II. RISK AND VULNERABLE GROUPS

There are two groups of people considered to be vulnerable research subjects. First, people lacking capacity to give informed consent are vulnerable because they depend on others to protect them, such as young children and adults impaired by trauma, illness, retardation, or dementia. Second, people who are likely to be coerced or manipulated are vulnerable because fear, ignorance, or pressure may account for their agreement to participate. Institutionalized persons, prisoners, members of the military, students, hospital staff, laboratory assistants, and pharmaceutical personnel are frequently cited as vulnerable to coercion or manipulation(U.S. Public Health Service; CIOMS). In addition, the indigent, uninsured, or desperate may be unduly tempted into study participation that they would otherwise reject by financial remuneration. Insofar as participation of vulnerable subjects is problematic, enrolling them in research protocols often requires special justification and safeguards (CIOMS; "Protection of Human Subjects," 45 C.F.R. 46).

Dilemmas of Inclusion vs. Exclusion

It is important to include all segments of society in research, including vulnerable people, so that everyone benefits from research studies. Yet dangers exist in both too many and in too few protections. When too few protections exist, vulnerable people may be exploited. When too many protections exist, however, it is hard to conduct research with vulnerable populations; consequently there is little information about how to diagnose, treat or understand their conditions. Without good research information, people from these groups become neglected. Doctors must then choose between using only modalities that have been tested on the group in question and risk undertreating these subjects, or using untested interventions and risk adverse effects. There are several ways to address this apparent dilemma, and many guidelines adopt some combination of them.

First, all research guidelines require studies to have a strict review by boards known by various names: institutional review boards (IRBs), research ethics committees (RECs), or ethical review committees (ERCs). These boards have discretion to disapprove or approve studies or to adopt suitable additional protections for studies with all or some vulnerable subjects. These committees should be sensitive to various forms of vulnerability (cognitive, environmental, institutional, deferential, medical, economic and social) and respond with appropriate and situationally-appropriate restrictions. The National Bioethics Advisory Commission (NBAC), which carefully distinguishes these different forms of vulnerability, relies upon review board discretion to protect vulnerable subjects because it is flexible (NBAC). But this option has been undermined by high-profile revelations of poor oversight or compliance by some of these review boards. Moreover, confidence in these boards varies according to people's perceptions of whether they represent the interests of vulnerable populations or are seen as favoring the research enterprise or commercial interests.

Second, others favor another approach with special regulatory requirements that must be met before enrolling some or all vulnerable subjects. Such regulations generally exist for infants, children, pregnant women, and prisoners. But critics argue that expanding regulations for other groups could become unwieldy since many, perhaps even most, people may be perceived as vulnerable in some situations (NBAC). If special regulations existed for all or most groups of people who might be vulnerable, it could become unreasonably difficult to approve or conduct research. Moreover, some competent persons, such as pregnant women, object to special restrictions placed upon their freedom of choice.

Third, some guidelines limit risk when subjects are deemed vulnerable. The Food and Drug Administration (FDA, 1997, 4.8.14) stipulates that, when people cannot give consent for trials that do not directly benefit them, risks must be low and other considerations must be fulfilled (i.e., the study cannot be conducted with consenting subjects, consent is obtained from subjects' legal representatives, IRB approval is gained, the negative impact on the subjects is minimized and low, and the study is not illegal). The Council for International Organizations of Medical Science (CIOMS) limits the risk of harm to vulnerable subjects to a "minimal risk," unless the study offers direct benefit to subjects; in some cases a minor increase over minimal risk is permitted in order to study vulnerable people's disorders or conditions. Critics argue that this policy unreasonably restricts people's choices and opportunities, especially when the vulnerable people are competent adults. As noted, NBAC objects to this proliferation of regulations and maintains that once review boards put safeguards in place, vulnerable subjects should be enrolled in studies on the same basis as other subjects.

