Human Challenge Studies to Accelerate Coronavirus Vaccine Licensure

by Marc Lipsitch, Nir Eyal, Peter G Smit

Controlled human challenge trials of SARS-CoV-2 vaccine candidates could accelerate the testing and potential rollout of efficacious vaccines. By replacing conventional phase 3 testing of vaccine candidates, such trials may subtract many months from the licensure process, making efficacious vaccines available more quickly. Obviously, challenging volunteers with this live virus risks inducing severe disease and possibly even death. However, we argue that such studies, by accelerating vaccine evaluation, could reduce the global burden of coronavirus-related mortality and morbidity. Volunteers in such studies could autonomously authorize the risks to themselves, and their net risk could be acceptable if participants comprise healthy young adults, who are at relatively low risk of serious disease following natural infection, if they have a high baseline risk of natural infection, and if during the trial they receive frequent monitoring and, following any infection, the best available care.


Jack Freer August 30, 2020 at 11:54 AM

Eyal, Lipsitch & Smith (ELS) propose a controlled human challenge trial of a SARS-CoV-2 vaccine as a means to accelerate the development of a vaccine for Covid-19. This is an alternative to a phase 3 study in which vaccinated/placebo groups are observed over time as they are exposed to viral infections in the natural course of their daily activities. In the challenge trial, those groups are actually exposed to the virus in a controlled fashion, generating outcome data in a compressed time frame. Obviously, this study design is ethically problematic because some participants will contract Covid-19, with its attendent morbidity and mortality risk, as a direct result of the intervention.

This study design demonstrates the fundamental difference between research and medical care. Participating in clinical research is an altruistic act, intended to generate new knowledge which will benefit others. The individual subject may derive some benefit from the study, but that (in principle) is not the reason to participate.

That said, it is still incumbent upon the study designers and researchers to minimize those risks. ELS strive for this by enrolling young healthy subjects, for two reasons. The first is that they are less likely to suffer the worst consequences of viral infection. While not a certainty, the likelihood of deadly outcomes are small (but not zero).

The second reason for a young healthy study population is that they are most likely to acquire the infection anyway. Arguably, this obscures the culpability of the researchers as much as it reduces the likelihood of study induced disease. In any case, these design details will tend to offset the ethical objections to infecting people intentionally.

The last caveat (that the article does not mention) is the meticulous avoidance of vulnerable populations. The most commonly mentioned groups are incarcerated "volunteers" or others that might not be giving their free and informed consent.


Jim Delaney September 1, 2020 at 12:54 PM

Jack, aren't there other motivations for people to participate in research besides just altruism? The most obvious one I can think of is compensation. My understanding is that compensation is usually modest, similar to something like jury duty in which the participant is just given what they would get if they were working rather than participating. But even so, people may enroll in studies even for such modest compensation. And in the abstract, if the situation is dire and the need for a vaccine is urgent, would we allow larger monetary compensation? There is a worry that the higher the compensation, the more pressure there will be for people (perhaps vulnerable people) to enroll: that is, an offer that is too good to refuse. Such offers it is thought could violate autonomy because the person can't pass it up. But we might be skeptical of this. At least an argument can be made that people can freely decide how much money makes a given risk worth it, and for some that amount would be quite high. Couldn't they still be acting autonomously?

Jack Freer September 1, 2020 at 6:42 PM

Jim, Yes, you are correct. Researchers often provide a stipend for participation -- the value varies (often corresponding to the amount of risk or nuisance involved with participation). This is obviously a transparent contractual reward.

The persistent inaccurate belief that the study intervention is medically beneficial is widespread and has been labeled "the therapeutic misconception" (Lidz CW and Appelbaum PS.Med Care. 2002 Sep;40(9 Suppl):V55-63). It endures despite lengthy detailed informed consent procedures (including mind-numbing 25 page informed consent documents.

Stephen Kershnar September 4, 2020 at 5:00 PM

The amount of money saved by challenge grants is likely very large.

(1) If it increases the GDP by 1%, it saves roughly $200 billion dollars.

(2) If it saves 5,000 lives at $5 million per life, it saves $25 billion dollars.

(3) It likely lessens business-destruction and further loss of liberty worth $10 billion.

Together the early roll out likely is worth $235 billion dollars. If there are 10,000 volunteers needed, we could pay each a million dollars to do the trial.

What is the objection to this?

Compare this expenditure of $235 billion to the $2.2 trillion dollars in welfare given to schools, universities, and other institutions whose spending is vastly inflated. This would be a bargain.

Stephen Kershnar September 5, 2020 at 8:36 AM



Consider this quote.

"Participating in clinical research is an altruistic act, intended to generate new knowledge which will benefit others. The individual subject may derive some benefit from the study, but that (in principle) is not the reason to participate."

I think this reasoning is similar to a major reason for the ban on organ sales (see, for example, the National Organ Transplant Act sponsored by everyone's favorites Al Gore and Orrin Hatch).

Here are the reasons to reject this claim.

(1) It is false in general. Are physicians, police officers, professors, soldiers, teachers, etc. motivated exclusively or primarily by altruism? Probably not.

(2) It is morally irrelevant. Unless you think someone's motivation makes a right-infringing act permissible or a right-respecting act wrong, it is hard to see why this is morally relevant.

(3) Even if this claim were not false and irrelevant, it is hard to see why it is not overridden in cases such as the challenge studies or organ transplants.

This reasoning is similar to the coercive drafting of young men to fight our wars because we don't want mercenaries fighting our battles. I think it is a common and unfortunate mistake to think in these terms. To paraphrase Milton Friedman, I prefer mercenary soldiers to enslaved soldiers. So should you.


