Republished by the LSU Medical Reseach Law Project

The Common Rule directs IRBs to ensure that research risks are minimized through careful study design and that risks are "reasonable in relation to anticipated benefits, if any, to subjects, and the importance of the knowledge that may reasonably be expected to result."177 Many commentators favor placing additional constraints on acceptable risks in research involving persons who, as a result of having certain mental disorders, may sometimes lack decisionmaking capacity.

This chapter discusses some of the conceptual and practical problems that arise not only for IRBs, but also for investigators and potential subjects who must make judgments about the acceptability of risk in relation to the prospect of benefit. It first discusses some of the difficulties inherent in defining risk, followed by an explanation of NBAC’s rationale for urging IRBs to evaluate research protocols involving this population under two categories: minimal risk and greater than minimal risk. Next, it discusses some of the difficulties in defining benefits. Finally, it comments on the difficulties of assessing research risks in relation to potential benefits. In particular, this discussion focuses on the protections that should be required for research involving greater than minimal risk that holds out the possibility of direct medical benefit to subjects, and for research involving greater than minimal risk that does not hold out the possibility of direct medical benefit to subjects. The final section of this chapter also proposes procedures to minimize risks to subjects.

The concept of risk is generally understood to refer to the combination of the probability and magnitude of some future harm. According to this understanding, risks are considered "high" or "low" depending on whether they are more (or less) likely to occur, and whether the harm is more (or less) serious. In research involving human subjects, risk is a central organizing principle, a filter through which protocols must pass; research evaluated by IRBs that presents greater risks to potential research subjects will be expected to include greater or more comprehensive protections designed to reduce the possibility of harm occurring. The ethical basis for this position was usefully summarized in the National Commission’s Belmont Report: "The requirement that research be justified on the basis of a favorable risk/benefit assessment bears a close relation to the principle of beneficence, just as the moral requirement that informed consent be obtained is derived primarily from the principle of respect for persons."178 In contrast, relatively little progress has been made in describing the criteria for assessing risk by IRBs.179 In large part, this is due to the multiple difficulties inherent in classifying risk judgments, including the difficulty associated with risk perception in general,180 and other aspects of objectively quantifying risk.181

The purpose of having multiple categories of risk is to trigger different requirements from IRBs, just as the "minimal" and "greater than minimal" risk categories trigger different types of minimal protections in the Common Rule. The Common Rule does not specify that IRBs use three categories of risk in making judgments about the acceptability of risks in relation to potential benefits, nor do the regulations specific to pregnant women or prisoners specify that IRBs use three categories of risk.182 Only the regulations pertaining specifically to children describe three categories of risk.183 Indeed, IRBs are free to use as many categories of risk as they deem appropriate in assessing risk in relation to potential benefit. The Common Rule categories are only for the purposes of establishing minimum protections. NBAC recommends that IRBs use their existing authority to determine whether to add protections above the minimal regulatory requirements for all research involving greater than minimal risk.

According to the Common Rule, a study presents minimal risk if "the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests."184 Although the concept of minimal risk remains controversial in academic and scholarly discussion, it is widely used to determine which set of protections are to be required for particular research protocols. For example, when a research protocol is determined to involve minimal risk, IRBs are given the latitude to waive certain consent requirements, so long as certain conditions are met. Moreover, IRBs may expedite the review of research protocols—relying on the review by a single member, such as the chair—that typically (though not exclusively) involve minimal risk.185 Still, the application of this concept can be difficult in practice. For example, a "typical" minimal risk encountered in everyday life or in clinical care may be perceived differently by some individuals with certain disorders. It is important, therefore, to establish a practical level of minimal risk against which IRBs can measure proposed research protocols in order to decide which protocols require additional protections. The level of minimal risk will change (in one direction or another) over time, as experience and additional knowledge alter the way the research community, IRBs, and research subjects perceive the acceptability of various research risks. Under the current system, IRBs have complete discretion to apply none or only some of the added protections to protocols that they believe to be of greater than minimal risk. Indeed, they are entitled to add additional protections for protocols involving minimal risk as well.

The DHHS addressed the issue of IRB latitude in its regulations on research involving children by permitting IRBs to approve research presenting no greater than minimal risk as long as requirements for parental permission and child assent are satisfied. Even in this case, an IRB could add further protections if it thought it was appropriate to do so. However, the regulations stipulate that studies presenting greater than minimal risk must meet additional requirements.

Like the DHHS regulations for children, many proposals on research involving impaired or incapable adults employ the concepts of minimal risk and minor increase over minimal risk. Indeed, many public comments suggested that NBAC group research protocols involving persons with mental disorders into three categories of risk: (1) minimal risk; (2) minor increase over minimal risk; and (3) greater than minimal risk (which encompasses risks greater than a minor increase over minimal risk). The ostensible purpose of this tripartite division is to allow protocols involving only a minor increase over minimal risks to proceed with only minimal additional protections. Three categories of risk, it has been argued, provide IRBs with more flexibility in requiring certain protections. Two categories of risk, it has been suggested, would prevent certain protocols from going forward since IRBs may believe that the additional protections would effectively bar research involving greater than minimal risk without the prospect for direct medical benefit. As evidence of this alleged difficulty, NBAC received correspondence from NIH describing examples of research that might be limited by retaining the Common Rule’s two-level categorization of risk. These examples included: affected sib-pair genetic linkage studies in Alzheimer’s disease, biological measures for the purpose of treatment response prediction, neuroimaging studies in developmental disorders, studies to improve understanding of Alzheimer’s disease, studies to improve diagnosis of Alzheimer’s disease, study of urinary incontinence in severely demented nursing home residents, studies of delirium in older patients, and studies of aphasia. In several of these examples, research was considered to involve minimal risk or a slight increment above minimal risk. NBAC did not find these concerns convincing. As explained above, the key point is that IRBs should focus on the need for a continuous range of protections that are related to the perceived level of risk, and whether there are two or more levels should make little difference. NBAC’s view is that IRBs should perceive their responsibility as requiring, where necessary, a set of additional protections beyond those minimally required, from a continuous spectrum that will be governed by the particular nature of the proposed protocol. In short, NBAC is not persuaded that three categories of risk are necessary for accomplishing the twin goals of providing protection for persons with mental disorders while allowing important research to go forward. At least one professional organization, the Alzheimer’s Association, has adopted as official policy the view that the three categories of risk are not necessary.186 

