A recent article in The New England Journal of Medicine shows that reports of clinical trials may not be all that they seem to be. Eric Turner and colleagues report that selective reporting of clinical trials and their outcomes can lead to unrealistic estimates of drug efficacy and of the cost-benefit ratio of using such drugs.
The team conducted a systematic study of reviews from from the US Food and Drug Administration (FDA) of 12 trials of antidepressants that involved more than 12 000 patients. They found that, among 74 FDA-registered studies, 31% were not published. This accounted for more than 3 000 participants. More importantly, it was the outcome of the trials that determined whether or not they were published. A total of 37 studies that were regarded as positive by the FDA were published - only one positive study was not published. Studies that were viewed as having negative or questionable results, were, with three exceptions, either not published or published in a way that suggested a positive outcome.
The authors are not sure if the selective bias came from manuscripts that were not submitted if the results were not as expected or indeed required, or whether the decision not to publish was from peer reviewers or editors, or both.
The bottom line is that selective publishing of positive trial results can result in unwarrented claims for drug efficacy.
