It was with interest that I read the excellent PLoS editorial of June, 2010 and the ensuing comments. As President of the commercial, independent IRB central to the article, I wish to add a few comments more from the IRB point of view.

Determinations about Human Subjects Research: A boundary issue

The Belmont Report’s first section describes the distinction between research and practice and emphasizes that their three enunciated principles are relevant to research. The regulations that followed defined the terms “research” and “human subject” in 45 CFR 46.102. Using these definitions, the regulations describe the regulatory boundaries of the IRB’s jurisdiction. There have been many impassioned articles and white papers decrying mission creep. Once we understand where the boundaries are we can apply the regulations appropriately.

FDA boundary definitions are different. Their regulations are in two parts: 50(informed consent) and 56 (protection of human subjects). Each of these has different applicability criteria and can be used independently.

Consent is a far larger concept; it occupies a place in etiquette, clinical care and research. Certainly, a finding that an activity is not research or not human subjects does not negate the need for consent.

Interestingly, few if any IRBs made this decision until journals adopting the Uniform Requirements for Manuscripts Submitted to Biomedical Journals started asking authors,- who believed that they had not involved human subjects - for evidence of their IRB review. IRBs are usually asked to make the decision after a journal has raised the question and thus, such decisions are generally retrospective and not at all the oddity implied in one comment.

Were there Human Subjects?

Our form is designed to parse the definitions used by both the Common Rule and FDA. The decision was that 23andMe was conducting research but that they had no human subjects. The words in these definitions are, indeed, arguable. (If they weren’t we would not need reviewers.) Whether there was intervention or interaction and whether the information available to the researchers was and would continue to be not identifiable to them, were core to the decision. Our decision was made based on the facts presented in the written application, followed by multiple questions, and a visit to the site.

Conflict of Interest and Commercial IRBs

All IRBs, whether independent or internal to an organization (aka, a dependent IRB), commercial or academic, have inherent conflicts of interest. We, in this private world, are very aware of our potential conflicts and work with them daily. In fact, in this case we could have charged more by determining that human subjects were involved in the study and conducting a full board review.

Consent Information

We have continued to work with 23andMe on other activities and have had long discussions regarding consent. We found, as did you and others before us, that what they are doing is often difficult to put into lay language. However, the people voluntarily responding to the surveys are their clients who have devoted time and considerable money to learning about their genetic makeup. Although “lay” they are likely to have more than a passing vocabulary of genetic terms.

In reviewing the consent document we have focused on the essential elements of consent; the fact that there is research, that there is no intent to provide individual benefit, and that participation is voluntary. The editors are right, of course; we agree that every IRB review can lead to an improved consent process. Certainly every consent reviewer believes he or she can write a better version.

Of course, even when IRB review is not required by a regulation, it can be voluntarily required by any institutional or company policy, as has been the case with 23andMe and others. IRB review is about more than consent as everyone involved in this debate knows.

IRBs and GWAS Studies.

In calling for standardization of a consent process for GWAS studies, there must first be clear understanding of the separation or overlap of customer service and research procedure. This is a fact-based decision. Only then would IRB jurisdiction over the research activity be appropriate. These decisions must be based on fact rather than belief.

We applaud your apparent conclusion; that the consent process “is more informed than most processes that have been formally reviewed.“ Consent forms have become long and complex. The more they attempt to explain everything, the farther they move from the essentials; that is, that this is research and participation should be voluntary.

Finally, it has been a pleasure working with 23andMe as we have worked through these issues. I was not surprised that there was heated response to the editorial but have been surprised by opinions dressed in the presumption of knowledge.