

Overall, 80.4% of papers reported positive outcomes, with no difference based on whether or not they cited industry employees as coauthors.

Looking solely at randomized controlled trials, 26.8% had at least one industry-employee authors. Two-thirds reported positive outcomes (66.7%). The presence of industry-employee coauthors on papers related to randomized studies was associated with an increased likelihood of reporting positive results (RR 1.49; 95% CI 1.23-1.81). This pattern was also seen in industry-sponsored randomized controlled trials (RR 1.33; 95% CI 1.07-1.66).

On multivariable adjustment, there were two independent predictors of whether a study would have positive outcomes:

• Being a randomized controlled trial (adjusted OR 0.20; 95% CI 0.10-0.41)
• Having at least one industry-employee author (adjusted OR 2.66; 95% CI 1.10-6.40)

Interestingly, whether a paper did or did not contain author-reported conflicts of interest had no bearing on positive outcomes (adjusted OR 0.86; 95% CI 0.43-0.17).

According to the researchers, “These findings add to evidence that industry collaboration is associated with study outcomes.”

To TCTMD, Weisz stressed that it cannot be assumed that the relationship between industry collaboration and positive findings is causative. “One potential interpretation of the results,” he suggested, “may be that when industry employees see that the study they’re involved in is positive, they say, ‘You know what? I want to be a . . . coauthor of that, because it’s nice to [put] on my resume and show I was associated with something successful.’ If the study is negative, then the employee says, ‘You know what? It doesn’t matter.’”

It can be difficult to tell the degree of a coauthor’s involvement in a paper, or to know whether someone who contributed to the work was left unmentioned, the researchers point out. “In some cases [they] may be ‘honorary’ authors who have not substantially contributed to the design of the study being described in a published report, to the acquisition or interpretation of the study data, or to drafting or revision of the manuscript and do not meet accepted authorship criteria,” Weisz et al write. “Conversely, many industry employees involved in studies may remain uncredited.”

Note: Weisz is a faculty member of the Cardiovascular Research Foundation, which owns and operates TCTMD.

Related Stories:

• A Ray of Sunshine: CMS Experts Give an Update on 2 Years of Public Disclosure
• Cardiologists Stand Out for Their Ties to Industry in US Government Database
• Stent Literature Rife with Financial Disclosure Problems