The Ethics of Self-Experimentation

Self-care is a civil right

AI conception of self-care

The term “self-care” gets a bad rap. My AI defines it as:

The ability to care for oneself through awareness, self-control, and self-reliance to achieve, maintain, or promote optimal health and well-being.

That sounds great, but if I follow up with the prompt “are there derisive meanings” I get:

Self-care is sometimes mocked as code for pampering, consumerism, and “treating yourself” (e.g., spa days, expensive products) rather than substantive efforts to improve health or functioning.

I confess that when I free-associate self-care the first thing that comes to mind is Goop. I’m not a subscriber.

The basic problem with the term self-care is the same as the problem with science in general. Science is a strong-link problem, not a weak-link problem. The scientific method gives us a way to identify strong links that don’t break under experimental testing. The trouble is the strong links are often hidden under an immense pile of weak links. It’s the classic “diamonds in a dung heap” formulation that Thomas Jefferson used to describe his efforts to find wisdom in the Bible.

In short, the existence of lots of useless crap in the self-care arena absolutely does not mean that everything in the arena is crap. Careful use of the scientific method can help us sort out the diamonds. By which I mean developing hypotheses that comport with available evidence and then doing experiments that test the hypotheses. Here, Thomas Jefferson missed a diamond. When the Declaration of Independence talked about unalienable Rights, it should have said, “Life, Liberty, and the pursuit of Experiments.”

All this may sound so obvious it doesn’t warrant spelling out, but this week I’ve been wrestling with the fact that an institutional ethics committee rejected a self-care proposal in which I would bring vaccine yeast home from the lab, brew beer with it, and then prick my own finger to see if I developed any antibodies. After the institutional veto, I resorted to setting up a shoestring molecular biology lab at home, re-creating the needed yeast using only my own personal resources, and drinking vaccine beer in the comfort of my kitchen. To my surprise, the ethics committee is now arguing that because the kitchen self-care experiment is scientifically “related” to my official duties it should be viewed as unauthorized human subjects research - a form of misconduct that can be grounds for termination.

The idea that an institution might have the legal authority to veto “work-related” menu planning in an employee’s home kitchen is prima facie ridiculous. How did we get into a situation where an ethics committee thinks this argument sounds even remotely reasonable?

My brother and I struggled mightily with this question when writing up our home vaccine beer projects. We think the answer boils down to a badly written section of a law called the Common Rule. After huge amounts of email and Facetime back-and-forth (my brother and I love to debate) we hammered out Appendix A - which presents our considered layman’s analysis of the Common Rule and the history behind it. I guess Appendix A comes off as a little bit stiff and scattered - you can kind of tell it was written by committee. I’ll reproduce the scholarly text in its entirety in the footer, but up here in the main post I’ll try to briefly summarize my individual view of the controversy in more conversational terms.

For anybody who’s unfamiliar with the concept of an Institutional Review Board (IRB), I recommend Scott Alexander’s epic tale, “My IRB Nightmare.” Even for people who are familiar with IRBs, Alexander’s classic post can help emphasize the point that the nightmare I’m describing here isn’t unique.

The basic idea behind the Common Rule is that if scientists are allowed to do whatever experiments we want then some of us might prioritize our own research interests over the individual best interests of study volunteers. A scary thing is that abuse of study subjects can be accidental - a scientist might simply fail to adequately inform study subjects about risks, leaving subjects poorly equipped to defend their own individual interests. It’s a serviceable description of what happened in the insidious Tuskegee Study of Untreated Syphilis - where scientists quietly withheld penicillin from black men with latent syphilis. The Common Rule addresses a real and very serious problem by applying the entirely reasonable approach of requiring scientists to get a second opinion from a neutral third party (an IRB) on research study designs involving human subjects. An central element of IRB review is to ensure that study subjects know what they’re getting into - meaning their consent to participation in the study is well-informed.

