Social value, clinical equipoise, and research in a public health emergency

The 2016 CIOMS International ethical guidelines for health‐related research involving humans states that ‘health‐related research should form an integral part of disaster response’ and that, ‘widespread emergency use (of unproven interventions) with inadequate data collection about patient outcomes must therefore be avoided’ (Guideline 20). Keywords: CIOMS guidelines, Ebola, equipoise, public health emergencies, research ethics, right to try.

Paper 6. Social value, clinical equipoise, and research in a public health emergency by A. J. London


Jack Freer July 5, 2020 at 10:35 AM

London argues against Schuklenk and Smalling's claim that there is no clinical equipoise between unvalidated medical interventions (UMIs) and the standard of care (with or without placebo). Specifically, the context is desperately ill patients in a widespread emergency involving a high-mortality disease.

I agree with London's conclusions on all counts (including his other argument that societal benefit of high quality evidence via clinical research outweighs any individual's right to receive UMIs). For now, I will address two equipoise related issues.

1. Clinical equipoise vs. personal equipoise.

Clinical equipoise is based on the uncertainty existing in the expert medical community regarding efficacy (not any individual clinician, researcher or patient). Leaving aside for the moment, how expertise is demonstrated and verified, the expert community is the relevant body to address because equipoise is primarily a research ethics concept. True uncertainty must exist in the group that will utilize the research findings--those clinicians and researchers who work in the medical domain being addressed. While some individual professionals might believe they have personal certainty about efficacy, that belief is generally private and inaccessible to outside scrutiny. This leads into my second point.

2. Anecdotal evidence, cognitive bias, and placebo effect.

Support for UMIs are often based on anecdotal evidence and testimonials. While not totally worthless, these types of evidence are generally the jumping off point for higher quality evidence. By themselves, they are close to worthless. Anecdotal evidence, however, is often persuasive to the layperson (reflecting the sad state of scientific literacy in our society). Clinicians and researchers are not immune to cognitive bias and placebo effect. Thus, it is all the more important to design and carry out double blinded randomized controlled trials, since this will tend to mitigate such biases.


Stephen Kershnar July 5, 2020 at 3:01 PM


When London says that a desperately ill patient has no right to a UMI, do you interpret him to mean he has no right to it - claim against another - or that it is merely not good policy?

If the latter, then why does he spend time talking about entitlements, stakeholders, and personal decisions?

If a master's degree in education were not to make a teacher more effective (see link below), would a school - whether privately or publicly owned - have no right to require a master's degree in order to be hired? Whose right would it infringe if it were to do so?

(See page 4)

I hope life is treating you well,

Steve K

Jack Freer July 5, 2020 at 6:27 PM


I think he believes both (Section 2 and 3 respectively). It sounds like he thinks it's bad public policy because (being a positive right), a right to UMI places obligations on other stakeholders. A right to unvalidated interventions is particularly poor public policy.


Stephen Kershnar July 8, 2020 at 5:45 PM


Thank you for your thoughts. My concern is that London is basing his right-based conclusion on his policy-based conclusion. I do not think this is the correct way to argue about rights (rights do not rest on maximizing the good or the interests of stakeholders). Nor do I see the policy analysis as distinguishing between public and private medicine-makers. I think this matters.


Steve K

Stephen Kershnar July 5, 2020 at 2:44 PM


Here is London’s reasoning.

(P1) If a desperately ill person has a right to a UMI-drug, then stakeholders do not have a

morally relevant interest regarding his getting that drug.

(P2) Stakeholders do have a morally relevant interest regarding his getting that drug.

(C1) Hence, a desperately ill person does not have a right to a UMI-drug. [(1), (2)]

Here are two ways we might understand the right to the UMI-drug.

(a) Negative Right. One person has a right to a UMI against a second person if and only if the first has a claim against the second that the second not interfere with his receiving it so long as he does not use (unjust) force, fraud, or theft to obtain it.