Steve K

Harvey Berman August 30, 2020 at 4:30 PM

Paul Offit says the question to ask is not whether natural immunization is better than vaccination, but is vaccination good enough. The answer depends on the dose of the antigen and whether it produces a memorialized response:

He uses measles as an example, during which natural immunization by contracting measles as a child generated 3-times as many antibodies as generated via the measles vaccine. But not being vaccinated he avoided the risk of pneumonia, encephalitis, and he didn’t die, some of the unintended risks associated with a vaccine.

Which raises the question of what the authors mean by providing appropriate care if the subjects develop COVID as one point, and second have they informed the subjects about the unanticipated effects that might occur, a la measles vaccine. Finally, assuming the investigators provide suitable informed consent for unintended but possible (not yet foreseeable?) affects, there is the likelihood that the vaccine might not provide immortal or long-lived immunity, affording a false sense of security and risk of re-infection. All of which is to remember that this is research on the unknown effects of this virus.

All of which leads me to believe that in spite of the risks, and if the research is serious, well-designed and well-intentioned, coupled with more than a hint of double effect, that the risks of not searching for a vaccine exceed those inherent in the study. Is this any different than men in WW II signing up for service in the military to fight an implacable foe?


Stephen Kershnar September 5, 2020 at 8:45 AM



Great point.

Here is your quote.

"All of which leads me to believe that in spite of the risks, and if the research is serious, well-designed and well-intentioned, coupled with more than a hint of double effect, that the risks of not searching for a vaccine exceed those inherent in the study. Is this any different than men in WW II signing up for service in the military to fight an implacable foe?"

There are several similarities between young adult volunteers in World War II and the human challenge studies if they were to be authorized: they are fighting evil, were they not to participate they would be in little to no danger, and they give valid consent.

I am not sure that WW II volunteers were that informed about what they were volunteering for (at least not from the government). Consider, for example, whether the earlier Marine Corps volunteers knew the bloody island hopping campaign.

Still, if that wasn't an issue there, I wonder if informed consent is relative to the information that is available and the cost of providing it (a cost-benefit analysis of informed consent).

In any case, great point,

Steve K

Jim Delaney September 1, 2020 at 12:46 PM

Lipstich, Eyal, and Smith (ELS) argue in support of human challenge studies for COVID-19. I’ve been working through this response and so I apologize if I repeat what others have said on earlier posts.

ELS suggest that by replacing the standard Phase 3 part of a clinical trial with a challenge study, we can potentially produce a vaccine more quickly, thus saving thousands or perhaps even millions of lives (1753-4). The major ethical issue is the injection of live virus into human beings. The challenge aspect of the study design would involve three groups of people directly exposed to COVID-19.

1. A group of volunteers for an initial dose-escalation to determine how much exposure is needed to produce a clinical response that is similar to (but not more severe than) that associated with natural infection.

Post the initial study when a potentially effective vaccine is developed:

2. A control group that is injected with a placebo and then injected with the virus.

3. An experimental group that is given the vaccine and then injected with the virus.

ELS identify two main areas of ethical concern: the research subjects’ autonomy and the minimalization of potential risks. For now, I’ll focus on the former.


Pat D September 2, 2020 at 9:44 AM

Jim, in addition to the ethical issue of inoculating subjects with the virus, there is also the potential for genetically modified vaccines to make the disease worse rather than better. There is a significant problem for experts to determine the actual risks involved relative to both the virus and the vaccine because of our limited knowledge - which makes the 'informed' part of informed consent questionable. Just how questionable is the problem. I say more about this in a separate post below.

David H September 3, 2020 at 2:51 PM


I didn't see your #3 in the article. I thought after the initial challenge experiment, an extended placebo controlled study would be conducted in the field using 3000 vaccinated and placebo people but they were not injected with the virus but just exposed to it in the natural environment. They write on page 1753 "If this human challenge study showed a vaccine candidate to be efficacious, an expanded placebo-control study would be conducted in the field, invovling at least 3000 vaccinated persons...This study, not involving a challenge)..."

Pat D September 4, 2020 at 11:01 AM

David, I think that Jim's list is correct. What you are referring to is a subsequent stage that ELS plan for that would compare outcomes in a larger, more diverse (with respect to age and overall health) test population (reference class) treated with a placebo vs. a fast-tracked vaccine.

Pat D September 4, 2020 at 12:14 PM

See the three black boxes in ELS's Figure 1 (p.1754):

Step 1 (Jim's group 1) involves preparation of a challenge system, including a "Good Manufacturing Practice (GMP) challenge stock with demonstrated stability and consistent infectivity" (9/3 NEJM article cited below). It does not involve a placebo group.

Step 2 (Jim's groups 2&3) involves involves a placebo and an experimental group drawn from a select, young, previously healthy test population. Both groups are inoculated with a specially prepared strain of the SARS-CoV-2 virus and are monitored in a controlled environment for the duration of the study (weeks).

Step 3 involves a placebo controlled study that David refers to that draws from a larger, more diverse population. Neither group is experimentally inoculated with SARS-CoV-2. This step tests whether the fast-tracked vaccine prevents naturally acquired infection. Participants go about their lives as normal for the duration of the study (months).

David H September 4, 2020 at 12:16 PM


I don't see any where in the text that someone is challenged with the virus without the vaccine. Is it the ELS page 1752 claim that in the initial dose selection a placebo is given? Is it the following sentence, "Volunteers, previously uninfected, would be required for an initial dose escalation study of the viral challenge to select a dose of virus exposure such that most placebo recipients become infected (for statistical reasons) and have a clinical response that it not more severe than one associated with natural infection (for ethical reasons.)"

Phil Reed September 4, 2020 at 2:05 PM

David, you’re challenging whether the challenge studies themselves include a placebo group? The article says, after the preparatory studies, "Volunteers...would be randomized to receive either the candidate vaccine under investigation or placebo" (1752-3), after which they get the virus and are then monitored.

David H September 4, 2020 at 2:33 PM

I see. I misinterpreted or overlooked the relevant passage.

