A new transparency study has identified Sanofi, Novartis and GSK as the companies that keep the most clinical trial results unreported.

TrialsTracker is a tool developed at the University of Oxford that automatically rates transparency on clinical trials by checking if there are results available in clinicaltrials.gov and PubMed. The results: globally, 45% of data on completed clinical trials by major institutions, i.e. those with more than 30 trials, is missing.

The new tool has pointed at Sanofi as the worst offender, with 65% of its trial results from 2006 to 2014 kept secret. Novartis and GSK are the other pharma companies listed in the top 5 of secretive institutions, with 38% and 22.6% of data missing respectively.

On the other hand, US-based Shire is setting an example with 100% of its trial results publicly published. Bristol-Myers Squibb, Eli Lilly, Roche’s Genentech and Johnson & Johnson are also among the most transparent companies.


Of 25,000+ trials completed by major sponsors since 2006, 45% of them haven’t been published (orange)

The development of the algorithm was backed by AllTrials, an international initiative to promote the publication of results of all clinical trials. The transparency data published online is intended to raise awareness and pressure pharma companies into publishing results in order to improve their scores.

Why is transparency so important? If results are not available, doctors and patients can’t make informed decisions taking into account all the benefits and risks of a treatment. The data is also necessary for regulators and ethics committees to see the full picture before taking action.

The results from TrialsTracker highlight that there’s a huge bias in the information available from clinical trials. According to its website, negative results are twice as likely to not be reported, even though these are actually vital to avoid wasting money in what doesn’t work. If the initiative succeeds, researchers will have it easier to design trials that use resources efficiently and save patients time and pain.

Featured image by Tashatuvango/shutterstock.com

Figure from TrialsTracker

Previous post

Meet the Mind behind the Artificial Pancreas that could change the Lifes of Diabetic patients

Next post

Green Light for Genfit to Tackle another Liver Disease

  • I don’t think the statistics reported by Trials Tracker are reliable enough to be meaningful. Their data come from a crude automated algorithm that does not search in anything like as thorough a manner as is necessary to find all trial results, so it systematically underestimates disclosure.

    As Sanofi is singled out as the worst performer, I tried to see how well the algorithm did by picking the first 10 “undisclosed” studies to see if I could find them. In just a few minutes of Googling I found results for 9 of them. The undisclosed one was for a drug that was discontinued in clinical development in 2008, so no patient is being denied information on a drug they are taking as a result of the non-publication of that trial.

    I don’t know if those 10 are typical of the “undisclosed” trials identified by Trials Tracker, but if they are, then it is overestimating the number of undisclosed trials by a factor of 10.

    • Hi Adam. Thanks for your strong interest in the topic! As the researchers say in the TrialTracker website, the method isn’t perfect. However, their criteria reflect the view that apart from being public, results should be easily found in ClinicalTrials.gov and PubMed. I think the point here is that a patient should not be forced to have the time or in-depth knowledge to research whether a trial has been published somewhere else. And the algorithm makes it easy for the companies to quickly fix any mistakes by reporting the results in those databases.

      In terms of global numbers, the researchers have published the TrialsTracker results in a paper in which they compare the algorithm with previous studies done manually. Interestingly, they found that for the same datasets, both methods report similar results.

      I hope this bit of extra info helps to clarify how the tool works!

      • Adam Jacobs

        Many thanks for your prompt reply, Clara.

        You are completely right, of course, that results of studies can sometimes be hard to find. I found them very quickly, but then I’m a clinical research professional, and I realise that not everyone has the training and experience to do that.

        But that’s a rather different claim from the claim made in the article that the trials are being “kept secret”. Perhaps you could consider editing the article to make it clear that trials are not being kept secret, but just being imperfectly indexed?