The principle of beneficence states that persons should be “treated in an ethical manner not only by respecting their decisions and protecting them from harm, but also by making efforts to secure their well-being” (National Commission 1979, 6). Therefore, the principle requires that investigators attempt to maximize possible benefits and minimize possible harms. In research, however, the process of gathering data to gain knowledge of benefit to society may expose some individuals to harm, and IRBs must determine, therefore, “when it is [ethically] justifiable to seek certain benefits despite the risks involved, and when the [potential] benefits should be foregone because of the risks” (National Commission 1979, 7).
In the simplest terms, IRBs assess risks and benefits.
However, risk refers both to the probability that harm may occur and its magnitude. Thus, the term small riskis ambiguous, for it could mean that the probability of the harm is small, or that the magnitude (i.e., seriousness) of the possible harm is insignificant, or both. On the other hand, benefitrefers only to the magnitude of the positive outcome, not to its probability. Thus, harms are properly contrasted with benefits, and risks (i.e., possible harms) with potential benefits.
Federal regulations incorporate the obligation of beneficence by requiring IRBs to ensure that risks are minimized to the extent possible, given the research question, and are reasonable in relation to potential ben-efits (45 CFR 46.111(a)(1)-(2); 21 CFR 56.111(a)(1)-(2)).
Such an analysis of risks and potential benefits often will be complex, because IRBs are called on to assess the bal-ance between any number and type of risks and potential
benefits. The current regulations do not further elaborate how risks and potential benefits are to be assessed, and little additional guidance is available to IRBs. The National Bioethics Advisory Commission (NBAC) discussed these issues at some length in two previous reports (NBAC 1998; NBAC 1999b) and has made specific recommen-dations in this area. Among the important considerations NBAC has raised is how IRBs and prospective participants might not evaluate risks and potential benefits in the same way. To account for this factor, in making assess-ments of risks and potential benefits, IRBs must find ways to be sensitive to the perspectives of prospective participants (NBAC 1998).
Types of Harms
Harms can be categorized broadly as physical, psy-chological, social, economic, legal, or dignitary; these categories are not mutually exclusive, however, and more than one type of harm might be present in a given research study. The terms harmand injury have distinct meanings in law, but as in the work of the Health Education and Welfare Task Force (DHEW 1977) and the President’s Commission (1982), the terms are used interchangeably in this report.
Physical harms include injury, illness, pain, suffering, or discomfort. This category encompasses such diverse harms as death in a cancer study, myocardial infarction related to a maximal exercise treadmill test, and discom-fort related to the requirement to lie still in an MRI machine for an extended period during an imaging study.3
Psychological harms include the research participant’s negative perception of self, emotional suffering (e.g., anx-iety or shame), or aberrations in thought or behavior.
Examples of such harms were found in Milgram’s
“Obedience to Authority” experiments (Milgram 1974).
Psychological harms also include distress, anger, or guilt related to the disclosure of sensitive or embarrassing information (NBAC 1999a) and distress and fear upon learning of one’s likelihood of developing a disease for which there is no treatment or cure (Glass et al. 1997;
Marteau and Croyle 1998; Meiser and Dunn 2000).
Psychological harms also could be associated with refusal to participate in a research study, such as the guilt that arises when a prospective participant does not wish to
donate allogeneic bone marrow to be used in an experi-mental manner to treat a patient with AIDS.4
Social harms involve the negative effects on one’s interactions or relationships with others.5 For example, research findings or even study participation itself may expose participants to insurance or employment discrim-ination or other forms of social stigmatization (NBAC 1999b). Research findings may also damage a participant’s relationship with others—for example, when paternity status is disclosed in genetic or child development studies.6 Economic harms involve the imposition, direct or indirect, of financial costs on participants.7For example, participants in clinical research may incur financial obligations for treatments that are higher than those asso-ciated with standard therapies. In some studies, partici-pants may need to take time off from work or pay for transportation or childcare to enable them to take part in the research. They also may be faced with loss of health insurance, or even loss of employment.8
Legal harms are legal actions that could be taken against the participant, such as arrest, conviction, incarcer-ation, or lawsuits. Such harms could occur, for example, in studies of possession or use of illicit drugs, sexual abuse, or workplace theft.9
Dignitary harms are those incurred when individuals are not treated as persons with their own values, prefer-ences, and commitments,10but rather as mere means, not deserving of respect.11 Such harms occur in failures to respect personhood or in violations of justice. These types of harms might be present, for example, in research studies in which informed consent is not obtained or in genetic research that results in the discrimination against the individual participant (NBAC 1999b, 45–46).
