Examining the Quest to Eliminate Discrepancies in IRB Decisions

Abstract

Beginning in 1990, investigators at Arizona State University (ASU) collected more than 200 blood samples from members of the Havasupai Indian tribe in an effort to describe genetic variants that might be contributing to the increased incidence of diabetes among tribe members. Over the course of the next 10 years, those blood samples were also used to examine DNA variants linked to schizophrenia, alcoholism, metabolic disorders, and the geographic and anthropologic origins of the Havasupai people. This work resulted in more than 20 academic papers, including several that might be considered stigmatizing for members of the tribe and others that were disruptive to beliefs concerning the tribe’s origins (Mello and Wolf 2010). The Havasupai people became aware of the research using their blood samples only after a tribal member attended a research presentation at ASU and asked whether permission had been obtained to use the blood samples for purposes other than diabetes research. Ultimately, a lawsuit was filed against the university. The case was eventually settled and included monetary compensation, an apology, and the return of blood samples so that they could be properly buried (Harmon 2010). The Havasupai research presents an interesting case study within which to consider the issues raised by Klitzman’s article in this issue. On its surface, the original Havasupai diabetes study appears to be minimal risk, and I suspect most institutional review boards (IRBs) would have agreed with that assessment. Yet it’s doubtful that members of the Havasupai Tribe would agree. Likewise, the decision to allow the use of blood samples for purposes beyond those intended by the original research would have been handled differently by different IRBs. Some would have considered the original consent form adequate to cover the new activities, others would have granted waivers of informed consent for the use of the existing specimens, and a few might have considered the new uses to be exempt from the federal regulations if the samples could be de-identified. Of course, the tribe would argue that even with individual deidentification, the community could not be de-identified. While most would agree that IRBs should consider these