More disturbing anecdotal evidence that everyone is using ‘science’ to their advantage:
Drug companies fund a growing number of the studies in leading psychiatric journals, and drugs fare much better in these company-funded studies than in trials done independently or by competitors, researchers reported Wednesday.
About 57% of published studies were paid for by drug companies in 2002, compared with 25% in 1992, says psychiatrist Igor Galynker of Beth Israel Medical Center in New York City.
His team looked at clinical research in four influential journals: American Journal of Psychiatry, Archives of General Psychiatry, Journal of Clinical Psychiatry and Journal of Clinical Psychopharmacology.
In the report, released at the American Psychiatric Association meeting in Toronto, reviewers did not know who paid for the studies they evaluated, Galynker says. There were favorable outcomes for a medication in about:
• Eight out of 10 studies paid for by the company that makes the drug.
• Five out of 10 studies done with no industry support.
• Three out of 10 studies done by competitors of the firm making the drug.
The findings don’t prove the companies are knowingly biasing studies, says co-author Robert Kelly Jr., also with Beth Israel. The report didn’t look at the evidence for bias in design of the studies.
Some quick caveats- it is important to understand that just because researchers have reports funded by drug companies that provided more positive findings, that does not necessarily mean they are fixing data or acting unethical in any way. In social science research, it is entirely possible to shape the outcome of a study simply through the research questions asked, and that is more than likely what is happening here. University researchers receive a grant to do a study, the research questions are provided for them and/or the design, they complete the study, and because of the nature of the questions asked, more often than not industry funded studies yield ‘positive’ results.
There is, in and of itself, nothing unethical about that, IMHO. What is troubling is that it would appear that the hard questions may not be asked at the rate they need to be, and that university researchers are forced to take money from people who have a vested interest in an outcome.
I want to be careful, here- this is my conjecture, and I think it will be unfortunate if people jump the gun and start claiming that ‘researchers are selling out.’ That more than likely is not the case, and what I have outlined is probably more accurate. Again, I am coming from this from a social sciences perspective, so I will leave it to Tim to fill in the blanks regarding research design and the hard sciences.
Having said all that, given the movements on the right and the left to politicize and denigrate science to achieve their political goals, it only seems to make sense that private corporations would themselves want to get in on the action. It makes sense, but it is still depressing.
*** Update ***
These comments are from Tim:
The story here is actually pretty simple. I don’t think that the studies themselves are corrupted, mostly because standard practices exist that almost eliminate researcher bias. Most of these studies at least claim to follow these practices, e.g. double-blind experimentation, agreeing on statistical methods prior to data collection, having the data itself analyzed by independent hands, so unless they are performing outright fraud (which I doubt) the studies are on the level.
The real disconnect happens after the research is done and before the work sees publication. Unlike government funding researchers who take industry money sign an agreement to publish the results only if the funding agency (e.g., Pfizer or Merck) approves. One of the reasons why drug companies have such massive budgets, then, is that they will fund five or six separate studies and only agree to publish the three or four that come out looking the best for drug X. Or to put it in a more generous light, imagine that you fund four studies and three come out looking pretty good for your putative drug. The fourth, though, looks terrible, performing as well or worse than the placebo. Maybe it’s an outlier, or the researchers screwed up somehow. As a competitive drug company you have huge, enormous pressures to let that outlier fourth study disappear into obscurity.
So in my view this result is entirely unsurprising. If you take two equal bell-shaped curves, A and B, and you cut off the left-side tail of curve A, then the average value for A will end up slightly higher than B. I think it is very likely that private and public-funded studies come out like curves A and B where the “real” distribution is more or less the same but the drug companies cut off the unfriendly “tail” of their curve before it reaches the public.
My guess is that the amount that gets cut off depends strongly on how competitive the drug is. If a company has an exclusive patent that they know gives them years of flex time to get it right then they may let research get through more or less unmolested. In a neck-and-neck race to release a blockbuster like Vioxx I would bet that the pressure to push the curve a little is practically unbearable.
None of this should surprise anybody. Free market rules control all publicly-traded companies and until somebody writes a law requiring them to publish every study that they fund (which would not be a bad idea) they will be within their rights to protect their interests.