Data from drug trials often not published
■ A federal law expands an existing clinical trial registry and would make additional information available online for researchers and clinicians.
By Susan J. Landers — Posted Oct. 13, 2008
Washington -- More than half of the clinical trials conducted for drugs approved by the Food and Drug Administration during a three-year period remained unpublished five years later, according to a study appearing Sept. 23 in the open-access journal, PLoS Medicine.
The researchers found published trial results for just 43% of the 90 drugs approved between 1998 and 2000. In addition, trials classified as "pivotal," meaning those in which the new drug performed better than an older drug, were more likely to be published than were those with negative findings.
Such selective reporting of results may lead to an inappropriately favorable record in the medical literature of a drug's risk-benefit profile compared with standard therapies and could persuade more physicians to prescribe the new and more-expensive treatments, researchers concluded.
"What we think we know is not necessarily true," said Kay Dickersin, PhD, director of the Center for Clinical Trials at Johns Hopkins Bloomberg School of Public Health in Baltimore. She also is director of the U.S. Cochrane Center, one of 12 centers worldwide that make up the Cochrane Collaboration, an organization that conducts systematic reviews of health research. She did not take part in the PLoS study.
Published literature is most physicians' route to knowledge, noted Dickersin. "If positive results are three times as likely to be published as are negative results, then we have a problem in that the published literature doesn't represent the truth."
The flaw in the system resides more with the researcher than the journals, she said. Researchers are less likely to submit negative findings.
However, Ken Johnson, senior vice president of the Pharmaceutical Research and Manufacturers of America, which represents the large pharmaceutical and biotechnology companies that sponsor most of the studies, contests this view. He said that PhRMA, for example, has, since 2004, sponsored a free Internet database that provides comprehensive summaries of trial results.
Regardless, he added, "no single publication of a study in the medical literature can take the place of critical analysis by the Food and Drug Administration. The information physicians need to make appropriate prescribing decisions is found in the FDA-approved drug label, which synthesizes key details, including important drug safety information that may not be included in the published paper."
A sea change is under way
However, such business-as-usual approaches may be about to change. A law passed last year expands an existing federal clinical trial registry and requires the posting of additional data about trials undertaken for drugs and devices that gain FDA approval. The information is to be placed on a federal Web site, which serves as a registry of federally and privately funded trials conducted in the U.S. and around the world (link).
The PLoS study provides a look at the publishing scene before the law went into effect, said the researchers. "We wanted to establish a baseline and if things either got worse or improved, we would know, since having the data is the first step to addressing a problem," said senior author Ida Sim, MD, PhD, associate professor of medicine at the University of California, San Francisco.
The requirement that trials be registered had been long in coming and was recommended by many health groups, including the American Medical Association. Ron Davis, MD, immediate past president of the AMA, advocated the creation of such a registry when he spoke to a congressional panel in 2004, "Quite simply, physicians need complete and unbiased information about the safety and effectiveness of the treatments they prescribe for their patients. A centralized clinical trials registry would improve physician and researcher access to this information and would facilitate patient enrollment in clinical trials," Dr. Davis said during that testimony.
With a registry established, Congress went even further. "It has really been groundbreaking," said Dr. Sim. "Ten years ago we would never have imagined being at this point."
"It's a huge step forward," said Wilson Pace, MD, professor of family medicine at the University of Colorado in Denver and director of the American Academy of Family Physicians' National Research Network. The trial registry provided evidence that there were trials under way that never came to light, he said.
But there is still some distance to cover, noted Dickersin. The new reporting requirement applies to very few drugs since very few are approved each year. Last year, only about 20 new drugs reached the market. "So we need to expand the law to apply to more than just approved drugs and devices."
Plus, physicians want to know more, she said. "They have questions way beyond drugs and devices."
She advises physicians to use with caution the information that will soon become available on single studies and to look instead to broad, systematic reviews for prescribing information. The new data will likely be useful to those doing such reviews, said Dickersin. "Already, clinicaltrials.gov has made a huge difference," she said.
"We don't know about the results portion yet because it just came out, but we are expecting it to be quite valuable."