USING STATISTICS TO DETECT BIAS IN CLINICAL TRIALS

Posted on August 24, 2009 by Jeffrey Newman

Researchers have found new statistical methods to see whether publishers of medical journals are biased about what they report or do not report about clinical trials. The new methods use as a “gold standard” newly released FDA data on anti-depressant drugs in the U.S. and what has been published about it. This data is considered free from biased and as such it is used as the basis of the algorithms to test for publication bias. The researchers are at the University of Leicester and an article on the subject can be found in the British Medical Journal by Santiago Mareno. the new methods are significant as policy makers base their decisions on what is published about the clinical trials. The researchers are not examining the clinical trial methods or use of statistics, just the phenominon of publication bias.