While these research approaches suggest some similar ways to protect vulnerable people, the moral and policy issues differ greatly for those who are not legally competent and those who are. NBAC describes these two groups' vulnerability as "intrinsic" and "situational," respectively, and CIOMS refers to them as "absolute" and "relative." Yet this language is misleading. First, many children and legally incompetent adults have the capacity to participate in some tasks but not others, so they are neither "absolutely" nor "intrinsically" incompetent; if they have the capacity to assent, which refers to their permission and is a notion different from legal consent, it generally should be sought for research participation. Second, it is misleading to call people either absolutely or intrinsically incapacitated when, for many members of these groups, capacity fluctuates or it grows, as it does for most for children. Third, legally competent persons may view additional protections and restrictions as unjustified paternalism that places obstacles in their path to gain participation in research on the same basis as others; they deny that their situation or relations make them vulnerable. Because the issues are so different for the two groups, their policy options are discussed separately.

Competent Adults Vulnerable to Coercion and Manipulation

There has been some consensus, at least in theory, about how to protect the rights and welfare of competent adults who are vulnerable to coercion or manipulation. First, since the right to consent is grounded in its utility, fairness, and the right of self-determination, studies should be reviewed to ensure that consent is voluntary and that the risks of research are not unfairly distributed to vulnerable groups (CIOMS; "Protection of Human Subjects," 45 C.F.R. 46). This evaluation should be conducted by IRBs, RECs, or ERCs.

Review boards bear a heavy responsibility in recognizing when competent persons may be vulnerable and need additional protection as research subjects. The views of reviewers, however, may differ from those of the potential subjects. One remedy is to assemble a group of prospective subjects and conduct a group consultation to learn their views.

Second, most review boards, investigators, and bioethicists agree that the greater the vulnerability and risk to competent adults, the more specific protections should be adopted; where it is difficult to supervise the voluntariness of vulnerable people's consent, it may be necessary to adopt special regulatory protection. For example, because prisoners live in settings that are inherently coercive and because of past abuses, most guidelines provide additional protections for this population (CIOMS; U.S. "Protection of Human Subjects," 45 C.F.R. 46 Subpart C). In general, research guidelines limit the risk of harm to prisoners to a minimal risk unless the study offers direct benefit to the prisoner subjects (individually or as a class) and requires demonstrated utility, special safeguards, experts' approval, and authorization from the U.S. Secretary of Health and Human Services. In some cases a minor increase over minimal risk is permitted to study their disorders or conditions. These restrictions make biomedical research with prisoners difficult to justify, because there are no diseases unique to them as a class. Given their extraordinary living conditions, however, social or behavioral studies might gain approval.

A third area of general agreement about protecting vulnerable competent adults from coercion or manipulation concerns the importance of avoiding interference with people's self-determination or unjustified paternalism. There is less consensus, however, on how to do this. Competent people may resent paternalistic restrictions of their liberties because someone views them as potentially vulnerable. People may deeply object to being denied options open to others, such as innovative or subsidized care for their illnesses in research programs. Impoverished people, including students, may willingly volunteer for risky studies that pay well. They may argue that if firefighters or fighter pilots receive high pay for taking risks, civilians, too, should have the choice to obtain high pay for taking research risks. They may object to the views of some that payment, other than expenses and tokens, constitutes undue influence and should be prohibited (CIOMS).

DISPUTES ABOUT INCLUDING WOMEN. Perhaps the greatest sustained debate has been over limiting the research options for women of childbearing years or women who are pregnant. One dilemma, as noted, is that if some group is excluded from studies, then it is hard to provide good treatment options for them. Pregnant women have diseases and conditions that need study for their own sakes as well as for the sake of their fetuses. At issue is who makes the decision, the woman herself or others.

In many but not all guidelines, women of childbearing years and especially pregnant woman are listed as "vulnerable" and sometimes denied opportunities open to others. In its instructions to the IRB, for example, the U.S. federal regulations state that: "When some or all of the subjects are likely to be vulnerable to coercion or undue influence such as children, prisoners, pregnant women, mentally disabled persons, or economically or educationally disadvantaged persons, additional safeguards have been included in the study to protect the rights and welfare of these subjects"(U.S. "Protection of Human Subjects," 45 CFR 46.111[b]). The U.S. Public Health Service's "Consultation on International Collaborative Human Immunodeficiency Virus" also includes pregnant and nursing women on their list of possibly vulnerable groups. The goal of these guidelines is to protect the fetuses, newborns, and pregnant and nursing women from research risk. Such policies are controversial since there is no uniform agreement about how to rank duties to protect the women and her fetus and duties to also honor women's rights of self-determination.