As noted above, providing clear meaning to these concepts, poses serious practical difficulties. The Common Rule’s minimal risk definition, which refers to the risks of everyday life and medical care encountered by the population as a whole, often is praised for its flexibility: "It is inescapable and even desirable that determinations of risk level (and its acceptability when balanced with benefit consideration) are matters of judgment rather than detailed definition, judgments which are patient-specific, context-specific, and confirmed after consideration and debate from many points of view."187 The concept’s reference to "risks of everyday life" also conveys a defensible normative judgment that the sorts of risks society deems acceptable in other contexts may be acceptable in research as well.188

In contrast to the minimal risk concept’s reference to the life and medical experiences of the overall population, the concept of minor increase over minimal risk is, in the case of children, tied to the prospective subject’s individual situation. Because persons with mental disorders often undergo treatment and tests involving some discomfort and risk, a study presenting similar procedures and potential for harm may qualify as presenting a minor increase over minimal risk to them.189 For subjects not accustomed to or in need of such medical interventions, however, the same study could present a higher level of risk.

In its Report on Research Involving Children, the National Commission defended this approach to greater than minimal risk research on grounds that it permitted no child to be exposed to a significant threat of harm. Further, the National Commission noted that the approach simply permits children with health conditions to be exposed in research to experiences that for them are normal due to the medical and other procedures necessary to address their health problems. An example is venipuncture, which may be more stressful for healthy children than for sick children who may be more accustomed to the procedure.

Commentators have criticized both the Common Rule’s "minimal risk" definition and the category "minor increase over minimal risk" in the children’s regulations. Loretta Kopelman provides perhaps the most detailed critique. First, she finds the notion of "risks of everyday life" too vague to provide a meaningful comparison point for research risks. Ordinary life is filled with a variety of dangers, she notes, but "[d]o we know the nature, probability, and magnitude of these ‘everyday’ hazards well enough to serve as a baseline to estimate research risk?" Second, though the comparison to routine medical care furnishes helpful guidance regarding minimal risk, it fails to clarify whether procedures such as "X rays, bronchoscopy, spinal taps, or cardiac puncture," which clearly are not part of routine medical care, could qualify as presenting a minor increase over minimal risk for children whose health problems dictate that they must undergo these risky and burdensome procedures in the clinical setting. Kopelman argues that the phrase "minor increase over minimal risk" should be replaced or supplemented by a clearly defined upper limit on the risk IRBs may approve for any child subject.190

Difficulties with the minimal risk standard may be due in part to a historical confusion. Some contend that the drafters of the definition of minimal risk deliberately dropped the National Commission’s reference to normal individuals, intending to make the relevant comparison to risks ordinarily encountered by the prospective research subject. This approach would allow classifying research risks as minimal if they were reasonably equivalent to those the subject encountered in his or her ordinary life or routine medical care. Using this approach with persons with mental disorders who face higher-than-average risks in everyday life and clinical care, a research intervention could be classified as minimal risk for them, but classified as greater than minimal risk for healthy persons. If this was the intention of the drafters of the regulations, it is not at all clear in the current Common Rule.

In August 1998, the major federal funding agencies in Canada developed a policy statement on "Ethical Conduct for Research Involving Humans" that explicitly adopts the standard of relativizing risk to the potential subject in question. It defines "normally acceptable risk" as "when the possible harms (e.g., physical, psychological, social, and economic) implied by participation in the research are no greater than those encountered by the subject in those aspects of his or her everyday life. . . ."191 The Canadian policy statement goes on to insist that therapeutic risks should be treated differently from nontherapeutic risks. Therapeutic risks can be considered as minimal for patient-subjects, since they are inherent in therapy and thus the everyday life of the subject.

In some cases, procedures presenting greater than minimal risks to people with mental disorders might be treated as such, while in other cases (e.g., in persons with special vulnerability to those procedures) they might not be. A procedure classified as minimal risk at one institution could be classified as higher risk at another, or even from one study to another in the same institution. Also needed is further clarification of acceptable risk in research involving incapable adults whose ongoing health problems expose them to risks in their everyday clinical setting. Because some persons with mental disorders who are accustomed to certain procedures may experience fewer burdens when undergoing them for research purposes, some would argue that it may be defensible to classify the risks to them as lower than would be the case for someone unfamiliar with the procedures.

We must guard against assumptions like these. The psychological context of illness may well make some research procedures, however familiar, more burdensome than they would be to someone who enjoys good health. These procedures must not be classified as lower risk for subjects who have had the misfortune of enduring them in the treatment setting.192 Like the level of minimal risk, the boundaries that separate particular risk categories can be expected to shift over time in response to many complex and interrelated factors. What is required is a focus on the "package" of reasonably interpreted risks, on the one hand, and a correspondingly appropriate set of protections, on the other.

One way to reduce variation in risk classification would be to provide examples of studies that ordinarily would be expected to present a certain level of risk to members of a certain research population. For example, the Maryland draft legislation includes in its definition of "minimal risk" research those "types of research that are . . . identified by the United States Department of Health and Human Services as suitable for expedited IRB review."193 The Maryland proposal effectively incorporates examples like venipuncture, magnetic resonance imaging (MRI), electroencephalography, and the study of existing biological specimens. This is consistent with federal regulations; however, it should be noted that while current federal regulations permit studies involving MRI to be reviewed on an expedited basis, this does not always imply that such studies always involve minimal risk.

