Ethics Roundtable: Ethical Issues in Randomized Clinical Trials
Response by Jason Karlawish
Jason Karlawish, M.D.
University of Pennsylvania
Center for Bioethics
Uncertainty justifies starting a clinical trial. This is the condition of equipoise a state of legitimate uncertainty as to whether any arm of the trial is better than another. When there is no equipoise, the trial should not be done.
This randomized clinical trial plans to compare three methods of speech-language therapy for patients with aphasia caused by a recent stroke. The investigators justify this study because centers use different approaches to treat aphasia. The research ethics question raised by this practice variation is who is in equipoise? The "who" refers to the community that treats patients with aphasia caused by a recent stroke. When presented data that describe this practice variation, will the therapists at the centers share the investigators' uncertainty, or will they remain all the more committed to the various standards of care that they may have fought long and hard to establish at their centers?
The answer to this question is critical. The variation in the standard of care may reflect the randomness of clinical practice that occurs when no one really knows the right way to take care of a problem. In this case, the study will enroll subjects and generate data that will change clinical practice. It will be a success. But what if this variation has little to do with uncertain knowledge about what is the right treatment and more to do with limited resources and skills available to provide treatment to patients with aphasia? In this case, a randomized clinical trial is not appropriate. In between these cases, there is the condition of legitimate uncertainty upon what is the right way to treat not all patients with aphasia after a first stroke, but only certain kinds of patients.
These issues are often mixed. Investigators cannot sort them out alone. What is needed is democratic dialogue within the community who will participate in the research and use its results. Are therapists who are committed to 3 sessions of family education and a pamphlet willing to surrender their ability to recommend this therapy and instead allow patients to be randomized to one of three treatments? What kind of evidence would sway them to change their commitment to family education? Answers to questions like these will suggest the sources of legitimate uncertainties in the community and the trial designs that can settle them.
To submit cases or to be added to the list of respondents please contact: Helen Sharp Department of Speech Pathology and Audiology, 307 WJSHC University of Iowa, Iowa City, IA 52242. Phone: 319-335-6596, fax 319-335-8851, e-mail: [email protected]