The ethical requirement of favorable risk-benefit ratio involves minimizing risks, enhancing potential benefits, and ensuring that the risks to the subjects are justified by the potential benefits to the subjects and/or society.18 Compared to the vast literature on informed consent, the scholarship on risk-benefit analysis in research ethics is comparatively thin.38,39 IRBs, however, tend to see a review of the risk-benefit ratio as one of the most important things they do.40,41
A persistent controversy exists in interpreting the federal requirement that the risks must be reasonable in relation to anticipated benefits, either to the subject directly or to society in the form of increased knowledge (45CFR46.111a). There are at least three relevant issues for neurotherapeutics research in this regard. First, when the research subjects lack decision-making capacity, it is not clear whether and how such risks can be balanced against the gain in scientific knowledge and benefit to future patients, i.e., when the protocol has no potential for direct benefit to the subjects or when the protocol is the very first human experiment whose main aim is to test the safety of the new intervention.2,42 Such persons need third party (“legally authorized representative”) permission for participation in research and usually it is not known what the subject would have wanted. Third party consent in the medical treatment context is routine because there is a long-standing tradition of accepting such permission given that the decisions will be made in the best interests of the patients (if the principal’s wishes are not known). In a research protocol with no potential for direct benefit to the subjects, however, the issue is how much risk and burden to the incapable subjects can be reasonably justified by the potential increase in knowledge (that will presumably benefit other patients). In research with children, the federal regulations impose the restriction that children may not be exposed to more than minor increase over minimal risk, when the research involves an intervention or procedure “that does not hold out the prospect of direct benefit” to the subjects (45CFR46.406). No analogous policy exists for research with incapacitated adults.43
Second, even for situations in which the subjects in neurotherapeutics research do have decision-making capacity, there have been prominent debates regarding whether certain types of research should be conducted at all.44–46 In randomized neurosurgical trials of tissue transplants for Parkinson’s disease, the sham control subjects were exposed to procedures that carry risk but without potential for direct benefit. Such a control condition, however, may be invaluable in controlling for the placebo effect and the natural fluctuations in PD symptoms.46 Because a false positive result from a clinical trial can lead to a large number of unnecessary and even harmful surgeries, the gain in scientific knowledge (assuming that the sham control condition has such incremental knowledge advantage over open controls) could be substantial and not merely academic. Does this gain justify exposing such subjects to nonbeneficial procedures that carry risks?
Even if we accept the current practice of sham controls for PD research, there are future possibilities that raise thornier questions. For example, if a randomized controlled trial of gene transfer therapy for PD is proposed, what is the appropriate control condition? The range of possibilities include drilling of burr holes without interrupting the dura, adding probe manipulation into the striatum (or whichever structure is targeted for gene transfer), adding injection of the same vector medium without the vector, injecting an “empty” vector, and so on, even to the addition of a “placebo” gene that will insert into the genome but, presumably, will not produce the gene product in question.
Third, what is the best way of balancing risks and benefits when thinking about early phase research? Is it appropriate to conduct human studies based only on rodent data?47 Most phase I studies are conducted in healthy volunteers. The most well established paradigm for phase I studies involving patients (rather than healthy volunteers) is in cancer dose-escalation protocols.48 That paradigm does allow a relatively separate evaluation of safety and efficacy. It does not however capture well the risk-benefit evaluation of interventions such as gene transfer or tissue transplants. For instance, if in a phase I trial of gene transfer protocol there were absolutely no suggestion of benefit from the intervention, is it the same situation as finding no suggestion of efficacy in a cancer phase I trial?
Finally, once an effective treatment is found, the testing of newer and, hopefully, better interventions can become ethically problematic, if the scientifically optimal design is a placebo control (for example, if the outcome variable is known to be susceptible to the placebo effect and tends to have wide natural fluctuations). Withholding of effective treatments for research purposes can become an issue in another mode, because the notion of neuroprotection is particularly relevant to neurodegenerative disorders and the testing of an effective agent for its putative neuroprotective effects will involve withholding of such a drug.
What kind of evidence is needed to aid policy in this area? It is worth reiterating that the most ethically relevant evidence is accurate, current scientific data. However, ultimately, what constitutes an acceptable risk-benefit tradeoff is a matter of social policy, not just of scientific expertise. Ideally, such policy should be informed by the views of the stakeholders. The most established method for eliciting such views is the survey. Surveys of potential research participants can give useful data that is distinct from any conclusions that can be drawn from the fact that patients do in fact enroll in research studies.49 Surveys of IRB members or scientists and scientific administrators regarding how they apply the risk-benefit analysis regulations can also illuminate how policies may in fact get applied.50 Surveys of the general public on the appropriateness of certain implied tradeoffs in research protocols can also inform the debate.
In general, however, the methodology for eliciting public opinion on complex policy questions is not well developed in bioethics research. Although surveys are useful, they provide a rather thin description. Many bioethics policy issues are not everyday issues for which we would expect informed or even formed opinions in lay persons. In this regard, bioethics needs to borrow methods from other social science disciplines that aim to elicit views that are richer than surveys and yet more generalizable than focus groups. Some social scientists now aim to elicit “constructed preferences” using methods that do not assume preformed ethical opinions.51,52 Political scientists have long discussed the notion of deliberative democracy53,54 in which policies are developed or informed by citizens deliberating together guided by certain normative principles of deliberation.54 This tradition has developed a well-articulated theory about how to address ethical controversies in a liberal democracy.55 Empirical research methods using the deliberative democracy framework are now beginning to be used in health care policy.53 The basic idea is that a group of representative citizens (or from the ethically relevant stakeholder groups) are brought together to engage in in-depth education, discussion, and argumentation (in the sense of giving and receiving varying viewpoints and reasons for and against various policies). Not only individual but group level outcomes can be measured.