The Common Rule grew out of an earlier analysis called the Belmont Report, which was commissioned in direct response to the Tuskegee Study. The Report establishes a clear distinction between routine individual healthcare and scientific research. Unfortunately, the line is blurry because individual healthcare sometimes uses the tools of scientific research. Here’s how the Report describes the problem:

The distinction between research and [healthcare] practice is blurred partly because both often occur together (as in research designed to evaluate a therapy) and partly because notable departures from standard practice are often called “experimental” when the terms “experimental” and “research” are not carefully defined.

So far, so good. Where stuff hits the fan is with the Common Rule’s un-careful definition of the term “research.”

Research means a systematic investigation, including research development, testing, and evaluation, designed to develop or contribute to generalizable knowledge.

Lawmakers intended for their specialized legal definition of research to reinforce the Belmont Report’s exclusion of individual healthcare from human subjects review, but they evidently failed to appreciate the fact that individual healthcare is often at least somewhat systematic and it sometimes contributes to generalizable knowledge (e.g. the scientific publication of case reports). Lawmakers also failed to foresee that scientists and institutional administrators might intuitively drift back to the dictionary definition of research:

Studious inquiry or examination; especially: investigation or experimentation aimed at the discovery and interpretation of facts, revision of accepted theories or laws in the light of new facts, or practical application of such new or revised theories or laws.

If my doctor did not apply the dictionary definition of research to my individual healthcare I’d find a new doctor, pronto! A consequence of the Common Rule’s sloppy definition of the term research is that many modern institutions have fallen into the groupthink belief that the IRB’s purview includes any form of research that in any way involves human subjects. Employees screwing around with beer on the weekends? You betcha! That’s human subjects research, alright! Watch out, America’s Test Kitchen - the bioethics police may be coming for you next.

When the problem is seen through the individual healthcare lens, it becomes clear that the Common Rule never intended to interfere with self-care. The Common Rule was actually intended to prevent research institutions from interfering with individual medical treatment, including self-care. The groupthink failure to understand the sloppy legal definition of “research” has put us into a situation where ethics committees are making bids to replicate the medical withholding abuse of Tuskegee - instead of their actual charter to prevent it.

It’s important to note that self-care includes well-being. If it makes me happy to drink vaccine beer, prick my own finger, and test my own blood then ethics committees have no business demanding an IRB review to determine what I can or cannot do with my own body. The law allows institutional ethics committees to argue that a given experiment isn’t a scientifically useful way to spend institutional resources, but it’s unethical (and arguably illegal) for them to demand veto authority over individual healthcare activities.

An additional risk is that bioethics authoritarianism can needlessly inhibit scientific progress. That’s immoral too.

I believe it’s my civic duty to tell bioethicists to butt out of my self-care. Even if it means they try to fire me [1].

---

Song of the day:

---

[1] There’s a special place in my heart for Sister Sally Butler: “This is my church. If they don’t like it, they should leave. I’m not going to leave just because of the hierarchy,” Butler says. “But if I stay, I have to try to clean house.”

---

Appendix A. Human subjects considerations.

State and federal laws mandate that research involving human subjects must be pre-approved by an Institutional Review Board (IRB). Section 45 CFR § 46.102(I) of the federal Common Rule offers a restricted legal definition of the term research:

Research means a systematic investigation, including research development, testing, and evaluation, designed to develop or contribute to generalizable knowledge.

There is ongoing public debate about how the terms “systematic” and “generalizable” (boldface ours) should be construed. At one extreme, some have argued that few investigations involving human subjects meet a strict definition of the term generalizable (Munger, 2025). The other extreme rests on the fact that the scientific method can succinctly be described as a system designed to develop generalizable knowledge. Both extremes are absurd - lawmakers can’t have intended that almost nothing requires IRB pre-review, and they can’t have intended that physicians must apply for institutional permission each time they use the scientific method in service of individual healthcare.

The Common Rule was developed based on the findings of the Belmont Report (NCPHSB, 1979), which describes individual healthcare practices as:

designed solely to enhance the well-being of an individual patient or client and that have a reasonable expectation of success.

