For more than a decade, the U.S. Preventive Services Task Force (USPSTF)has published a ranking of clinical evidence. The rankings, however, are based on what clinical investigators have known for decades about the validity of scientific research in general. In brief, they are as follows:
Grade I (strong): Evidence from at least 1 properly designed RCT.
Grade II (moderately strong): Evidence from properly designed, controlled trials without randomization; cohort or case-control studies; a time series; and uncontrolled study.
Grade III (weak): Evidence from opinions of respected clinical authorities; descriptive studies; case reports; and opinions of expert committees.
In nearly all cases for the justification of equipoise, we are dealing with either weak evidence or no evidence used by the physicians/investigators. In other words, we use nonscientific or no evidence to claim equipoise between 2 treatments. This is because we are comparing a standard of care that has been established with a minimum of 1 clinical trial with an experimental drug about to be tested.
Is There Equipoise in Clinical Research -- Phase 1 Trials?
Freedmanrejects Fried's concept of "theoretical equipoise" because Fried claimed that "theoretical equipoise" depends on data from "the literature, uncontrolled experience, consideration of basic science and fundamental physiological processes, and perhaps a 'gut feeling' or 'instinct.'" Freedman describes "theoretical equipoise" as a "fragile" event. But then he reports that the concept of "clinical equipoise" is caused by 2 existing conditions: 1) a conflict among the clinical community over the preferred treatment because of some new evidence, but not true clinical trials; and 2) the clinical community's use of 2 different standards of treatment based on its own experience, meaning that no consensus exists. The 2 reasonings by Freedmanand Friedare similar. Nevertheless, neither suffices. Freedman's explanation is not scientifically valid, and Freedman's clinical equipoise is equally fragile to Fried's theoretical equipoise. The lack of consensus is based on lack of definitive clinical trial(s). Freedman states:"[W]e need to recall the basic reason for conducting clinical trials. There is a current or imminent conflict in the clinical community over what is preferred for patients in a defined population." Freedman substitutes the illogic of "theoretical equipoise" with the uncertainty of the clinical community at large in different groups of a clinical trial. He goes on to say: "Good medicine is the product of a professional consensus."
The clinical community should base its conclusions on evidence-based medicine and not "gut feelings" or "hunches." Who decides what size and number constitutes the clinical community? Who declares, on behalf of the clinical community, that they have reached consensus? Freedman's statements reflect confusion about the scientific method. In 2003, Shamoo and Resnik stated in their textbook Responsible Conduct of Research: "Private intuitions, hunches, faith, introspection or insight can play an important role in generating new ideas to test, but they do not constitute rigorous proof."
Researchers should use scientific evidence to determine the validity of clinical trials. The primary valid scientific evidence (grade I, strong) is obtained from RCTs. Grade II and weaker evidence should be considered temporary until grade I evidence is obtained. In certain circumstances and at times for ethical reasons, we cannot obtain grade I evidence. Therefore, we can use weaker evidence (grades II and III), but we should not use it to establish equipoise between 2 clinical treatments. This is because we are comparing a standard of care established with grade I evidence with weaker evidence. More important, there is no equipoise between weak evidence of 2 varying degrees, and it is not really scientifically valid. This acrobatic justification of clinical equipoise is illustrated by the statement: "At the outset of a trial there exists a state of honest professional disagreement in the community of expert practitioners as to the preferred treatment." It is difficult to imagine how, who, and which community of professionals would settle this disagreement.