Many regulations view pregnant women and those of childbearing years as "vulnerable" and favor more regulatory protections, even if they restrict women's options. These restrictions include limiting the array of studies in which they can participate to those designed to benefit them or their fetus, or those having low risk and requiring their husbands' consent as well as their own. In the United States, for example, research with pregnant women designed to benefit the fetus requires the consent of the father (unless he is unavailable, incompetent, or incapacitated, or the pregnancy resulted from rape or incest). (U.S. CFR 45 46.203 [e]).

Pregnant women's illnesses need to be studied, and it is not in their interests if regulations make this difficult. For example, without research pregnant women are denied the benefits of learning about how drugs affect them as a group. Second, participation in research may be a woman's only or best means to gain access to subsidized care or to investigational drugs or therapy, so restrictions deny them options or direct benefit that are available to others. It may be an unfair denial of benefits to rule that women cannot be considered as subjects. Third, it seems unfair that men of reproductive age are not similarly excluded from drug studies; yet many drugs cause changes in male germ cells that are mutagenic.

A consensus is developing that where there is a conflict between the health needs of the mother and that of the fetus, the mother should be at liberty to resolve the conflict herself (CIOMS; U.S. 45 C.F.R. 46 Subpart B). Restrictions to protect the fetus sometimes rest upon poorly founded assumptions about what might cause harm to the fetus. Informed consent from any woman, however, presupposes that she is informed of the likely harms or benefits, including those that effect her fetus. Pregnancy and nursing make women neither incapable of consent, like children, nor vulnerable to coercion or manipulation, like students and prisoners.

CIOMS does not automatically include women as vulnerable subjects, separating guidelines for vulnerable groups and those for women. It states that vulnerable subjects are those "incapable of protecting their own interests … they may have insufficient power, intelligence, education, resources, or other needed attributes to protect their own interests" (CIOMS, Guideline 13). For vulnerable persons or groups, CIOMS limits the risk of harm to a minimal risk unless the study offers direct benefit to them; in some cases a minor increase over minimal risk is permitted to study vulnerable people's disorders or conditions. In Guideline 16, CIOMS states, "Investigators, sponsors or ethical review committees should not exclude women of reproductive age from biomedical research. The potential for becoming pregnant during a study should not, in itself, be used as a reason for precluding or limiting participation." In CIOMS's commentary on these guidelines, the committee notes that the general presumption should be to include women and that past practices of excluding them is unjust, but that "it must be acknowledged that in some parts of the world women are vulnerable to neglect or harm because of …social conditioning to submit to authority.…" CIOMS also takes a stand on seeking consent from husbands: "In research involving women of reproductive age, whether pregnant or non-pregnant, only the informed consent of the woman herself is required for her participation. In no case should the permission of a spouse or partner replace the requirement of individual informed consent … A strict requirement of authorization of spouse or partner, however, violates the substantive principle of respect for persons." Thus, CIOMS favors women's rights of self-determination and their needs to have drugs and other interventions tested on them. NBAC agrees, and even goes beyond CIOMS, arguing that once review boards put safeguards in place, women and other potentially vulnerable subjects should be enrolled in studies on the same basis as other subjects.

Thus, when vulnerable people are competent, disagreements abound concerning what specific restrictions on their choices are fair, promote their well-being, and respect their self-determination. Too little protection risks their exploitation; too much protection risks unjustified paternalism. Before limiting the liberty of competent people, reviewers and researchers should use community consultation with members of the potentially vulnerable group to consider if they want such protection, if the probability and magnitude of harm warrants constraints, and if the restrictions are the least invasive to secure their well-being.

