July 7, 2005
Clinical Trials Registration: Overdue yet Elusive
Perhaps half of the 60,000 clinical trials ongoing in the United States will never be published in a scientific journal, according to Kay Dickersin, PhD, the new director-designate of the Department of Epidemiology’s Center for Clinical Trials. Patients as well as the scientific community are harmed when deprived of all these findings.
Kay Dickersin, PhD
At the Bloomberg School on July 1, Dickersin reported her investigations about exactly what becomes of research results in the United States. When she asked a group of investigators why they didn’t publish some of their findings, 22 percent said they had never finished the work; 11 percent said publicity had not been their aim; and a full 30 percent said the results were “uninteresting” (meaning, in most cases, “negative”).
“In any case,” said Dickersin, “[not publishing our results] is not really what we scientists are about.” She went on to quote Einstein: “The right to search for truth implies also a duty; one must not conceal any part of what one has recognized to be true.”
Why did Einstein use the word duty? Dickersin, currently professor of Medicine and director of the Center for Clinical Trials and Evidence-based Health Care at Brown University, listed several well-known cases where inaccessible information from unpublished studies could have reduced patient risk—or even saved lives.
- In 1986, John Simes tried to determine which treatment was better for advanced ovarian cancers: single-agent or multi-agent chemotherapy. He found that, whereas the combined results from all published trials showed multi-agent chemo to be more effective, the combined results from published plus unpublished studies revealed that multi-agent chemo was in fact no more effective. Multi-agent therapy has more side effects than single agent therapy: important information for patients.
- In the 1980s and early 1990s, anti-arrhythmic drugs were used extensively in heart patients with arrhythmias to ward off a second myocardial infarction. But then researchers discovered an unpublished study that proved these drugs caused real harm to cardiac patients. Doctors stopped prescribing anti-arrhythmic drugs for these patients, but not before an estimated 50,000 to 70,000 heart patients had died in the U.S. per year from the drugs.
- In the 1990s there was eveidence from 27 published randomized trials that hormone therapy for menopausal women might harm cardiac health, but these results were not statistically significant. Seven unpublished studies, however, were unearthed—and when their results were combined with the others, the studies revealed (six years before women were warned of this unintended consequence) that the hormones could cause heart trouble in these women.
Said Dickersin, “Failure to publish has definitely had harmful effects for patients.”
But how is science hurt when results are buried? She ticked off a number of downsides: expensive duplication of effort; lost opportunities for collaboration; underreporting of negative results—even ethical problems, since “we are breaking a covenant with the participants in clinical trials” when we assure them they will be contributing to new knowledge but then fail to publish results.
What can scientists do? First, said Dickersin, the scientific community must create a standardized, all-inclusive, unified registry of every clinical trial initiated, wherein all results must be entered. She added that many partial registries of clinical trials currently are up and running but that none is comprehensive and they contain few industry trials.
“It seems like a no-brainer to do trials registry,” said Dickersin, “but we’ve spent the last 20 years talking about it. Who’s going to take charge?” she asked. Some people are concerned that the local institutional review boards (IRBs) are too busy already, but scientists worldwide feel this may be the best approach.
Help with a genuinely global registry, however, finally seems at hand: The World Health Organization has pledged to take on the responsibility of this mammoth recordkeeping task. Add to this good news the media outcry last summer about antidepressants being linked to teen suicidal behavior, and people are starting to ask why such important results don’t surface.
Already, the members of the International Committee of Medical Journal Editors have declared they will no longer accept articles unless the studies are registered—and they are giving researchers until September 2005 to get all their past studies registered as well. Moreover, Congress is considering a bill (the Fair Access to Clinical Trials (FACT) Act, S. 470) that will amend the Public Health Service Act to expand the clinical trials drug data bank and would require the National Institutes of Health to refuse further funding if information is not submitted to ClinicalTrials.gov, the U.S. registry of drug and device trials.
Dickersin urged researchers to get involved, talk to their colleagues, write editorials and even, for instance, find out whether all trials at the Bloomberg School are registered. She asks all interested scientists to call their federal representatives in Congress about the new FACT Act. “One person can change a representative’s mind,” she said. “They’re grateful for your opinion and guidance.” —Rod Graham