A DECADE or so ago, there was a big push to get all clinical trials listed on a public register before they started enrolling patients.

The move was designed to increase transparency, prevent researchers and trial sponsors from secretly changing endpoints, and make it harder for those with vested interests to hide trials that didn’t produce the outcomes they wanted.

Major journals came on board, with the International Committee of Medical Journal Editors announcing that member journals would only consider an article for publication if the trial was registered before the recruitment began, including information on the outcomes to be measured, sample sizes and funding sources.

Did it work? Well, yes and no, as I wrote last year.

The number of pre-registered trials has certainly grown dramatically. The world’s largest trial registry, clinicaltrials.gov, listed 12 000 trials at the start of 2015. It is now hovering just under 230 000.

But, as this 2014 survey showed, not all journal editors have rigorously enforced the “no-registration-means-no-publication” policy.

Around half of the responding editors said that they did not require registration, two-thirds said that they would accept retrospectively registered trials (completely defeating the purpose), and a meagre 18% bothered cross-checking submitted papers against the details in the original registration.

It gets worse.

The new registration regime has also not had the desired impact in ensuring that trial results are made public, according to data released in October 2016.

The TrialsTracker tool developed by researchers at the University of Oxford shows that 45% of registered trials run by major sponsors over the past 10 years have not had their results fully published.

Topping the list of non-disclosers is Sanofi, with 285 of its 435 trials (66%) not having fully reported results.

Other pharmaceutical companies in the top ten for numbers of trials with missing results are Novartis Pharmaceuticals (201 trials), GlaxoSmithKline (183) and Novartis (127).

If you combined the two different arms of Novartis listed there, the Swiss company might edge out Sanofi for the top spot, though of course there could be a myriad of other commercial relationships between companies on the list to further complicate matters.

It’s not just the pharmaceutical companies that are failing to report their trial results: the top 10 (or should that be bottom 10?) also includes the US National Cancer Institute, the Paris hospital system, the Mayo Clinic, two South Korean universities and the Alliance for Clinical Trials in Oncology.

Some of the apparently unreported trials may be clerical errors. Perhaps researchers or organisations have failed to enter the correct registration number when publishing their results.

But, whether it’s sloppiness or a more sinister desire to hide unpalatable findings, the failure to share the results of a total 11 714 trials, enrolling an estimated 8.7 million people over the decade, is a major betrayal of trial participants.

People may have many reasons to volunteer for a clinical trial, but one of them must often be the desire to help others, to contribute to the sum of knowledge about the condition being studied.

Patients take on the risks inherent in receiving an unproven, unlicensed treatment in the hope that they may, in some way, make a difference.

The least they can expect is that the results of the studies they have contributed to should be made available to other researchers and to the public.

Maybe it’s time for the journals to up the ante and start rejecting articles from sponsors who have failed to report on, say, 10% or more of the trials they have undertaken?

Jane McCredie is a Sydney-based science and health writer and publisher.


Latest news from doctorportal:



I welcome the inclusion of a mandatory self-reflective 40-point CPD requirement by the RACGP
  • Strongly disagree (37%, 39 Votes)
  • Strongly agree (20%, 21 Votes)
  • Agree (18%, 19 Votes)
  • Disagree (17%, 18 Votes)
  • Neutral (8%, 9 Votes)

Total Voters: 106

Loading ... Loading ...