The MIT Press
Howard Mann - Therapeutic Beneficence and Patient Recruitment in Randomized Controlled Clinical Trials - The American Journal of Bioethics 2:2 The American Journal of Bioethics 2.2 (2002) 35-36

Open Peer Commentaries

Therapeutic Beneficence and Patient Recruitment in Randomized Controlled Clinical Trials

Howard Mann, University of Utah

Therapeutic Beneficence and Enrollment in Clinical Trials

Miller and Brody (2002) offer a critique of the argument from therapeutic obligation advanced by proponents of the role of clinical equipoise in clinical trials by proposing a sharp distinction between the objectives in clinical care and in a randomized controlled trial (RCT): individualized treatment recommendations expressive of the principle of therapeutic beneficence, and accumulation of scientifically valid data, respectively. They imply that these disparate objectives cannot simultaneously be pursued in the context of an RCT and point to "conflicts between patient welfare and scientific investigation, inherent in clinical research." as a source of role-based conflicts of interest when a physician contemplates offering trial enrollment to his patient. With regard to an RCT in which clinical equipoise is determined to be present, these arguments are not convincing.

An RCT is proposed when there is reasoned uncertainty in the clinical community about the relative net therapeutic effectiveness of the interventions under consideration. Uncertainty may be manifest at three levels: the community of "expert" practitioners (this level is also relevant to an institutional review board's (IRB) decision about trial approval); the individual physician who has to decide which treatment to recommend to his patient; and the patient who is uncertain which of the trial arms is better (Sackett 2001b). For the patient, who has to make the final treatment decision, equipoise implies that the expected size and probability of health improvement balance the size and probability of harmful effects of comparator treatments. When a patient is in equipoise with respect to an approved trial, randomization offers the best odds (50%) of getting the treatment revealed to be superior at the conclusion of the trial. In this situation the "best treatment is the trial" (Ashcroft 2000). Thus, under conditions of uncertainty that commonly attend treatment decisions, offering the option of trial enrollment is clearly an example of personalized decision making in service of therapeutic beneficence.

When a physician has a preference for one of the treatments included as a trial arm, he resolves to recommend that treatment to his patient, but may acknowledge that other physicians, so situated, may be genuinely uncertain concerning the relative merits of the treatments being compared. Here he should recall that the doctrine of informed consent requires that reasonable alternatives to the proposed course of action be provided to the patient, which means information about the availability of an RCT (Marquis 1999). I share Marquis's observation that offering trial enrollment in this situation may not be morally optional. Thus, the process of informed consent is employed in service of therapeutic beneficence as the patient's values and preferences are given appropriate weight in the deliberations surrounding uncertainty about comparator treatments.

Miller and Brody state that "the process of treatment in RCTs differs radically from routine clinical practice" but offer no empirical evidence that patients treated in the context of RCTs are systematically or predictably disadvantaged when medical outcomes are considered. An RCT may be considered nothing more than randomization in service of uncertainty, and an associated process of data gathering to generate knowledge. Application of Weijer's (2000) method of risk analysis to proposed protocols will ensure that the risks of harm from any nontherapeutic components, if any are included, are stringently limited. In addition, there is accumulating evidence that patients treated in the context of trials do better, as a class, than those treated in the context of routine clinical practice (Sackett 2001a; Braunholtz, Edwards, and Lilford 2001; Vist et al. 2001). Such evidence adds further weight to the argument that trial enrollment should be considered in the calculus of therapeutic beneficence and that the notion of therapeutic obligation is applicable to clinical research in the form of RCTs.

At this point the reader may aver that the foregoing is pertinent to RCTs involving active interventions in both trial arms but not to an RCT involving the use of a placebo. However, a placebo may be utilized other than as a replacement for known effective therapy. Consider, as one example, a placebo-controlled trial evaluating an investigational agent as add-on therapy. Here the notion of equipoise at trial inception is indeed pertinent. There may be an expectation that the investigational agent is better than placebo, but this outcome is uncertain. All the participants are given an established treatment, and each participant has an even chance of also getting the agent believed to be superior. If the investigational agent is subsequently shown to be no better than placebo, the enrolled subjects randomly assigned placebo will not have been disadvantaged [End Page 35] and will also have been spared any harmful adverse effects produced by the investigational agent. Patients and physicians should better understand the utility of a randomized clinical trial as a treatment decision under conditions of uncertainty; but this will predictably engender confusion when a trial employs a placebo in place of known effective treatment.

