Is centralizing IRB duties the wave of the future?

There are different models for centralization

As IRB staff and members find their workload increasing more quickly than their resources, some may consider using a centralized IRB model as a way to allow them to spend their time and energy where it is most needed.

The idea is that when protocols are submitted for research that will be conducted at multiple sites across a region or the country, then the best way to handle these could be to cooperate with one IRB who will review the initial protocol and all continuing reviews. The other IRBs that are involved could reserve the right to disagree with a review or make changes to suit their community.

"Central IRBs usually operate in an environment in which there are no local IRBs, and the model here is we are involving all of these institutions that have local IRBs," says Jacquelyn Goldberg, JD, project officer on the Central IRB Initiative of the National Cancer Institute (NCI) in Bethesda, MD. The initiative is a pilot project sponsored by NCI in consultation with the Office of Human Research Protections (OHRP). "People are looking for new ways to protect the rights of research participants, and I think this model is one of several that will be out there."

For some large multisite trials, there could be hundreds of IRBs reviewing the same protocol, and there is no evidence that this provides better human subject protection than if the same protocol were reviewed by a central IRB that works with all of the local IRBs, Goldberg adds.

Also, a centralized IRB review reduces both IRBs’ and investigators’ administrative burdens and frees time for IRBs so that they may concentrate on actual conduct of research at their institutions and staff education.

"We have created a model where we have a central IRB and local IRBs that share regulatory responsibility," Goldberg explains. "The central IRB handles initial and continuing review protocols, and the local IRB’s function is primarily a consideration of the local context and oversight of the local performance."

In another model for centralizing the IRB process, an IRB could have a cooperative arrangement with other IRBs in the geographical area so that, for certain multisite protocols, several different IRBs could rely upon one IRB for the main review process, says Karen Hansen, director of the Institutional Review Office of Fred Hutchinson Cancer Research Center in Seattle. Goldberg and Hansen describe how these two centralized IRB models work and what might be gained from becoming part of a centralized IRB process:

1. One IRB reviews a protocol before it is distributed nationally. In the case of the Central IRB Initiative, the central IRB reviews a protocol and makes its decision and all protocol materials available on-line to local IRBs, Goldberg explains.

"The local IRB sees everything the central IRB utilizes in making a decision," she says. "So when the local IRB has all that stuff, they don’t have to do a full board review."

For instance, the local IRB could assign the chair or a committee to review the central IRB’s documents for local concerns and to decide whether it will accept the central IRB’s decision. If the local IRB accepts this facilitated review, then IRB staff notify the central IRB via e-mail, and the central IRB becomes the IRB for life of the protocol, conducting all continuing reviews, amendments, etc, Goldberg says. "The local IRB remains in the information loop when the central IRB makes decisions."

However, if the local IRB decides not to accept the central IRB’s review, then the local IRB can conduct a full board review of that protocol, she notes.

2. A group of IRBs cooperate and more efficiently handle multisite research protocols. "I support a centralized IRB review provided that all parties involved are truly engaged in the spirit of the process," Hansen says.

For example, the IRB at the Fred Hutchinson Cancer Research Center has developed a cooperative arrangement with other IRBs so that in cases of certain multisite research protocols, one IRB could be the IRB of record, she explains.

"You can avoid duplicate reviews by permitting another IRB to be the one of record," Hansen says. "These types of arrangements are documented in an assurance if you receive federal funding."

Then the IRBs that have allowed another IRB to be the IRB of record will provide an administrative review of the protocol that will be fully reviewed by the selected IRB. This is not the same as a full review or expedited review, and is listed only as an administrative review that either approves or does not approve of the full review done by the IRB of record, she explains.

"Typically, your agreement with the other IRB states that we always reserve the right to conduct our own full review," Hansen adds.

3. Work toward developing trust between central IRB and other IRBs. For the model in which a group of IRBs share responsibilities for reviewing multisite protocols and also share opportunities to be the IRB of record, it’s important that IRB members know how each of their partner IRBs operate, Hansen notes.

For example, the Fred Hutchinson Cancer Research Center IRB has exchanged IRB members with a partner IRB. Members can sit in on this other organization’s IRB meetings and gain firsthand knowledge of their processes and procedures, she adds. "It’s a good way to get a sense of the review process, and seeing how the IRB is conducting its review helps to promote understanding. We would send two visiting IRB members a packet, and they were welcome to come and share in deliberation with the rest of our IRB reviewers."

This trust-building process continued for six months, and the Fred Hutchinson Cancer Research Center IRB members also would receive their packets from the other IRBs and, likewise, attend their meetings, she reports.

"That was one step to lead us toward coming into an arrangement of IRB acceptance," Hansen says. "I went to these meetings also, and I can only say that both IRBs gained from it."

NCI’s central IRB now has about 150 local IRBs involved in the initiative, and there is potential for the initiative to include 1,500 IRBs, Goldberg notes.

All of the IRBs involved in the Central IRB Initiative review adult, Phase III, NCI-cooperative group trials, which are the only kind that the central IRB reviews, she says.

So far, the anecdotal evidence suggests that the IRBs who have used the central IRB’s facilitated review have been satisfied with the process, Goldberg adds. "Investigators appreciate the ease of use, and the local IRBs think the quality of the review is really high," she reports. "Now, we’re beginning an evaluation plan to measure this more formally, so we’ll have objective qualitative data."

4. IRBs can focus on their strengths. When a group of IRBs work together to share reviewing responsibilities, there is potential for improved quality and a streamlined process, saving resources, Hansen says.

For example, one IRB would be the lead IRB and conduct the review of a protocol because that IRB’s institution would be the main site for actual intervention with human subjects, she explains.

Then another IRB could oversee the laboratory and research analysis of the same study because that site has greater expertise in these areas and its site is where lab work and analysis will take place, Hansen adds.

Since the IRBs involved in a cooperative arrangement have developed trust and understanding, it’s easy to keep communication lines open.

"Because of the dialogue and communication we have with other IRBs, if there is anything we notice that might need more attention, then we can just give the IRB office a call and share that information," Hansen says.

5. Improve technology and infrastructure to facilitate central IRB processes. NCI’s Central IRB Initiative includes experts from across the country, so some IRB members have to travel long distances to attend IRB meetings, Goldberg says.

To make the board meetings more efficient and convenient for IRB members, the initiative plans to make better use of electronic technology, including having board meetings conducted through the Internet, teleconferences, and hybrids of both, she adds.

For instance, IRB members would access a specific web site at the appointed meeting time, and while they are on the telephone in a conference call with the other board members, they could scroll to places in the web site where protocol information is available, Goldberg explains.

For IRBs involved in a cooperative agreement, the process might be slower than desired if some of them don’t have access to the Internet.

"We use all hard copy at this point," Hansen notes. "However, we do meet fairly regularly with two other academic IRBs because we’re all part of a consortium, and as time goes on there could be more blending of processes electronically."

Some IRBs have been using enhancement grant money to update their electronic capability, and when all of the participants have made this technological progress, there will be potential for coordinating electronic submissions, she adds.