On this account, a person’s right to access UMI is similar to his right to receive a book. He has no right to another’s book. If a second person is willing to give him the book, though, a third party may not prevent the transfer.

(b) Positive Right. A person has a right to a UMI against a second person if and only if the first has a claim against the second that second give him the drug.

On this account, a person’s right to a UMI is similar to a person’s right to be given another person’s house, kidney, or labor. There is no such natural right. If there were a non-natural right to such things, it would depend on right-waiver (for example, gift or sale) or forfeiture.

If the right in question is a negative right, (a), then premise (P2) is false. Stakeholders – for example, academics and government labs – have no right-related interest in a transfer of private property, whether a book or a drug.

If the right in question is a positive right, (b), then the argument succeeds but is irrelevant. No one has natural positive rights. For example, David has no right to Neil’s house, kidney, or labor. A positive right derived from negative rights depends on waiver (that is, promise or consent) – see, for example, U.S. Constitution – not on stakeholders’ interests. That is, (P1) is false.

Hence, we end up with the following argument.

(1) London’s argument focuses on a negative or positive right.

(2) If it focuses on a negative right, then it fails [because (P2) is false].

(3) If it focuses on a positive right, then it fails [because (P1) is false].

(4) Hence, London’s argument fails.

If people have rights, then they are negative and depend on shareholders (owners) not stakeholders. When someone makes an argument about what rights demand and supports it with stakeholder-related claims, we should be wary.


Phil Reed July 7, 2020 at 2:43 PM


Helpful reconstruction as always.

Two related comments. First, why think that a UMI is a private good or private property? I would guess London is thinking of it as a public good, which no private person "owns,” like water or air.

Second, I assume London would say that the relevant right is negative. Do you think premise 2 in the original argument is even false under conditions of scarcity? The context for his argument is scarcity/emergency.

Stephen Kershnar July 8, 2020 at 5:50 PM


Great comments as always.

On your (1), I don't see how a medicine can be a public good. Following Nozick, if you and your chemistry colleagues make an valuable drug (cancer treatment) out of household chemicals in your kitchen, I don't see this as being a public good like a road or clean rivers. I think Johnson & Johnson's drugs are similar to what you and your colleagues made.

On your (2), if the right is negative, then why do stakeholders' interests count. Imagine that Apple wants to maximize the profits for its phones, computers, etc. In deciding what right is has to do so or not do so, I don't see why it need consider the interests of its employees or government workers who regulate it. They have to protect their interest through contract. That is, I don't see that the negative-rights theory fits with stakeholder theory. Do you disagree?


Steve K

Stephen Kershnar July 5, 2020 at 2:45 PM


Hypothetical. New York’s government (governor and legislature) decides that students who want to go to college in New York should be randomly distributed to universities and colleges there. This includes private and public universities (for example, Cornell and Old Westbury). The government reasons that relevant experts – including diversity officers and education professors – disagree about whether ability-matching has benefits that outweigh the costs of losing out on socially valuable knowledge and, also, providing role models, reducing stereotypes, promoting integrative friendships, etc. that random assignment would bring about.

An outstanding student claims that he has a right to go to a private university (for example, Cornell) so long as the school permits him to attend.

The government reasons as follows. “Without credible evidence of ability-matching students to universities, there can be no duty to provide an ability-matching university to a student. In the presence of uncertainty about the benefits of the best available educational alternative, it might be permissible to provide an ability-matching university. But it would also be permissible not do so If it is permissible to offer students access to an ability-matching university and permissible not to do so, then it must be permissible to allow students to be randomized to universities. Allowing students to be randomly distributed to universities facilitates the generation of socially valuable knowledge without depriving anyone of an education to which they are entitled because of their human rights and, also, federal and state law.”

Even if the government were to have the right to do this with SUNY and CUNY schools, it might not do so – at least if we are to focus on people’s rights – with private universities.