Perhaps over time, if there is adequate communication and disclosure, it will become evident to the IRB community that protocols tend to cluster in certain ways. For example, one author proposes that lumbar punctures and PET "can be reasonably viewed as having greater than minimal risk for persons with dementia because (1) both procedures are invasive, (2) both carry the risk of pain and discomfort during and after, and (3) complications from either procedure can require surgery to correct."194 The Maryland draft legislation designates research as presenting more than a minor increase over minimal risk if, as a result of research participation, the subjects would be exposed to more than a remote possibility of "substantial or prolonged pain, discomfort, or distress" or "clinically significant deterioration of a medical condition."195

One proposed list of minimal risk procedures for dementia patients includes "routine observation, data collection, answering a questionnaire, epidemiological surveys, venipuncture, and blood sampling," as well as neuropsychological testing.196 Though some reportedly classify lumbar punctures and bone marrow biopsies as presenting a minor increase over minimal risk, Keyserlingk and colleagues suggest that such procedures may present "greater risks for some patients with dementia who are unable to understand or tolerate the pain or discomfort" accompanying the interventions.197

NBAC’s review of a series of research protocols turned up a good example of an IRB that sought expert assistance in assessing risks. The protocol involved a challenge study which entailed a higher than standard dosage of the challenge agent, although the investigator described the study as minimal risk in the consent form. The expert evidently advised the IRB that the risks were in fact greater than minimal due to the increased dosage and that the dosage should be reduced and properly identified in the consent form. An IRB that seeks expert opinion, where necessary, can dramatically improve both research design and the bases for subjects to provide informed consent.

The debate about the meaning of minimal risk will surely persist because of the philosophical and practical difficulties of defining it precisely. But this does not mean that research involving persons with mental disorders cannot be conducted. Rather, it means that research procedures that would entail minimal risk for a general population must be assessed in light of the specific research population. In no case, however, should procedures classified as greater than minimal risk for the overall population be classified as minimal risk for this population. Therefore, research proposals should be more highly scrutinized if they involve persons with mental disorders, and special care may be required to understand particular risk levels for this population. NBAC believes that these special considerations are important and should not prevent the most valuable research from continuing within such constraints.

Strictly speaking, risk assessment is a technique used to determine the nature, likelihood, and acceptability of the risks of harm.198 In actual practice there is always a great deal of controversy about how such assessments should occur. Moreover, few IRBs conduct formal risk assessments, and there may be good reasons for this: First, reliable information about risks or potential benefits associated with the relevant alternative interventions is often lacking. As a result, highly accurate risk assessment is difficult and in many cases impossible. Second, each component of risk assessment—identification, estimation, and evaluation—involves time and requires particular kinds of expertise.199 Even at the conceptual level, it is a matter of both scientific and philosophical debate as to whether risk assessment should involve purely objective or purely subjective factors (or both). The "objectivist" school argues that quantitative risk assessment should be a value-free determination limited only by the technical ability to derive probability estimates.200 In contrast, the "subjectivist" school argues that the values of those who conduct the assessment, those who interpret the results, and those who bear the risks should play a role in the overall assessment of risks.201 It is reasonable to hold that both schools of thought ought to influence IRB decision making, the former because risk judgments should be empirically based insofar as possible, and the latter because many who have an interest in research can contribute to these assessments despite the lack of formal quantitative data.

The National Commission’s Report on Research Involving Children advised IRBs to assess risks from the following points of view: "a common-sense estimation of the risk; an estimation based upon investigators’ experience with similar interventions or procedures; any statistical information that is available regarding such interventions or procedures; and the situation of the proposed subjects."202 Evaluating risks to subjects with mental disorders requires familiarity with how such subjects may respond, both generally and individually, to proposed research interventions and procedures. What may be a small inconvenience to ordinary persons may be highly disturbing to those with decisional impairments. Thus, for example, a diversion in routine can, for some dementia patients, "constitute real threats to needed order and stability, contribute to already high levels of frustration and confusion, or result in a variety of health complications."203 Similarly, as the National Commission observed, some subjects institutionalized as mentally infirm may "react more severely than normal persons" to routine medical or psychological examinations.204 Because of the special vulnerability to harm and discomfort that particular subjects may have, risk assessment should anticipate the range of reactions subjects may experience to certain proposed study procedures. Difficult as it may be, careful risk assessment is the key to deciding on the appropriate level of protections.

Research involving adult subjects can yield three types of potential benefit: (1) direct medical benefit to subjects; (2) indirect benefit to subjects; and (3) benefit to others.

Particular research protocols may hold out the prospect of direct medical benefit to the subjects themselves, even though such benefit can never be assured. The potential direct medical benefits to the subjects include health improvements which may or may not be related to the disorder responsible for the subject’s incapacity.205 For example, the National Commission stated that research offering potential direct benefits to persons institutionalized as mentally infirm could include:

studies to improve existing methods of biomedical or behavioral therapy, or to develop new educational or training methods. The studies may evaluate somatic or behavioral therapies, such as research designed to determine differential responsiveness to a particular drug therapy, or to match particular clients with the most effective treatment. Studies may also assess the efficacy of techniques for remedial education, job training, elimination of self-destructive and endangering behaviors, and teaching of personal hygiene and social skills.206

According to the National Commission, "[t]o be considered ‘direct,’ the possibility of benefit to the subject must be fairly immediate [and t]he expectation of success should be well-founded scientifically…"207 A more recent statement on dementia research limits direct medical benefit to:

a short- or long-range improvement, or a slowing of a degenerative process, in the specific medical condition of the relevant subject, whether in the patient’s condition of dementia, a medical symptom associated with dementia, or another physical or mental condition unrelated to dementia. Such direct benefits include those resulting from diagnostic and preventative measures.208

