Protecting human subjects from experiments that might put them at risk is necessary.
IRBs, as presently constituted, are a menace to navigation (even when they aren’t simply an excuse for the research censorship).
One root problem is simply that if you give a bunch of people the job of meddling with other people’s research, they’re going to figure out ways to meddle just to avoid feeling useless. Another is the fetishization of “consent.”
If I’m going to do something that puts you at risk, I ought to have to get your informed consent before doing so. (Though note that, under this standard, all journalism would be forbidden.) But if I’m going to do something that doesn’t put you at risk — for example, institutijng a checklist to make sure that your medical providers observe some basic hygiene and safety precautions — then the question of whether you “consent” to be protected doesn’t really arise.
Here’s my radical proposal: Once a protocol has been submitted to an IRB, that body may not deny the application, delay it, or ask for more information unless it can specify some way in which the proposed activity might put some actual human being at actual risk of actual damage. I would add additional procedural protections for researchers: a time limit on IRB consideration; the elimination of the arbitrary “clinical hold” under which an IRB can stop a research effort at any time for no reason; having at least two IRBs at each institution, giving scientists the option of choosing a forum; the right for any scientist whose research is not approved by the IRB when first considered to appear in person to make the case; and a requirement for a written, and appealable, statement of cause whenever an IRB denies approval.
Respect for “science” — teaching evolution, taking and listening to honest scientific advice, allowing embryonic stem cell research — has become something of a political issue. I think it’s a winner for liberals, because even people who don’t know anything about science mostly revere it. (The hostility to science from the Right bears all the marks of anti-clericalism in a country with an established church.)
But there’s been no substantial attempt to mobilize scientists as a political force by appealing to their professional interests, despite the fact that scientists are fairly numerous and prosperous, with some capacity to act as opinion leaders. If I were a Democrat running for President, I would propose a doubling of the NSF research budget (cheap, since the NSF gets less than $5 billion a year to spend on research), increasing the NIH budget by 50% (much more expensive, but with NIH now funding only the top 6% of the proposals it gets, there’s obviously plenty of good science going undone for lack of money), and reforming the IRB process.
Indeed, I might do the same thing if I were a Republican hoping to get some non-Neanderthal vote in November.
Update Nick Burbules and colleagues wrote a long white paper on this point and a ran a shorter editorial in Science. Key observation: it isn’t the case that the more obnoxious the IRB is to scientists, the better the protection it provides to subjects.
Researchers increasingly think of IRBs as the “ethics police.” In fact, all researchers must take primary responsibility for professional, ethical conduct. Our systems should reinforce that, not work against or substitute for it; the IRB should be a resource, not the source, for ethical wisdom. All compliance systems require the buy-in and collaboration of the regulated, and it will be a sad day if scholars come to see human protection in research as the source of frustrating delays and expensive paperwork.
[An] obsession with paperwork and mechanical monitoring may undermine protection of human subjects. IRB members spend too much time editing documents, marking typos, and asking for more details. One researcher, 10 years into a longitudinal study, was asked by an IRB to remove the term “anemia” from consent forms because participants might not understand it. Such actions, about which we hear frequently, carry a serious risk: They reduce trust in the guidance of IRBs and may alienate some researchers enough to turn them into scofflaws.