Randomized trials with industry-employee authors are more likely to report positive outcomes

Randomized clinical trials of interventional cardiology devices with industry-employee authors were significantly more likely to report favorable findings compared with randomized trials that did not have industry-employee authors, according to an analysis of prospective studies.

Lead researcher Nathaniel R. Smilowitz, MD, of New York University Langone Medical Center, and colleagues published their results online in JAMA Internal Medicine on March 28.

“Industry employees, academic coauthors, and journal editors should consider the potential implications and perceptions of [industry-employee authors] of scientific manuscripts,” they wrote. “Specific oversight of [industry-employee authors], with detailed disclosure of their role in a publication, may be warranted to ensure the reporting of evidence without industry bias.”

Giora Weisz, MD, one of the authors of this study and a cardiologist at Shaare Zedek Medical Center in Jerusalem, reported that he served on the medical advisory boards of Angioslide, AstraZeneca, Calore, Corindus, Filterlex, Medtronic, Medivisor, M.I. Medical Incentives and Vectorious. He also received research grants from Angioslide, Corindus and Mitrazyme. The other two co-authors did not report any conflicts of interest.

For this analysis, the researchers searched PubMed for prospective studies of cardiovascular devices published in nine journals between Jan. 1, 2010 and Dec. 31, 2012. They included studies that evaluated the clinical performance, efficacy or safety of interventional cardiology devices.

Of the 357 studies that met the inclusion criteria, 21.8 percent listed at least one industry-employee author and 4.0 percent reported industry employment. Further, 11.2 percent reported two or more industry-employee authors and 6.2 percent reported three or more industry-employee authors. Studies were more likely to have an industry-employee author if they were funded by industry, had a funding source in the U.S. or registered with clinicaltrials.gov.

In addition, 80.4 percent of the studies and 66.7 percent of the randomized trials reported positive study outcomes. The authors mentioned there was no difference in positive study outcomes in articles with or without industry-employee authors (87.2 percent versus 78.5 percent), although the difference was statistically significant between randomized trials (87.8 percent with industry-employee authors and 58.9 percent without industry-employee authors).

Of the 101 industry-sponsored randomized trials, 87.5 percent of those with industry-employee authors and 65.5 percent of those without industry-employee authors had positive outcomes. After adjusting for multiple variables, industry-employee authors were independently associated with positive study outcomes, according to the researchers.

They also mentioned that 66.1 percent of manuscripts reported a conflict of interest. However, the proportion of studies with positive outcomes was similar in studies that reported a conflict of interest (80.9 percent) compared with those that did not report a conflict of interest (79.3 percent).

The researchers mentioned that the role of industry-employee authors is not always disclosed. They added that industry employees might only be interested in authoring manuscripts reporting positive outcomes.