Patient Recruitment and the Use of Placebo in Place of Effective Treatment

Miller and Brody correctly explicate the ethical concerns that arise when physician-investigators enroll in their trials individuals with whom they have an ongoing patient-physician relationship. However, further elaboration is in order. RCTs are typically multicenter trials involving the recruitment of physician-investigators for the purpose of patient enrollment. The local physicians are not involved in the process of trial conception and design. These physicians or their agents may also recruit potential participants from other local physicians' practices. Patients are not typically recruited for clinical trials by unrelated third parties but are recruited within the context of an established patient-physician relationship. When any physician or his agent offers trial enrollment to a patient so situated, cumulative ethical concerns arise as each of the following becomes applicable: a placebo is used in place of effective treatment; the trial protocol utilizes a washout period during which established therapy is withdrawn; the protocol contains potentially harmful or burdensome nontherapeutic components; and when the physician is paid a capitation fee for patient enrollment. The latter financial conflict of interest may be significant and is not addressed by Miller and Brody. In my experience it is not uncommon for all of these elements to be present in a placebo- controlled trial. Thus, physicians, and IRBs evaluating an application for research, should seriously consider the extent to which such an offer should be associated with a formal and explicit notification to be provided to the patient, clarifying that the physician offering trial enrollment is doing so as a paid scientist-investigator rather than a caregiver. This notification should accompany information concerning:

1. why a placebo is being used in place of effective therapy;

2. whether any established treatment has to be withdrawn; and

3. the possible consequences of both actions.

Grant for a moment that such a trial ought to be considered and reviewed solely according to the standards governing nontherapeutic research as Miller and Brody describe and that the involved physicians are bound solely by the principle of nonexploitation. Given the known parameters and circumstances within which patients are currently recruited for clinical trials as described above, how do the authors envisage the actual process whereby individuals for such trials are to be ethically recruited? The authors' consideration of such placebo-controlled trials as nontherapeutic research is plausible, but many vexing ethical issues related to the process of patient recruitment remain to be debated. n

Howard Mann, M.D., is a Program Associate in the Division of Medical Ethics, University of Utah School of Medicine. He has proposed A Standard for the Scientific and Ethical Review of Trials by research ethics committees (http://www.assert-statement.org).

References

Ashcroft, R. 2000. Giving medicine a fair trial: Trials should not second guess what patients want. British Medical Journal 320(7251): 1686-87.

Braunholtz, D. A., S. J. L. Edwards, and R. J. Lilford. 2001. Are randomized clinical trials good for us (in the short term)? Evidence of a "trial effect." Journal of Clinical Epidemiology 54:217-24.

Marquis, D. 1999. How to resolve an ethical dilemma concerning randomized clinical trials. New England Journal of Medicine 341(9): 691-93.

Miller, F. G., and H. Brody. 2002. What makes placebo-controlled trials unethical? American Journal of Bioethics 2(2):3-9.

Sackett, D. L. 2001a. How do the outcomes of patients treated within randomized control trials compare with those of similar patients treated outside these trials? Available from: http://hiru.mcmaster.ca/ebm/trout.

Sackett, D. L. 2001b. Uncertainty about clinical equipoise. Canadian Medical Association Journal 164(13): 1831-32.

Vist, G. E., K. B. Hagen, P. J. Devereaux, and A. D. Oxman. 2001. Outcomes of patients who participate in randomized controlled trials versus those of similar patients who do not participate (Protocol for a Cochrane Methodology Review). In The Cochrane Library, Issue 4. Oxford: Update Software.

Weijer, C. 2000 The ethical analysis of risk. Journal of Law, Medicine, and Ethics 28:344-61.



Share