IRBs and other reviewers should carefully scrutinize investigators’ assertions that research offers the prospect of direct medical benefit to subjects. Furthermore, the protocols reviewed by NBAC reflected some confusion about the definition of direct medical benefit. One protocol referred to the challenge procedure as the "treatment phase." The consent form that accompanied this protocol included in the benefits of the assessment phase "a thorough psychological evaluation at no cost, the results of which will be the basis for a treatment recommendation either within or outside of the treatment phase of the study. Benefits of the treatment phase may include decreases in the . . . severity of . . . symptoms." Unless the distinctions between direct medical and indirect benefits are identified, and their relative significance explored carefully, there is a danger that investigators may construe the concept of direct medical benefit too broadly.209

Finally, potential direct medical benefits to the subjects participating in the research protocol not only must be carefully evaluated but may not, by themselves, justify experimental interventions that present risks to a subject population. Instead, these possible benefits must be considered in relation to the risks involved. Even though a research protocol may offer potential direct medical benefits to individual participants, it cannot be justified by the possibility of that benefit alone.

Indirect Benefit

Subjects may obtain other forms of benefit from research participation. As the National Commission noted, "[e]ven in research not involving procedures designed to provide direct benefit to the health or well-being of the research subjects . . . there may be incidental or indirect benefits."210 Examples of indirect benefits are "diversion from routine, the opportunity to meet with other people and to feel useful and helpful, or . . . greater access provided to professional care and support."211 NBAC agrees with the view expressed by one group that an indirect benefit may be acknowledged, but should not be assigned as heavy a weight as direct medical benefit in the IRB review and discussions with prospective subjects and their representatives.212

There is a continuing debate about whether the reimbursement subjects receive for their time and inconvenience constitutes a direct or indirect benefit of research participation. The benefits of financial incentives for the subject are indirect in the strict sense that they do not stem from the research interventions themselves, but the subject may view them as very important. A secondary concern here, as with research on other potentially vulnerable populations, is who actually receives and controls the funds: the subject or a third party who authorizes research participation?

The principle that financial incentives should not exceed "reimbursement" for the subject’s time and expenses, so as not to establish undue motivation to participate, is well established but not always easy to apply. The problem is complex because both healthy volunteers as well as some who are ill may agree, for example, to pharmaceutical testing as an important supplement to their income (if not their sole income source) as their main reason for participating. Remuneration must be appropriate to justify their commitment of time and their submission to discomfort, but not be so great as to lead them to take unreasonable risks. Similarly, some who are suffering from an illness, especially those who are uninsured, may be tempted to join a study if it appears that the ancillary medical care will be superior to what they can otherwise obtain.

This category encompasses benefit to subjects’ families or other caregivers, to persons with the same disorder as subjects, and to persons who will suffer from the same disorder in the future. When such research is invasive and presents no realistic possibility of direct health benefit to the subject, it "poses in the most dramatic form the conflict between the societal interest in the conduct of important and promising research and the interests of the potential subject."213 Thus, it is necessary to examine carefully how risk and potential benefits—especially to others—can be balanced; and what protections should be in place to minimize risk to subjects with mental disorders that may affect decisionmaking capacity.

The National Commission was fully aware of the problems inherent in making risk-benefit assessments when it wrote that:

It is commonly said that the benefits and risks must be "balanced" and shown to be "in a favorable ratio." The metaphorical character of these terms draws attention to the difficulty in making precise judgments. Only on rare occasions will quantitative techniques be available for the scrutiny of research protocols. However, the idea of systematic, nonarbitrary analysis of risks and benefits should be emulated insofar as possible.214

This chapter has described some of the difficulties with defining risks and benefits in research; the following describes the difficulties with evaluating their relationship to each other in order for IRBs, as required by current regulations, to assess the ratio of risks to benefits involved in individual research protocols. Most researchers and IRBs take the position that adults who lack decisionmaking capacity may be involved in studies presenting little or no risk, as long as requirements for third party consent are met and the research protocol offers a reasonable prospect of advancing knowledge or benefiting the subject, or both. There is substantial support, however, for adopting additional restrictions and review requirements for studies presenting higher risk, particularly for higher-risk studies that fail to offer subjects a reasonable prospect of direct benefit.215

Research presenting greater than minimal risk to subjects is generally classified into one of two categories. The first category is research offering subjects the prospect of direct medical benefit. The second category is research that is not designed to offer the prospect of direct medical benefit to subjects. Although these categories may seem to imply a distinction between "therapeutic" and "nontherapeutic," that is not the case and, in fact, is a serious misconception, particularly if studies are described as beneficial on the basis of a therapeutic component. Rather, it should be understood that some research may hold out the prospect of direct medical benefit for some individuals while some research may not.

NBAC recognizes that there are both practical and philosophical difficulties with applying the concept of minimal risk. However, NBAC’s view is that research protocols that involve minimal risk present fewer ethical issues for IRBs, researchers, and potential subjects than those protocols involving greater than minimal risk, where, in the latter case, the harms or discomforts may be expected to occur more often or are more serious. As NBAC recommends in Chapter 5, an IRB may approve protocols that present only minimal risk using several different routes for subject enrollment: when the potential subject gives informed consent (unless consent is waived); if the potential subject gives permission in advance (which NBAC describes more fully below as "Prospective Authorization") and a surrogate decision maker has given permission; or if a surrogate decision maker has given permission under certain conditions.