Determinations concerning the probability of physi-cal harms are often easier to make than those involving the probability of nonphysical harms. For example, the magnitude and probability of harms associated with a blood draw are well known and can be objectively quan-tified. This is generally not the case for psychological, social, economic, and legal harms. Although IRBs are able to identify such nonphysical potential harms, it is often difficult to determine the probabilities of their occurrence. IRBs, therefore, can err in either direction, by assuming a higher probability and recommending
unnecessary protections or preventing research from being conducted or by assuming a lower probability and allowing research to occur without all the appropriate protections. Although a good deal of information has been gathered about some nonphysical harms—for example, the risks from disclosures associated with trans-mitting or storing certain types of information—the pos-sibility of such harms is not widely appreciated. Efforts should be made to ensure that IRBs are sensitive to the possibility that nonphysical harms might occur and that they better understand how to consider such risks in their assessments.
Risks to Those Other Than Participants
Risks may accrue not only to research participants, but also to persons not directly involved in research, such as to the participants’ family, loved ones, other contacts, social groups, and to society in general (National Commission 1979; NBAC 1999b). For example, in a pre-vious report, NBAC recommended that to the extent that risks of harm to groups can be anticipated, investigators should design their research studies to minimize these risks and should consult with representatives of relevant groups regarding study design (NBAC 1999b). This rec-ommendation was directed to investigators as one impor-tant measure that could be taken without revising the federal regulations. However, encouraging investigators may not be sufficient. Both individuals and other groups experience research risks, so it is appropriate for IRBs to be given explicit authority and guidance in how to consider them.
All of the types of risks described above could accrue to others affected by the research. For example, in a study of a new live virus vaccine, there may be risks to partici-pants’ family members or other contacts who could con-tract the attenuated disease. In some states, there may be legal risk to parents whose minor children participate in a study of illegal activity.12Genetic research may result in certain groups being associated with certain diseases, thus exposing members of those groups to the possibility of stigmatization or discrimination in insurance or employment.13Society in general may incur harm related to research involving procedures, such as xenotransplan-tation or studies of viruses in which there may be some potential of releasing pathogenic organisms.14
National Bioethics Advisory Commission
Current regulations focus only on individual partici-pants. However, other individuals and communities not directly involved in research can also bear risks that are associated with a research study, but that are not acknowledged or mentioned in the regulations. If IRBs focus only on risks to individual research participants, they fail to apply fully the principle of beneficence, because the scope of this principle can extend beyond the individual participants in the research.15
Minimizing Risks
Once risks have been identified, IRBs and investigators must ensure that they are minimized to the extent possible within the limitations imposed by the nature of the research study. Risks may be reduced in a variety of ways—for example, by assuring that the study design is valid; the investigator and research study personnel are qualified; the necessary infrastructure is in place to con-duct the research and deal with any harmful sequelae;
participant privacy and confidentiality are adequately protected; participants are properly monitored; criteria for participant enrollment and withdrawal are appropriate;
a timely treatment plan is in place; and prospective participants at undue risk of harm are excluded.16
Types of Potential Benefits
IRBs should identify potential benefits associated with a research study. Generally, IRBs consider potential bene-fits as accruing either to society or to participants. Just as research studies might involve risks to individuals other than the participants, they might also provide potential benefits to others, especially the participants’ community.