People Lacking Capacity to Give Consent

As with the competent people, the ethical basis for research policy with persons lacking capacity to give informed consent concerns promoting their self-determination, fair treatment, and well-being. There are four important policy options that were adopted in the twentieth century, and each offers different approaches to ranking what is fair, most protective of incompetent people's well-being, and most respectful of whatever self-determination they have or may develop. These four policies represent different regulative ideals because they rank these primary values differently, and because they offer different authority principles (stating who decides) and guidance principles (substantive directions about how decisions should be made). The remaining discussion will focus on these options.

THE "SURROGATE" OR "LIBERTARIAN" SOLUTION. The oldest policy adopts no special regulatory protection for people lacking the capacity to consent, and allows the same sort of research with them as with other subjects, if their legal guardians consent. Since guardians have the authority to choose the mode of care, religion, and schooling for their dependents, then, according to this view, guardians should determine whether their charges participate in research.

Critics argue that guardians have no authority to volunteer another for studies that are hazardous or that do not hold out benefit for them (Ramsey; Levine; Kopelman, 1989). Guardians have authority insofar as they promote the well-being of those under their care and prevent, remove, or minimize harms to them. They have discretion about how to do this. In nonresearch settings guardians can allow their children or wards to participate in dangerous activities, such as football, presumably because in their judgement there are also direct benefits to them. This differs from volunteering them for risky research, however, when there are no direct benefits to compensate for the risks and where others benefit from that information. Volunteering to put oneself in harm's way to gain knowledge may be morally admirable. But volunteering to put another in harm's way is not admirable, and violates the guardian's protective role. Critics argue that allowing guardians to enroll their charges in potentially harmful experiments wrongs the charges, sets a dangerous precedent, and has a brutalizing effect upon society.

THE "NO CONSENT-NO RESEARCH" OR "NUREMBERG" SOLUTION. Another policy forbids enrolling people as research subjects without their consent. This view is maintained in the first international research statement, the Nuremberg Code. Its first principle states, "The voluntary consent of the human subject is absolutely essential." It goes on to define consent—in a way that has become fairly standard—as requiring legal capacity, free choice, and understanding of "the nature, duration, and purpose of the experiment; the methods and means by which it is conducted; all inconveniences and hazards reasonably to be expected; any effects upon his health or person which may possibly come from participation in the research" (Germany Military Tribunals).

Composed at the end of World War II, the Nuremberg Code stands as an international response to the horrible, involuntary medical studies done by Nazi physicians in which many unwilling subjects and prisoners were killed or permanently maimed (Proctor). It is uncertain if it was intended as a comprehensive code for research (McCormick). If it is taken as a general policy, however, subjects who lack capacity to give informed consent cannot serve as research subjects.

Critics argue that this policy would cripple evidencedbased medicine for people who cannot give consent, turning them into "therapeutic orphans" (Shirkey; McCormick; Levine). Children, retarded persons, and those incapacitated by mental illness have unique medical problems; thus, studies with normal adult volunteers may be inapplicable. Normal adults cannot serve as subjects in studies comparing treatments for schizophrenia, bipolar illness, or lung disease in premature infants. To test the safety and efficacy of many standard, innovative, or investigational treatments for distinctive groups, and give them due consideration, some members of the groups have to be subjects in controlled testing.

THE "NO CONSENT-ONLY THERAPY" SOLUTION. A third policy permits persons who lack the capacity to give informed consent to be enrolled as research subjects if the studies are therapeutic and offer at least as much direct benefit to subjects as other alternatives, and if guardians consent. This view was represented in the next major international code for research to follow the Nuremberg Code, the World Medical Assembly's Declaration of Helsinki, written in 1964 and revised in 1975, 1983, and 1989. (The Declaration's 2000 revision abandoned the "no consent-only therapy" stance after many years.)

