Information Bias in Research, Uh-huh

While we’re on the subject of patient studies, here’s another thing: “regular” research with real liveFrink patients is not without its faults, no sir. By that I mean to say that even though studies using real, live people may seem quite preferable to computer simulation, things are not always as forthright as they seem.

This month’s Diabetes Voice magazine from the International Diabetes Foundation (IDF) features an extremely interesting piece about information bias in research: why it happens and how to avoid it.

Essentially, the article notes, “the pharmacuetical industry has much to gain by the selective publication of positive studies.” Studies showing no significant impact, and/or disprove the effectiveness of particular products rarely get publicized, meaning the stuff highlighted on Google News and elsewhere for easy publication pick-up are invariably those that happen to make specific treatments look good.

The real information gaps, the article notes, “are in the core issues of health-related quality of life, diabetes complications, and mortality, which have never been investigated in in high-quality, randomized controlled trials (RCTs).” MRGGGIFF!!! (that’s my supressed scream, due to recognizing just how difficult these far-reaching issues are to study in any reasonable way).

“The industry spends over $10 billion USD per year on funding around 90% of the 40,000 to 80,000 RCTs being conducted around the world at any given time,” the DV article states. “That such a high proportion of RCTs is funded by industry is of concern. Industry has a vested interest: the trials focus on patentable and therefore profitable drugs.”

Yes, but what’s to be done? This article calls for a more broad registration and publication of ALL research results. It also hightlights the work of the Cochrane Collaboration, an independent non-profit organization that “promotes the search for evidence in clinical trials and other studies.” In other words, the Cochrane group gives you the skinny on open questions in research on their web site. Worth checking out.

The so-called “Cochrane Reviews” are “systematic assessments of evidence of the effects of healthcare interventions, intended to help people to make informed decisions about health care, their own or someone else’s.”

For example, see their synopsis of recent studies on Inhalable Insulin. Scroll down to read the “plain-language summary,” which concludes, “We need longer studies to see if there are any side-effects in the lung. More insulin has to be given by inhaled than by injection to achieve the same effect, and the cost-effectiveness remains to be assessed.” Gotcha, thanks!


5 Responses

  1. Mona Johnson
    Mona Johnson September 5, 2006 at 11:25 am | | Reply

    Amy, thanks for writing about the potential for bias in research! I’m planning a post about problems with research on Alzheimer’s and dementia treatments. These problems may be especially bad because it’s sometimes difficult for dementia patients to communicate their symptoms.

  2. Thomas
    Thomas September 6, 2006 at 1:16 am | | Reply

    Had to smile when I read this article: last week your “Weekly Nuggets” section highlighted research that suggested that a vegan diet was good for diabetics.
    The same study was reproduced uncritically on the ADA website, without noting that the research had been conducted by a group that campaigns in favour of vegetarian/vegan causes and appears to have been funded by PETA in the past.
    Of course, all conducted without bias ;o)

  3. Scott
    Scott September 6, 2006 at 6:09 am | | Reply

    The so-called “diabetes” articles that are routinely published in the popular press are unfortunately also biased heavily towards the research done by the pharmaceutical industry. I also believe that far too much research is done which merely repeats studies or conclusions that have already been reached. Earlier this year, The Washington Post ran an article on this topic asking “What part of ‘yes’ do researchers fail to understand?”, but they noted that some journals, most notably The Lancet, now require researchers to conduct a meta analysis for research on the topic before they will consider publishing the results, which goes a long way towards eliminating repetitive studies. The bottom line is that we need to scrutinize research, and I have tried to do that with my annual “dumb diabetes research” awards, meant to highlight wasteful or unproductive diabetes research done at the expense of diabetes patients without delivering knowledge which can help us. Keep a lookout for voting on this later this year!!

  4. AmyT
    AmyT September 6, 2006 at 7:10 am | | Reply

    Good point, Thomas, thanks!

    And thanks for your efforts, Scott.

  5. Mitch
    Mitch September 12, 2006 at 1:25 pm | | Reply

    Hi Amy,
    I think that it’s important for everyone to keep in mind that just because there is a potential for something, that doesn’t mean that it will actually occur. Yes, there is a potential for research to be tilted in favor of the research sponsor; but if the scientist that does the work and the peer-reviewed journal that publishes the results do their jobs the way they were trained, studies are designed to minimize the bias that actually occurs. All research is performed by someone that “cares” about the results or it wouldn’t get done. No one deliberately sets out to do research in which they have no interest. And, just because someone has an interest in the outcome of a research study, that does not automatically make the results of the study suspect.

Leave a Reply