Is it ever OK for scientists to experiment on themselves?

A virologist named Beata Halassy recently made headlines after publishing a report of successfully treating her own breast cancer by self-administering an experimental treatment. Having previously undergone a mastectomy and chemotherapy, Halassy informed her doctors that she wanted to treat her tumour by injecting it with viruses known to attack cancerous cells. This sort of approach is called oncolytic virotherapy (OVT). OVT has not yet been approved for the treatment of breast cancer, but it is being studied as an experimental approach.

Halassy is a success story of self-experimentation in medicine. She joins other examples, like Barry Marshall, who won the 2005 Noble prize in medicine following his work ingesting the Helicobacter bacterium to prove its role in gastritis and peptic ulcers. This work is estimated to have saved millions of lives. Yet, self-experimentation is often viewed with suspicion. Concerns about self-experimentation are growing because it is no longer solely the domain of professional scientists.

The availability of biotechnologies, and the prevalence of open-source science has led to the development of “bio-hacking” communities engaging in various forms of self-experimentation. Does selfexperimentation raise ethical concerns? To answer this question, it is useful to return to first principles of research ethics:

Autonomy

The informed consent process is a crucial protection in medical research. Researchers have to employ rigorous methods to ensure that individuals make a voluntary choice to participate in trials, and that they understand the risks and benefits of the experimental intervention (and alternatives). Clearly, some selfexperimenters can make informed choices to self-administer unproven interventions. Halassy's report makes clear that she gave informed consent. And her expertise in virology enabled her to develop a clear scientific rationale for her approach. Still, there can be unknown serious side-effects with experimental treatments, so a second set of eyes is desirable.

Also, not all self-experimenters will be as well informed. There is a legitimate concern that unregulated forms of self-experimentation may not involve this important safeguard when members of the public experiment on themselves.

Reasonable risk to the participant

Informed consent is not the only important safeguard in research ethics. It is often claimed that even consenting adults should only be exposed to “reasonable risks” in research. Ethicists often debate how to understand reasonable risk. But it is widely agreed that reasonable risks must be minimised to those that are necessary. However, it is more complex to work out what counts as reasonable, when someone has a serious illness and is undertaking experimental treatment that might (or might not) benefit them.

Where an individual stands to benefit from receiving an experimental intervention, “proportionality” partly relates to how the experimental intervention compares to other therapies that doctors might use as the “standard of care”. This is a relevant question in Halassy's case. Notably, Halassy had already undergone a mastectomy and chemotherapy over the course of her treatment.

What's more, the measles virus and the vesicular stomatitis virus she used in her experimental OVT had a good safety record.

In contrast, where the participant does not stand to benefit from receiving the intervention, it is sometimes claimed that it can only be proportionate to expose the participant to minimal risk. Others have argued that greater risks could sometimes be proportionate if the benefits of the research are sufficient.

One problem here is that self-experimentation often involves only one participant. That might mean that it is hard to generalise (or that it might even be misleading). But as Barry Marshall's case shows, selfexperimentation involving just one person can sometimes generate incredibly useful findings. However, we also have to consider potential harms.

Harm to others

In Hollywood, the self-experiments of rogue scientists often go devastatingly wrong – think Jeff Goldblum's portrayal of an eccentric scientist in The Fly. Sci-fi stories like these are often wildly implausible. But this should not blind us to the possibility of more credible harmful effects.

One concern is that other patients might be tempted to follow in Halassy's footsteps and attempt an unconventional therapy, perhaps before using other standard therapies. To forestall this, it is crucial to be clear about the limited generalisability of her case, and to ensure that patients understand the tried and tested benefits of existing therapies.

A different concern is that adverse publicity from very risky experiments might make it more difficult to conduct important research. At least eight early self-experimenters died from their research, including the 29-year-old doctor William Stark who died from scurvy in the 18th century after severely restricting his diet. There are other concerns to consider in self-experimentation more broadly. Self-experimenters can now easily access powerful technologies like Crispr-Cas9 gene-editing tools.

In 2017, a biohacker named Josiah Zayner injected himself with a DIY Crispr–Cas9 gene therapy aiming to enhance muscle growth. Crispr-Cas9 has the power to significantly benefit society, but it could also cause significant harm if misused through misunderstanding or malice. The worry about selfexperimentation here is not just about the direct harm that misuse might cause. Cases of misuse might also undermine societal acceptance of the regulated efforts to safely develop this important technology.

It can be ethical for scientists to experiment on themselves. Such studies should at least sometimes be permitted, and certainly should be published so that others can learn from them. But it is a mistake to assume that self-experimentation only ever affects the individual involved. Halassy embarked on her selfexperiment without any ethical oversight. Things ended well for her, but that won't always be the case.

The writers are researchers at University of Oxford