Equipoise Trumps Randomization

摘要

Ensuring the adequacy of informed consent is a cornerstone of human subjects’ protection in clinical research. A reasonable axiom regarding what must be disclosed and what ought to be understood relative to informed consent suggests that such information as is relevant to one’s ability to make a decision about participation is required. Wendler (2009) has examined whether understanding randomization is a prerequisite to valid informed consent for participation in a randomized clinical trial. Clearly, data suggest that randomization is not well understood by research subjects (Featherstone and Donovan 1998). However, the failure to grasp the implications of randomization, as explored by Wendler (2009), ultimately reflects the therapeutic misconception (Appelbaum et al. 1987). Subjects in clinical trials must understand that they are participants in a research protocol, and the nature of the research question (i.e., safety or efficacy), and what (if anything) is being compared. They also must understand the risks of the various treatments and the safeguards that are in place to protect them. However, in a randomized clinical trial, it is far less important that subjects understand randomization than that they understand there is a legitimate research issue in play. Arguably, it is less important that subjects understand the specific question than that they are assured, to whatever extent they can appreciate, that a condition of equipoise exists. In this respect, it is far more important that the investigator is convinced that a condition of equipoise exists. Failing that, it may be difficult to justify the investigator’s participation in the research protocol, i.e., their recruitment of research subjects. It is important to consider how clinicians become investigators, and how patients become research subjects, and the nature of the research question, as these issues have far greater impact on the likelihood of the therapeutic misconception than does the subject’s appreciation of randomization.