There can be confusing overlap between the Common Rule definition of research and individual healthcare, in the sense that individual healthcare sometimes involves systematic investigations aimed at evaluating whether an experimental treatment protocol will be of benefit for a particular individual. The design of individual healthcare can also accommodate the development of limited amounts of generalizable knowledge and can even result in scientific publications (e.g., case reports). Another layer of confusion is that colloquial usage of the word “research” can refer to activities that are neither systematic nor generalizable.

To illustrate the problem, we can consider author CBB’s (Chris’s) efforts to share mint family herbs with friends, family, and colleagues during the covid pandemic, in the hope that eating mint might prevent or mitigate SARS-CoV-2 infection (C. B. Buck, 2023). Chris instructed volunteers to systematically implement a scaled-up version of a treatment protocol Jan and colleagues developed using an animal model system (C. Buck, 2023; Jan et al., 2021). Everybody was encouraged to compare notes on their experiences, for the designed purpose of contributing to generalizable knowledge about whether mint might fight covid (Buck, 2024).

To understand whether Chris’s mint study meets the legal requirements for IRB review, it’s important to consider the history behind the Common Rule. The ethical roots of the law were initially laid out in the 1949 Nuremberg Code, which was formulated in response to Nazi medical atrocities during World War II. Nuremberg Core Principles, such as informed consent and the minimization of medical risk, remain central to the modern conduct of IRB review (Alexander, 2017).

The 1979 Belmont Report, which builds upon the foundation laid down by the Nuremberg Code, was commissioned in response to the disastrous Tuskegee Study of Untreated Syphilis in the Negro Male (TSUS), in which penicillin was abusively withheld from men with syphilis (Clinton, 1997; Tharrington, 2024; White, 2000). The medical-withholding abuse arose, in part, from a conflict between institutional research priorities and the right to access individual healthcare.

In the 1940s, it was unclear whether penicillin would be safe and effective for treating syphilis. A pioneering study by Mahoney and colleagues used an experimental protocol, scaled up from an animal model, to treat four syphilis patients (Mahoney et al., 1943). If we imagine a scenario in which Mahoney and colleagues paused the investigation to seek IRB pre-approval, it would have resulted in the risky withholding of penicillin from four syphilis patients. This cannot be the type of outcome the framers of the Common Rule intended to produce because medical withholding is exactly what the Common Rule was intended to prevent.

Clinical investigations designed to produce highly generalizable knowledge systematically include control groups. FDA guidance makes extensive use of the concept that control groups are a hallmark of generalizable research (FDA et al., 2013). A central lesson of the TSUS disaster is that studies with negative control groups (where treatment is withheld) are especially deserving of institutional pre-review.

Individual healthcare can include crossover controls (where a patient first tries one treatment and later tries another), but it never includes a control group. Therefore, we propose that study designs that lack a control group and meet the Belmont Report’s description of individual healthcare are not legally required to solicit IRB review. Under this interpretation, Chris’s mint study and the Mahoney study are both insufficiently systematic/generalizable to meet the legal requirement for IRB review.

The purpose of the treatments in the current study is to immunize ourselves against BK polyomavirus, with the hope that it might safely minimize our risk of polyomavirus-induced disease. When designing the study protocol, we anticipated that the experiments might produce limited amounts of generalizable knowledge. We do not view this as a valid standard for triggering IRB review because individual healthcare, which the Belmont Report explicitly excludes from the definition of human subjects research, sometimes results in published data and case studies. Additionally, delaying an individual healthcare investigation while soliciting IRB review would constitute a form of medical withholding. The lack of a control group in the current study means there is no risk of the type of medical withholding abuse the Common Rule is designed to prevent. In support of our interpretation that individual healthcare experiments are insufficiently systematic/generalizable to meet the restricted legal definition of human subjects research, most IRBs judge these types of experiments as outside their purview (Hanley et al., 2019).

The control-based standard cuts both ways. Recent covid vaccine policy has systematically assigned millions of Americans to a de facto negative control group. The controlled natural experiment will contribute to generalizable knowledge by experimentally testing the hypothesis that the vaccine is of no net benefit to the involuntary study subjects from whom it is being withheld. This poorly designed withholding trial should have been pre-reviewed by an IRB.