Stephen Kershnar July 5, 2020 at 2:49 PM


London says the following, “A necessary condition for any right or strong moral claim to access UMIs is that UMIs have a reasonable expectation of meaningful benefits.” See p. 329.

Here is a statement of this claim.

(1) If a desperately ill person has a right against the government to a publicly provided UMI, then the UMI has a reasonable expectation of meaningful benefit.

Consider it as a claim about rights in general.

(2) If one person has a right against a second to a thing, then the thing has a reasonable expectation of meaningful benefit.

This is false. One person has a right to a thing if either he came to own that thing and it was previously unowned or the previous owner validly consented to transfer to him. Neither requires that the gaining of ownership or transfer be reasonable.

Consider it as a claim about a right against the government.

(3) If one person has right against the government to thing, then the thing has a reasonable expectation of meaningful benefit.

This is false. The government has permissible to provide a thing even if it does not have a reasonable expectation of benefit so long as it follows the procedure to which its owners – citizens – have validly consented. Intuitively, one might have a right against the government that it return his racing motorcycle – taken away in an arrest involving mistaken identity – even if the thing (racing motorcycle) does not have a reasonable expectation of meaningful benefit.

Even if a right were to focus on what is reasonable, what is meant by reasonable?

Reasonableness is not efficiency because the latter is not a feature of rights in virtue of efficiency depending on facts about third parties, being forward-looking, and so on. Reasonableness is not just or permissible because we are trying to determine what is just or permissible.

Do not handle “reasonableness” weasels. They bite hard.

David H July 5, 2020 at 3:02 PM

Indeterminacy and Uncertainty about the Indeterminacy and Uncertainty of Equipoise:

First and Third Person Equipoise: If I were an expert on an infectious disease and recognized the other experts were split down the middle between UMI and standard care and couldn’t determine which side had more to say in favor of it, then I could, in good conscience, put a patient in a randomized trial on the basis of equipoise. But what if I thought the preponderance of evidence favored the UMI? Could I still, in good conscience, put my patients in the randomized trial? If not, then there is a difference and tension from the third person researcher equipoise and first-person researcher equipoise. Why favor third person equipoise?

Who is an expert? Let’s assume that it is true, that on one sense of “expertise”, the field is split. But there are experts and then there are experts. The first may be everyone in the field with an MD or MS or PhD as contrasted with laypeople, the second sense may be the elite of those in the field. It is not as clear me, that if I side with the elite in the second sense, that there is equipoise in that context.

A Reasonable Minority and Equipoise: Leaving aside the vexing problem of determining who belongs to the elite, equipoise as understood by London, allows a minority of experts to be grounds for randomize. “Each patient is guaranteed to receive a level of care that would be recommended for them by a reasonable minority of experts” (333). Thus, equipoise doesn’t require the field be evenly split, just some experts are required. Now I may be guilty of looking for precision where it can’t be found, but image that the split between UMI and standard care is 70/30. Is that difference of opinion grounds for claiming one is not “consigning them to care that is known to be inferior” (333)

Failing Treatment and Equipoise: Maybe there is no reason, in general, to favor the UMI against standard care, given the reason for pessimism based on intervention ensembles. So, someone just diagnosed with Ebola has a reason to favor the standard fluid replacement over UMI. But London and his opponents may sometimes be talking past each other. The latter may have in mind scenarios where the patient is not responding to standard care, perhaps is on his deathbed, standard care is just comfort care or a placebo, at best standard care will only slightly delay death. There may not be equipoise then as he has nothing to gain from standard care and so would prefer the Hail Mary Pill


David H July 5, 2020 at 3:53 PM

Tyop correction:

Perhaps it is evident from the context, but I meant to write in the last passage "So, someone just diagnosed with Ebola has NO reason to favor the standard fluid replacement therapy over the UMI"

Stephen Kershnar July 9, 2020 at 4:13 PM


An individual expert might have an opinion that conflicts with a reasonable minority of experts. I am unsure what work 'reasonable' does. If it means that the experts' opinion is likely true or supported by evidence, then I am unsure why it matters how many experts there are. If it does not mean this, then I am not sure why a minority of experts simpliciter will not do.