Jim Delaney September 1, 2020 at 12:46 PM

With respect to autonomy, ELS say that it “may seem impermissible to ask people to ask people to take on risk of severe illness or death” (1754). However, they argue, we ask people to take risks in many other activities for the sake of others (firefighters, organ donors, and even other challenge studies). The reason why it is permissible to allow people to take these risks, which would also extend to COVID-19 challenge studies, is because they can choose to do so autonomously.

I should say that I am largely sympathetic to ELS’ position. Nevertheless, I raise some concerns that will perhaps generate some discussion.

It is true that when one elects to enroll in a clinical trial, they are doing so in order to determine whether or not a given intervention is effective and not for personal benefit. Subjects realize that they may benefit and it is likely that a good many enroll in the trial with just this hope. Despite efforts to ensure that subjects understand this, my understanding is that they still sometimes have inaccurate beliefs about why they are in studies. That is, a patient might state that they understand they are in the study to further knowledge, but nevertheless lack an “appreciation” for this (where appreciation is the ability to acknowledge one’s condition and the likely consequences). Perhaps others with more clinical experience than I can speak to this phenomena in more detail. If this raises a concern with clinical trials in general, I wonder if the additional risk of injecting live viruses in any sense makes it even more worrisome.

I wonder a bit about the analogy to risk-taking in practices outside of medicine. There is a duty of beneficence on the part of the researcher toward the subject. So even if the goal of the clinical trial is knowledge and not therapeutic, at least the clinical trial does not knowingly put the subject in a worse off position than she would otherwise be. If the subject is getting standard treatment or is in the control group, we don’t know that she is better off than if she is in the experimental group. In a large phase 3 trial, this seems to be the case. The vaccine might be effective and so might make one better off, but it also might not be effective so we don’t know this and thus don’t know that those in the control group are better or worse off either. But the challenge study raises issues. In the dose escalation study and the control group, can we really say that we don’t know if these subjects are worse off? They are after all given live virus and no vaccine. It seems to me the only way to say this is similar to a phase 3 trial would be to say that there is a good chance they would become naturally infected and not get as good of care as they would in the study. I am not sure how plausible this is. And if the duty of beneficence means that we can’t make people worse off in a study, then this could be a problem for challenge studies and would limit what people could autonomously authorize. There is no such duty of beneficence on the part of professionals to protect people from things like being a firefighter, so perhaps this is a case where medicine is a unique sort of practice. And as for other challenge studies, maybe those violate beneficence as well.


Neil Feit September 2, 2020 at 2:35 PM

Jim, this is really interesting and I don't quite know what to think of it (sorry, I didn't see this post yesterday when I was writing my comment, below...). There might be questions of informed, autonomous consent from those participating in a challenge trial, but ignoring those, consent might relieve the researchers of any duty not to put subjects into a worse off state -- and even more clearly, and duty not to have subjects take on that risk. Other duties of beneficence might remain in force. Some of these are noted by ELS, like excellent care, priority for life-saving resources, sequential challenges so that later ones could be stopped if a vaccine candidate appears to enhance disease, etc. In any case, I'm inclined to think there is no duty of beneficence requiring that researchers cannot make, or risk making, participants worse off in a study.

Jim Delaney September 3, 2020 at 9:33 AM

Neil, I tend to agree with you. Ultimately I tend to agree with ELS about challenge studies. I recall our previous discussion of equipoise in one of the earlier blog discussion. So I suppose I was thinking of something along those lines. If one wanted to push this kind of objection against challenge studies (the idea that medicine is a different sort of practice with different norms), what about the following hypothetical. My understanding is that if a study must be stopped for ethical reasons if an alternative effective treatment becomes available. So suppose there is a clinical trial investigating whether an experimental treatment A is better than the standard treatment B. But then a different treatment C is shown to be more effective than B. I believe that on ethical grounds, the trial is stopped because we would knowingly be making the subjects worse off (by knowingly giving them sub standard care). Suppose there was some other reason that researchers nevertheless thought the trial might result in valuable knowledge, and further suppose some of the subjects wanted to continue in it (perhaps they are being compensated). Even if they were fully informed and consented, I think many would think continuing the study would be unethical. In nonmedical practices, we might be more liberal about allowing people to autonomously consent to being made worse off. I feel like I didn't articulate this point as well as I could, but I suppose I was thinking about something along these lines in my earlier post!

Stephen Kershnar September 5, 2020 at 8:56 AM



Consider these quotes.

"I wonder a bit about the analogy to risk-taking in practices outside of medicine. There is a duty of beneficence on the part of the researcher toward the subject. So even if the goal of the clinical trial is knowledge and not therapeutic, at least the clinical trial does not knowingly put the subject in a worse off position than she would otherwise be." ...

"And if the duty of beneficence means that we can’t make people worse off in a study, then this could be a problem for challenge studies and would limit what people could autonomously authorize. There is no such duty of beneficence on the part of professionals to protect people from things like being a firefighter, so perhaps this is a case where medicine is a unique sort of practice."

The duty of beneficence strikes me as a mistake.

(1) NO GROUND. There is no reason to think that the duty of beneficence applies to medical experimentation alone or, alternatively, that it is particularly stringent in medical experimentation without further argument. Unless there is something about a power or information asymmetry that is especially strong in medicine (and not applicable to the Marine Corps) or role-based obligation (see, for example, social understanding or role-occupation), the claim of an additional duty or an especially strong duty is implausible.

(2) SATISFIED. If there were a duty of beneficence, why wouldn't it be satisfied if we paid the volunteers tens or hundreds of thousands of dollars. We often pay military officers six figures (including the generous retirement and other benefits for those salary watchers) and the medical volunteers are likely providing a noticeably valuable service.

(3) WAIVER. If the medical researchers have a duty of beneficence and the volunteers have a correlative claim, the latter can waive it unless you think they don't own their own rights. I think denying this is a mistake and an affront to liberty, but not everyone loves liberty.