This policy option distinguishes clinical or therapeutic research (studies seeking generalizable knowledge and intending to provide medically acceptable therapy for the individual) from nontherapeutic biomedical research (studies seeking generalizable knowledge and not intended as therapy to benefit the individual directly). Therapeutic studies attempt to benefit the person through prevention, diagnosis, or treatment of disease. Thus, drawing the line at therapeutic research for people who lack capacity to give informed consent seems to defenders to be a good solution to the problem of when to permit incompetent people to serve as subjects (Ramsey).

One difficulty with this third option concerns the difficulty of classifying studies as therapeutic or nontherapeutic in a way that is not arbitrary or misleading. Therapeutic studies often have features that are not a part of routine therapy, such as extra tests, inflexible research protocols, and additional hospitalizations, or visits to the doctor. If these nontherapeutic features increase costs, risks, or inconvenience to the patient, classifying the study as therapeutic may be arbitrary and misleading. Moreover, this classification can be misleading if people assume therapeutic studies are always safe or beneficial. They may have a "therapeutic misconception" based on such labels. Labeling something as "therapeutic" may mask risk, inconvenience, costs, or nonbeneficial features, creating an inappropriate bias for participation.

A second problem is that it seems unreasonable to prohibit important low-risk research especially when it offers nontherapeutic direct benefits to subjects or allows progress for these groups. Subjects would be neither harmed nor wronged if they gained from the experience, liked participating, and were not at risk of harm. Children may enjoy and learn from participating in nontherapeutic studies in which they are asked to do such things as stack similar blocks or identify animals from sounds they make. Yet these nontherapeutic studies could be important for establishing criteria of normal vision and hearing. Adults who are not legally competent may also enjoy and learn from serving as research subjects in nontherapeutic studies. For example, they might like an outing to a research facility, meeting the investigators or learning about the study. In addition, they can benefit indirectly from nontherapeutic studies.

Because this option rules out even low-risk studies it seriously impedes medical progress for these groups including the formation of standards about children's typical growth and development. Such standards presuppose carefully tested criteria distinguishing people with developmental delays or impairments from those with normal growth and development. Establishing such norms requires collecting and analyzing data on the growth and development of large numbers of healthy children. Such safe but important research, however, is forbidden under this policy because it is not therapeutic. Even though these studies establishing norms for growth and development are safe, they are nontherapeutic because they are designed not to benefit the subjects directly but to gain generalizable knowledge. If children stack blocks at play, it is not research; if people test views about how they stack blocks, it is research but may be no more burdensome to the child. Thus, when nontherapeutic studies are needed to promote the well-being of incompetent people as a group, and involve little or no risk of harm or inconvenience to them, it is hard to understand how critics can make the case that the subjects are always harmed or wronged by participation.

This option also prohibits epidemiological studies and the investigation of the natural history of disease when there are no therapies. These are among the most important methods for collecting information, so this policy is flawed.

The initial justification for excluding persons who lack the capacity to give informed consent from nontherapeutic research was to honor their rights and protect their welfare. Safe, nontherapeutic research, however, seems neither unfair nor a violation of the rights or welfare of people who lack the capacity to give consent. Failing to do safe but important studies might be unfair and violate their rights and welfare, since it fails to consider all their needs. The Declaration of Helsinki (2000) now permits nontherapeutic studies with guardians' consent and if other subjects cannot be used; no upper level of risk is given unlike the next option.

THE "RISK-BENEFIT" OR "U.S. FEDERAL REGULATION" SOLUTION. A fourth approach allows research on procedures or interventions with incompetent persons when the research holds out direct benefit to them or does not place them at unwarranted risk of harm, discomfort, or inconvenience. Defenders of the fourth option should clarify what risk is unwarranted. This policy uses risk assessment to set priorities between the social utility of encouraging studies and the protection of people's rights of self-determination and well-being. To try to set priorities between the social utility of such studies and respect and protection of incompetent people, this option stipulates that the greater the risk, the more rigorous and elaborate the procedural protection and consent requirements. The U.S. federal regulations(U.S. "Protection of Human Subjects," 45 CFR 46 SubpartD) reflect this fourth policy option in the codes for research with children adopted and those proposed (in 1978 but never adopted) for institutionalized people with mental impairment or retardation. The Council for International Organizations of Medical Science has adopted a similar standard (CIOMS). Under this fourth option, therapy is one of the intended direct benefits that should be taken into account in a risk analysis. Whenever possible, the incompetent persons should give their assent to participate. Assent means affirmative agreement, not just lack of objection.

