Reports of clinical trials published in scientific journals contain less information than the summary of results from the same trial on the register ClinicalTrials.gov. In a new paper in PLOSMedicine researchers randomly selected 600 trials that were completed and had summary results uploaded to the register. After searching the scientific literature they found that around half of these trials had a journal article published about them. The researchers compared the information contained in the articles with the information in the summary results on the register and found that the summary results contained more efficacy results than the published article (79% vs. 69%), adverse events (73% vs. 45%), serious adverse events (99% vs. 63%) and the flow of participants (64% vs. 48%).
Based on these findings, the researchers recommend those wanting to make informed decisions about drugs should consult trial registry information. They also suggest the use of reporting templates to guide standardised reporting of trial results in journals, along with broader mandatory registration of results for all trials.
Ben Goldacre, one of the founders of the AllTrials campaign said “This paper shows that academic journals are failing doctors and patients, because they are often very bad places to report clinical trial results. They permit information on side effects to be withheld, and the main outcomes of the trial to be switched, which exaggerates the apparent benefits. This is why we ask for full reporting of the main outcomes of the study, and for the full methods and results in the Clinical Study Report to be made publicly available, where one has been written.”
An editor’s summary of the paper can be found here.
Nature, Half of US clinical trials go unpublished
BMJ, Half of drug trials with results on ClinicalTrials.gov are not published in journals