Plus, I wonder if there are cases when a reasonable minority of experts could recommend a Hail Mary UMI because it is at least as good as any other treatment or non-treatment available. That is, could the expertise take into account the badness of the other options.


David H July 5, 2020 at 3:03 PM

Self-Defeating Appeal to UMI

London claims that the critics of equipoise are putting forth a self-defeating position (332). He seems to interpret a quoted passage from Caplan et al as ruling out a randomized trial while permitting either UMI or standard care. He then claims that if both are permissible then so should be randomizing. But I don’t see the prohibition. Is it inferable from the sentence “…it is morally problematic to insist on randomizing patients to a control arm in the context of an ineffective standard of care”? But the “to insist” might be meant in that context to only allow UMIs in randomized trials. It might not be morally problematic to have trials while also allowing some UMIs outside of such trials. Anyway, the main charge of self-defeating is that allowing UMIs outside of trials doesn’t help any stakeholders as it doesn’t produce useful information about treatment options. London seems to be speaking of classes of stakeholders and ignoring the individual stakeholder who are dying and standard care has not helped. They are not going to live long enough to be a stakeholder benefitting from the trial. So, while the group of patients who will have the disease will, as a whole, benefit from the study and thus as a group it would be self-defeating to slow down those gains, certain diseased individuals will not benefit from that research and may benefit from the UMI.


Phil Reed July 7, 2020 at 2:47 PM

I don’t think the quoted passage from Caplan does explicitly prohibit the randomized trial. I think London infers it, both from the passage you quoted and from the beginning of that passage: “equipoise...breaks down when available conventional care offers little benefit.” If there is no equipoise, then the randomized trial shouldn’t proceed. All that means, though, is that a clinician shouldn't give substandard of care if he or she is convinced that one treatment is superior. It doesn't mean there is no basis for the trail at a social level, so I think you're right that London is wrong to read this as prohibiting a trial. See also my comment below.

David H July 5, 2020 at 3:04 PM

Question Begging on Rights and Stacking the Deck:

Rights trump Interests: London writes that “the decision to use UMI’s is not solely a personal choice, and there is thus not prior right to use of the UMI.” He then goes on to discuss all the other stakeholders and claims that the onus is on the advocate of the UMI outside of the trial to claim why an individual patient’s “Second, efforts to salvage the view that desperately ill patients have a right to access the UMIs would have to establish that the interests of those individuals are so morally weighty, and the value of access to UMI is so high, that access trumps or would outweighs the costs and consequences enumerated here” (329). London seems to be assuming what he has to show – there are no rights that trump the interests of others. But that is just what rights are, trumps against the majority, trumps against utility, trumps against aggregate interests, trumps against costs and benefits and the like.

Rights are Prior to Personal Decisions: London seems to think rights are just personal interests and so where there are other interests of stakeholders, then there are no rights. But rights will always affect the interests of other stakeholders. One doesn’t have to establish something is only a personal choice to be a right. Pace the quote “the decision to use UMI’s is not solely a personal choice, and there is thus not prior right to use of the UMI”, it is because someone has a prior right that it becomes a personal decision. It is not that there is first a personal decision that doesn’t affect others, so we then recognize rights.

Stacking the deck: London mentions all the costs of providing UMIs outside of a trial. But he doesn’t mention the costs of standard care. And then he worries that anecdotal good results from UMI’s may lead to wasteful expenditures on UMIs and poor anecdotal results may sabotage a good research project. But it could work the other way as well. The UMI’s may lead to good research projects being started and bad ones prevents. Anyway, the decisions of financing of research could just have higher standards and so be impervious to UMI anecdotes


Stephen Kershnar July 9, 2020 at 4:21 PM


I wonder if London's reasoning generalizes. On your interpretation, here is his reasoning.