The general view is that it is permissible to include persons with mental disorders that may affect decisionmaking capacity in a research protocol that involves greater than minimal risk, but does not offer the prospect of direct medical benefit to subjects as long as the research presents a balance of risks and expected direct benefits similar to those available in the normal clinical setting.216 The ACP guidelines allow surrogates to consent to research involving incapable subjects only "if the net additional risks of participation (including the risk of foregoing standard treatment, if any exists) are not substantially greater than the risks of standard treatment (or of no treatment, if none exists)." In addition, they suggest that there should be "scientific evidence to indicate that the proposed treatment is reasonably likely to provide substantially greater benefit than standard treatment (or no treatment, if none exists)."217

In a similar vein, the Maryland draft legislation deems "research involving direct medical benefit" permissible if an agent or family member or friend acting as surrogate, or an IRB-designated proxy, "after taking into account . . . treatment alternatives outside of the research . . . concludes that participation in the research is in the individual’s medical best interest."218 Likewise, with the permission of an individual holding a DPA, or court-appointed family guardian, the NIH Clinical Center permits greater than minimal risk research offering a prospect of direct subject benefit if there was an ethics consultation to ensure that the third-party decision maker understands the relevant information. For subjects without a DPA or court-appointed guardian, this form of research is permitted "if the situation is a medical emergency, when a physician may give therapy, including experimental therapy, if in the physician’s judgment it is necessary to protect the life or health of the patient."219

The ACP and other groups take the position that greater than minimal risk research offering incapable subjects no reasonable prospect of direct medical benefit should be permitted only when authorized by a research advance directive220 or after review and approval at the national level, through a process resembling that set forth in the current regulations governing research involving children.221 The National Commission also recommended a national review process for studies that could not be approved under its other recommendations on research involving persons institutionalized as mentally infirm.222 However, others see this position as either too liberal or too restrictive. In NBAC’s view, the proposal of national review has considerable merit, and is discussed below and in Chapter 5.

On the one hand, based on the Nuremberg Code’s and the Declaration of Helsinki’s convictions that vulnerable unconsenting individuals should not be put at undue risk for the sake of patient groups or society, some favor an absolute prohibition on moderate- or high-risk research offering no benefit to subjects but great promise of benefit to others. Supporters of this position contend that when these documents were created, "[i]t was presumably well understood that a price of that prohibition would be that some important research could not proceed, some research answers would be delayed, and some promising therapies and preventive measures would, for the time being, remain untested and unavailable."223 Some explicitly label this stance the most ethically defensible position.224

On the other hand, a position paper representing federally funded Alzheimer’s disease centers adopts a somewhat different view: "Research that involves potential risks and no direct benefit to subjects may be justified if the anticipated knowledge is vital and the research protocol is likely to generate such knowledge."225 This group also believes that a national review process is not necessarily the best way to decide whether to permit research presenting no potential direct benefit and greater than minimal risk to incapable subjects. While acknowledging that "there may be some advantages" to national review, but contends that "immediate and direct monitoring of such research and on-site assurance of its humane ethical conduct are at least as important as the process of evaluation and approval of any proposed research."226

The regulations governing research involving children subjects provide for a special review process to address studies that offer the subjects no prospect of direct benefit and that would pose greater than a minor increment over minimal risk. The process begins by requiring that, an IRB determine that a study in this category "presents a reasonable opportunity to further the understanding, prevention, or alleviation of a serious problem affecting the health or welfare of children."227 Upon such a finding, the Secretary may convene a panel of experts in pertinent disciplines to review the study and should provide the public an opportunity to review and comment on the study.228 After the panel’s review and the public comment period, the study may be approved if the Secretary has determined: (1) that the study actually falls into a category of research that the IRB could have approved on its own,229 or (2) that the research does present the "opportunity to further the understanding, prevention, or alleviation of a serious problem affecting the health or welfare of children, that the research will be conducted in accordance with sound ethical principles, [and] that adequate provisions are made for soliciting the assent of [the] children and the permission of their parents or guardians."230

This type of process, if modified, offers an additional route for assessing some protocols involving persons with mental disorders. In NBAC’s view, however, a more flexible and accessible process is required. It would be appropriate for the Secretary of DHHS to establish a Special Standing Panel (SSP) which would have the authority to review and approve particular protocols, involving greater than minimal risk and do not offer the prospect of direct medical benefit to potential subjects, that could not otherwise be approved by IRBs. These protocols would be forwarded to the SSP by the local IRB. The SSP, which is described more fully in Chapter 5, would also have the authority to establish guidelines over time for delegating to local IRBs the authority to approve certain types of protocols in this arena.

NBAC urges the Secretary of DHHS, when constituting the SSP, to be mindful of the reports NBAC has received from the research community asserting that a significant amount of important research may fall into this category and that medical progress for many suffering persons may very well depend upon the ability of IRBs to approve appropriate protocols in this area. In designing the SSP, the Secretary also should take into account the experience of the current panel mechanism identified in 45 CFR 46.407(b). In particular, NBAC believes that it is essential for the SSP to conduct its work in a timely and efficient manner. Although the mechanism for special cases in research with children has been used only twice, NBAC believes that a similar but modified mechanism could fulfill its intended purpose in research involving persons with mental disorders that may affect decisionmaking capacity.

Therefore, NBAC strongly encourages the Secretary to ensure sufficient administrative support of, and financing for, the SSP and to adopt procedures for the operation of the panel such that will: (a) make case-by-case decisions using a process that is easily accessible to IRBs; and (b) develop as appropriate, and issue, guidelines for categories of research that then can be used by IRBs without having to utilize the SSP. In addition, the process should include public participation and a public written decision that explains the SSP’s rationale for each of its decisions. The system of review recommended here will serve several important functions. Most important to NBAC is that it will increase protections for a subject population believed to have been historically underprotected. Second, it will permit research to go forward that has passed uniform expert and public review of risks and benefits.

Some have expressed concern that IRBs, if limited to two categories of risk when making judgments about the acceptability of risks in relation to potential benefits, may be inclined to consider all projects involving greater than minimal risk to require the most comprehensive protections.