There are advantages to focusing directly upon the likely benefits and harms of procedures or interventions being studied. First, there are benefits other than therapy that may play a role in deciding if it is reasonable to serve as a subject. A safe, nontherapeutic study that increases a child's understanding of a sibling's chronic illness, for example, might have important lessons about empathy for the child. Those giving consent need to know, of course, the nature and magnitude of the intended benefits (such as education or therapy) or risks of harms associated with the study. Second, calling something "therapeutic" can create the unwarranted idea that participating in the study is in a person's best interest. Risk assessment can reveal hazards, inconveniences, and costs in therapeutic studies that some reasonable people would prefer to avoid.

Using a likely-harms-to-benefits calculation, the U.S. regulations specify four categories of research for children(U.S. "Protection of Human Subjects," 45 CFR 46 SubpartD). IRBs can approve research that they judge to be in the first three categories, and all three generally require the child's assent, if possible, parental approval, and other safeguards such as minimizing risks of harm, having competent investigators and suitable background studies. The first category permits studies with no greater than a minimal risk; the second allows studies with higher risks as long as they are likely to have at least as much direct benefit to subjects as other available therapies; the third category allows research involving a minor increase over minimal risk and no likelihood of direct benefit to each individual subjects, if it is likely to yield vitally important knowledge about the children's disorder or condition. U.S. policy is unique in allowing studies having more than a minor increase over a minimal risk and that do not hold out benefit for the subjects but approval is needed from the federal government. As in the case of the guidelines for prisoners cited earlier, procedural safeguards increase with risk.

There are no final guidelines in the United States for research on those institutionalized as mentally infirm, but there is a proposal about how to treat those institutionalized with impairments like mental illness, senility, psychosis, mental retardation, or emotional disturbances (U.S. Department of Health, Education and Welfare, 1978a, 1978b). It is similar to that proposed for children, except that it allows incompetent adults more authority to decline to participate in studies. The consent or assent of those institutionalized with such impairments must be sought. Those who refuse may not be enrolled in any study that does not hold out direct benefit, without authorization from the courts. CIOMS has a policy that permits enrolling incompetent adults if the study has no more than a minimal risk, others cannot be subjects, the consent of the person or the permission of a responsible family member is obtained, and the research goal is to study the person's disorder or to benefit them.

Unfortunately, this fourth, popular policy option has difficulties. Key terms are either undefined or have vague definitions, permitting broad interpretations about what risks of harm are warranted and what constitutes a benefit. For example, the pivotal concepts of "minimal risk" and "a minor increase over a minimal risk" are problematic (Kopelman, 1989; 2000). The regulations state "'minimal risks' means that the probability and magnitude of harm or discomfort anticipated in the proposed 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" (U.S. "Protection of Human Subjects," 45 CFR 46 102 I). Many other countries and organizations have adopted a similar definition (Kopelman, 2000).

The first part of the definition is vague because daily risks include dangers from drive-by shootings, playing in traffic, flying in airplanes, terrorists attacks, and weapons of mass destruction. Can one know the nature, probability, and magnitude of these "everyday" hazards well enough to serve as a baseline to estimate research risk? And if one can, what reason exists for regarding them as a morally justifiable baseline? It seems easier to determine whether asking a four-year-old to stack blocks is a minimal-risk study than to determine the nature, probability, and magnitude of whatever risks people normally encounter in their daily lives. Moreover, it is unclear if it is the "everyday risks" refer to those some encounter (called the relative standard) or all of us encounter (called the absolute standard). It is also unclear why everyday risks should be a proper baseline to determine that research risk is minimal. Some people have terrible risks in their daily life, but it would seem unfair to use that to justify higher-risk studies for them than for other people.