(1) X is not solely a personal decision (because it affects the interest of stakeholders).

(2) If (1), then there is no prior right to X.

Consider the following activities. Using recreational drugs, dropping out of school, having children out of wedlock, joining a polygamous religion, and being a member of a religion whose members on average take a large net amount of taxpayer-provided benefits. Arguably, these activities affect the interests of stakeholders. Hence, they are not something to which participants have a right.

I wonder why London thinks that what rights we have depend on how rights relate to stakeholders' interests.


David H July 5, 2020 at 3:05 PM

Rights and Informed Consent, Confusion about Coercion:

The Extension of Informed Consent: Now I doubt that there is a right to use every experimental drug on shelf in the academy or private laboratory. So I am in agreement with London. But one should make that claim independently of all the discussion of costs to other stakeholders. And it shouldn’t be based upon informed consent and coercion. Patient autonomy will just sometimes be overridden or not apply to scarce and/or experimental drugs. So informed consent doesn’t apply, or is overridden. I think the former is better interpretation. If one doesn’t have a right to something, then its denial doesn’t count as failing to obtain informed consent. One’s right to give informed consent to treatment or refusal of a treatment doesn’t extend to treatments one has no right to. So, there is no override of a right to informed consent, it just doesn’t apply. London should just say denial of UMI’s is not a case in which informed consent can be met or denied. And the reason informed consent is not denied is not because there wasn’t coercion as conceived by Schuklenk and Smalling.

Confusion about Coercion: Schuklenk and Smalling claim that offers are not coercive whenever two conditions are met (1) someone is offered something that could or even would make them better off and (2) that they are not entitled to receive that something. But surely there is something X that is good for me that I am not entitled to but don’t want. If you threaten me with force if I don’t choose X surely that is a coercive offer but it meets the two conditions that it makes me better off and I am not entitled to it. I understand that there are many conceptions of coercion, but the classic case of “money or your life” suggests that the absence of coercion requires the absence of a threat as a condition. My not agreeing to a drug being withheld is not coercion even if force will be used to keep me from taking the drug. That is true of virtually every offer and it doesn’t make such offers coercive. If you want my car and I won’t sell it to you for the price that you want to pay, your then taking my car will be met with force but that doesn’t my offer a coercive offer. So not receiving a drug that you want and force being used to keep you from taking that drug is not coercion.


David H July 7, 2020 at 4:45 PM

London just needed to make a clear argument that patients don't have a right to UMIs and then there is no need to go into everything else in the article about informed consent, coercion, optimism and pessimism, risks and gains, self defeating aims, opportunity costs, various other costs and benefits, overrides, countervailing stakeholder interests, scarcity, threats to funding appropriate research projects and wrongly funding poor research programs etc., to limit UMIs to controlled randomized studies. Or at least none of the other discussion is needed if it doesn't serve to support equipoise between UMIs and standard care.

Stephen Kershnar July 9, 2020 at 4:39 PM


Like you, I wonder whether a coercive offer is even possible.

One might think that an offer does not worsen one's position whereas a threat does so.

If coercion is purely psychological - no reasonable person would avoid it - then really good deals become coercive offers. This seems incorrect. If coercion is moralized - no reasonable person would avoid it and it infringes on the right of the person to whom the proposal is given - then it is hard to see how there could be a coercive offer. After all, the offer has to at least as good from the offeree's position as he would have been in were the offer not made.

I agree with you that London should have limited himself to discussing the right to a UMI as his real argument. The rest was not merely unneeded, but irrelevant.


Steve K

David H July 5, 2020 at 3:07 PM


Would it ruin the study if some people who wanted an UMI were put in a trial in which they were given the UMI rather than just randomized? Would their knowing that and someone, perhaps not their doctors, knowing it as well, skew the study? Could the study compensate for satisfying their preference by giving people with similar profiles the standard care and placebo?


