Positive results of clinical trials for drugs or devices have a higher chance of getting published in the medical literature than negative trials, according to an investigation into the publication status of the trials submitted to the US Food and Drug Administration (FDA) as part of the 90 new drug approval applications approved by the agency between 1998 and 2000. Notably, when the authors of the study split up their analysis by trial type, they found that clinical trial sponsors publish the results from pivotal trials (an industry term that refers to those trials that show whether a drug or device really works) only 76% of the time (PLoS Med., doi: 10.1371/journal.pmed.0050191; 2008).

Overall, of the 909 trials they found related to the 90 drug applications, only 43% of all the trials conducted were published within five years after FDA approval of the drug or device.

The studies that found a statistically significant difference were more than three times as likely to be published. According to study co-author Ida Sim, director of the Center for Clinical and Translational Informatics at the University of California, San Francisco, this leads to a phenomenon called 'positive publication bias', which is a serious problem, because it can make a drug or device appear in the literature to be more effective than it really is.

Sim explains, “We have this idea of practicing evidence-based medicine, which is predicated on having a full and complete evidence base. But when the evidence base is skewed, we can't really do this.”

In a paper published this year, experts suggest that the FDA Amendments Act of 2007 has improved transparency, because the law mandates that sponsors or primary investigators of clinical trials for approved drugs post a summary of their results in a national open-access database (Science 319, 1340–1342; 2008).

The lead author of the report, Deborah Zarin, oversees the ClinicalTrials.gov registry at the National Library of Medicine of the National Institutes of Health and is in charge of ensuring the results are posted in compliance with what the new law. According to Zarin, “for the trials that are covered by this law, the results database should have a big impact on disseminating medical knowledge, because the results have to be publicly available.”

But not every type of clinical trial is covered by the legislation, nor does it directly affect medical journals. Although Sim applauds the FDA Amendments Act of 2007, she adds that it “doesn't address the issue of not publishing trials in medical journals. They remain one of the most influential and biased sources of information.”