Potential Benefits to Society
The goal of research is to develop knowledge that is beneficial to society. In considering whether to approve a proposed research study, IRBs must make judgments about the importance of the knowledge that is likely to result from research studies. The importance of the knowledge to be gained may increase when significant new findings are expected; when it may result in new products, treatments, or cures; or when it is applicable to many different social groups.
The assessment of whether risks are reasonable in relation to the potential gain in knowledge is difficult to
make, because those who participate in research and who will be exposed to the risks are often not the same indi-viduals who stand to benefit from the research. Thus, the IRB must be able to clearly identify what might be learned from the research and decide whether the gain in knowledge justifies the exposure of participants to harm.
For studies involving little risk, this assessment is not so difficult. However, when the research involves significant risks to human participants—for example, death or disability—the IRB must carefully weigh the risks and the potential gain in knowledge.
Potential Benefits to Participants
Individuals may directly benefit from participation in certain types of research, such as studies designed to offer interventions or procedures that offer a prospect of benefit. For example, potential benefits include receiving clinically significant information that could be used to influence the care provided, receiving standard treat-ments or interventions as part of the research, such as counseling and testing, or gaining access to experimental therapies that may improve the participant’s health status. These types of benefits are to be contrasted with unplanned or unanticipated benefits that are secondary to the objectives of the study. Sometimes, unanticipated direct benefits can result that are relevant to the research objectives. For example, in a research study on treatment for injection drug users, participants were asked to keep daily diaries as a data collection strategy, which served to raise self-awareness about the effect of daily habits on drug use. As a result, some users sought treatment (Singer 2000; Singer et al. 2000). When identifying potential direct benefits to participants, however, IRBs should consider only those that might result strictly from study participation.
Individuals might also benefit indirectly from partici-pation in certain types of research, from experiencing increased social contact, sharing information with another person, or gaining personal satisfaction from participating in the research (NBAC 1998). Indirect benefits typically are not planned by investigators in their research design and do not relate to the objectives of the research study. In addition, they are likely to vary among research participants. Although these benefits
should be acknowledged, they should not weigh in the judgment of IRBs regarding the balance of risks and potential benefits to the participant. To the extent that indirect benefits can be anticipated, however, investiga-tors should design research to increase them.
For purposes of IRB review, incentives for and payments associated with participation in research (e.g., remuneration for transportation, childcare, or lost work time) should not be considered potential benefits to research participants (OPRR 1993). Such potential bene-fits to research participants make it feasible to participate, but they do not result directly from study procedures in which the individuals participate. Including such incen-tives as potential benefits in the IRB assessment would inappropriately skew judgments concerning risks and potential benefits, because nearly any level of research risk could be offset by such gains if they were significant enough—for example, if participants were promised large sums of money for participating in research.
Although IRBs should not consider such incentives or payments as benefits, they should recognize that prospective participants might consider them as such.
Thus, IRBs should determine whether incentives or pay-ments are set so high that they induce prospective par-ticipants to enroll without carefully considering the risks involved in participation.
Potential Benefits to Others
Research institutions, social groups, or communities also can benefit from research. Institutions benefit, for example, from enhanced capacity to conduct research or by receiving resources to improve a program as a result of a research study. Communities can benefit through improved access to programs or through the emergence of programs targeted to specific groups within the com-munity. Research can result in the preservation of local history, customs, or practices that might otherwise be lost. IRBs should acknowledge that these potential bene-fits might be present and should consider ways to increase their likelihood as part of the research design.
However, they should not weigh them in assessing risks and potential direct benefits to the participants.
Relation of Risks to Potential Benefits
Once risks and potential societal and direct benefits have been identified and ways to minimize risks have been considered, IRBs must judge whether the risks are reasonable in relation to potential benefits. These can be difficult judgments to make, because IRBs might have to compare the significance of different types of risks and potential benefits, which are often difficult to quantify (Martin et al. 1995). Further, IRBs have little guidance on how to classify and compare risks and potential benefits.
