Institutional Review Boards in Crisis

OVERWHELMING VOLUME OF DATA

The classic focus of the IRB is to protect human subjects in clinical trials by evaluating the protocol for risks to patients, balanced against the potential for advancing knowledge, says Ross McKinney, Jr., vice dean for research at the Duke University School of Medicine.

The model was created when most trials were small and conducted at a single site. Times have changed. Over the last two decades, clinical trials have become bigger and more likely to include multiple centers. "We are grappling with how best to handle complex multi-site trials in an effective and efficient manner," says Baldwin.

William Hendee, senior associate dean and vice president of the Medical College of Wisconsin, reviews AEs reported locally, but says he has no context or way of evaluating AEs from other sites. "We're dying out here because of AEs coming at us from everywhere," he says. "It is an overwhelming volume of data that we don't know what to do with." In testimony to the FDA, Hendee showed a slide of an 11-foot stack of AE reports he had received.

The Association of Clinical Research Organizations (ACRO) surveyed its members and found "extraordinary variability in the structure and content of just the simple AE collection form," said Paul Covington, in testimony to the FDA on behalf of the group. It has developed a standardized form that will be presented to the federal agency.

The Association of Clinical Research Professionals has a detailed and far reaching proposal for a national AE database that would stretch from clinical trials to the postmarketing functions of the FDA's MedWatch program. It would be web-based, open to the public as well as healthcare professionals, and financed through a trust fund similar to the one used by the national air traffic safety system.

"The best outcome will come from mining a comprehensive database of timely, reliable information. It can and must serve many goals and masters, recognizing that not all parties need the same things," says Greg Koski, the former director of the US Office for Human Research Protections, who is now affiliated with Massachusetts General Hospital.

Baldwin believes that a centralized data safety monitoring board (DSMB) or its equivalent is essential for multicenter trials. At one time, IRBs could perform some of the duties of a DSMB, such as an independent audit of data at certain time points or triggers within the trial. However, as trials grew in size, complexity, and to multiple sites, the IRBs simply did not have the skills or the resources to perform this function.

IRBs can determine if a protocol has an adequate system for data safety monitoring in place, and to exercise oversight in reviewing the output/recommendations from such a structure. However, local IRBs simply do not have the resources to physically monitor data for safety, Baldwin says. What they do add to "the web of protection" for human subjects is a sense of the local issues that may affect patient safety.

Simeon Margolis sits on the scientific review committee of the clinical research center at Johns Hopkins University. His committee conducted an audit of federal regulations concerning the use of DSMBs in clinical trials. The best they could come up with was a list of "here are the kinds of things that seem to require a DSMB and here are the ones that don't seem to require it," he says.

One principle investigator at Johns Hopkins recently proposed that her mentor chair the DSMB for her protocol. That didn't seem appropriate to Margolis. He was dismayed to find that there are no guidelines as to who is eligible to serve on a DSMB.

FINANCIAL CONFLICTS OF INTEREST

Potential financial conflicts of interest have gotten big play in the media, but they may not be as big a risk to patient safety as other factors, says McKinney. Money is "the easy, the quantitative thing," which is why it gets most of the attention. However, financial conflicts of interest are "not the driver that has me most worried."

He points to the 1999 death of Jesse Gelsinger in a University of Pennsylvania gene therapy clinical trial. "There was money involved in the Gelsinger case, but it wasn't about money, it was ego. How do you quantify ego?"

Foisting review of financial conflicts off on the IRB is another example of mission creep and it's a mistake, says Baldwin. "If that happens you start to say, 'we really need someone with knowledge of investments on this IRB'. No you don't. Deal with that elsewhere, not on the IRB, because conflict of interest is broader than human subjects."

While informed consent is central to the IRB mission, there is a sense that institutions' legal liability is now the bigger concern. The informed consent document is 12 pages long, Margolis says. "They are so overblown in size and complexity that it is almost a waste of time. It's ridiculous to think that most people will understand them."

The Johns Hopkins clinical center has a full-time staff member, a position funded by NIH, whose sole responsibility is to monitor the consent process for comprehension. Margolis says, "She can't possibly do that for every subject on every protocol but she does sit in on at least one such interview for each protocol."

OVERSIGHT FOR INDUSTRY RESEARCH

Part of the job of IRBs in reviewing industry-sponsored trials is "to keep the situation honest" in terms of the trial design, says McKinney. Often a company will propose comparing their drug against an arm that is not optimal in terms of the existing standard of care, he says. Or maybe they will propose dosing "at the low end of the therapeutic range that the FDA has approved. As a result, their new drug will look better because the comparison group is in fact not fair."

However, McKinney adds, investigators and administrators who have professional and financial reasons for wanting to take part in a particular trial can put tremendous pressure on the IRB. "So, we have to protect the IRB's independence."

There is increased interest in "independent" IRBs, which are nonprofit or for-profit groups that are not affiliated with an academic medical center but are hired for the task. However, opinion is divided on the subject.

One way to solve the problems may be accreditation of IRBs. "Having an external independent body that helps lay out standards lets you test them against regulations, gives you mentoring, that can help," says Baldwin. She serves on the governing board of the Association for the Accreditation of Human Research Protection Programs.

Finding faculty members to serve on the board can be tough. Participation on an IRB generally is not directly compensated by the institution, and the workload has expanded and become more burdensome. Duke University has tried to address that issue by proposing to count one year of full-time service on the IRB as the equivalent of writing a first authored paper in a major journal. "We have said that we will explicitly reward you toward your promotion if you participate in the IRB, because it is a community activity that we value," says McKinney. The proposal is before the board of trustees.

"If the IRB can stay focused on the ethical principles underlying the construct of the study and the conduct of the study, they will have done a tremendous service," says Baldwin. "That's a big challenge."