In particular, NBAC recognizes the concern expressed by some that if research involving what are normally relatively benign interventions (such as PET scans or MRIs) were categorized as greater than minimal risk, this could result in restrictions that might substantially delay or otherwise limit research. NBAC believes, however, that the most appropriate way to address this issue is not to focus on an arbitrary line, which cannot be reliably established, but rather to focus attention on improving the quality of IRB judgments generally, and on the unavoidable responsibility of IRBs to ensure an appropriate balance between risks and benefits, but also an appropriate balance between risks and protections. One possible strategy is for IRBs individually and collectively to develop "research ethics case law," by publicly disseminating the results of their deliberations so that others may benefit.

The purpose of having a set of risk categories is to establish certain minimal protections and to enable individuals (in this case, IRB members) to discriminate more precisely when making judgments about whether adequate protections are in place, as well as when making judgments about risk in relation to potential benefits. But since risks will vary along a continuum that involves a number of factors, and since IRBs currently have the authority to require a variety of additional protections for persons involved as subjects (even in minimal risk research), NBAC was not persuaded by the assertion that an additional category of risk is needed to assist in these decisions. However, by limiting the categories of risk to two, NBAC is not intending for IRBs to require all available protections when they determine that a research protocol poses greater than minimal risk.

A few empirical studies indicate that there is substantial variation in how IRBs and investigators classify protocols using the current federal risk categories. For example, a 1981 survey found differences in how pediatric researchers and department chairs applied the federal classifications to a variety of procedures commonly used in research involving children.231 Similarly, there was substantial disparity in how the nine members of a special NIH review panel applied the federal classifications to a trial of human growth hormone in which healthy, short children were subjects.232 A survey asking research review committee members and chairs in Canada to classify four different dementia studies "confirmed that there is considerable disagreement and uncertainty about what risks and benefits mean, and about what is to be considered allowable risk."233

NBAC recognizes the difficulty that IRBs may face when making precise risk judgments, particularly about non-physical harms. For this reason, IRBs may find it useful to collect data on the types of protocols they review involving persons with mental disorders, and to assess whether any patterns emerge in which certain types of protocols fall along a spectrum from the most benign to the most dangerous. This could be accomplished within the context of NBAC’s proposed guidance regarding audit and disclosure (see Chapter 5).

In the initial review process, IRBs evaluate a research proposal’s risks and expected benefits based both on study design and on predictions of subject response, and it is widely acknowledged that part of that overall evaluation will include plans for safety and data monitoring. The Common Rule directs IRBs to ensure that "[w]hen appropriate, the research plan makes adequate provision for monitoring the data collected to ensure the safety of subjects."234 After evaluating human subject protections in schizophrenia research conducted at UCLA, OPRR required the institution to "establish one or more independent Data and Safety Monitoring Boards . . . to oversee [DHHS]-supported protocols involving subjects with severe psychiatric disorders in which the research investigators or co-investigators are also responsible for the clinical management of subjects."235 The institution was directed to submit to federal officials a proposal on creating and operating such monitoring boards.

It is necessary to distinguish the process of monitoring data and safety for the study as a whole from monitoring an individual subject’s safety. Data and Safety Monitoring Boards (DSMBs) are well established devices, particularly for multi-site studies, and often recommend the early termination of a study because of evidence that one arm of the study is safer or more efficacious than the other.236 But a major question is how and when to implement individualized subject monitoring, and whether such monitoring should be conducted by someone who is independent of the research team. For example, detailed provisions on monitoring are included in Loma Linda University IRB guidelines on psychopharmacology research in which placebos are administered. Investigators must specify how often subjects will be assessed for deterioration or improvement during studies. The most appropriate quantitative instruments must be used for assessment, and subjects must be withdrawn if their condition deteriorates to a level "greater than that expected for normal clinical fluctuation in a patient with that diagnosis who is on standard therapy;" if they exhibit previously specified behaviors indicating possible danger to self or others; or if no signs of improvement in their condition are evident after a specified time.237

Some have suggested that it would be appropriate to assign monitoring responsibility to the incapable subject’s representative as well. According to the Belmont Report, the representative "should be given an opportunity to observe the research as it proceeds in order to be able to withdraw the subject from the research, if such action appears in the subject’s best interest."238 In this spirit, the Maryland draft legislation directs subject representatives to "take reasonable steps to learn whether the experience of the individual in the research is consistent with the expectations of the legally authorized representative at the time that consent was granted."239 Similarly, the U.S. Army requires a "medical monitor" to "serve as an advocate for the medical safety of volunteers."240

An important policy question is whether research team members and subject representatives can provide sufficient protection to impaired or incapable subjects. On the one hand, research team members may face a conflict between protecting subjects and maintaining the study population.241 On the other hand, it is unlikely that subject representatives will be present during every part of an incapable subject’s research involvement, and lay persons might not recognize every indication of increased risk to subjects. In these circumstances, IRBs would benefit from guidance on possible approaches to monitoring harms and benefits to individual subjects and on criteria for determining when the involvement of an independent health care professional is needed.242 NBAC believes that, at certain risk levels in research using persons with mental disorders that may affect their decisionmaking capacity, independent monitoring is essential, and that such monitoring should be an ongoing process. In its review of protocols, NBAC noted that some lacked sufficient, ongoing monitoring. Although one study involved assigning both clinical and home monitors to subjects, the protocol included insufficient information for an IRB to evaluate how monitoring was actually to occur. More frequently the protocols failed to mention either monitoring or the risks of certain procedures like drug washouts, during which a subject’s condition is likely to deteriorate. IRBs should expect investigators to describe in their research proposals (particularly in proposing research that involves greater than minimal risk) how potential harms to subjects will be monitored.

The first four chapters of this report examined several critical scientific and ethical aspects of the research involving subjects with disorders that may affect their decisionmaking capacity. The final chapter presents NBAC’s recommendations for appropriate protections for this population and the summary justifications for recommendations.

177 45 CFR 46.111(a)(1) and (2) (1998).

