Introduction

It is uncontroversial that clinical researchers should not instrumentalize patients as mere objects for study. It is assumed that any research interests involving human subjects must be weighed carefully against the interests of the subjects themselves. Out of these considerations was born a widely accepted principle of research called clinical equipoise. Put simply, clinical equipoise requires that, during a randomized control trial (RCT), there not be an obviously ‘better’ or preferable intervention given to either the control or the experimental group. The reasons for this are twofold: if one intervention is a clear winner over the other, then (1) the study itself is possibly redundant because it aims to demonstrate something which is already known and agreed upon, and more importantly, (2) the researchers are actively and knowingly depriving one of the participant groups of a more beneficial intervention or treatment. Thus, clinical equipoise aims to establish ethical patient enrollment on the grounds that neither the control nor the experimental arm of a trial is being denied a better treatment or standard of care. 

Despite general consensus on the importance of clinical equipoise, questions remain concerning how we ought to implement or interpret the stipulations of clinical equipoise. How can or ought we determine whether one treatment is better than another for the purpose of designing studies? Do we poll clinical researchers? What level of consensus is sufficient to sanction the study, or what level of dissent to bar it? What if decisive evidence exists, but is not available or widely known? And finally, there is the inherent tension in claiming belief in the null hypothesis (that equipoise is present and neither arm superior to the other) while having set out for the express purpose of investigating or demonstrating the opposite (that equipoise is not present and one of the arms is, in fact, superior). How we answer these questions will determine our research ethics norms and how permissive our regulatory bodies are when it comes to trial design and implementation. 

Assigned Readings

Goldsby, Michael and Kabasenche, William, “Uncertainty, Bias, and Equipoise: A New (Old) Approach to the Ethics of Clinical Research” Theoretical and Applied Ethics 3(1) (2014): 35-59. 

Thesis: Michael Goldsby and William Kabasenche follow in the footsteps of Benjamin Freedman’s foundational work on equipoise, which introduced two fundamental approaches to thinking about equipoise: clinical and theoretical. Clinical equipoise, defined by Freedman, represents “a situation in which there exists (or is pending) an honest disagreement in the expert clinical community regarding the comparative merits of two or more forms of treatment for a given condition” (Freedman, 1992). Theoretical equipoise, by contrast, represents “A situation in which there is insufficient evidence for the individual clinician-researcher to determine with reasonable confidence whether one arm of a trial is better than the other” (Goldsby and Kabasenche, 2014). The authors reject Freedman’s conclusion that clinical equipoise is preferable to theoretical equipoise and offer a third, novel definition which builds on theoretical equipoise called epistemic equipoise. They define epistemic equipoise as “a state of intolerable uncertainty about whether one of the two competing treatments under consideration is better for a particular patient” (Goldsby and Kabasenche, 2014). 

Discussion Questions

  1. Do you find Goldsby and Kabasenche’s proposal of epistemic equipoise a convincing alternative? Is it clear, conceptually? Does it seem practical and actionable? How might an IRB apply epistemic equipoise to their evaluation? 
  1. Goldsby and Kabasenche reject clinical equipoise on several grounds, one of them being that it remains unclear exactly how much dissent or contention there need be within the medical community to generate a state of clinical equipoise, but their own epistemic equipoise relies on gauging what a state of “intolerable uncertainty” is, a question they leave open. What distinctions might favor intolerable uncertainty as the relevant metric or conceptual anchor? Or are there no relevant distinctions that make intolerable uncertainty in epistemic equipoise an improvement on the ambiguity plaguing what constitutes clinical equipoise? 
  1. The authors point out that two major objections to theoretical equipoise are the problem of clinical complexity and the problem of fragility/accumulating data/the knife’s edge problem. What do you think of these problems? Are they irresolvable for the theoretical equipoise position? How ethically hazardous are they if unaddressed? 
  1. Goldsby and Kabasenche assert that epistemic equipoise can obtain by either lack of evidence or the possibility of bias in evaluating the evidence. Does the latter condition capture too much? Does it imply that an investigator is always in (or should always be in) a state of equipoise given that at least some marginal bias is always possible (or likely)? 

Additional Assorted Readings

Freedman, Benjamin. “Equipoise and the Ethics of Clinical Research.” New England Journal of Medicine 317, no. 3 (July 16, 1987): 141–45. https://doi.org/10.1056/NEJM198707163170304