Search This Blog

Thursday, October 28, 2010

Meta Analysis, Measuring Measurments

In the early chapters of Dr. Ben Goldacres "Bad Science" he begins to unfold the concept of meta analysis. In statistics, a meta-analysis combines the results of several studies that address a set of related research hypotheses. Simply follow this link, and type meta analysis into the search box. I know I'm being lazy but no one says it better than Bad Science.

Bad science can be done in a clinical trial ran by a homeopathy company or a big pharma company. Both groups can work around almost any set of data. They practice an advanced form a cherry picking statistical analysis that is hard to detect. Meta analysis however has a way of measuring the measurements of clinical trials. In essence, it is a clinical trial of the scientific reasoning behind a particular area of interest.

Bring in Dr. Ionnidis (YO NEE DEES). In the November issue of The Atlantic David H. Freedman has an article titled "Lies, Damned Lies, and Medical Science". The main character is Dr. Ionnidis. He is a meta-researcher and he's one of the world's foremost experts on the credibility of medical research. Alas the CCS is just a lowly blogger but this guy has introduced science into the equation.

Dr. Ionnidis: "There is an intellectual conflict of interest that pressures researchers to find whatever it is that is most likely to get them funded".


Indeed!

Dr. Ionnidis sees a pattern (the as does the CCS) and he has a method of testing his hypothesis. Of course this means that he has his own bias. As they point out in the article, it would not be very interesting if he did the analysis and found that most clinical trials are more or less correct. Dr. Ionnidis claims that roughly 90% of the medical research that doctors rely on is flawed. By merely pointing out the bias however, he has risen above the norm.

No comments: