Drug controversies prompt call for clinical trial registry
■ Advocates hope to build on existing databases to create one mega-source, including published and unpublished research.
By Victoria Stagg Elliott — Posted July 5, 2004
- ANNUAL MEETING 2004
- » Our coverage
- » Archive of proceedings
- WITH THIS STORY:
- » AMA calls for more federal funding for medical research
- » External links
- » Related content
When David Fassler, MD, a child and adolescent psychiatrist from Burlington, Vt., was called to review data -- much of it unpublished -- about antidepressant use by kids and teenagers for a February Food and Drug Administration hearing, he was stunned.
The information suggested that several of these drugs were not effective for depression in this age group and could actually lead to an increased risk of suicidal behavior.
"This data clearly had an impact on our own analyses, our comments and our recommendations and on the subsequent public debate," said Dr. Fassler.
The fact that the findings had not previously been accessible has focused new attention on efforts to open up the clinical trial process. Adding to this momentum are other related events, including the FDA's initial June 2003 caution regarding the antidepressant paroxetine, its expanded advisory issued in March, and a New York State lawsuit that charged the drug's manufacturer, GlaxoSmithKline, with concealing information.
Registries are one of the mechanisms currently gaining favor.
Delegates to the AMA's Annual Meeting in Chicago last month endorsed a policy that would urge the Dept. of Health and Human Services to establish a comprehensive registry for all clinical trials and require every trial to have a unique identifier.
Similarly, a proposal being considered by the International Committee of Medical Journal Editors would require all clinical trials to be listed in a registry as a requirement for publication.
Meanwhile, the AMA has recommended that all trial data be made available through publication or an electronic data repository. The Association also will study methods to enhance public access to data considered by the FDA as part of the drug approval process.
"What we want to do is eliminate the problem of information just disappearing," said John Schneider, MD, MPH, a member of the AMA Council on Scientific Affairs.
Let the sun shine in
Clinical trial registries are not a new concept. More than 300 exist. The largest, run by the National Institutes of Health, is available online (link). GSK, which has posted on its Web site all its paroxetine research findings, announced in June that it would create a public registry listing protocols and results of GSK-sponsored trials.
But physicians and others pressing for more transparency say that while these kinds of registries are a step in the right direction, much more progress is necessary.
"There are lots of registers," said Kay Dickersin, PhD, a professor of epidemiology at Brown University and director of the U.S. Cochrane Center, an independent nonprofit organization based in Providence, R.I., that produces systematic reviews of health care interventions. "They're hard to use, and they're not comprehensive." Advocates say the sheer number of registries makes the process of finding trials too cumbersome for patients and scientists alike. And many trials aren't registered anywhere.
Also, no repositories include unpublished data resulting from trials. The AMA policy attempts to address this point by calling for findings to be published or put into a database -- but not simply left to perish. Experts estimate as much as 50% of clinical trial results never see the light of day, and supporters argue that the result is unethical because it leads to buried data that could otherwise have had an impact on patient care. It can also slow the progress of science because researchers unknowingly repeat lines of inquiry that may have already been proven unsuccessful.
"Studies that do not make it into journals should be available to clinicians and physician researchers," said Samuel Blackman, MD, PhD, an AMA Council on Scientific Affairs member. "A comprehensive clinical trial registry is crucial. In addition to reducing publication bias, a comprehensive registry will allow physician researchers to know what research questions have already been asked and potentially will allow researchers to refine their research questions and improve study design."
Those who run clinical trial registers agree that what the AMA and others have in mind is feasible. The NIH collection, for instance, which only lists studies related to severe and life-threatening diseases and links to published findings, could be expanded.
There is significant concern, however, about making unpublished data, which have not yet gone through the peer-review process, more publicly available.
"From a technical point of view, the infrastructure is in place," said Alexa McCray, PhD, director of biomedical communications at the National Library of Medicine and of the NIH's registry, ClinicalTrials.gov. "But there is the whole question of quality control. We don't want people putting data into ClinicalTrials.gov that has not been validated or peer-reviewed in one way or another. But as a community we need to have a discussion about what counts as validated data."
While scientists are worried about how lay people would interpret such information, pharmaceutical companies are worried about another angle -- trade secrets.
"It's very important that proprietary data be protected," said Jeff Trewhitt, a spokesman for the Pharmaceutical Research and Manufacturers of America. "And there is strong potential for confusion with a registry unless it's done right."
Still, advocates say that much of the information companies are seeking to protect is already out there. A universal trial registry would simply make it easier to find. Such a repository would also make companies less likely to be accused of hiding things.
"As a scientist, I feel that, if there's scientific data, there's no reason to hide it," Dr. Schneider said. "That information is valuable, and it needs to be available. I don't think the drug companies are interested in hiding information either. By [creating registries], it probably protects them."