Twenty-Six Percent of Boxes Go Unchecked

Approximately 74 percent of U.S. institutions with federalwide assurances apply all or part of 45 CFR 46 to all research regardless of support, down from more than 90 percent of such institutions a decade or so ago, according to OHRP officials.

[Carol Weil, Lisa Rooney, Patrick McNeilly, Karena Cooper, Kristina Borror, and Paul Andreason, "OHRP Compliance Oversight Letters: An Update," IRB: Ethics & Human Research 32, no. 2 (2010): 1-6.]

The authors write that since their previous review of determination letters, published in 2003 and covering the period 1998-2002,

more institutions have decided not to extend their FWA to research not supported by HHS, which means that OHRP has jurisdiction over fewer research studies than in the past. Based on an informal review of a sample of institutions, approximately 74% of domestic institutions currently holding an FWA that formerly held a Multiple Project Assurance (MPA) apply either subpart A or subparts A, B, C, and D of 45 CFR 46 to all research regardless of support. On the other hand, greater than 90% of those same institutions applied 45 CFR 46 to all research regardless of support when they held MPAs.

Along with that finding, the authors "describe [their] review of 235 compliance oversight determination letters that the Office for Human Research Protections (OHRP) issued to 146 institutions between August 1, 2002, and August 31, 2007." Such a large dataset would seem to be a rich source of information for those interested in how researchers and institutions go wrong.

Unfortunately, determination letters are generally opaque documents that fail explain the events leading up to a complaint. For example, OHRP's March 2007 letter to New Mexico State University tells us a bit about office procedure at New Mexico State, but nothing about the design of a project called "The Impact of Education in Navajo Nation Border Community Public Schools on the Hearts, Minds, and Spirits of Navajo Students," or whether that project violated any research ethics.

Because of this opacity, a review of determination letters can tell us that 56 percent of institutions were cited for problems with the "IRB initial review process" while only 3 percent were cited for misapplication of exempt categories of research. But it can't tell us what percentage of citations concerned ethnographic fieldwork, what percentage involved graduate students, or what percentage involved questions of whether data were publicly available. Nor can it tell us whether any one of these 235 letters documented actual harm to a research participant.

(I am puzzled by the finding that OHRP issued 260 citations for deficient "IRB-approved informed consent documents/process" but only 19 for "Failure to obtain informed consent of subjects." Does that mean that deficient processes still work 93 percent of the time?)

OHRP is missing an important opportunity. When courts rule on cases, they take care to explain not only the legal principles involved, but also the underlying facts, so that other courts can understand which principles should be applied to certain fact patters. By failing to document the substantive concerns that led to its investigations, OHRP has failed to educate researchers and IRBs about what kinds of research are most likely to lead to trouble.