The second disjunctive of the definition seems to set a standard for physical interventions that have a minimal risk, especially if it is understood as referring to the routine examinations of healthy persons. The test is whether the risk in the research activity is like that of a routine examination. Accordingly, review boards may not approve as minimal risk research such procedures as X rays, bronchoscopy, spinal taps, or cardiac catheterization because they are not part of routine examinations, at least for healthy persons. Review boards, however, can approve studies that have a minor increase over minimal risk, and some of these procedures have been approved as having a minor increase over a minimal risk if their goal is to study a child's "disorder or condition." The terms "minor increase over minimal risk" and "condition" are undefined and vague with no definition for the crucial upper limit of risk that can be approved by review boards, considerable variation exists in how they are understood (NBAC; Kopelman, 2000; 2002).

Finally, this definition of "minimal risk" offers little guidance about how to assess psychosocial risks such as invasion of privacy, breach of confidentiality, labeling, and stigmatization. In routine visits, doctors and nurses ordinarily encounter discussions of family abuse, sexual orientation, and diagnoses that could affect reputations or the ability to get jobs or insurance.

Without clear standards for risk assessment, how effective are these guidelines? A 1981 survey of pediatric department chairs and pediatric research directors (Janofsky and Starfield) found considerable differences of opinion about whether procedures such as venipuncture, arterial puncture, and gastric and intestinal intubation are hazardous. For example, most regarded arterial puncture to have a "greater than minimal risk"; but between 8 and 24 percent thought it had less than a minimal risk, depending on the child's age. An editorial in the Journal of Pediatrics found such variation "cause for concern" and said that better standards of risk assessment are needed (Lascari). Two decades later, similar concerns remain (NBAC; Kopelman, 2000, 2002).

In short, this fourth policy is vague and open to very different interpretations. For example, in 1992 the National Institutes of Health appointed a nine-member review board to assess whether a study of the safety and efficacy of synthetic growth hormone (hGH) was in compliance with federal research guidelines. Eighty children whose adult height was projected to be at or below the first percentile would participate with their parents' consent. The children would receive injections three times a week for four to seven years (600 to 1,100 injections), half getting hGH and, for comparison, the other half receiving salt water, an ineffective placebo. Neither the doctors, the parents, nor the children would know who got water and who got the growth hormone. Each year all the children would come to the National Institutes of Health to undergo a variety of tests, including physicals, X rays, nude photographs, and psychological evaluations. Of the nine panelists, a majority held there was a minor increase over minimal risk, but this risk was offset by the health benefits of being in the study. Two others judged there was no benefit to offset the risks, inconvenience, and discomfort to those getting water rather than hGH, but the study was important enough to be justified. One panelist (this author) argued that a study of a terrible disease might justify these risks for the group getting hundreds of water injections; but shortness is no disease, and so the risk is unwarranted.

If there is any consensus that the fourth approach represented by the U.S. rules and others is the best way to set priorities between the need to protect the rights and welfare of people who lack the capacity to give informed consent with the need to encourage research, it may mask different understandings of what constitutes an acceptable risk of likely-harms-to-benefits ratio. There is a lively debate in the literature about how to clarify these thresholds and, not surprisingly, some favor more restrictive definitions than others (NBAC; Kopelman, 2002).

Conclusion

IRBs, ERCs, and RECs should continue to play an important role in protecting vulnerable subjects while making it possible to continue important research, but there are sharp differences about whether additional regulatory protections are needed (NBAC). Without safeguards, vulnerable subjects risk exploitation. Excessive restrictions, however, have dangers as well. They can thwart the advance of knowledge needed to improve medical care for the groups they seek to protect. Where potential subjects are capable of giving legal consent but are vulnerable to pressure or manipulation, their consent should be monitored to see if it is coerced or manipulated, and regulations should be sought only when they can be justified. There is general agreement that competent adults should serve as research subjects whenever possible, and that when people who lack capacity to give consent are enrolled as subjects in biomedical research, the study should be related to their healthcare needs. The guardian's consent should be obtained; and, if possible, the assent or permission of the person lacking capacity to consent should also be sought. Since there are difficulties with each of the four policies regarding subjects lacking capacity to give informed consent, IRBs, ERCs, and RECs will have to consider issues of utility, fairness, and protection without entirely satisfactory guidance.

loretta m. kopelman (1995)

revised by author

BIBLIOGRAPHY

Council for International Organizations of Medical Science (CIOMS). 2002. International Ethical Guidelines for Biomedical Research Involving Human Subjects. Geneva: Author.