Here is my pithy response: There is no duty of beneficence, but if there were it could easily be met and, in any case, would be waived.

Thanks for the interesting point and great article,

Steve K

Neil Feit September 1, 2020 at 4:36 PM

Jack and Harvey, good points. If conducted properly (for example, as Jack points out, not coercing vulnerable populations among other things) then it seems to me there are no moral worries about the challenge trial itself. One possible concern is that with young, healthy people making up the entire challenge trial population, that results might not transfer straightforwardly to older people and those at higher risk for other reasons. The authors realize this and propose a "phase 2" placebo controlled trial (not involving a challenge) that would include more vulnerable people. Is this just as safe for such people as the more traditional phase 3 field efficacy trial? I don't see any reason to think that it isn't -- especially since all the vaccine candidates will have gone through phases 1/2. So, it seems to me to come down to a cost-benefit analysis, which I suppose is largely going to have to consider the relative time frames. I wish the authors had said more about this.


Jack Freer September 1, 2020 at 6:58 PM

Neil, exactly right--exclusive use of healthy young volunteers in clinical research is historically why such research was deemed ageist. In the ELS model, however, more traditional inclusive and extensive studies are still anticipated by ELS. Because they will take so much more time to yield results (since they wait for naturally occuring infection to occur), the challenge study would be started immediately as well.

Harvey Berman September 1, 2020 at 9:28 PM

Jack, I do not know what you mean. I think they have to do the no-challenge phase-2-placebo study first in order to see if their results are generalizable fro younger to older and even more frail subjects. That is they have to extend the phase-1 safety trial with a larger number of subjects more diverse in age in order to evaluate the immunogenic response to the vaccine.

Pat D September 2, 2020 at 10:13 AM

Jim, thanks for picking this article. It raises some great questions. I agree with others that human challenge studies can be justified in some cases. As I mentioned above, there are scientific issues related to our limited knowledge of the long term effects of SARS-Cov-2 (even in young people) and the fact that genetically modified vaccines (which make up nearly half of Covid 19 vaccines currently under development) have never been used in humans before. There are also economic and political issues related to funding studies of various vaccine candidates, transparency and willingness to participate in comparative vaccine studies, and equity in the international approval and distribution of select vaccines. There is an 19 page WHO position paper that discusses these issues at some length available at:

Peter Smith, one of the co-authors of the ELS paper, was a member of the working group that drafted this position paper.

There is also a good 3 page Scientific American article (June 2020) that describes the different types of Covid vaccines under development.


David H September 3, 2020 at 3:29 PM

INFORMED CONSENT: I don’t know the details of the informed consent process in challenge studies but I have some concerns that the following might be in play. 

1) The authors write (p. 1752) “multiple measures would be put in place to ensure that prior to consenting, potential participants fully comprehend the unusual risks involved in the study.” But do the researchers themselves fully comprehend the risks given the disease is so new? There aren’t long term studies of morbidity. One does hear about discharged patients with all sorts of residual problems for months. We know the risks of organ donation, which was one of their justificatory analogies and we probably know the odds and prognosis of firemen being injured from smoke or burns 

2) One reads often about the role of the degree of exposure – the magnitude of the dosage of virus – determines if and how sick one gets. The ELS guidelines involve participants to be likely to be exposed during or soon after the trial period given where they reside but do we have reason to believe the dosages will be the same? Will the participant know whether his or her exposure is the same or higher or lower in the challenge study than through the baseline risk of natural exposure? 

2B) Authors seek to control the risk of the challenge study by enrolling only “people with an especially high baseline risk of getting exposed during or soon after the trial period should be recruited, (e.g. people residing in areas with high transmission rates” (pp. 1754-55). Does the volunteer know or have good reason to believe that the exposure in the challenge study won’t be greater than the natural exposure? Would the exposure be worse for someone in a challenge study who then is exposed further by other challenge patients if they are isolated together or when returned to society? Perhaps RC MDs Jack or Pat could comment upon whether the interval between challenge and return mean the virus is fully out of his system and so no greater cumulative exposure will then occur due to the injection coupled with natural exposure? 

2C) Wouldn’t it be fairer if the participants weren’t just from the area of high transmission but jobs and social activities that make it more likely they will be infected –factory jobs, service jobs, dense housing, regular use of crowded public transportation, health care workers themselves or do they fear charge of exploitation and unfairness and perhaps ingratitude in the latter? (Perhaps I should have included Trump rally participants without masks or social distancing as good candidates for challenge studies.) See below post about social justice. 

3) The authors emphasize (p. 1753) the importance of sequential exposures to limit the dangers that vaccine induces more severe disease following infection. Will the challenge participants know that where they are in the order of the study as that affects their safety? 

4) The challenge studies are justified and the young volunteers are motivated by the threat of great number of mortality and morbidity. But aren’t there vaccines that are not as effective with the elderly with their comorbidities but are effective with the young and healthy? If so, then the consent to a challenge study of just healthy young people is less informed if participants believe their successful protection will extend to millions of others who are at risk of death 

5) The participants are promised “priority for any scarce life -saving resources...” (p. 1753) and the best available care available, if needed…notwithstanding the possibility of severe shortages during the evolving pandemic” (p. 1755). Does that mean they will get priority over others – health care providers, other sick patients with better chances of recovery - when there are shortages? Do the participants understand that shortages may mean that they won’t get the care that they would receive if there weren’t shortages? 

6) I just want to repeat Harvey’s concern above that challenge study participants need to know that the vaccine may not provide long lasting immunity


Stephen Kershnar September 4, 2020 at 5:06 PM


Consider the following question.

