'We've been deceived': Many clinical trial results are never published
Canadian universities and research hospitals are among the worst offenders, according to new online tool
Every year, thousands of Canadians sign up to participate in clinical trials, offering their bodies to further the development of important medical advances like new drugs or devices. But the results of many of those trials never see the light of day.
A new online tool aims to put pressure on some of the companies and institutions behind the problem. TrialsTracker maintains a list of all the trials registered on the world's leading clinical trials database and tracks how many of them are updated with results.
Amid pharmaceutical companies and research bodies from around the world on ClinicalTrials.gov, maintained by the U.S. National Institutes of Health, nine Canadian universities and institutions rank in the top 100 organizations with the greatest proportion of registered trials without results.
"It's well documented that academic trialists routinely fail to share results," says Ben Goldacre, who was part of the team from the University of Oxford that developed TrialsTracker. "Often they think, misguidedly, that a 'negative' result is uninteresting — when, in fact, it is extremely useful."
The University of Toronto's David Henry says "publication bias," as it's called, is robbing the medical community and patients of important information.
"We've been deceived about the truth about treatments that we've used widely over a long period, in very large numbers of individuals, because of the selective publication of results that are favourable to the product," says Henry, a professor of health systems data at U of T's Institute for Health Policy Management and Evaluation.
- Hidden drug safety data could be unlocked with new key
- Medical journal aims to stamp out sexism in clinical trials
But Henry adds that publication bias isn't the only reason results aren't being made public. He says many institutions haven't made it a priority.
"If you leave it to the trialists, they've often moved on to the next trial," he says. "At the end of the day, I don't think they give enough weight to it."
Henry notes there has been progress as the scientific community begins to recognize the importance of making all results available.
The U.S. Food and Drug Administration (FDA) now requires most clinical trials to register and post results on ClinicalTrials.gov.
But studies show many organizations are ignoring the rules. In a paper that accompanied the launch of TrialsTracker earlier this month, Goldacre noted that approximately half of registered clinical trials fail to publish their results, and studies with negative or non-significant results are twice as likely to be unpublished.
TrialsTracker's real-time data supports these findings. An algorithm scours for results within ClinicalTrials.gov and also among the available scientific literature.
Using this method, the researchers found that between 2006 and 2014, 45 per cent of the clinical trials registered on ClinicalTrials.gov — or nearly 12,000 studies — are missing results.
Both the University of British Columbia and University Health Network — the two Canadian institutions with the highest number of missing results on TrialsTracker — point out, in statements sent to CBC News, that the site's algorithm will miss some results.
Goldacre acknowledges the method isn't perfect, but says trialists must take responsibility for ensuring results are easily accessible.
"Research that is hard to discover is not transparently reported," he writes.
Both institutions said they continue to work on ways to ensure that research participants are better informed about the results of studies they participate in.
One possible solution might be found at the Ottawa Hospital Research Institute (OHRI), where the organization's first "publications officer" was hired about a year ago, to help researchers navigate the often daunting process of publishing their results.
David Moher runs the Centre for Journalology at OHRI, which studies the science of academic publication. He advocated for hiring the publications officer, and says part of their job is to explain there are more ways to publish results than in traditional academic journals.
"The important point is to make results available, and there are many ways to do that in 2016," he said, pointing to the open-access repositories that are available at several Canadian universities as an example.
Moher hopes to study the effect of the publication officer at his institution and, if it's effective, see the model replicated at institutions across the country.