178 National Commission, Belmont Report, 6.

179 Thomas A. Shannon and Ira S. Ockene, "Approving high risk, rejecting low risk: the case of two cases," IRB 7, no. 1 (1985): 6–8; Eric M. Meslin, "Protecting human subjects from harm through improved risk judgments," IRB 12, no. 1 (1990): 7–10.

180 Paul Slovic, "Perception of risk," Science 236 (1987): 280–85.

181 Kristen Schrader-Frechette, "Values, scientific objectivity and risk analysis: five dilemmas," in Quantitative Risk Assessment, James M. Humber and Robert F. Almeder eds. (Clifton NJ: Humana Press, 1986): 149–170.

182 DHHS is the only federal agency that has adopted regulations governing research involving pregnant women and prisoners. See 45 CFR 46 Subpt. B and C (1998).

183 DHHS and ED have adopted regulations governing research with children. See 45 CFR 46, Subpt. D; 34 CFR 97, Subpt. D (1998).

184 45 CFR 46.102(i) (1998).

185 45 CFR 46.110 (1998).

186 Alzheimer’s Association Position Statement, "Ethical Issues in Dementia Research," Approved by the Alzheimer’s Association, Board of Directors, May 18, 1997.

187 Keyserlingk et al., supra, 329.

188 Benjamin Freedman, Abraham Fuks, and Charles Weijer, "In Loco Parentis: Minimal Risk as an Ethical Threshold for Research Upon Children," Hastings Center Report 23, no. 2 (1993): 13–19, 13, 17–18. According to the National Commission, "where no risk at all or no risk that departs from the risk normal to childhood (which NBAC calls ‘minimal risk’) is evidenced, the research can ethically be offered and can ethically be accepted by parents and, at the appropriate age, by the children themselves." Report on Children, 137.

189 The DHHS regulations governing research involving children provide that studies may be approved as presenting a minor increase over minimal risk as long as the risks and experiences "are reasonably commensurate with those inherent" in the child subjects’ actual or anticipated medical or other situations. 45 CFR 46.406(b) (1998).

190 Loretta M. Kopelman, "Research Policy: Risk and Vulnerable Groups," in Encyclopedia of Bioethics, rev. ed., ed. Warren T. Reich (New York: Simon and Shuster, 1995); Loretta M. Kopelman, "When Is the Risk Minimal Enough for Children to Be Research Subjects?," in Children and Health Care: Moral and Social Issues, eds. Loretta M. Kopelman and John C. Moskop (Boston: Kluwer Academic Publishers, 1989). See also Berg, "Legal and Ethical Complexities," 24 (noting possible interpretations of minimal risk and concluding that "[i]t clearly does not mean only insignificant risk, but its exact scope is unclear"). The Maryland draft legislation adopts a definition of minimal risk similar to that in the Common Rule. It also refers to minor increase over minimal risk, which is defined as "the probability and magnitude of harm or discomfort anticipated in the research, including psychological harm and loss of privacy or other aspects of personal dignity, are only slightly greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests." Office of the Maryland Attorney General, supra, A-5.

191 Medical Research Council of Canada, Natural Sciences and Engineering Research Council of Canada, Social Sciences and Humanities Research Council of Canada, Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (Ottawa, 1998), 15.

192 Prior exposure to procedures could actually increase the fear and anxiety for some incapable subjects. Incapable adults with memory impairment may not recall undergoing procedures; for them, each procedure will be experienced as a new one.

193 Office of the Maryland Attorney General, supra, A-5.

194 Evan DeRenzo, "Surrogate Decision Making for Severely Cognitively Impaired Research Subjects: The Continuing Debate," Cambridge Quarterly of Healthcare Ethics 3 (1994): 539–548, 540.

195 Ibid., A-17.

196 Keyserlingk et al., supra, 330.

197 Id., 330. In 1980, the President’s Commission issued a paper on the Swedish system for compensation of subjects injured in research. Harry Boström, M.D., "On the Compensation for Injured Research Subjects in Sweden," in Vol. 2: Appendices, Compensating for Research Injuries: The Ethical and Legal Implications of Programs to Redress Injured Subjects, The President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research (Washington, D.C.: GPO, 1992).

198 Richard Wilson and E.A.C. Crouch, "Risk assessment and comparisons," Science 236 (1987): 267–70.

199 Eric M. Meslin, "Protecting human subjects."

200 W. Haefle, "Benefit-risk tradeoffs in nuclear power generation," in Energy and the Environment, eds. H. Ashely, R. Rudman, and C. Starr (New York: Pergammon Press, 1981).

201 See Ashley, Energy and the Environment; Schrader-Frechette, "Values, scientific objectivity and risk analysis." 

202 National Commission, Report on Children, 8–9.

203 Keyserlingk et al., supra, 324.

204 National Commission, Report on Those Institutionalized as Mentally Infirm, 8–9.

205 Keyserlingk et al., supra, 327.

206 National Commission, Report on Those Institutionalized as Mentally Infirm, 31.

207 Id., 13. Berg also emphasizes the need to weigh the likelihood of direct benefit to subjects. In clinical trials, for example, "the benefit calculation must take into account how probable it is that a particular subject will get the experimental medium as well as the probability that, once received, the intervention will help." Berg, "Legal and Ethical Complexities," 25.

208 Keyserlingk et al., supra, 327.

209 This problem was of concern to the intermediate appellate court in the T.D. litigation.

210 National Commission, Report on Those Institutionalized as Mentally Infirm, 31.

211 Keyserlingk et al., supra, 327.

212 Thus, indirect benefit ought not be deemed sufficient to enter an incapable subject in studies presenting more than a "minor increment over minimal risk." Id., 333–34. Keyserlingk et al., characterized indirect benefits as "by nature difficult to predict with any accuracy and . . . often very person-specific." Id., 327.