2C) Wouldn’t it be fairer if the participants weren’t just from the area of high transmission but jobs and social activities that make it more likely they will be infected –factory jobs, service jobs, dense housing, regular use of crowded public transportation, health care workers themselves or do they fear charge of exploitation and unfairness and perhaps ingratitude in the latter? (Perhaps I should have included Trump rally participants without masks or social distancing as good candidates for challenge studies.) See below post about social justice.

Do you think that claims of fairness persist after consenters have given valid consent? If yes, the concern is not right-based. I don't see what it could then consist of. If no, then I do not see why the fairness issue has not disappeared.


Steve K

Stephen Kershnar September 4, 2020 at 5:09 PM


Consider this question.

3) The authors emphasize (p. 1753) the importance of sequential exposures to limit the dangers that vaccine induces more severe disease following infection. Will the challenge participants know that where they are in the order of the study as that affects their safety?

Here is my parallel question. Do 18-year-olds know the risk that they will be involved in wars or how bloody the wars would be? If not, then why would this be a consideration here.


Steve K

David H September 4, 2020 at 10:25 PM


One of the authors' criterion for "containing risk" (pp. 1754-55) was that "only people with especially high baseline risk of getting exposed during or soon after the trial period should be recruited." So I was suggesting that they should have fine tuned the criteria. It was an internal critique. If they shared the views of the WHO standards that Pat D linked to, then I suspect that they might not have wanted to narrow the target participants because of social justice concerns.

Fairness is a different issue. In response to your first question, I think fairness issues could persist if you don't mean by valid consent what is meant by informed consent in the medical setting. Because of what you might have said at a Romanell working dinner months or years ago, I suspect you don't believe medicine is special and doctors aren't different from car salesmen. But my memory may be failing me about what you mean by valid consent - is it roughly nothing more than competency and the absence of fraud? Is your view that it valid consent in the medical setting doesn't involve a fiduciary duty?

In regards to your very interesting second question, I suspect the nature of war is not as difficult to know and doesn't require the expertise on the 18 year old to grasp as does the nature of Covid 19. I also suspect that since you are not talking about a wartime draft but a peacetime volunteer, there may be less knowledge of risk of injury to convey than in the medical case. Such facts aren't known or if they are, perhaps it won't be hard for the lay person to discover them. Hollywood and news reporting does a good and graphic job of letting young movie watchers know how dangerous war is. On the other hands, perhaps enlistees should be informed about high rates of PTSD and suicide rates of returning soldiers. So I not dead set against demanding 18 year olds know more and the recruiters having a duty to inform them. But these facts might still be easier to discover than Covid vaccine facts. Our authors (ELS) believe that the "wide news coverage should keep it clear that the exposure to the virus is no small matter" (p. 1754). But the news coverage is so politicized - compare Laura Ingram's show with a Don Lemon CNN show - and one may be sympathetic to the idea that young lay people may not know what to think about the risk.

Stephen Kershnar September 5, 2020 at 9:26 AM



As always, great response.

I do think that medicine is not morally special and that physicians are morally similar to car salespeople.

Let us consider some reasons to think they are different and see if they apply to challenge studies.

(1) METAPHYSICS. There is a metaphysical difference between the transaction of a physician and car salesperson and a buyer. Implausible.

(2) MORE IMPORTANT OUTCOME. Medicine has a prudentially more important product. I do not see how this affects people's ability (Hohfeldian power) to waive their rights.

Even if it did affect the ability, it is an odd view that decisions that have enormous consequences for someone's life - marriage and home purchases - require little to no disclosure, but exposure to COVID, if it were the case - which has been reported - that COVID is less dangerous than the flu for people under 50 require greater information.

(3) ASYMMETRY OF POWER OR KNOWLEDGE. Asymmetry of power or knowledge. The medical experimenters in challenge studies do not have an asymmetry in power. Perhaps it is an asymmetry of knowledge.

But unless this is justified in terms of a risk of exploitation, harm, or right-infringement, this is implausible.

If the focus is on risk, then I'm not seeing how this distinguishes those running human challenge studies from those, transacting with Marine Corps volunteers or employers for the fishing and logging industry.

The real concern here is for autonomy and exploitation (solved by high pay), not over autonomy/consent. Hence, this is the wrong area to bring in these factors.

In any case, the problem with a risk-based analysis is that this would make very little information needed if experimenters wanted to do challenge studies on individuals 18 or younger (very small risk from COVID), whether the information would be given to parents or the teenagers.

(4) ROLE. Perhaps medicine has a different level of informed because physicians have a unique role or because of our social understanding of their role (see, for example, Michael Sandel or Michael Walzer).

This involves role-based obligations - not themselves justified by consent, promise, or utility - that I think both you and I find implausible.

Even if there were such role-based obligations, why think they would require such a high level of information in cases in which (a) the subjects are in less danger than the flue, (b) the stakes are so high, and (3) their expected outcome is enormously good because of the high pay. That is, this is new territory and applying vague role-based duties requires an argument rather than applying duties that work one way in ordinary areas of medicine.

Here is my response, then, in slogan form. There is no reason to think that a physician have duties that are not also owed by car salesperson. Even if a physician were to have such a duty, it would likely be satisfied given the low risk to young patients and the great expected benefit to the volunteer and society.

As always, thanks for the very interesting comments,

Steve K

Stephen Kershnar September 5, 2020 at 9:40 AM



Great response. Thank you.

Here is my question.

Is fairness a claim-right that someone may waive?

(1) Horn #1: Yes. If so, then fairness is not a distinct issue from valid consent.

(2) Horn #2: No. If no, then fairness is either an unwaivable right or not a right at all.

The notion that it is an unwaivable right intuitively seems to conflict with other rights we have (for example, to body and property) and, also, why it would be unwaivable. This is particularly true if autonomy, rather than interests, justifies rights. See the interest in informed consent in this context.

If the duty of fairness does not correlate with a claim to it, then it is free-floating. A free-floating duty is disturbingly mysterious in existence (What justifies it?) and content (What is fairness that is not right-based?). It is not tied to rights or desert and is, thus, a phantasm.

