72. The Costs of Decentralized Human Subjects Review Boards: A Case Study from a Multi-site VA Research Project
K Humphreys, VA HSR&D Center for Health Care Evaluation; J Trafton, VA HSR&D Center for Health Care Evaluation; T Wagner, Health Economics Resource Center
Objectives: Protection of human subjects is of paramount concern in health services research. At the same time, different mechanisms for ensuring such protection vary in their efficiency. In the VA, human subjects procedures for cooperative studies are approved through a national review process that applies at all participating sites. However, other multi-site VA studies have a decentralized institutional review board (IRB) process, in which each site follows its own procedures for protecting human subjects. The purpose of this case study was to provide qualitative and quantitative information on the impact of this decentralized approach.
Methods: Qualitative and quantitative data were drawn from a process evaluation of the first three months of a newly established, 8-site VA HSR&D outcome research project (The MOST Study). All human subjects procedures for the MOST Study were approved in advance by the VA Palo Alto, where the study is based; costs were calculated based on what occurred after this initial approval. Data for this case study came from qualitative comments made by staff at each site, and from cost calculations based on staff time and other resources expended by the MOST study team and by participating sites.
Results: Every site (n=9) approached for participation required that all IRB review procedures be repeated at their own medical center. One site considered this so onerous that it refused to participate, and another site was persuaded to participate only when research staff agreed to complete all required IRB paperwork. Each site had unique IRB procedures and required unique information for approval. All site-base changes in human subjects procedures had to be re- approved by the original IRB. To date, these procedures have cost the VA a total of $32,048, divided between $11,384 in costs to the MOST study and $20,664 in costs to the participating sites. Also to date, over 9,000 pages of printed material have been generated by IRB procedures. We estimate that ultimately, between 15-25% of the first year project budget for the MOST study will be consumed by IRB procedures. Further, subject recruitment will be delayed at all 8 participating sites, which will increase costs to an unknown extent over the course of the project. In no case did the additional IRB reviews at each site result in significant changes in any study procedure, suggesting that the added costs had no apparent benefits in terms of increasing protection of human subjects.
Conclusions: Although only a single-case study, this project's experience suggests that the VA could improve the efficiency and quality of multi-site research and reduce its costs if it the national cooperative study IRB procedure was employed in all multi-site studies. An additional advantage of such of an approach is that it would reduce a potential source of bias in outcome studies, namely sites deciding to participate based on the perceived difficulty of local IRB procedures.
Impact: If replicated, the results of this study could lead to modification of IRB procedures that would facilitate the conduct of high quality, multi-site VA research. This would increase the ability of research to improve care quality in the VA.