Academic independence is essential to the proper conduct of clinical trials and to safeguard patients’ interests, Martine Piccart of the Jules Bordet Institute in Brussels, Belgium, and Aron Goldhirsch of the University of Bern in Switzerland write in their commentary in the March 8 Nature. The authors belong to the Breast International Group, a consortium of 38 groups of breast cancer researchers who oversee clinical trials involving 60,000 women, and the article is coauthored by eight other international researchers.

Given that large trials require financial and other support from pharmaceutical companies (including the provision of the drugs themselves), academia and industry are obligated to collaborate to bring new therapies to patients. However, the authors note a “push” by pharmaceutical companies for greater control of clinical trials and data, which marginalizes the role of academics. This trend, they write, does a disservice to patients and may actually deter some from participating in trials. When companies control the studies and the data, study design can suffer. Company control of data can also result in withholding or delayed reporting of unwelcome findings, as has happened in several recent cases.

The alternative, Piccart and coauthors write, is a more equal partnership, where the clinical trial database is held by independent researchers while the trial is ongoing, with limited access by the sponsoring company. Once the primary endpoints have been reached, the data is transferred to the company to use in their application for marketing approval. They cite four large and prominently published studies on breast cancer recently conducted in this way (Goss et al., 2003; Piccart-Gebhart et al., 2005; Thurlimann et al., 2005; Coombes et al., 2004), but the model would clearly be applicable to other diseases and treatments.

In their view, this academic independence and openness with the data is the key to successful clinical investigation. “Only by ensuring untrammeled access to long-term data, both good and bad, can we conduct clinical trials in a credible manner,” the authors conclude. “This access will ensure that those patients who consent to participate in them maintain faith in the clinical trial process.”

That theme of the importance of open access to data is echoed in the second report, from Aaron Kesselheim from the Brigham and Women’s Hospital and Michelle Mello of the Harvard School of Public Health, both in Boston, Massachusetts. In the March/April issue of Health Affairs, they argue that the FDA practice of treating all clinical data as confidential may not be appropriate, especially where safety data is concerned.

When companies conduct a trial, all the safety and efficacy data is submitted to the FDA. The primary data stays under wraps, with the results presented in a summary report, which is often written by the applicant company. The presumption behind this scheme is that releasing the raw data could harm the business interests of the company by helping their competitors.

The case of nonsteroidal anti-inflammatory drugs, where safety issues emerged long after drugs were on the market (see ARF related news story and ARF news story), and the incomplete disclosure of dangers related to antidepressants, have raised concerns about the confidential treatment of safety data. Kesselheim and Mello question whether the FDA’s default policy of keeping safety data secret is consistent with public’s interest in disseminating the information, which not only protects people using existing drugs, but also can help efforts to make future drugs safer.

While it may be true that the release of some kinds of information, such as data on chemical composition and formulation, and sometimes efficacy, can put companies at a disadvantage, this is rarely the case with safety data, the authors write. “The legal question is whether the information will give other drug companies an unfair advantage,” Mello said in a press release accompanying the study. “But it is strange to argue that evidence that a drug is harmful will enable others to develop a similar drug.”

In their analysis, the authors cite several examples where the FDA had data indicating safety issues with drugs in clinical trials or on the market, but had no way to act on the confidential information. For example, when approved drugs are tested in unapproved indications, new safety issues can sometimes emerge. By existing policy, these discoveries cannot be revealed to people who are using the drug as labeled.

The data is not always lost to public scrutiny. Consumers can sue to have data released under the Freedom of Information Act. Sometimes this approach is successful, and sometimes not, but getting data this way is always inefficient. This mechanism may not satisfy the public interest in a full accounting of safety data. Lawsuits are expensive and take a long time.

Kesselheim and Mello suggest changes in current policy that would both protect pharma’s trade secrets and allow a fuller evaluation of drug safety. As a first step, they propose that sealing safety data should not be the default mode: companies should have to show why the data should be kept confidential by demonstrating how its release would harm their interests. They suggest replacing the summary document with a full accounting of safety data for all new drugs. If the FDA will not act to change these policies, Congress may need to step in and pass new laws requiring the disclosure of safety data. The goal, they write, is to let external researchers better help the FDA, physicians, and consumers make appropriate medication choices, while balancing the commercial interests of drug companies.—Pat McCaffrey.

References:
Piccart M, Goldhirsch A, Wood W, Pritchard K, Baselga J, Reaby L, Kossler I, Kyriakides S, Norton L, Coates A. Keeping faith with trial volunteers. Nature. 2007 Mar 8;446(7132):137-8. Abstract

Kesselheim AS, Mello MM. Confidentiality laws and secrecy in medical research: Improving Public Access to Data on Drug Safety. Health Affairs. 2007 March; 26(2): 483-491. Abstract