In summary, if there were a duty of fairness, it would correlate with a right and, thus, can be waived. If it did not correlate with a right, its existence and content would be disturbingly mysterious. Its mysterious nature becomes more extreme once we recognize it is distinct from rights and desert.


Steve K

David H September 5, 2020 at 1:25 PM

FAIRNESS AND RIGHTS. Imagine that someone is beneficially exploited (a sweat shop job or a pre-nup is better than no job and no marriage), that is treated unfairly but both parties benefit. [I suspect you will just deny that such a concept exploitation is coherent] They have waived any right to be treated better. Is it not still unfair given some theories of distributive justice? So unfairness is independent of rights. Waiving a right to better treatment doesn't seem to remove the ideal of fairness that has not been met. That is unfairness isn't removed by consent.

David H September 3, 2020 at 3:30 PM

SHOULD CHALLENGE STUDY PARTICIPANTS RECEIVE PRIORITY IN CARE? I think Eyan, Lipsitch and Smith (ELS) believe they should be prioritized and likewise for the WHO guidelines on Covid 19 challenge studies “Key Criteria for the Ethical Acceptability of COVID-19 Human Challenge Studies” that Pat forwarded. Why should they get benefits over front line health care workers or even benefits over those who took a vaccine in stage 3 study? The authors (ELS) even justified challenge studies by analogy to asking citizens to volunteer as Emergency Medical Services. Given the small number of people who may be needed in a challenge study (100 people was mentioned on p. 1753), some of those in the stage 3 vaccine study may have even been willing but not needed for the Challenge study. If the baseline risks are virtually the same for those in the challenge study and those not, then that provides even less reason to prioritize the sick challenge patients from sick volunteers in vaccine studies or just sick people in the same area.


Phil Reed September 4, 2020 at 2:21 PM

This isn't an ethical reason, but maybe the authors believe that receiving priority in care is necessary to incentivize volunteers for the challenge study.

David H September 4, 2020 at 2:38 PM


Maybe that is the reason. I just think the justifactory analogies - e.g. it is like asking EMS workers to take risks of exposure - aren't morally equivalent if those EMS volunteers or those in the equally risky baseline group don't WARRANT equal prioritization.


David H September 3, 2020 at 3:31 PM

FALSE ADVERTISING OF BENEFITS OF CHALLENGE STUDY WITH YOUNG? The Author’s main justification for a challenge study is that we want to speed up the production of a safe vaccine during an epidemic because thousands or millions of lives are at stake (pp. 1753-54). I wonder about the likelihood that vaccines for the young don’t work as well with those more at risk because of age and co-morbidities. Perhaps our Romanell Center MDs, Pat and Jack, could weigh in on the likelihood of this being the case. Of course, the at risk groups are still benefitted by the vaccine working on others when it doesn’t work on them. WHO pg 19 footnote 29 of their “Key Criteria for the Ethical Acceptability of COVID-19 Human Challenge Studies.”


Jack Freer September 4, 2020 at 7:54 PM

David, I think there are many unknowns here, owing to the many atypical and surprising features of this disease. The generalizability of data from young healthy subjects is one. The urgency of the situation forces compromizes in research approaches. Still, the biostatisticians embedded in these groups generally agree on most things and can, at least, explain the tradeoffs to the team. The study design must be informed by the biostatistical framework from the start. Retrospectively drawing conclusions from outcome measures that weren't among the primary endpoints is considered flawed.

David H September 3, 2020 at 3:32 PM

AUTONOMOUS AUTHORIZATION: ASKING VS. ASKING AND KILLING: The authors justify the risk in challenge studies of severe illness and death by other examples of asking people to take risks like volunteer firemen, organ donation to relatives, volunteers in drug and vaccine toxicity studies, volunteers for Emergence Medical Services (EMS) and point to Moderna’s candidate vaccine bypassing animal studies and the margin of safety that might apply (p 1754). The authors don’t distinguish directly killing someone by imposing a pathology that one knows is a pathology from asking and allowing someone to take a lethal risk from smoke inhalation and burns, or put oneself in a risky situation (EMS), give an organ to a loved one rather than take a risk for a population group that may include loved ones, giving someone a drug that one doesn’t think is pathological, and using a perhaps unjustified safety shortcut (Moderna) to justify other risks. [The latter reminds me of Leon Kass’s story that people put forth defenses of cloning on the basis of our already creating new lives in test tubes while such test tube baby making procedures were accepted at the time with reassurances that they would not lead to a slippery slope of more extreme reproductive technologies.]

The authors may be working with a loose sense of causality as making a difference but killing probably involves causality with oomph – billiard ball transmission of energy and the like that is present when one injects someone with a virus; it is not present when one asks someone to enter a risky profession. Difference making causation (counterfactual dependency etc) is not sufficient for killing as unplugging someone from life support is an act that makes a difference, receives CF support etc. Now I might be unfair to the authors as their point may just be that we accept requesting people to take on risks to their health but some of the examples do lack a potentially morally relevant feature of killing as involving proximal causation in the sense of pushing and pulling and those that do have it (organ donation and toxicity) are done in the former for more personal gain (loved one) and in the latter the intention is not to impose a pathology.

I would also like to see the comparative probabilities of morbidity and mortality in the different scenarios used to justify Challenge Studies. There is also Jim’s point above about health professionals standing in a different professional role relationship which obliges them towards some benevolence considerations that others who advocate to impose risks are not under.

David H September 3, 2020 at 3:34 PM

CHANNELING STEVE KERSHNAR: Perhaps pay volunteers and definitely compensate them if things go south. The former to increase numbers and the latter for fairness or desert. See the organ sales literature (Radcliffe-Richard, James Taylor, Gerald Dworkin) for worries about exploitation, coercion, autonomy. The WHO guidelines do mention compensation for harm (page 12 and page 13). ELS don't mention it. Perhaps it is taken for granted.


