Skip to content

The investigation of conflicts of interest

The Daily talks to a McGill graduate student about uncovering bias in pharmaceutical studies

Studies of studies – meta-analyses – are important to medical research: they help physicians, clinical policy makers, and the keen consumer by consolidating large amounts of information done in separate trials. Clinical guidelines are considered more valid when they are based on a meta-analysis rather than on just a few individual case studies. But Michelle Roseman, a McGill graduate student in Psychiatry, found that in the aggregation of information, something critical is lost. When a study is published, its authors are required to include their source of funding and any ties they might have to the pharmaceutical industry. However, unlike trial results and methods, this information about the individual trials is not often included in the meta-analysis. Roseman looked at 29 meta-analyses published in high-impact medical journals, and found that only two reported who funded the trials, and seven included only trials that were industry-funded. Not one discussed the author’s personal financial ties. The Daily sat down with Roseman to discuss her role in the investigation of conflicts of interest, and the importance of making their existence known.

The McGill Daily: Is the presence of conflicts of interest in these studies something that’s of concern in the research world?

Michelle Roseman: It wasn’t on the minds of the authors of the meta-analyses. We sent an email to survey the authors of the meta-analyses that we included: we asked whether or not they had extracted this information as part of their protocol – regardless of whether or not they ended up publishing it. Most said that they thought that was a great idea, but they hadn’t thought of it.

MD: How did you decide to do this particular study?

MR: My supervisor was looking to do a meta-analysis, and he was in contact with one of the major journals, and was discussing whether or not they would be interested in this. In their discussions, the journal said, “Actually, too many of the studies that you would want to include are industry-funded.” This prompted him to want to study this systematically. I was a research assistant in his lab at the time – he discussed the project with me, and it sounded really interesting. I became a grad student in the lab and continued to work on it as my thesis project.

MD: What are you hoping people who read your study will take away from it?

MR: We’re hoping that researchers will start including the funding sources of the individual trials. There are guidelines that already exist on how to write up a meta-analysis, called the Prisma guidelines –  the standard in the industry on what information should [be] included in a meta-analysis. What we suggested is that these guidelines need to be updated.

MD: How might this change the way physicians operate?

MR: In many cases, there might only be industry-funded studies of a drug – the industry currently funds about 70 per cent of the clinical trials that occur. But certain meta-analyses could be looked at with a more critical eye.

– Compiled by Shannon Palus