You might think that the answer to this question is obvious. Clearly, it’s your business, and yours alone, right? I mean, sure, maybe it would be considerate to discuss the potential ramifications of this activity with your partner. And you might want to consider the welfare of the bee. But other than that, whose business could it possibly be?
Well, as academic empiricists know, what others can do freely, they often require permission to do. Journalists, for instance, can ask potentially traumatizing questions to children without having to ask whether the risk to these children of interviewing them is justified by the expected knowledge to be gained; academics, by contrast, have to get permission from their institution’s IRB first (and often that permission never comes).
So, too, with potentially traumatizing yourself — at least if you’re an academic who’s trying to induce a bee to sting your penis in order to produce generalizable knowledge, rather than for some, um, other purpose.
Yesterday, science writer Ed Yong reported a fascinating self-experiment conducted by Michael Smith, a Cornell graduate student in the Department of Neurobiology and Behavior who studies the behavior and evolution of honeybees. As Ed explains, when, while doing his other research, a honeybee flew up Smith’s shorts and stung his testicles, Smith was surprised to find that it didn’t hurt as much as he expected. He began to wonder which body parts would really smart if they were stung by a bee and was again surprised to learn that there was a gap in the literature on this point. So he decided to conduct an experiment on himself. (In addition to writing about the science of bee stings to the human penis, Ed is also your go-to guy for bat fellatio and cunnilingus, the spiky penises of beetles and spiders, and coral orgies.)
As Ed notes, Smith explains in his recently published paper reporting the results of his experiment, Honey bee sting pain index by body location, that
Cornell University’s Human Research Protection Program does not have a policy regarding researcher self-experimentation, so this research was not subject to review from their offices. The methods do not conflict with the Helsinki Declaration of 1975, revised in 1983. The author was the only person stung, was aware of all associated risks therein, gave his consent, and is aware that these results will be made public.
As Ed says, Smith’s paper is “deadpan gold.” But on this point, it’s also wrong.Most obviously, were Cornell to lack a specific policy about self-experimentation, that would not mean that it possesses a policy exempting self-experimentation from IRB review. As Cornell’s Office of Research Integrity and Assurance correctly notes, although
[f]ederal regulations are silent on the matter of researchers who want to participate in their own studies. . . . , the regulations do not distinguish between self-experimentation and research on people who are recruited for a specific project.
And in fact Cornell, like many other institutions (including Johns Hopkins, the University of Wisconsin-Madison, and Georgia Tech), does have a policy (last updated in October of 2013) requiring researchers to obtain IRB approval before experimenting on themselves:
As part of its commitment to the protection of the rights and welfare of individuals participating in research, Cornell’s Human Research Protection Program requires investigators who wish to act as participants in their own studies to submit for review and approval following standard procedures outline in the IRB policies.
Though investigator self-experimentation may not raise the conventional ethical concerns outlined in the Belmont Report, all human research projects should undergo ethical review to assure the safety of people involved and the integrity of the research at the university. While researchers may be aware of the risks of self-experimentation, they may also be more willing to accept risks that are ill-advised. Application for review with the IRB office allows a neutral third party to raise concerns and/or propose measures to promote the welfare of researchers.
According to his paper, Smith’s research “is based on work supported by a United States National Science Foundation (NSF) Graduate Research Fellowship.” But even if it weren’t federally funded, it wouldn’t matter. Cornell filed a FederalWide Assurance (FWA) with DHHS’s Office for Human Research Protections (OHRP) promising to apply the same federal regulations to all human subjects research, regardless of funding (see #5 here).
Of course, to come within the IRB’s jurisdiction, an activity not only has to involve human subjects; it also has to constitute “research” — “a systematic investigation . . . designed to develop or contribute to generalizable knowledge.” Did Smith’s bee sting adventure fit the bill? Here’s Ed again:
Smith was methodical. He collected bees by grabbing their wings “haphazardly with forceps” and pressing them against the body part of choice. He left the stinger there for a full minute before removing it, and then rated his pain on a scale of 1 to 10. . . . He administered five stings a day, always between 9 and 10am, and always starting and ending with “test stings” on his forearm to calibrate the ratings. He kept this up for 38 days, stinging himself three times each on 25 different body parts.
Sounds systematic to me. What about “designed to contribute to generalized knowledge”? Ed again:
Now, clearly, these data are very subjective, and they all come from one person. Smith is clear that his anatomy of pain can’t be generalised to everyone else. “If someone else did this, they’d probably have different locations that they felt were worst”, he says, although from talking to his colleagues, he feels that the rough shape of the map would be similar. “I didn’t see a lot of merit in repeating this with more subjects,” he says.
But Smith’s pain map being identical to everyone else’s is hardly the only way that his experiment could lead to generalizable knowledge. And anyway, “generalizable knowledge” is a notoriously fuzzy concept in IRB review. In determining whether an activity will contribute to generalizable knowledge, Cornell’s IRB, like many, uses the proxy of asking whether it will “be used in a publication, presentation, or achievement of a degree.” Smith’s study pretty clearly passes that test.
My point in noting Smith’s error about his institution’s IRB policy is not to cause trouble for him or to impugn his character. To the contrary, I think requiring IRB review for all self-experimentation is absurd. And it’s worth highlighting absurd laws and policies.
Does Cornell have a point when it notes that researchers “may . . . be more willing to accept risks that are ill-advised”? In discussing the list of body parts he tested, Smith told Ed that he “had originally had the eye on the list, but when I talked to [my advisor], he was concerned that I go blind. I wanted to keep my eyes.” But it’s not surprising that researchers might pose risks to themselves. All people have the potential to do that. The question is whether academics need special protection from themselves in the form of mandatory, prospective risk-benefit review. It’s one thing to require, as Cornell puts it,”a neutral third party” to examine a protocol when there are information asymmetries between investigator and subject, and when the protocol’s risks are externalized onto subjects who may not share much or any of the expected benefits. Mandatory review of self-experimentation takes IRB paternalism to a whole other level.
Incidentally, the penis shaft turns out not to be the most painful place in which one can be stung (and as suggested above, the testicle wasn’t even close). Check out Smith’s paper or Ed’s post on it to see which two body parts won that particular contest.
Cross-posted at The Faculty Lounge.