Thursday 25 February 2010

One-sided skepticism

I'd earlier noted that worst of all is one-sided skepticism of research based on funding source: total skepticism of things industry funded and zero skepticism of government/NGO funded work being the usual case.

PLoS Medicine goes for worst of all.
This month also marks the implementation of a new policy on tobacco papers at PLoS Medicine.

While we continue to be interested in analyses of ways of reducing tobacco use, we will no longer be considering papers where support, in whole or in part, for the study or the researchers comes from a tobacco company. As a medical journal we do this for two reasons. First, tobacco is indisputably bad for health. Half of all smokers will die of tobacco use [4]. Unlike the food and pharmaceutical industries, the business of tobacco involves selling a product for which there is no possible health benefit. Tobacco interests in research cannot have a health aim—if they did, tobacco companies would be better off shutting down business—and therefore health research sponsored by tobacco companies is essentially advertising. Publication is part of tobacco company marketing, and we believe it would be irresponsible to act as part of the machinery that enhances the reputation of an industry producing health-harming products.

Second, we remain concerned about the industry's long-standing attempts to distort the science of and deflect attention away from the harmful effects of smoking. That the tobacco industry has behaved disreputably—denying the harms of its products, campaigning against smoking bans, marketing to young people, and hiring public relations firms, consultants, and front groups to enhance the public credibility of their work—is well documented. There is no reason to believe that these direct assaults on human health will not continue, and we do not wish to provide a forum for companies' attempts to manipulate the science on tobacco's harms.
Imagine the second paragraph re-written:

"Second, we remain concerned about the anti-smoking lobby's long-standing attempts to distort the science and exaggerate the harmful effects of smoking. That the anti-smoking lobby has behaved disreputably - exaggerating the harms of tobacco products, campaigning in favour of smoking bans, exploiting young people in anti-tobacco advertising campaigns, and hiring public relations firms, consultants, and front groups to enhance the public credibility of their work - is well documented. There is no reason to believe that these direct assaults on personal freedom will not continue, and we do not wish to provide a forum for these individuals' attempts to manipulate the science on tobacco's harms."

As good a case can be made against Glantz and ASH as can be made against tobacco industry funding. How many anti-tobacco public health researchers would be able to continue getting grants from Ministries of Health if their research found that smoking isn't as bad as the Ministry might have thought?


  1. Interesting. Government 'regulatory capture' as a source of pressure to do bad science. Government funded climate science is, alas, apparently an example of this trend. Another example, that I just stumbled across a couple of days ago is some of the info contained in the book' Fat Politics' about how the official US gov. definitions of 'overweight' and 'obese' were arrived at.

  2. Not to mention how the funding from "big pharma" biases their research towards nicotine patches and against other valid mechanisms of quitting.

  3. @TGGP: Yup, I assign that piece :>
    @dlr: Will have to pick up that book.
    @rsw37: ASH in the UK is particularly guilty on that one.

  4. Do you honestly not see a difference between research sponsored by a for-profit company whose sole purpose for existence is to profit from the sale of the product in question, and who have been found repeatedly in the past to have published misleading results, and research sponsored by not-for-profit organizations whose stated mission is (for example) raising public awareness about the product in question?

    Yes, I understand that individual researchers have incentives--they always do--but do you honestly believe the situations are comparable?

    Having just read the explanation for the masthead, a similar question comes to mind. Yes, there are offsetting effects, and attaching a spike to the steering wheel of a car will make people drive more carefully. But do you honestly believe that this would make driving safer overall? Do you honestly believe that seatbelts make driving more dangerous?

  5. @Rob, regardless weather you think a company trying to increase sales is comparable to an organisation trying to increase awareness, the fact is that both have a real financial incentive to nudge the results in their own direction. Each study should be judged on its merits, financial incentives should be kept in mind but don't prove falsity. Have a closer look at how organisations like ASH behave if you don't believe a not-for-profit organisation could be just as corrupt as the tobacco industry.
    With regards to seatbelts, there have been many studies into the effects of seatbelt legislation. From what I have read it appears there is some evidence that overall deaths increased.

  6. @Rob: Of course there's a difference: the former will tend to understate harms in an environment where it knows anything it does will be checked over in detail by folks suspicious of their every motive, so they have to be at least somewhat careful; the latter will massively exaggerate harms in an environment where their funding level is proportionate to the degree to which they scare the public and where their findings are largely taken uncritically. So they're not comparable at all. The latter is worse, but only because at the current equilibrium there's far too much skepticism of tobacco funded research relative to anti-tobacco research.

    As for steering wheel spikes: I don't advocate that they be put on steering wheels. It's a pointy reminder that regulations have offsetting effects. Mandating spikes would greatly reduce pedestrian and cyclist deaths while massively increasing motorist deaths; the latter effect would dominate. The best research I've seen on seatbelts suggests folks drive in riskier fashion while wearing them, leading to more accidents but a lower fatality rate. Pedestrian deaths will be up as consequence; I've not been convinced yet on overall deaths.