If your profit margins depends on a supposedly “gold standard” objective scientific process, guess how long it will take before you start imagining ways that process could be manipulated?
If you answered, “Not very,” you’d be right.
The process is peer-reviewed journal articles, of course, which are the “gold standard” for health care research. The theory is that if researchers review other researchers’ work and vet it before publication, only the good stuff will make it to publication (and if things need clarification or further disclaimers, it often comes out in the review process).
So how do you manipulate such an objective process? Well, you could start at its foundation. The study data itself.
The problem with today’s study data and research is that a lot of researchers don’t keep their hands wet in statistics — they hand off the work to statisticians or researchers with a specialty in statistics. That means that, as a researcher, you may never even look at the raw data in a huge multi-center clinical trial. A statistician or graduate student handles all of that, manipulates it with different analyses, and presents the research with the results of the analyses.
But most researchers choose their own statisticians, colleagues they often work side-by-side with for years or even decades.
What happens when you take the heavy lifting off of a researcher’s shoulders and just give them the final, cleaned data results?
What if the summary data were provided by one of the drug companies whose drug you were studying? Hmm… See a possible conflict of interest here?
CL Psych has the full, sordid story describing how GlaxoSmithKline (GSK) apparently manipulated researchers studying the link between suicide and one of its popular antidepressant medications, Paxil: Key Opinion Leaders and Information Laundering: The Case of Paxil.
We found CL Psych’s analysis insightful and incriminating, and well worth the read of researchers trying to defend their choice of not looking into the raw data themselves:
If you are an academic researcher, and you simply take data tables from drug companies then reproduce them in a report and/or publication, you are not doing research — you are laundering information. People think that you have closely examined the data, but you have not, and you are thus doing the public a disservice.
Calling it like it is…!
The peer review process breaks down the minute researchers stop doing their job and rely on others to do a part of their job for them. And we believe journals shouldn’t be publishing articles where the researchers don’t have a clear “chain of custody” with regards to the data they are studying. That is, researchers should always be able to know and defend exactly how their data was collected, stored, processed, analyzed, and compared.