213 Melnick et al., supra, 535.

214 National Commission, Belmont Report, 7.

215 State of New York Department of Health, Recommendations on the Oversight of Human Subject Research; Citizens for Responsible Care in Psychiatric Research, public comments in a letter to NBAC, September 4, 1998; Nathaniel S. Lehrman and Vera Hassner Sharav, "Commentary, Ethical Problems in Psychiatric Research," The Journal of Mental Health Administration 24:2 (Spring 1997): 227–250.

216 Benjamin Freedman, "Equipoise and the Ethics of Clinical Research."

217 American College of Physicians, "Cognitively Impaired Subjects," 845.

218 Office of Maryland Attorney General, supra, A-26–A-28. Other commentators take a similar position. See, e.g., Berg, supra, 25 (approving this category of research if "no alternative treatment is available of at least equal value, and the experimental treatment is not available through any other source").

Much of the recent controversy over trials involving medication withdrawal for persons with serious psychiatric disorders concerns whether sufficient potential direct benefit exists to justify allowing subjects of questionable capacity to enter or remain in such trials. See Appelbaum, "Drug Free Research"; Patricia L. Gilbert et al., "Neuroleptic Withdrawal in Schizophrenic Patients," Archives of General Psychiatry 52, no. 3 (1995): 173–78. The Loma Linda IRB Guidelines for use of placebos in studies involving persons with psychiatric illness present specific exclusion and inclusion criteria for such studies. Enrollment is limited to persons whose use of standard treatment has produced responses or side effects deemed unacceptable by the patient or an independent psychiatrist. Orr, "Guidelines for the Use of Placebo Controls," 1263. Similarly, Appelbaum endorses a requirement for an independent clinician to screen prospective subjects with the goal of excluding those facing a high risk of harm from psychotic deterioration. Appelbaum, supra, 4.

219 NIH Clinical Center, supra.

220 However, the American College of Physicians would rule out research that "would unduly threaten the subject’s welfare." American College of Physicians, "Cognitively Impaired Subjects." The Maryland draft legislation would permit research presenting more than a minor increase over minimal risk and no reasonable prospect of direct benefit only when subjects appointed a research agent and "the research is unambiguously included in the [incapacitated] individual’s research advance directive." Office of Maryland Attorney General, supra, A-32. Berg proposes that high risk research offering little or no prospect of direct subject benefit should be prohibited unless there is clear evidence that a subject’s competent preferences would support participation. Berg, supra, 28.

221 American College of Physicians, supra, 846. See also Melnick et al., supra, 535 (advising national ethics review prior to any decision to permit studies in this category).

222 National Commission, Report on Those Institutionalized as Mentally Infirm.

223 Keyserlingk et al., supra, 334.

224 Id., 334. The group would accept this form of research for a small group of incapable subjects who previously consented to it in an advance directive, however. Annas and Glantz also contend that without previous competent and specific consent, incapable nursing home residents should not be enrolled in "nontherapeutic experimentation that carries any risk of harm with it." Annas and Glantz, supra, 1157. See also Shamoo and Sharav, supra, S:29 (calling for "moratorium on all nontherapeutic, high risk experimentation with mentally disabled persons which is likely to cause a relapse"); Thomasma, supra, 227–28 (incapable persons should not be involved research failing to offer direct benefit if study presents more than "very mild risk").

225 The group representing the Alzheimer’s disease centers does not explicitly address whether limits on risk should be applied to this form of research. High et al., supra, 72–73. Two other commentators recently argued in favor of permitting incapable persons to be involved in research offering no direct benefit if the risk is no more than a minor increment over minimal risk. Kathleen Glass and Marc Speyer-Ofenberg, "Incompetent Persons as Research Subjects and the Ethics of Minimal Risk, Cambridge Quarterly Healthcare Ethics 5 (1996): 362–72.

226 High, et al., supra, 72. Another statement from the Alzheimer’s centers’ group questions the assumption that a national review body would be particularly qualified to determine "whether the research in question is indeed extremely important to society or to a class of patients—sufficiently so that standard research norms could be put aside." High and Doole, supra, 335.

227 45 CFR 46.407(a), 34 CFR 97.407(a) (1998).

228 45 CFR 46.407(b), 34 CFR 97.407(b) (1998).

229 45 CFR 46.407(b)(1), 34 CFR 97.407(b)(1) (1998).

230 45 CFR 46.407(b)(1)(i)-(iii), 34 CFR 97.407(b)(1)(i)-(iii) (1998).

231 Jeffrey Janofsky and Barbara Starfield, "Assessment of Risk in Research on Children," Journal of Pediatrics 98 (1981): 842–46.

232 See Carol A. Tauer, "The NIH Trials of Growth Hormone for Short Stature," IRB 16, no. 3 (1994): 1–9, 1.

233 Keyserlingk et al., supra, 326.

234 45 CFR 46.111(a)(6) (1998).

235 Office for Protection from Research Risks, supra, 22.

236 See, e.g., Appelbaum, supra, 4 (noting importance of close monitoring to detect early symptoms of relapse so that medication can be resumed to minimize deterioration); Keyserlingk et al., supra, 324 (researchers "must have in place at the start the needed mechanism to monitor subjects, not only as regards the research question, but also in order to identify and prevent unanticipated complications and harms, both physical and psychological").

237 Orr, supra, 1263.

238 National Commission, Belmont Report, 6.

239 Office of Maryland Attorney General, supra, A-25.

240 U.S. Army Regulation 70-25, Research and Development, Use of Volunteers as Subjects of Research, Sec. 2-9(e) (January 25, 1990).

241 In the UCLA schizophrenia research, subjects received clinical care from psychiatrists who also were coinvestigators for the study. There was concern that such a conflict of interest could lead psychiatrists to be insufficiently responsive to signs of possible relapse in patient-subjects.

242 See Shamoo and Sharav, supra, S:29 (researchers and IRBs should be held accountable for monitoring to ensure welfare of subjects protected; physician not associated with research or institution where research conducted should help decide whether subjects’ interests served by continued participation).

Return to Table of Contents