Jack Freer July 5, 2020 at 6:51 PM

Allowing a research subject choice in which study arm they are placed would diminish the value of the findings (not randomized, not fully blinded) and is unlikely to be taken seriously. Individuals who are contemplating participating in research should (but don't necessarily) understand that they are performing an altruistic act to help future patients. If they achieve some benefit (from the study drug), or dodge some serious side effect (by being in the placebo or standard treatment group), that is a bonus.

Stephen Kershnar July 8, 2020 at 5:53 PM

There is something funny about the medical field explaining that UMI need not be given out because the drugs are not empirically validated. Much of what is considered in medical school admissions (for example, essays, shadowing doctors, and interviews) is not empirically validated and yet in part guides the distribution of medical school and, perhaps, other medical positions.

I always wondered how the field thinks about this seeming inconsistency.

Phil Reed July 7, 2020 at 2:49 PM

This was an interesting paper and a good choice for a pandemic-related discussion. My thoughts on this are inchoate. I drafted this without first reading the other posts, but I have updated it since.

I’m inclined to agree with the CIOMS guideline, which I take as the conclusion of the paper, that it is appropriate to avoid the administration of widespread emergency UMIs in public health emergencies.

London considers two objections to this guideline. First, it is unethical to restrict access to UMIs to the clinical trial context because desperate patients have a right to them. I find London’s argument in section 3 against this objection wholly convincing: the use of UMIs is not merely a private decision for the reasons he lays out.

The second objection to the guideline is that equipoise does not exist. Here, I think London gets into some problems. I believe his way of rebutting this objection is to argue that there is clinical equipoise in the clinical trial context. I think this is a mistake. His argument should instead be that equipoise is not relevant to this situation (i.e. where the patient demands a UMI).

I understand the purpose of equipoise as protecting patients against substandard interventions for the sake of the progress of scientific/medical knowledge. London’s insistence that equipoise obtains upends the purpose of equipoise. That is, London wants equipoise to work for the advancement of “socially valuable information.” But his use of equipoise in this way does not respect the patient in question who wants the UMI. That patient judges that his last best hope is to receive the UMI. According to the patient’s own judgment (and his clinician’s, who is presumably willing to administer the UMI), the patient who is randomized in a clinical trial is undertaking the risk of receiving a substandard intervention. Thus equipoise is not protecting this patient in the way it is designed to do. It’s still true that there can be equipoise in a general sense in the medical community. But the patient’s judgment is a necessary (though not sufficient) component of subjecting him to a randomized treatment. (I think this point is related to Steve’s analogy of randomizing universities.)

This criticism doesn’t affect the conclusion of the article; I agree with London that the patient isn’t entitled to a UMI and that in a public health emergency it is permissible to limit the use of UMIs to clinical trials. But the grounds for this latter claim have nothing to do with equipoise. Equipoise only justifies giving patients an experimental treatment when the clinician and patient/subject concur that there is uncertainty between the effectiveness of it and the standard of care. This is just not the case here. Instead, the patient can be denied access to a UMI on the basis of the reasons London gives in sections 4 and 5: the UMI may not be his last best hope (contrary to his and his clinician’s judgment) and even if it is he is not entitled to it in the context of a public health emergency.

It’s not that the equipoise breaks down, it’s that the equipoise isn’t relevant if the patient does not consent to the uncertainty.

London, in his other writings, has a tendency to justify wide application for equipoise. That tendency is on display as well in this article. This worries me, from the standpoint of protecting research subjects from being used as a means.


Stephen Kershnar July 9, 2020 at 4:29 PM

Dear Dean Reed:

You state the following, "I find London’s argument in section 3 against this objection wholly convincing: the use of UMIs is not merely a private decision for the reasons he lays out."

