[Blogger’s Note: I am very pleased to share this post by my colleague at Seton Hall Law, Carl Coleman. This post was cross-posted at Health Reform Watch.]
By Carl H. Coleman
With well over 5,000 global deaths from Ebola already reported, drug developers are working fast to begin human clinical trials of promising experimental treatments. Earlier this month, US government officials announced plans to launch a study of multiple Ebola interventions at the NIH Clinical Center, Emory University, and the University of Nebraska. Shortly thereafter, the international relief organization Médecins sans Frontières (MSF) announced that it would soon begin testing of three experimental interventions at its treatment centers in West Africa, in collaboration with a coalition of European partners and the World Health Organization.
As predicted in an earlier blog post, a major area of contention in these trials involves the ethical acceptability of using placebo controls. Plans for the US study are to give some participants the experimental drugs and others placebos, with everyone receiving the best supportive care available, such as fluid replacement and medications to fight off other infections. In the MSF trials, by contrast, none of the participants will be given placebos; instead, everyone will receive one of three different experimental interventions.
From a methodological perspective, it is easy to see why the designers of the US study have chosen to use placebos. Placebo-controlled trials are widely considered the “gold standard” of clinical research. Using placebos makes it possible to identify the extent to which observed outcomes in participants are the result of the experimental intervention, as opposed to factors such as access to better health care facilities, receipt of supportive care, or psychological expectations (the so-called “placebo effect”).
Historically, the primary concern with placebo-controlled studies has been the possibility that participants will be given a placebo when a proven intervention for their condition already exists. Thus, international ethical guidelines state that, with narrow exceptions, placebo-controlled trials should not be used if there are existing interventions for the condition being studied that have already been established to be safe and effective. The thinking behind these guidelines is that, in the absence of an existing intervention, participants who receive the placebo are not being harmed as a result of participating in the study, as they are not being denied access to an intervention that has already been established to work.
This framework is reflected in the World Medical Association’s Declaration of Helsinki, which states that it is acceptable to give placebos to members of a control group when “no proven intervention exists” for the condition being studied or, in exceptional cases, when the use of a placebo is necessary “for compelling and scientifically sound methodological reasons” and participants who are given placebos “will not be subject to additional risks of serious or irreversible harm.” The Declaration – which is generally considered to reflect a conservative position on the use of placebos – would therefore appear to support the use of placebos in clinical trial of Ebola medications, given that no “proven intervention” for Ebola currently exists.
Yet, opponents of using placebos in trials of Ebola medications argue that, given the known high death rate of untreated Ebola, it would be unfair to require some study participants to receive a placebo when it would be possible to give them an experimental drug that offers some possibility of benefit. They argue that, from the perspective of an individual who is suffering from Ebola, the risks of the unproven drug are likely to pale in comparison to the risks of Ebola itself. As a prominent group of ethicists recently argued in the Lancet, “When conventional care means [as high as 70%] probability of death, it is problematic to insist on randomizing patients to it when the intervention arm holds out at least the possibility of benefit.”
Of course, if we assume the perspective of the individual Ebola patient, it is difficult to argue with this risk-benefit analysis. Imagine an even more extreme situation: A new disease is known to be fatal in 100% of cases even with the best supportive therapy, and an experimental drug has shown promise in pre-clinical studies and has successfully passed Phase 1 safety testing. Who in their right mind would not want to take such a drug?
The problem with this line of thinking is that it ignores the ultimate purpose of conducting a clinical trial, which is to promote the best interests of patients and the public health over the long term. Certainly, clinical trials should never be conducted in a manner that affirmatively harms the interests of the individual study participants – that is why, for example, it is considered unethical to use placebo controls in situations where a proven treatment for the condition being studied is already available. But, particularly in situations involving widespread disease outbreaks that threaten the health and welfare of entire populations over an extended time period, it would be equally wrong to prioritize current patients’ desire for access to unproven medications at the expense of generating the kind of reliable evidence that will benefit the public in the long run.
It is therefore important not to lose sight of the long-term risks associated with foregoing placebo controls when conducting Ebola clinical trials. First, because all participants in these trials will receive the best supportive care possible – something that most Ebola patients have not been fortunate enough to receive – it is likely that the overall death rate in the studies will be considerably lower than it has been historically. Without a placebo control, it will be difficult to determine the extent to which participants’ improved survival is due to the experimental intervention as opposed to the enhanced supportive care. As a result, there is a real danger that a suboptimal, or even useless, intervention will mistakenly be deemed successful, ultimately causing more harm than if the study had never been conducted.
Second, even if it may appear that no experimental intervention could possibly cause a worse outcome than a disease with a mortality rate of up to 70% if left untreated, we don’t in fact know if that is the case. As Annette Rid has argued, given that most experimental drugs are ultimately proven not to be safe and effective, “it is more likely than not that the interventions [for Ebola] will not improve symptoms for patients, and might even weaken them as they battle a life-threatening disease.” Again, without a placebo control group it will be difficult to determine the extent to which negative outcomes experienced by trial participants are due to the experimental interventions as opposed to Ebola itself.
In response to these concerns, the group of ethicists writing in the Lancet suggested that there are other ways to gather reliable scientific evidence. For example, they argue that a “viable approach would be to try different treatments in parallel and at different sites, following observational studies that document mortality under standard care.” Although such an approach would clearly be inferior to a placebo-controlled study from a methodological perspective, it might be a reasonable compromise if it were possible to ensure that the observational data were limited to past patients who had received the kind of state-of-the art supportive care that will be given to participants in the clinical trials. But if the observational data are simply general assessments of the average level of care that most patients have been receiving, virtually any experimental intervention will look better by comparison – even if the improved outcomes are due to ancillary factors such as improved supportive care. Moreover, to the extent the observational studies delay the initiation of the clinical trials, the result would be that some patients would still be denied access to the experimental intervention. The only difference would be that this would happen before the trials begin, rather than as part of a simultaneous control.
Ultimately, perhaps the strongest argument in favor of using placebo controls has to do with the limited supplies of Ebola medications currently available. Given these limits, the question is not whether some Ebola patients who want experimental medicines will have to forego them; rather, it is whether those who forego the medicines do so inside or outside of a clinical trial. In a situation of scarcity, clinical trials in which everyone receives the experimental intervention would have to be smaller than studies involving a placebo control, as enrollment would have to stop once available supplies of the test articles have been exhausted. From the perspective of an individual who will not have access to the interventions because of limited supplies, it would be preferable to participate in a larger clinical trial as a member of the control group, where at least there will be access to high-quality health care facilities and the best supportive care possible. When we add in the societal interest in generating the most reliable evidence possible, placebo-controlled trials appear to be the best solution for everyone, at least as long as the availability of experimental medications is insufficient to meet the demand.