Alternative Models of IRB Review

By Amy Davis, JD, MPH, Senior Director for Publications and Programs at PRIM&R


There was a lot of talk about alternative models of institutional review board (IRB) review at PRIM&R’s 2011 Advancing Ethical Research (AER) Conference:

  • There is the Advance Notice of Proposed Rulemaking’s (ANPRM) request for comments on a proposed requirement that all domestic, multi-site research rely on central IRBs;
  • There are National Institutes of Health (NIH) research grants that favor central IRB models;
  • The National Cancer Institute (NCI) is piloting a central IRB model; and
  • The Department of Veterans’ Affairs relies on central IRB review for some multi-site research. 

So, why all this talk about centralization? Efficiency, reduction of duplication of effort, harmonization, speed of review, reduction of institutional conflicts of interest, improved availability of reviewers who have expertise in the research area, were just some of the reasons that I heard.

Few people would challenge the idea that multi-site IRB review can be cumbersome, time consuming, and perhaps, redundant.  Based on what I heard at two breakout sessions at the 2011 AER Conference, there is much work to be done to figure out a better system.

At session C5, ‘Two Roads Diverged in a Yellow Wood:’ The NCI Central IRB (CIRB) Considers a Model Change’, the presenters discussed the NCI’s initiative to pilot an independent CIRB system.  In recent years NCI has been using a facilitated IRB review model that relies on a partnership between local IRBs and the CIRB. Under this model, local investigators are allowed to download protocols and informed consent documents that have been pre-approved by the NCI CIRB. The local investigator then submits the pre-approved documents to his or her local IRB for a review of local context concerns. Designated individuals from the local IRB decide whether they are comfortable relying on the CIRB as the primary reviewing IRB for the protocol, or if local concerns are an issue, they may submit the protocol to the local IRB for full board review.  The goal of facilitated review is to reduce the number of local IRBs conducting independent review of a single protocol. 

NCI is now exploring further refinement of this process in an effort to eliminate local review, so that the CIRB becomes the single IRB of record. NCI is conducting a pilot study of this process in collaboration with approximately 30 institutions. This model will require all participating institutions to rely fully on the CIRB. Local issues would be managed through the use of forms that prompt investigators to submit information about local context to the CIRB.

At D5 session, ‘Independent IRB Models’, there was a discussion of how independent review boards external to the institution review research.  This model is distinct from the NCI CIRB model in that it is a review system that is independent of the researching organization, and may review research for single-site or multi-site research. However, the discussion during this session was focused on how to set up a central review system for multi-site research.

The questions that came up in both sessions related to how to review unanticipated problems that involve risks for subjects. How should these be reported and managed? Similarly, how should local issues related to conflicts of interest be managed? Other questions related to the management of informed consent documentation were raised. In all cases it seems that local concerns and operations cannot be ignored.

A key lesson I took away from these sessions is that institutions and researchers who rely on a central or independent review system must always remember that the research enterprise is conducted by a human research protections program that involves the researcher, the IRB, (regardless of where it is located or with whom it is affiliated), the sponsor, the sites where the research takes place, and of course, the research subjects. No matter what model is used for the review of human subjects research, no entity that is part of the HRPP relinquishes responsibility for the protection of subjects.