Over half of the clinical trials used to obtain drug-approvals from the FDA never get published, according to Californian researchers.

Kirby Lee of UC San Francisco and colleagues identified 909 trials in applications submitted in support of 90 eventually approved drugs. Only 43% every made it into journals within five years, the researchers report in PLOS Medicine.

Unsurprisingly there is also a bias towards favourably results, as fellow study author Ida Sim told Bloomberg:

We found that there was indeed a pattern that favourable studies were more likely to be published than unfavourable trials. This is something that is essentially structural in the way clinical trial information is disseminated to the public

As the researchers note this is an important figure to establish as new FDA rules will soon make it compulsory to list the basic results of clinical trials on a National Institutes of Health. However Sim told Boomberg this might be a doubled edged sword (much like a lancet) as drug companies might have even less incentive to submit studies that don’t favour their drugs, arguing that the NIH summaries provide full disclosure.

Another point raised by this post on Respectful Insolence and Science-Based Medicine (slightly different posts):

A depressing additional observation was that the percentage of trials published in the peer-reviewed literature ranged from 0-100%, with a mean of 55%. That’s right; there were drugs that had 0% of the trials supporting their FDA approval published over five years later.

On a similar note, see this recent column from Ben ‘Bad Science’ Goldacre regarding this trial:

After downloading the contents of the National Institutes of Health’s ClinicalTrials.gov registry, we used key words to identify trials in oncology. We then evaluated the proportion of trials that had been published in journals listed in PubMed.gov.

…

Less than one in five studies in cancer that are registered with clinicaltrials.gov have been published in peer-reviewed journals.