It would be impossible to provide IRBs with a single algorithm for making such decisions that would do jus-tice to the large variety of cases that might come before them. However, the counsel provided in the Belmont Report should guide IRB deliberations:
It is commonly said that potential 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 of making precise judgments. Only on rare occasions will quantitative techniques be available for the scrutiny of research protocols. However, the idea of systematic, nonarbi-trary analysis of risks and benefits should be emulated insofar as possible. This ideal requires those making decisions about the justifiability of research to be thorough in the accumulation and assessment of information about all aspects of the research, and to consider alternatives systematically. This procedure renders the assessment of research more rigorous and precise, while making communication between review board members and investigators less subject to misinterpretation, misinformation and conflicting judgments (National Commission 1979, 16–17).
An IRB may approve a research proposal only if it judges that the risks are reasonablein relation to potential benefits (National Commission 1979). This judgment may be an IRB’s single most important and difficult deter-mination, because it ensures that when research partici-pants voluntarily consent to participate in a research study, they are offered a “reasonable choice.”17 Research with significant risks could be approved after such an analysis if the potential benefit is significant. However, this does not imply that there is no upper limit to the determination of acceptable risk. For any research study involving significant risk, not only should the local IRB
National Bioethics Advisory Commission
find the risks reasonable in relation to the potential bene-fits, but also so should a larger community of investigators, IRBs, and the public.
Historical Perspective: The Foundation Provided by the National Commission
A coherent conceptual framework is needed for the analysis of risks and potential benefits, and yet there is evidence that this is not provided by current regulations.
For example, it has been suggested that the National Commission espoused three distinct views on the analy-sis of risk: risk assessment based on an analyanaly-sis of the whole protocol, analysis of risks for whole protocols based on the combined analysis of each of the particular components within a protocol; and distinct analyses of risks for individual components of research protocols without judging the whole protocol as a single unit.18 Each of these views is reflected in the current federal regulations.
In its first report, Research on the Fetus, the National Commission presented a framework for risk analysis that relies on categorizing research studies into two types, therapeutic and nontherapeutic (National Com-mission 1975) and made separate recommendations con-cerning the evaluation of risks for each type of research.
Current DHHS regulations concerning fetal research (45 CFR 46 Subpart B) incorporate this framework.19
According to some commentators, this “whole proto-col” approach, which requires that entire research studies be classified as either therapeutic or nontherapeutic, is flawed (Levine 1986).20 The National Commission char-acterized therapeutic research as “designed to improve the health condition of the research subject by prophy-lactic, diagnostic, or treatment methods that depart from standard medical practice but hold out a reasonable expectation of success” (National Commission 1975, 6).
Levine has argued that the concept of therapeutic research is a contradiction in terms:
There is, of course, no such thing as a ‘systematic collection of data or observations…designed to improve the health condition of a research subject…
that departs from standard medical practice.’ Thus, the [National] Commission developed recommendations for the conduct of a nonexistent set of activities…
(Levine 1986, 298).
Because the purpose of research is to acquire knowl-edge, no research study as a whole can be accurately classified as therapeutic research. Indeed, for this reason the National Commission later rejected this whole protocol framework. NBAC has also recognized the difficulty of such a framework (NBAC 1998, 46). As a result of this conceptual flaw, the whole protocol framework allows any research study to be classified as therapeutic as long as it contains at least one component that offers the potential of direct benefit to the participant. However, such research might include nontherapeutic components involving significant risk. For example, a pharmacokinetic study might involve the administration of a standard therapy followed by extensive nontherapeutic testing. In the whole protocol approach, the entire protocol would be classified as therapeutic and there would be no limit to the number of nontherapeutic procedures that could be administered to participants with the protocol still classified as therapeutic. Indeed, IRBs might fail to con-sider the risk associated with the nontherapeutic proce-dures, because the conceptual framework encourages them to think about the entire protocol as offering the prospect of some direct benefit. Thus, it is difficult to envision research participants being adequately protected under such a framework.21
In its reports Research Involving Prisoners (1976) and Research Involving Children (1977), the National Commission rejected the distinction between therapeutic and nontherapeutic research, recognizing that an entire research study could not be considered fully therapeutic, but that, instead, protocols might contain therapeutic or nontherapeutic components. Thus, the National Commission eliminated the conceptual problem that stems from the category of “therapeutic research.” How-ever, the National Commission did not avoid the ethical problems described above. It adopted the language of components, but retained an analytical model based on assessing the protocol as a whole. For example, in its report on research on children, the National Commission endorsed one set of recommendations for research involving no more than minimal risk, another for research involving therapeutic components that exceed the minimal risk threshold, and yet another for research involving nontherapeutic components that exceed this threshold. It has been argued that this framework also
suffers from several weaknesses.22 All research studies containing components that offer the prospect of direct benefit to research participants also contain components that do not offer the prospect of benefit and only answer the research question(s). Thus, two different sets of recom-mendations for assessing risks would apply in a research study involving children that contains both therapeutic and nontherapeutic components in which both compo-nents involve more than minimal risk. Because the rec-ommendations that comprise this framework refer to a research study as a whole involving a particular type of component, it is unclear how multiple recommendations could be applied simultaneously to an entire study.