If I am reading London correctly, and perhaps I am not, he argues that there is no right to a UMI because it does not involve a personal decision. It does not involve a personal decision because it affects stakeholders.

I wonder what you and London think there is a right to do. That is, what sort of things are personal decisions or do not affect stakeholders. Because we are looking at policies, I fail to see why anything in the aggregate (education, food, sex, or reproduction) does not affect stakeholders.

If we are looking at individual acts, then what is a personal decision for one person might not be for another.

In short, why do you think that London's argument is plausible?


Steve K

Phil Reed July 7, 2020 at 2:50 PM

Self-Defeating Appeal to UMI (see also David's comment above)

More specifically, here is where I think AJL goes wrong. He says that pro-UMI position is “self-defeating”: “It permits clinicians who favor the provision of UMIs to provide them to patients and it permits clinicians who favor standard treatment to provide that, but it prohibits research that would allocate patients to a UMI or the standard of care in a way that would generate the evidence necessary to establish the relative merits of those practices.” Their position doesn’t prohibit this research in general. It prohibits _a particular patient_, who wants the UMI, from being randomized and so subject to a substandard treatment against the patient’s wishes.

Now, some would regard this argument is moot. Since my own view doesn’t really help the patient: (he’s not entitled to a UMI, on my view, so the only thing he can do to try to get it is to enroll in the clinical trial, where he has a shot at getting it). This is true. I don’t think my view helps the patient get what he wants. But I do think my view protects equipoise from corruption for the sake of the advancement of science.


Phil Reed July 7, 2020 at 2:52 PM

As a related remark, I think the arguments from sections 5 and 6 are contradictory (or at least in tension). Section 5 wants to convince us that the UMI is very unlikely to produce a meaningful net clinical benefit. In that case, contrary to section 6, then there really isn’t equipoise - the standard of care is preferable.

I think he can get out of this objection, though, by insisting that his conclusion from section 5 is more tepid - that it’s merely reasonable to believe that the UMI is very unlikely to benefit, and it’s also reasonable to believe that it will benefit. However, this is part of the danger of equipoise - we see what we want to see so that the clinical trial can go forward (see David’s comment above about indeterminacy and uncertainty). The thrust of section 5, if sound, seems to point against equipoise.

Harvey Berman July 11, 2020 at 9:40 PM

Equipoise is easy to define in principle. How it is enacted in practice is a different matter.

I express two concerns: 1) about clinical equipoise in practice; 2) the meaning of “unvalidated."

1) In randomized clinical trials, the area of London’s concern, it is clinicians who recruit patients into these trials. How clearly such clinicians communicate the risk-benefit equation is key in achieving clinical equipoise. That is, how do they satisfy the requirement for equipoise in their entreaties to patients?

An irony in seeking equipoise is, as London comments with appropriate pessimism, is that fewer than 10 % of drugs in the “pipeline” are workable. Yet, physicians continue to express to patients, likely in desperate straits, that experimental drug X may be helpful, knowing all the time the odds against it.

2) What is the meaning of unvalidated? Is it a new drug entity that is absolutely unknown? Is it a “me, too” agent based on an already known agent? Or is it like hydroxychloroquine (HCQ) where anecdotal (and problematic is correct, too) knowledge is available?

Clearly HCQ merits study. But why shouldn’t a free man not in a formal RCT, not be allowed to try such an unvalidated drug if there are no other acceptable standards of care? Is this not too dissimilar from off-label prescribing? It is debatable if should RCTs be the standard on which validations are based for all conditions and all clinical situations?)

3) One final point, a caution really, about the politicization of drug studies as occurred during Trumps over-selling of HCQ. Had it been another president, perhaps Obama, the media would have been calmer and more deferential to him and to the use of HCQ. It has been forgotten that quinine, the parent drug of HCQ, was effective (in high concentrations) in treating some patients of the Spanish 1918 flu. (Yes, a flu, and not a coronavirus, pandemic.)