Institutional Review Boards for Everyone

Rebecca Tushnet

The Code of Virginia states: "Every person engaged in the conduct of human research or proposing to conduct human research shall affiliate himself with an institution or agency having a research review committee, and the human research which he conducts or proposes to conduct shall be subject to review and approval by such committee in the manner set forth in this section." Human research is defined as "any systematic investigation, including research development, testing and evaluation, utilizing human subjects, that is designed to develop or contribute to generalized knowledge," the same as the definition used by the federal government for programs it funds.

The government can decide to fund only research meeting its standards (setting aside differences between government speech, which can express the government’s viewpoint, and government-funded public forums, which can be topic-limited but not viewpoint-discriminatory). But can it impose such restraints on anyone conducting research at all? Where the question is medical research involving a physical intervention, the answer is likely yes; such restraints surely have a rational basis. But what about when the research consists of asking people questions, as pollsters and sociologists, among others, do? Then it seems we might have a First Amendment problem.

This seems like a content-based restriction on speech: If you are doing human research, you have to get approval from an IRB. If you are doing nonsystematic investigation, you don’t need approval. This matters because (1) IRBs are expensive, thus raising the costs of speaking to research subjects and of producing the ultimate research, (2) the necessary affiliation with an IRB-having institution may be difficult to achieve for an independent scholar, deterring and even foreclosing speech.

If the government isn’t funding the research, it seems as if strict scrutiny would apply. The government may well have a compelling interest in preventing harm to research subjects. Such harm doesn’t just come from side effects of medications; it can be harm from being asked to revisit traumatic subjects, or from invasion of privacy. (If the law does apply to political pollsters, as it appears to, the government may not have a compelling interest in preventing people from being confronted with unpleasant political topics – but I’ll assume it does, and say that the political nature of some of the affected speech goes to the countervailing concerns for allowing people to be exposed anyway.)

The big problem with Virginia’s method of addressing these harms is that we have some conventional torts to deal with them, like invasion of privacy and intentional infliction of emotional distress. If those are for some reason insufficient, we could try rigorous informed consent requirements to ensure that people know what they may be risking when they participate. In other words, there are a range of less restrictive means the government could use to target the harm. 

Relatedly, the law is underinclusive in odd ways. Nonsystematic investigation can also harm privacy and revisit old trauma. Indeed, journalism is far more likely to do so than sociological research. Normally, legislatures can choose to address whatever portions of a problem they wish to; half measures are constitutionally legitimate. But when it comes to speech, substantial underinclusiveness is evidence that the government has blundered and that its intervention will cause unacceptable distortions in the marketplace for speech. Here, it doesn’t seem like a good idea to make systematic research more difficult to conduct than anecdotal reportage.

Nonmedical research should, of course, be conducted according to appropriate professional norms. But those norms aren’t always about avoiding any individual harm. Historians, for example (who are increasingly subject to claims by university IRBs that their research requires IRB approval), do harm to reputations when they write critical histories. Virginia’s law is an example of a necessary idea – oversight of clinical research – taken too far, and one of the casualties is free speech.

Comments

[Dean C. Rowan]

If I read it correctly, Virginia's definition of human research specifies a limitation: "Human research shall not be deemed to include research exempt from federal research regulation pursuant to 45 C.F.R. § 46.101(b)." Among the variety of exemptions in (b) is "Research involving the use of . . . survey procedures, interview procedures or observation of public behavior. . . ." Exceptions to this exemption are provided to prevent damage--from publicizing names of survey respondents, for example--to subjects of this kind of research as well.

So, it appears that political pollsters are safe. In any case, and even if my cursory reading of the statute and regulation is incorrect, an apt elaboration of the problem identified in this post--the failure of a FA challenge to have arisen against what appears to be a state restriction on speech--is Frederick Schauer's The Boundaries of the First Amendment: A Preliminary Exploration of Constitutional Salience, 117 Harv. L. Rev. 1765 (2004). There, Schauer discusses the range of regulated behaviors involving speech that simply do not (yet) command FA attention. He also discusses the "magnetism" of the FA, and its potential usefulness as a defense against charges of violations of regulations such as Virginia's. The upshot is that "nondoctrinal" factors (i.e., cultural, social, symbolic factors and the like) can help us to account for why it is sometimes acceptable (so far...) to regulate speech we would otherwise imagine protected.

[Rebecca Tushnet]

I think Schauer's work is great; I'm using it heavily as I try to think through why advertising law hasn't really felt the First Amendment's sting yet. I'm not sure that anything in his theory suggests that Va's law would be safe if it were interpreted to cover historical-type research; the fact of the matter is that it's simply unlikely to be enforced at all. I should note that the exemptions to which you refer have not stopped university IRBs from insisting that their authority covers history, sociology, psychology, and so on, and claiming justification in the federal regulations. That's not to say they're right -- indeed, they seem quite wrong -- but their imperialism within institutions is pretty stunning.

[Dean C. Rowan]

The Schauer piece in fact suggests that the law would succumb to the FA once it is perceived to have fallen under the FA's purview. The FA's "magnetisim" generates a good deal of extra-doctrinal force, according to Schauer.

I gather a fear of some sort of liability prods the imperialist universities to cover their collegial butts. As you point out, there are less restrictive -- and less expensive -- means to do this than more bureaucracy embodied in review boards and their attendant procedures. Why don't they retreat in the face of those costs?

[Rebecca Tushnet]

I think it's a matter of power within the university. Once the IRB
exists, it tries to expand its mandate. And once it gets the idea that
someone other than the researcher should be reviewing research
protocols to protect subjects, it's easy enough to expand that to other
areas. The problem is when the specific means of protecting research
subjects useful for medical research get generalized to areas where "do
no harm" is not the rule, like the writing of history. The IRBs do have
a reasonable story -- it's bad to hurt people unnecessarily -- but they
don't usually have sufficient representation from nonmedical
specialties to modify their requirements for specific fields.

[Jerry Menikoff]

I don't claim to be a First Amendment scholar, but I do know a bit about how we regulate the world of IRBs. Many of the complaints relating to what IRBs are doing with regard to social science research stem from institutions and their IRBs not understanding the rules. In my non-FA-expert opinion, if the rules are applied correctly, the system actually provides minimal hindrance to the conduct of non-medical research, and those minimal effects are extremely unlikely to lead to FA problems.

In any event, for more on this question (including from people who are FA experts), you might want to check out the output from the conference that took place at Northwestern University back in April, Censorship and Institutional Review Boards. Conference co-organizer Philip Hamburger started the whole thing going by having earlier written a paper claiming that the system does indeed violate the FA. The full set of papers from the conference is going to be published in February 2007 in Volume 101, Issue 2, of the Northwestern University Law Review. I know that Jim Weinstein's paper, for one, will deal with Schauer's views (and Weinstein concludes there isn't a substantial FA problem).

By the way, at least in the world of medical research, there are good reasons to think that the protections being afforded research subjects are not as strong as they should be. I describe this in a recent book, What the Doctor Didn't Say: The Hidden Truth about Medical Research (Oxford 2006). Richard Epstein, who provided a blurb, found the state of affairs described in the book to be "disturbing."

[Rebecca Tushnet]

Thanks for the comments -- excellent points all, though I note that First Amendment jurisprudence cares about implementation as well as theory. That's the source of a number of speech-protective doctrines that require us to err against regulation, which we wouldn't need to do if regulators could reliably sort truth from falsity, dangerous speech from nondangerous speech, etc.

I look forward to reading the conference papers.

I absolutely agree that medical research is a different animal, which is yet another reason for IRBs not to apply the same standards across the board.