The control or healthy subject's participation in phase 1 clinical trials rests on the premise that the potential benefits should outweigh the potential risks. The participants in the control group in phase 1 trials encounter no risk, nor do they have a chance of benefit. Control participants may receive psychological benefit through the notion of helping others and from the physical examination. However, in reality this is a tenuous benefit at best. The volunteer does receive financial payment. But, both the National Institutes of Health and the Food and Drug Administration, for good reasons, do not allow the inclusion of payment in the benefit calculation.This is consistent with the federal regulation, which states: "In evaluating risks and benefits, the IRB [institutional review board] should consider only those risks and benefits that may result from the research." (45 CFR46.111(2)).The risk to subjects, however, is real in terms of toxicity, which cannot -- and is not claimed to be -- a minimal risk. Moreover, is it really worth the effort to determine 2 groups of a disagreement where the data for one arm of the trial are weak or nonexistent? This method is illustrated by the claim that phase 1 agent risk-benefit should be compared with no treatment. This stems from the flawed concept that in research, the researcher has a "therapeutic obligation" to subjects. A recent survey indicates that 83% of physicians will use clinical guidelines with evidence when comparing one drug with another. The researcher's primary obligation is to acquire new knowledge for the public good. This method of equipoise determination will certainly contribute to the confusion and variation in institutional review board decisions.
Emanuel and coworkersadvance the notion that ethical research must meet 7 requirements. The first requirement is that research ought to have "social, scientific, or clinical value." This is clearly a utilitarian calculation. While their second requirement is scientific validity, with which I agree, they support the concept of equipoise to justify scientific validity. As discussed, the concept of equipoise is scientifically invalid. Moreover, Miller and Brodypoint out their use of the flawed concept of equipoise to justify their argument, again conflating the scientifically invalid notion of equipoise to justify research with human subjects.
So, where is the claimed "equipoise" between the volunteers in a phase 1 clinical trials and the nonparticipant, ordinary human being? The only ethical justification for the healthy volunteer in a phase 1 is the potential opportunity to benefit society. In other words, it can never be justified with an autonomy-based argument without ethical contortions and obfuscations. Shamoo and Resnik, in their discussion of the justifications of a phase 1 trial, mention that it is due to "strong utilitarian reasoning."
Moreover, this type of argument is consistent with John Dewey's philosophy (surely not a "pure" utilitarian), who would probably agree with Immanuel Kant's admonition that: "One must act to treat every person as an end and never as a means." A patient may agree to procedure "X" out of a sense of duty, not because he or she expects a benefit. This, I think, in many ways demonstrates Dewey's method, which consists of taking many theories, histories, needs, etc., into consideration.
Tishler and Bartholomae, in a meta-analysis of literature of phase 1 trials, found that financial incentives were the dominant feature mentioned in recruitment materials. In a subsequent publication, Tishler and colleaguesfound that 50% of volunteers in a pilot study had a clinically significant psychopathology. This is not surprising, considering that there are no mental health screenings for healthy volunteers. Furthermore, Kass and coworkers, in the first empirical study on healthy volunteers, found that the volunteers were "disproportionately of disadvantaged social class" and that 55% were enrolled because of money. Elliott,in a narrative article in The New Yorker, cites many heart-wrenching examples of healthy volunteers in phase 1 clinical trials.
It appears that payment to subjects is common in phase 1 trials.[7,32]Such payment could cast doubt on the validity of informed consent (ie., autonomy) because of undue influence, as well as to the participants' perception of risk and benefits. This could undermine the whole concept of equipoise, if one ever existed.
This is extremely troublesome -- especially in a society in which there is so much poverty: It seems to constitute coercion. Not only will the poor necessarily be attracted by payment (and often by at least an assured bed and food), it also frees the well-to-do from volunteering. This is similar to selling your kidney in India to enable a hungry family to live and perhaps send a child to college.
Clinicians are obligated to provide therapy to individuals. Research is definitely not therapy. The concept of equipoise diverts attention from the moral conflict inherent in the physician's duty for treatment and enhances therapeutic misconception. Treating physicians are concerned with the best interest of the individual and not the community or the public good. If, as a by-product, new truth is discovered, so much the better.
On the other hand, the research physician's first obligation is to produce "truth" -- if, as a by-product the patient benefits, so much the better.
Equipoise is fragile, momentary, and subjective. It connotes ambivalence and is not scientifically valid. It is time to remove the term "equipoise" from our justification of research with human subjects -- at least in phase 1 clinical trials.
Medscape J Med. 2008;10(11):254 © 2008
Cite this: The Myth of Equipoise in Phase 1 Clinical Trials - Medscape - Nov 05, 2008.