Institutional obstruction of individual access to healthcare – particularly if it qualifies as self-care – fails to fulfill the Common Rule’s intent of preventing the central abuse of TSUS. Instead, it replicates the abuse by withholding medicine from people who might experience net benefit. In cases where such abuse appears to be happening, investigators are within their rights to instead consult family, friends, colleagues, and physicians when choosing which scientific experiments to do with their own bodies.

In addition to being an individual human right, self-experimentation can also be valuable to the collective interests of humanity by enabling rapid testing of new ideas. There is a long tradition physicians and scientists using self-experimentation to speed the process of making lifesaving discoveries (Hanley et al., 2019; Marshall et al., 1985). For these reasons, we view institutional obstruction of self-experimentation as unethical.

Alexander, S. (2017). My IRB Nightmare. Slate Star Codex. https://slatestarcodex.com/2017/08/29/my-irb-nightmare/

Buck, C. (2023). Mint vs Covid: A Culinary Perspective. Viruses Must Die.

Mint vs Covid: A Culinary Perspective

Buck, C. (2024). Viruses are Supervillains, Herbs are Superheroes. Viruses Must Die.

Viruses are Supervillains, Herbs are Superheroes

Buck, C. B. (2023). The mint versus Covid hypothesis. Med Hypotheses, 173. https://doi.org/10.1016/j.mehy.2023.111047

Clinton, B. (1997). Apology For Study Done in Tuskegee [Speech]. https://clintonwhitehouse4.archives.gov/New/Remarks/Fri/19970516-898.html

FDA, Evaluation, C. f. D., Research, Evaluation, C. f. B., Research, Safety, C. f. F., & Nutrition, A. (2013). Investigational New Drug Applications (INDs): Determining Whether Human Research Studies Can Be Conducted Without an IND. Guidance for Clinical Investigators, Sponsors, and IRBs. Rockville, MD: U.S. Food and Drug Administration Retrieved from https://www.fda.gov/files/drugs/published/Investigational-New-Drug-Applications-(INDs)-Determining-Whether-Human-Research-Studies-Can-Be-Conducted-Without-an-IND.pdf

Hanley, B. P., Bains, W., & Church, G. (2019). Review of Scientific Self-Experimentation: Ethics History, Regulation, Scenarios, and Views Among Ethics Committees and Prominent Scientists. Rejuvenation Res, 22(1), 31-42. https://doi.org/10.1089/rej.2018.2059

Jan, J.-T., Cheng, T.-J. R., Juang, Y.-P., Ma, H.-H., Wu, Y.-T., Yang, W.-B., Cheng, C.-W., Chen, X., Chou, T.-H., Shie, J.-J., Cheng, W.-C., Chein, R.-J., Mao, S.-S., Liang, P.-H., Ma, C., Hung, S.-C., & Wong, C.-H. (2021). Identification of existing pharmaceuticals and herbal medicines as inhibitors of SARS-CoV-2 infection. Proceedings of the National Academy of Sciences, 118(5), e2021579118. https://doi.org/doi:10.1073/pnas.2021579118

Mahoney, J. F., Arnold, R. C., & Harris, A. (1943). Penicillin Treatment of Early Syphilis-A Preliminary Report. Am J Public Health Nations Health, 33(12), 1387-1391. https://doi.org/10.2105/ajph.33.12.1387

Marshall, B. J., Armstrong, J. A., McGechie, D. B., & Glancy, R. J. (1985). Attempt to fulfil Koch’s postulates for pyloric Campylobacter. Med J Aust, 142(8), 436-439. https://doi.org/10.5694/j.1326-5377.1985.tb113443.x

Munger, K. (2025). Experiments as Performance Art.

Experiments as Performance Art

NCPHSB. (1979). The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research. https://www.hhs.gov/ohrp/regulations-and-policy/belmont-report/read-the-belmont-report/index.html

Tharrington, C. (2024). The Lost Lesson of Tuskegee.

The Lost Lesson of Tuskegee

White, R. M. (2000). Unraveling the Tuskegee Study of Untreated Syphilis. Arch Intern Med, 160(5), 585-598. https://doi.org/10.1001/archinte.160.5.585