For years, doctors and academic researchers have complained about the medical industry’s hiding and manipulation of data from human medical trials, with harmful and even deadly consequences for untold numbers of patients worldwide.
Suicides linked to Paxil. Cancers associated with Infuse. Heart attacks tied to Vioxx. A litany of benefits overstated and risks underacknowledged, too often enabled by systems and rules of medical discovery that allow scientific claims without the hard data to prove them.
“There’s no doubt about this, that patients have suffered and died because of this practice,” said Sir Iain Chalmers, a British health-services researcher who has studied the problem.
Now, some major companies are starting to turn it around. The medical-technology company Medtronic just paid university scientists to review all records of its human testing of Infuse, a bone-growth protein. The pharmaceutical giant Roche has promised to release internal study data on its influenza drug Tamiflu. The pharmaceutical and biotechnology company GlaxoSmithKline is publicly posting more than a decade’s worth of study records.
The sudden move toward transparency is a clear win for patients. It’s so sudden, in fact, that university researchers, rather than their corporate counterparts, might soon find themselves portrayed as the bigger obstacle to the free flow of information in medical research.
“It’s time for us not only to be looking at industry, but to be asking why academia’s not doing more,” said Harlan M. Krumholz, a professor of medicine who leads a center on medical-outcomes research at Yale University.
In a two-year project that ended in June, Medtronic handed Dr. Krumholz all its human-subjects test data for Infuse and allowed him to commission two separate evaluations by other universities. The scrutiny was prompted by patient lawsuits and data indicating that the risk for adverse effects was 10 to 50 times higher than that suggested by publicly available reports on human clinical trials of the product.
Both study groups hired by Dr. Krumholz affirmed suspicions that publicly reported evaluations of Infuse, in which researchers financed by Medtronic were often the authors, gave inflated assessments of the product’s value.
One group, from Oregon Health and Science University, found that Infuse had no proven clinical advantage over bone graft for spinal fusion and that it might be associated with some increased risk of cancer. The Oregon group also identified biases in earlier medical-journal articles on Infuse, including the practice of repeatedly publishing selected findings that cast Infuse in a more positive light.
Under less scrutiny, however, universities are struggling to improve the transparency of their own human-trial practices. Some of the apparently biased articles in medical journals have been written by researchers working in corporate labs, but many have been done by university researchers.
They include Thomas A. Zdeblick, a professor and chairman of the department of orthopedics and rehabilitation at the University of Wisconsin at Madison, who promoted Infuse while receiving a reported $34-million in consulting fees and other payments from Medtronic. (Dr. Zdeblick has said his research wasn’t influenced by his ties to the company).
More broadly, however, university researchers have been struggling to meet calls by medical journals and federal officials to cooperate with clinical-trial registries. Such registries consist of two main parts: a public declaration of planned procedures before the start of testing on human subjects, to deter researchers from reporting only favorable results afterward, and a posttrial filing, so other researchers and the public can better assess the conclusions.
But 16 years after Congress first passed a law requiring pre- and posttrial filings, and eight years after an association representing major medical journals approved its own mandate for pretrial registration, much confusion remains.
The 13 publishers in the International Committee of Medical Journal Editors, which approved the group’s pretrial requirement in 2005, have largely been following it, with rare exceptions in cases of genuine mistakes about the requirements, said Christine A. Laine, editor in chief of the Annals of Internal Medicine.
It’s less clear whether, and to what degree, the thousands of other scientific journals are following suit, said Dr. Laine, a clinical associate professor of internal medicine at Thomas Jefferson University. And data from the government’s own clinical-trial registration system also gives a murky picture.
“It’s a mixed bag,” said Kay Dickersin, a professor of epidemiology and director of the Center for Clinical Trials at the Johns Hopkins University, who has studied efforts by university researchers to comply with registration. “A lot of investigators really don’t know what registration is and how they’re supposed to do it.”
To encourage the registration process, the National Institutes of Health created ClinicalTrials.gov as a site to handle both pretrial and posttrial filings. It’s not the only legally acceptable registration site, though it has become the most popular option.
ClinicalTrials.gov now has more than 9,000 posttrial entries, said Deborah A. Zarin, the site’s director. As a marker of the site’s value, Dr. Zarin said, nearly half of those trials are not described in any published articles, meaning the site is the sole source of publicly available information on them.
Over all, though, a series of studies in the past year has shown low rates of compliance. One analysis, published last year in the Journal of the American Medical Association, estimated that the findings of fewer than a quarter of all clinical trials have been posted to ClinicalTrials.gov within a year of publication in a journal. And clinical trials sponsored by industry were three times more likely to have results reported at the Web site than trials financed by the NIH, the study found.