Although the model avoids the linguistic confusion asso-ciated with a “whole protocol” approach, it still fails to analyze risk and potential benefits separately for thera-peutic and nontherathera-peutic components.23 Nevertheless, current DHHS regulations concerning research involving prisoners (45 CFR 46 Subpart C) and children (45 CFR 46 Subpart D) incorporate this framework.24
Finally, in its report Institutional Review Boards(1978) the National Commission began to move toward a frame-work that involved the analysis of components of a research study rather than the study as a whole; however, it did not adopt specific recommendations for performing such an analysis of risk. Requirements relating to analysis of risks and potential benefits in the Common Rule (45 CFR 46 Subpart A) seem to be based on this frame-work.25 The lack of a single coherent, fully developed conceptual framework hinders the efforts of IRBs to assess and evaluate risks and potential benefits of research studies. At best, the terms therapeuticand non-therapeutic are confusing. By labeling research or proce-dures as therapeutic, it implies that these activities are beneficial when, in fact, the very purpose of research is to make that determination. NBAC has not used these terms in its previous reports, and it does not do so here. Rather, components should be assessed on the basis of whether they offer the prospect of direct benefits to the individuals being studied or general, more indirect benefits to society.
A Framework for the Analysis of Risks of Harms and Potential Benefits
Charles Weijer has proposed a framework based on a
“component analysis” that he believes provides a better
ethical analysis of research that contains a mixture of components—some that offer the prospect of direct benefit to research participants and others with the sole intent of answering the research question(s).26 In the context of a research study, allcomponents are designed to answer the research question(s). Components may include one or more procedures. However, some compo-nents also offer the prospect of direct benefit to research participants, while others do not; the latter solelyoffer the potential societal benefit of knowledge. For example, a research study designed to test whether aspirin or aceta-minophen reduces fever more effectively uses a number of procedures, including the administration of aspirin to one group; the administration of acetaminophen to another group; the randomization of participants to these two groups; and the measurement of fever. All of these procedures are included in the study design because they are needed to answer the research question(s). However, some of these procedures are also designed to offer the prospect of direct benefit to participants—for example, the administration of aspirin or of acetaminophen.
Others, such as randomization, temperature taking, and chart comparisons, do not offer the prospect of such direct benefits; their soleintent is to answer the research question(s). Although a research study must be examined in its entirety, the two types of components should be judged differently.
The component-based approach to the analysis of risks and potential benefits requires IRBs to sort research study procedures into these two types of components to determine their ethical acceptability. The first type con-sists of those components containing particular proce-dures that may offer the prospect of direct benefit to participants. The second type includes procedures that do not. Most procedures in a research study are easily clas-sified into one of these two types of components.
However, in some studies, it might not be clear into which component the procedures best fit—for example, a survey involving a questionnaire and measurements of health in which clinically meaningful results (e.g., blood pressure) are reported back to participants, but no treat-ment is offered. Further, in any research study using diagnostic procedures that might provide useful health information to the participant in addition to the informa-tion provided for the study—such as the use of CAT