Stephen Kershnar September 4, 2020 at 5:12 PM


Some of this also appears above. If we pay volunteers very handsomely, it is hard to see what the objection is to this. Below is a back of the envelope calculation.

I wonder if the people concerned with this - and most bloggers here seem to approve of it - would withdraw his concern if the pay were high enough.

If a blogger were not to withdraw his concern, I wonder why we should think that he is looking after the interest of study participants.

The amount of money saved by challenge grants is likely very large.

(1) If it increases the GDP by 1%, it saves roughly $200 billion dollars.

(2) If it saves 5,000 lives at $5 million per life, it saves $25 billion dollars.

(3) It likely lessens business-destruction and further loss of liberty worth $10 billion.

Together the early rollout likely is wort $235 billion dollars. If there are 10,000 volunteers needed, we could pay each a million dollars to do the trial.

What is the objection to this?

Compare this expenditure of $235 billion to the $2.2 trillion dollars in welfare given to schools, universities, and other institutions whose spending is vastly inflated. This would be a bargain.


David H September 3, 2020 at 3:37 PM

WHO RESTRICTIONS ON CHALLENGE STUDY ENROLLMENT: The WHO's “Key Criteria for the Ethical Acceptability of COVID-19 Human Challenge Studies” states that we shouldn’t enroll those at risk for the disease due to background social injustice (p. 12-13). “Those whose background risk is high as a result of social injustice should be excluded from participation because their inclusion could be considered unethical” (p. 13) Of course, the latter is more contentious than WHO document lets on. But it also seems paternalistic. How does limiting opportunities of those with already limited opportunities benefit them or respect their autonomy? If those who made them less well off were then giving them an offer they wouldn’t take if better off, e.g. (fired someone illegitimately from their job so they would then give the former a better deal for their house or the wronged would then sell the wrongdoer one of their kidneys) then it is coercive. But that isn’t the case with Covid-19 challenges


Phil Reed September 4, 2020 at 2:49 PM

I think this one should be called "CHANNELING STEVE KERSHNAR PART TWO."

The autonomy of vulnerable people is already compromised. Abuse of their situation is a real concern. Preventing classes of people who are vulnerable (the sick, the imprisoned, the mentally ill) from certain options we extend to the fully autonomous respects their diminished autonomy by trying to prevent abuse.

David H September 4, 2020 at 3:44 PM


I think the WHO had in mind, young, healthy, competent folks who are victims of a contentious notion of social justice and thus are poorer people who are living and working in high risk regions which would be service jobs, congested factories,crowded homes and using public transportation. The competent would exclude the mentally ill, the challenge studies avoid the sick but seek healthy folks ages 20-40 or so. Prisoners may be considered by progressives and liberals to be victims of social injustice even if they are justly punished. But the WHO quote and context left me thinking they had in mind poorer people who live and work where the disease is more common. I didn't see why they couldn't volunteer to be in a challenge study. I don't even think they were talking about challenge studies with financial incentives other than compensation for harm.

Stephen Kershnar September 4, 2020 at 5:16 PM


Phil, I am not sure I see your point. It seems that people who are cognitively disabled or vulnerable could not give valid consent. For example, prisoners can't consent to sex with correctional officers. Ditto for 17-year-olds consenting to have sex with 30-year-olds in some states - such as California and Florida.

Why think anything like this is happening here?


Steve K


David H September 3, 2020 at 3:39 PM

CONSCIENTOUS OBJECTION: Maybe I am guilty of nitpicking ELS since they penned a brief article with other concerns in the forefront, but I think the authors and the WHO guidelines should have mentioned a right of conscientious objection of health care practitioners who don’t want to participate in inducing pathologies in the healthy, even with their consent? There is a big difference between being the proximal cause of death by injecting a pathology that one knows is a pathology and the other examples of asking or allowing firemen to take risks etc.


Pat D September 4, 2020 at 11:41 AM

David, can you give an example where this might become an issue? To my knowledge no one is expected to participate in these studies as researchers or experimental subjects against their will or better judgment.

David H September 4, 2020 at 2:26 PM


I am glad to hear that is the case. I just assumed that there might be hospital, academic, and institutional staff, rather than principle researchers, who might be expected to take part in challenge studies that were uncomfortable with doing so. Perhaps it wasn't mentioned in the article because it was just assumed that no one is involved against their will. Since conscientious objection is a huge topic due to people like Emmanuel, Schuklenk (editor of bioethics), Savulescu (editor of journal of medical ethics) advocating that there is no right to conscientious objection, I assumed it would appear in challenge experiments as well

Pat D September 5, 2020 at 10:50 AM

I see. This might be an issue if there was a rush to do these studies without thorough vetting of risk vs. benefits from multiple perspectives. If there was a known rescue treatment for Covid 19 and a well-defined challenge strain of the virus, would you still be concerned about a conscientious objection option for those providing institutional support?

David H September 5, 2020 at 1:14 PM

I think the nature of medicine is pathocentric (you are not a doctor or a nurse or pharmacist if you just facilitate cosmetic surgery, abortions, euthanasia, sterilizations, contraception, state executions etc. and never fight or prevent or ameliorate pathologies) and so health care personnel should not be forced to do something contrary to the goals of medicine. Inducing a pathology, even if it is legal, expected, standard, in the patient's interests etc., is still at odds with the nature of medicine that doctors profess allegiance to. Such medical personnel are being asked to do something contrary to the nature of medicine, not just do something NOT required by medicine (cosmetic enhancements). Incidentally, I actually think refusals are more akin to civil disobedience than conscientious objection because the former holds up the norms of the society or institution and demands everyone live up to and extend those norms to all - think of civil rights protests- while Conscientious objection usually involves an appeal to a morality that is not shared by the society or institution (e.g. pacificism). Anyway, whatever the proper taxonomy of the refusal, inducing a pathology, even if wanted, is something no one in medicine should be forced to do. Now all surgery induces pathology (cutting the skin) but it is for the sake of reducing pathology. Perhaps there is wiggle room in the challenge experiment in that the inducing a pathology is akin to some permissible cases of pathology inducing (surgery, vaccines?)in the service of fighting pathology but I suspect the analogy breaks down and it is more like taking an organ for one person to give to another.

