The amount of clinical data that drug companies must share with the public could soon vastly expand under a U.S. regulation proposed today. Trial sponsors would need to report summary results for drugs and devices that are never approved—and not just for products that reach the market—under the proposal.
The goal of the plan, released by the U.S. Department of Health and Human Services (HHS), is to ensure that summary results for drugs that fail in trials, or are dropped for other reasons, still make it into a public database: ClinicalTrials.gov. Sharing these results should not only be useful for researchers, but also “helps fulfill society’s ethical responsibility” to people who volunteer for trials, said National Institutes of Health (NIH) Director Francis Collins during a press teleconference. “We owe to our patients, to our participants in these trials, the explanation of what happened.” Sharing the data may also help researchers avoid duplicating failed trials, officials say.
Launched at Congress’s request in 2000, NIH’s ClinicalTrials.gov now contains registration data for more than 178,000 trials; all sponsors of trials regulated by the U.S. Food and Drug Administration (FDA) are supposed to register. As a result of a 2007 law, within 1 year of the completion of a trial, drug companies must also submit summary results that include information such as the number of participants, their age and gender, outcomes, and adverse events. These results have been submitted for more than 15,000 trials.
But the summary results requirement only applies to drugs and devices approved by FDA. Under the new rule proposed by HHS, companies will also need to report results for unapproved products. That should add another 100 to 150 reports to the 100 summaries the database now receives each week, said Deborah Zarin, director of ClinicalTrials.gov.
Trial sponsors can ask for an extension of up to 3 years if they are still seeking approval for a product. And results still won’t be required for early trials, known as phase I, which test a product’s safety.
NIH also plans to require that results for all clinical trials that it funds, no matter how early or small, be submitted to ClinicalTrials.gov. At the moment, only investigators conducting trials of products regulated by FDA must register the trial and report results. A draft policy released today will mean that results for another 650 trials per year will be added, NIH officials say.
Both the NIH and HHS draft rules also clarify such details as who should submit data and when it must come in; trials must be registered within 21 days of enrolling the first patient, for example. This will make it easier to enforce compliance with the rules, for example by withholding funding for NIH grantees or imposing fines on companies regulated by FDA, said Kathy Hudson, NIH’s deputy director for science, outreach, and policy. The public may eventually be able to check the database to see if researchers have met the deadlines, providing a “wall of shame,” NIH officials joked.
One advocate of clinical data sharing, Peter Doshi of the University of Maryland School of Pharmacy in Baltimore, says he welcomes the new requirements. Some NIH-funded trials “are not published for years and years,” he notes. And often negative results never appear in a journal. “That’s not good from an evidence perspective and it’s embarrassing from a policy perspective. There’s an interest in trying to close these gaps.”
Requiring that trial results for unapproved drugs be shared is also a positive step, Doshi says. “You’re reducing the chance that somebody will redo experiments that were already done and put people in harm’s way because the research wasn’t shared,” he says.
However, the summary results in ClinicalTrials.gov can be confusing—they sometimes conflict with results reported in journals—and they lack details that could help researchers understand the data, Doshi notes. By contrast, the European Medicines Agency plans to make detailed clinical data reports publicly available and to provide data on individual patients to researchers—albeit only for approved drugs.
“The white elephant in the room is that FDA sits on more data, across more drugs, across more therapeutic indications than anybody else on the planet,” Doshi says. “And their attitude is, ‘Great idea [to make the data public], let somebody else take care of this.’ ” Doshi is an associate editor of The BMJ, which is part of a group called AllTrials that is pushing for release of detailed clinical trial data.
NIH’s Hudson said the agency is looking at whether to require access to data on individual patients without compromising privacy. The proposal also asks for input on whether trial sponsors should be required to submit their protocol to help the public understand the results.
NIH and HHS are taking public comment on the two proposed policies until 19 February.
*Correction, 20 November, 7:56 a.m.: The number of additional weekly submissions the new rule is expected to generate has been corrected. The correct numbers are 100 to 150 addtional submissions, not 200 to 250.