In many cases, university researchers appear unsure of what types of trials need to be included in the registry. They’re also finding the ClinicalTrials.gov site confusing, raising questions about whether the effort is worthwhile.
Rules Delayed
NIH officials agree that the site is hard to use, and say they’re working to improve it. It may take the NIH even longer to resolve the uncertainties over exactly what types of trials need to be registered. Congress used fairly broad language when it wrote the laws mandating clinical-trial registration, leaving the NIH and the Food and Drug Administration—the main agencies responsible for financing research and approving drug safety—to hash out the details through the process known as rule making. But because of complicated subject matter and the involvement of two separate agencies, rule making has now dragged on several years since Congress passed an updated version of the trial registration mandate in 2007.
That’s worrying some in Congress, leading to new legislative proposals to better define the requirements for trial registration without waiting any longer for the rule-making process. “Unreported results and missing registrations leave trial participants, doctors, and researchers vulnerable,” said the author of one such bill, Senator Edward J. Markey, Democrat of Massachusetts.
Even when the final rules are put into effect, researchers will be required to file only summaries of their findings to the ClinicalTrials.gov site. But for some uses, such as Yale’s in-depth review of Infuse, researchers might also want to see data for each person in the trial.
For that possibility, the FDA posted a public notice in June asking for comments on how it could make available to researchers the patient-level data it gets in applications for new drugs and medical devices without revealing patients’ identities.
Again, industry may be leading the way. Glaxo, as part of its promise this year to publicly post all clinical-trial summary data that date back to the company’s 2000 merger with SmithKline Beecham, has proposed also offering patient-level data. To do that, the company is working to establish an independent entity that would review requests for individual patient data in ways that would protect identities and ensure publication of any resulting analyses.
And the industry’s two main lobby groups, the Pharmaceutical Research and Manufacturers of America and its European counterpart, last week announced a set of principles committing their member companies to register trial results and share patient-level data when requested by qualified researchers.
Glaxo, Medtronic, and other companies nevertheless face persistent skepticism about whether they are truly making a fundamental commitment to openness and abandoning old strategies that appeared designed to build consumer demand through deception.
Medtronic, for instance, agreed to share its data on Infuse only after facing heavy public criticism and legal troubles. Glaxo promised to publicly post its clinical-trial data back in 1998, then retreated, then revived the idea after pleading guilty last year in the nation’s largest health-care fraud case.
GlaxoSmithKline’s current senior vice president for science and innovation, Perry Nisen, said he could not explain why the company had abandoned its 1998 pledge. But as for the company’s attitude now, Dr. Nisen said, “it’s fundamentally the right thing to do, for society, and scientifically—and it doesn’t get a lot more complicated than that.”
Charles D. Rosen, a spine surgeon at the University of California at Irvine who has seen patients in pain from bony overgrowth after getting Infuse in an unapproved manner, said he doesn’t think such companies will really change their stripes.
Medtronic agreed to the Yale project only “because they got clobbered” with lawsuits and other criticisms, said Dr. Rosen, a clinical professor of orthopedic surgery and co-founder of the Association for Medical Ethics. “Why on earth does anybody think that because they say this,” he said of the companies’ stated commitment to openness, “they’re going to all of a sudden do it?”
The chief scientific, clinical, and regulatory officer at Medtronic, Richard E. Kuntz, said payments to researchers like Dr. Zdeblick at Wisconsin represented legitimate fees and royalty payments. But Dr. Kuntz said the company has now recognized that such payments can give at least the impression of a financial conflict, and it has forbidden the hiring of journal authors with other financial ties to the company.
The Yale initiative was approved by the chief executive of Medtronic, Omar S. Ishrak, shortly after he arrived at the company in June 2011. While Medtronic is still assessing the full ramifications of that decision, the company has at least two other products in development for which it plans to allow a similar treatment, Dr. Kuntz said. “Our view is that transparency has left the barn—this is going to happen,” he said. “It’s just part of how our society is working, it’s part of the democratic process, and we’re anxious to be able to be part of the designers, to make this thing work for everybody, including ourselves.”
Regardless of what industry now does, universities and their faculty must ingrain skepticism into every medical-school graduate, said Mark Helfand, a professor of medicine and of medical informatics and clinical epidemiology at Oregon Health and Science University who led one of the Infuse reviews. Medical professionals need to see new products and say, “it sounds great—we’re not going to take your word for it, we have to have an independent assessment of the data in order to even believe it,” Dr. Helfand said.
Dr. Krumholz said he’s also hoping for a cultural shift on campuses, so that university researchers also stop trying to hoard their trial data. Rather than try to milk a set of results to generate years of journal articles, he said, researchers should immediately release their data and see what others can do with it.
Producing a widely used set of data, rather than a long list of self-written articles, he said, “should be the highest honor in our field.”