Pat D September 5, 2020 at 9:26 PM

While I don't agree with your ethical analysis, I do respect your right to conscientious objection. One way to defend this is in terms of double effect, as Harvey suggested early on, and which Steve has trashed in other contexts. Maybe worth discussing with respect to another article in another round of this blog.


David H September 3, 2020 at 3:44 PM

PUBLIC TRUST: While it is true that challenge studies are a safer way to discover that vaccines don’t have bad side effects than larger field studies in which more people are vaccinated and thus there are greater numbers of possible victims, still, if a challenge study leads to the vaccine enhancing a disease, this may increase the anti-vax sentiment and limit the usage of any eventual test a vaccine than large. Given polls that suggest widespread skepticism of fast-tracked forthcoming vaccines, bad publicity from increased morbidity or mortality of a challenge study - with or without the vaccine being a disease accelerator - might make people later more suspicious than they would be of a safer better tested vaccine. We are going to have a lot of problems in the near future getting enough people to take the vaccine. Should the authors have focused more on the dangers of this public perception? The WHO document seemed more aware and concerned with the public perception and public involvement. Perhaps this is accounted for their having a different audiences and stakeholders than ELS.


Pat D September 4, 2020 at 10:52 AM

Another article about accelerating development of a Covid 19 vaccine using “Controlled Human Infection Models” (CHIMs) was published yesterday (9/3) in NEJM. It is available open access at

While recognizing a role for CHIMs in vaccine development, the authors recommend caution in rolling out such studies. For instance, they say: "Currently, we lack sufficient knowledge of SARS-CoV-2 pathogenesis to inform inclusion and exclusion criteria for a SARS-CoV-2 CHIM. A single death or severe illness in an otherwise healthy volunteer would be unconscionable and would halt progress. A challenge virus with potentially attenuating mutations may mitigate that risk, though such modification…would increase development time. Critically, a rescue therapy should be available, since even well-established CHIMs have resulted in unexpected severe illness" [e63(2)].

If proper precautions, such as those suggested here, were taken, I’m not sure that a single severe illness in a previously healthy volunteer would be “unconscionable,” although it might well “halt progress” in terms of public trust.


Phil Reed September 4, 2020 at 3:08 PM

Jim, thanks for picking such a short article. Of course, it has the negative effect that there is not a lot of (any?) philosophical meat.

It is striking to me that this article was pushed out in March 2020. We knew very little about the novel coronavirus then, and still don't know a lot (as Pat has pointed out). The authors were willing to justify challenge studies on the grounds that global death tolls "will be enormous," including the possibility of "20 million this year." 6 months in and we're still far from a million. I'm not congratulating myself for hindsight, but pointing out that their justification seemed to depend on a near-worst case scenario.

Obviously I do not mean to say that the coronavirus is not a serious situation. It certainly is. But it is not one that will cause 20 million deaths in a year. I worry that the authors were quick to argue for an ethically dubious experimentation strategy on strongly utilitarian grounds with little merit that the consequences would be as great as they feared. It introduces a worrying experimental precedent for any viral or contagious public health concern that has the mere threat of enormous death tolls.

This point is consistent with the ethical permissibility of challenge studies, but in more restricted cases.


David H September 4, 2020 at 3:25 PM


I agree. I was also wondering what the rough criteria was between epidemics and lesser threats to justify challenge studies. The authors seem to leave it at there are likely terrible consequences to be avoided and then a passing reference without much nuance to consent authorizing the risks. One couldn't help think the utilitarianism was doing all the work and the consent was damage control or just to forestall some objections. Not only would I have liked a lot more about what the patient must know to give informed consent in conditions of such uncertainty, but regardless of the nature of consent, I would have liked a bit more in the way of principle to when a challenge study is justified and when it is not.

Jack Freer September 5, 2020 at 4:41 PM

Phil and David, The death toll is a critical element in the utilitarian argument. Also, on the other side of the ledger are the normal volunteers who will participate in the standardization phase of the study which will have to identify the optimal dose of virus to administer in the actual vaccine vs. placebo trial. That standardization phase would entail a cohort of subjects at each of the viral doses, as it is incrementally raised. Incidentally, this phase would expose these subjects to the infectious agent with no offsetting hope of protection from a vaccine. Admittedly, the intervention study to follow would provide placebo vaccine to half the subjects. Still, I suspect some weight is given to the mitigating factor of vaccine administration in the intervention study.

As Phil points out, the numbers matter when making a consequentialist argument for the plan. The standardization phase adds a substantial number of intentionally infected subjects to those in the main vaccine trial.

Stephen Kershnar September 4, 2020 at 5:21 PM



Great article.

This is an easy case.

(1) If a test does the following three things, then it would seem morally permissible.

a. Satisfies all non-consequentialist side-constraints (rights and, perhaps, other Kantian considerations such as degradation, exploitation, objectification, etc.).

b. Maximizes the good simpliciter.

c. Maximizing individuals’ average well-being.

(2) Human challenge studies do the following three things.

(3) Hence, human challenge studies are permissible.

I am not sure why valid consent is not morally sufficient, but even if it were not, the combination of (1) through (3) seems to make this an easy call.

My question is what sort of reasoning would oppose this sort of experiment but support involuntary lockdowns, involuntary tracing, and involuntary vaccination. I cannot see how someone who values aggregate well-being, average well-being, or autonomy would oppose this type of trial but support the state use of force for these other things.


Steve K