Research Disclosure Rules Raise Questions

Recently enacted legislation calls for federal agencies to devise a system in three years that will expand public access to data from clinical trials. That may not be fast enough for some policy makers, researchers, and drug safety advocates. A long delay in releasing less than stellar results from studies on the anti-cholesterol drug Vytorin (ezetimibe/simvastatin) has raised an outcry. And a recent article in the New England Journal of Medicine¹ has reignited concerns about the loss of valuable information from never published trials on antidepressants—evidence of the so-called "waste basket" effect.

Jill Wechsler

Similar issues have been fueling demand for greater access to study data to increase research transparency, reduce reporting bias, and better inform patients and practitioners about more safe and effective treatments. The study disclosure bandwagon took off four years ago when a group of leading medical journals agreed to reject articles discussing results from clinical trials that were not registered in a public database. The journal editors sought to deter sponsors from hiding unfavorable clinical data and to give reviewers access to protocols and full results as a check on accuracy.

The policy also aimed to rebuild public trust in the biomedical research system and to provide reassurance to patients that "their altruism would see the light of day," commented Christine Laine of the Annals of Internal Medicine, at a January conference sponsored by ExL Pharma.

Expanded registration

To further these goals, the Food and Drug Administration Amendments Act (FDAAA) enacted last September requires sponsors to register most clinical trials on the www.ClinicalTrials.gov Web site operated by the National Library of Medicine (NLM) at the National Institutes of Health (NIH). Previously, FDA registration was required only for drugs to treat serious or life-threatening conditions and did not include medical devices. Now more products are covered, and the listings have to provide more detailed information on each study.

Implementing this requirement has been an enormous task for FDA and NLM. In the last weeks of 2007, new trial registrations doubled from about 250 to nearly 500 a week, as sponsors raced to meet a December 26 deadline. Some 12,000 registered studies also had to be modified to include additional data elements, reported Nicholas Ide, chief architect of the NLM Web site, at the ExL Pharma conference. The updates involve adding more information on studies (primary and secondary outcomes, start and completion dates); subject recruitment criteria; and study sites and contacts. FDA still exempts most Phase I studies from registration, but the World Health Organization does not—one of the many inconsistencies among proliferating registration requirements around the world.

While drug trial sponsors have to post required data within 30 days of FDA study approval, the rules are very different for medical device makers. Those manufacturers convinced Congress that it would stymie innovation to disclose device study details during the investigative period. FDAAA thus prohibits posting medical device study information prior to market approval. Some large medical device companies now would like leeway to permit earlier registration, but that would involve amending FDAAA.

Linking to results

There are lots of new requirements governing clinical trial listings, but the more contentious FDAAA requirement is to link study registration to resulting safety and efficacy data. Sponsors have been posting some research results voluntarily in response to the rising clamor for more transparency in study outcomes, many through a Web site launched by PhRMA (Pharmaceutical Research and Manufacturers of America) in 2004.² The new legislation codifies and expands this activity.

The first step, which already is in effect, links www.ClinicalTrials.gov registrations to public information on the resulting product. This includes FDA's drug approval package, documents discussed by advisory committees, public health advisories, NLM Medline citations of published articles, and the product's approved label in the NLM's DailyMed database.

By September 2008, the results database also has to include "basic results" from clinical trials supporting approved products. A sponsor must provide baseline characteristics of study participants, primary and secondary outcomes, point of contact for scientific queries, and information on sponsor agreements with investigators that could restrict their ability to discuss or publish trial results. This disclosure must be made within 12 months of trial completion, which occurs when investigators stop seeing patients, not when all the analysis is done.

The next requirement is to establish a process for including adverse event reports in the results database. FDA has to propose new rules by March 2009 for listing information on serious and frequent adverse events. If the agency fails to devise a new policy, the law reverts to a default position that requires sponsors to post most adverse events.

The most problematic requirement is for sponsors to provide summary information on study results. By September 2010, FDA has to define how technical and nontechnical ("plain language") summaries should be constructed so as not to mislead the public and to avoid any appearance of being promotional.

Questions, questions

Implementing all these requirements is a "huge task," observed Terry Toigo, director of FDA's Office of Special Health Issues, at a November conference on FDAAA sponsored by the Food and Drug Law Institute (FDLI). There are different posting requirements for ongoing studies and for newly launched clinical trials, she noted, adding that a unique identifier system for FDA IND submissions and for clinical trial registration is critical to being able to link all the information.

Regulations will have to spell out what protocol information is needed in providing full results, how quickly this additional information has to be submitted after product approval, and what kinds of disclosures would be considered false or misleading. Informed consent rules have to be revised to advise subjects about future data results disclosure, and guidance is needed on how to deal with study outcomes that differ from initial protocols.

Although registration does not apply to most Phase I studies, it might cover early trials in patients, as opposed to healthy volunteers. It's also unclear whether bioequivalence studies for generic drugs should be registered and whether sponsors have to post study results for products that never come to market. FDA and NIH will hold public meetings and sponsor a pilot study to examine best ways to verify the accuracy and quality of submitted results data.

Certification of compliance is a major headache. The law says that companies have to certify compliance with trial registration and results reporting requirements in applications and submissions to FDA, as well as in NIH grant applications. This is generating a vast volume of paperwork because sponsors are interpreting the policy to cover all IND supplements—they want to err on the side of caution to avoid the stiff penalties attached to violation of disclosure rules.

Journal publication is another thorny issue. The law provides little leeway for sponsors to delay disclosure pending publication of articles stemming from a clinical trial, and journal editors are not sure how the tighter rules will affect them. Meanwhile, sponsors continue to fear that detailed disclosure will divulge proprietary research information and that the release of data out of context could raise unnecessary public alarms. But too slow disclosure in the past and attempts to sweep unfavorable results under the rug have undermined the credibility of such objections.

Disclosure advocates believe that access to prior research data and negative results will benefit all parties and open the door to more and better meta-analysis. Innovator firms find the new requirements for establishing better trial tracking and monitoring systems may be burdensome but that such systems can improve companies' corporate compliance with global research requirements.

Expanded federal disclosure requirements also promise to strengthen FDA pre-emption of diverse state registration requirements. However, this protection does not go into effect until national disclosure standards are fully implemented in three years, making it necessary for industry to comply with local rules already in effect, such as in Maine.

Broader goals

Clinical research disclosure policy clearly has moved far away from its original purpose back in 1997: to help seriously ill patients enroll in clinical trials. The broader aim now is to ensure that research outcomes data is available to help patients and health care professionals make treatment decisions. But while transparency is a worthwhile goal, Toigo commented last fall, it's important to consider the potential public health consequence of providing this vast amount of information.

Jill Wechsler is the Washington editor of Applied Clinical Trials, (301) 656-4634 jwechsler@advanstar.com

References

1. Erick Turner et al., "Selective Publication of Antidepressant Trials and Its Influence on Apparent Efficacy," New England Journal of Medicine, 358 (3), 252–260 (January 17, 2008).

2. Pharmaceutical Research and Manufacturers of America Web site: www.clinicalstudyresults.org.