Five years after Congress required the open publication of the results of human medical trials, to help both patients and researchers, a comprehensive analysis has found widespread noncompliance.
The study, looking at relatively large trials—a total of 585 with 500 or more participants—found that 29 percent had not been published in any journal five years after completion, and the results of nearly four-fifths of those were unavailable on public websites.
The analysis was limited to large trials because of the correspondingly large potential of unpublished data to distort medical practices, said Christopher W. Jones, a co-author of a report on the study and an attending physician in the department of emergency medicine at the Cooper Medical School of Rowan University, in New Jersey.
Failure to publish in such cases also “violates an ethical obligation that investigators have toward study participants,” Dr. Jones and his co-author, Timothy F. Platts-Mills, an assistant professor of emergency medicine at the University of North Carolina at Chapel Hill, wrote in their report, which was published by BMJ.
The federal government opened ClinicalTrials.gov to the public in 2008, after Congress approved legislation requiring the results of most human medical trials to be posted to it or to alternative websites.
Previous studies of trials of all sizes have found that a quarter to a third of them remain unpublished, Dr. Jones said. A well-established cause is publication bias, he said, meaning the tendency of researchers and journals to avoid reporting “negative” results—an outcome that doesn’t show a benefit for the drug or intervention being tested.
But that reluctance to publish is a problem for the overall research enterprise because hiding negative results can skew the record for doctors and researchers who want to know what does and does not work.
Many of the decisions not to publish are apparently being made consciously by the scientists, rather than the medical journals, Dr. Jones said, since journal editors probably would be interested in negative results when the studies are as large as those included in the BMJ review.
The director of ClinicalTrials.gov, Deborah A. Zarin, said this year that the site had accumulated more than 9,000 reports of completed trials, with nearly half of them not described in any published articles, meaning the site was the sole source of publicly available information on them.
That’s not perfect, but it’s an impressive accomplishment in a few years, Dr. Zarin said. “It’s absolutely true that the culture has changed in terms of trial registration,” she said.
The study by Dr. Jones and Dr. Platts-Mills found industry-sponsored trials were less likely to be published than publicly financed trials, though the results of industry-sponsored trials were more likely to be available at ClinicalTrials.gov.
It’s a continuing problem, despite all the attention placed on it in recent years by Congress and academe, said Joseph S. Ross, an assistant professor of general internal medicine at Yale University who has carried out his own studies of clinical-trial registration.
“What remains important is that we as a field need to do better,” Dr. Ross said. “The results from all trials need to be available in some way or form.”