Germany (Territory Under Allied Occupation, 1945–1955: U.S. Zone) Military Tribunals. 1947. "Permissible Medical Experiments." In Volume 2 of Trials of War Criminals Before the Nuremberg Tribunals Under Control Law No. 10. Washington, D.C.: U.S. Government Printing Office.

Janofsky, Jeffrey, and Starfield, Barbara. 1981. "Assessment of Risk in Research on Children." Journal of Pediatrics 98(5): 842–846.

Kopelman, Loretta M. 1989. "When Is the Risk Minimal Enough for Children to Be Research Subjects?" In Children and Health Care: Moral and Social Issues, edited by Loretta M. Kopelman and John C. Moskop. Dordrecht, Netherlands: Kluwer.

Kopelman, Loretta M. 2000. "Children as Research Subjects: A Dilemma." The Journal of Medicine and Philosophy 25(6): 745–764.

Kopelman, Loretta M. 2002. "Pediatric Research Regulations Under Legal Scrutiny: Grimes Narrows Their Interpretation." Journal of Law, Medicine & Ethics 30: 38–49.

Levine, Robert J. 1986. Ethics and Regulation of Clinical Research, 2nd edition. Baltimore: Urban & Schwarzenberg.

McCormick, Richard A. 1974. "Proxy Consent in the Experimental Situation." Perspectives in Biology and Medicine 18(1): 2–20.

National Institutes of Health, Human Growth Hormone Protocol Review Committee. 1992. Report of the NIH Human Growth Hormone Protocol Review Committee. Bethesda, MD: Author.

Proctor, Robert. 1988. Racial Hygiene: Medicine Under the Nazis. Cambridge, MA: Harvard University Press.

Ramsey, Paul. 1970. The Patient as Person: Explorations in Medical Ethics. New Haven, CT: Yale University Press.

Shirkey, Harry C. 1968. "Therapeutic Orphans." Journal of Pediatrics 72(1): 119–120.

U.S. Code of Federal Regulations Title 21—Food and Drug Administration Title 21 "ICH (International Conference on Harmonization Guideline) for Good Clinical Practice" as adopted by the FDA, published in the Federal Register May 9,1997.

U.S. Department of Health, Education, and Welfare; National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. 1978a. Research Involving Those Institutionalized as Mentally Infirm: Report and Recommendations. Washington, D.C.: Author.

U.S. Department of Health, Education, and Welfare; National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. 1978b. Appendix to Report and Recommendations, Research Involving Those Institutionalized as Mentally Infirm. Washington, D.C.: Author.

U.S. Protection Human Subjects, 45 Code of Federal Regulations 46.

U.S. Public Health Service. 1991. "Consultation on International Collaborative Human Immunodeficiency Virus (HIV) Research." Law, Medicine and Health Care 19(3–4): 259–263.

World Medical Assembly. 1989. "The Declaration of Helsinki: Recommendations Guiding Medical Doctors and Biomedical Research Involving Human Subjects." Law, Medicine, and Health Care 19(3–4): 264–265. (Adopted by the 18th World Medical Assembly, Helsinki, Finland, in 1964; amended in 1975, 1983, 1989, and 2001.)

INTERNET RESOURCE

National Bioethics Advisory Commission (NBAC). 2001. Ethical and Policy Issues in Research Involving Human Participants Volume I: Report and Recommendations. Bethesda, MD: Author. Available from <http://georgetown.edu/research/nrcbl/nbac/pubs.html>.

Research Policy: II. Risk and Vulnerable Groups

views updated