Funding as a source of bias

January 12, 2007 at 7:45 am (Sarah's Posts)

Beverage Study

This was an interesting article that I came across that speaks to something that we implicitly understand but that is hard to measure.  This is one of the first papers that I’ve found that addresses in a systematic manner how funding can influence the outcomes of studies. 

It makes me wonder how this comes about (very few researchers would openly admit to biasing their studies, even to themselves) and how to abate the soft pressure that comes from sponsorship of research.  Where is the role of public health folks in advocating for unbiased science, particularly when it has a large impact on policy development?  This is one of the reasons that I think that a solid foundation in science is supremely important for public health people, be it in biological sciences, epidemiology or any area that gives a basis on how to evaluate the validity of scientific research.

Advertisements

6 Comments

  1. msoule said,

    This article should be read by all. What you’re saying is right: we do intuitively understand that industry funding probably biases studies but there’s also an assumed amount of denial or hedging from the industry people on this account. After all, they are just providing the money, they’re not actually doing the tests. Well, if the researchers doing the tests want to continue getting the funding, they’ll say what the money wants to hear. This is obviously a problem. It happens if research is not totally independent of private interests. The only way I can think of amending this is by creating an independent fund that industry people can donate to if they really want to have research done but that they have no control over. Whether anyone would donate is doubtful. We do have government funding for this purpose but with the current administration, the life is being drained from research funding. Hopefully regime change helps that.
    We should also be careful about what we read. Ideally, there is an up-front statement of possible bias (many papers funded by big biotech or pharma interests state these biases in order to avoid the devalidation of their conclusions after publication) but if this goes by the wayside, we’re in trouble if we don’t dig up the dirt ourselves. It’s extra work but it’s worth it.

  2. misarita said,

    I think that a lot of the bias is really subtle and underneath the awareness even of the researcher him/herself. When you come into a study with some idea of what the outcome will be (which all researchers do. People don’t just embark on completely blind experiments very often) there is a lot pushing in the direction of proving that point. A lot of it is in experimental design. One of the biggest flaws that I’ve seen in this area is not setting up valid negative controls. The very idea of negative controls is fascinating to me, because the way that you set them up and how you do it is the premise for having a good study. I had to throw away many an experiment that *really* looked like it was proving what I wanted because in retrospect the controls weren’t solid enough. It is tempting not to do that, particularly when you feel in your gut what the data is right. That is the part that is the most alarming to me–not that there is bias, but that it might be under the awareness of the researcher himself.

    Good scientific training should come with this. It should train people in skill of disproving their ideas, of pursuing explanations other than what they think is going on.

  3. sstolper said,

    Sarah, in what situation were you throwing away ‘many an experiment’? Was this a job or a class? And are you referring to the choice of control or the quality of the result?

  4. monicaey said,

    Failure to disclose conflicts of interest in scientific research is an issue of major concern in the public health world – all I had to do was search for “science industry disclose” in PubMed and it turned up links to tons of articles on the subject. This sort of bias is not so subtle, either; I don’t remember the exact details, but recently an extremely famous and well-respected scientist came under heavy scrutiny and public disgrace for failing to disclose the industry money that had funded his research for several decades. There’s an interesting article by Barrow and Conrad in the Feb. 2006 edition of Environmental Health Perspectives, if anyone wants to read more.

  5. msoule said,

    Monica,
    Can you link the article or post it on here? That would be awesome.

  6. misarita said,

    Sam–I was thinking of my job at a biotech. I did a lot of assay design work, which might be part of the reason. If you’re running set experiments (i.e., if yuo bought the ELISA kit from R&D Systems or something like that) then the controls are already set up. But, every time that I was trying to design a new assay, there were always issues of how the controls were carried out and whether they were measuring what we wanted them to. It could also be because I was a new and inexperienced lab rat and my boss gave me tons of room to make as many mistakes as possible. But, it certainly forced me to think about what my controls were measuring and why.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: