Social and Ethical Considerations
The process of clinical research takes place in a social, ethical, and legal context that shapes and constrains the pursuit of science. The governments, foundations, and corporations that provide financial support for scientific research also provide ethical guidelines for its conduct. Negotiation and maintenance of boundaries between scientific and nonscientific spheres is constant and often problematic. It is not clear, however, that the goals of each sphere are necessarily incompatible; good science and good public policy can be pursued in tandem, although not without some effort. NIH's legislative mandate to ensure greater gender equity in clinical studies can be viewed as an explicit attempt to create a balance between the goals of science and those of public policy.
This chapter examines the ways in which the social and ethical constraints affecting the way science operates present challenges to the achievement of equity in clinical studies. The "legal" world—as it is more narrowly construed to mean federal policies and liability issues—will be discussed in the following chapter.
The social context may explain the differential treatment of men and women as scientific research participants, which raises the question of whether science is truly "value neutral" with respect to gender. Similarly, in a society composed of different races, ethnicities, and economic backgrounds, unconscious biases may render those of lesser power and status "invisible" to those of greater power and status. This invisibility can affect the conduct of scientific research in important ways, such as a lack of investigation into
diseases that are pervasive among certain racial and ethnic groups, or a tendency to focus on the reproductive capability of women while ignoring that of men. Finally, these influences have significant practical implications for the design of research studies, for recruitment and retention strategies in diverse populations, and for logistical and cost issues. There is a particular concern for the potential exploitation of disadvantaged individuals and communities who are ascribed lower status and power in our society; the principal protection against this possibility is the informed consent process.
SCIENCE AND OBJECTIVITY
As in all scientific endeavors, clinical studies seek to achieve results that are unbiased by the preconceptions or preferences of investigators, staff, or study participants. This does not mean that the conduct of science is necessarily objective. Many of the principles and methods of the conduct of science can be viewed as precautions to protect the objectivity of the results from a process that is inherently subjective (NAS, 1989). This issue is at the core of many debates about the underlying biases within scientific processes.
Does science reflect the society and culture in which it evolves? For example, is it permeated by the biases related to gender, race, or ethnicity that operate in the broader culture? For some the answer would be no, that science produces a higher, universal "truth" that will always be triumphant (Nagel, 1961; Popper, 1981). These commentators rely on a rational model of the scientific process in which deductive reasoning and observational refutation provide the basis for "objective" results (Popper, 1981).
By contrast, other scholars have argued that science does reflect the culture in which it evolves, including its biases. In the literature on the sociology of knowledge and the history and philosophy of science, for example, Mannheim (1936) and Kuhn (1970) posit that we can only relate to any experience on the basis of our social and historical understanding. In other words, some subjectivity is unavoidable:
Observation and experience can and must drastically restrict the range of admissible belief, else there would be no science . . . but an apparently arbitrary element, compounded of personal and historical accident, is always a formative ingredient of the beliefs espoused by a given scientific community at a given time [Kuhn, 1970].
Kuhn (1970) maintains that "normal science" tends to present a conservative, status quo approach toward knowledge, not because scientists are entrenched in powerful social positions, but because "fundamental novelties . . . are necessarily subversive of its basic commitments." He sees dramatic revolutions in scientific thinking as the result of a profound shift in the
basic assumptions and body of knowledge considered applicable to specific scientific problems, labeling this process a "paradigm shift." So, for example, the works of Einstein and Copernicus were revolutionary precisely because they forced such paradigm shifts: scientists had to discard the basic assumptions that formed the foundation of their work, substituting new ideas and examining new evidence.
Similarly, Mannheim (1936) argues that individual "thinkers" operate in an inherited context, seeking to elaborate on that context or to substitute other contexts that deal more adequately with new challenges. Scientific observation is not an isolated act but one that is "colored by values and collective-unconscious, volitional impulses" that permeate not only science but politics, justice, religion, journalism, division of labor, and language.
This perspective on science has led some observers to argue that scientific knowledge is fundamentally biased by context and social values. That this view currently represents a consensus is suggested by the conclusions of the National Academy of Sciences Committee on the Conduct of Science (NAS, 1989):
Scientific knowledge emerges from a process that is intensely human, a process marked by its full share of human virtues and limitations. . . . Many people think of scientific research as a routine, cut-and-dried process. They associate the nature of scientific knowledge with the process of deriving it and conclude that research is as objective and unambiguous as scientific results. The reality is much different. Researchers continually have to make difficult decisions about how to do their work and how to present their work to others. Scientists have a large body of knowledge that they can use in making these decisions. Yet, much of this knowledge is not the product of scientific investigation, but instead involves value-laden judgment, personal desires, and even a researcher's personality and style.
These debates about the observational biases, processes, and the conservatism of science are basic to many of the legal and ethical issues regarding inclusion of women in clinical studies. We turn next to these more specific applications.
QUESTIONS OF GENDER
Bem (1993), writing on the issues of male bias and male norms in scientific and political discourse, offers a useful analogy to explain the power and pervasiveness of biases. The basic, hidden assumptions involving gender—so basic as to be transparent—are "lenses" through which people perceive, conceive, and discuss social and scientific reality. It is important to look at these lenses in order to understand their consequences, instead of looking through them. Two types of gender-based assumptions—or lenses—seem particularly relevant to this report:
- Male bias refers to the "observer bias" present in scientific inquiry when scientist-observers adopt male perspectives.
- Male norm refers to the tendency to perceive men's identity and experience as the characterization or standard of what it is to be a person and to portray female differences, where they occur, as "deviant."
These lenses can operate independently or simultaneously and can affect observations regardless of the gender of the scientist-observer (women scientists can adopt the cultural values of male biases).
The observer effect is a well-studied and accepted phenomenon in the social sciences, but it is more invisible in the biological and clinical sciences, which are viewed as value neutral. In all scientific fields, however, the theories proposed, questions posed, and data collected are subject to this observer effect. As male bias affects interpretation of information about females, so too do other biases affect the interpretation of information about other less powerful groups, including racial and ethnic groups, the poor, the elderly, and homosexuals.
As an historical example of male bias, Bern (1993) cites the nineteenth century scientific theories that were used to justify the lack of education for women. Arguing from the physical law of conservation of energy, Edward Clarke reasoned that the reproductive capacity of women, especially women during menstruation, would be irreparably harmed by education because thinking would detract from the energy needed by the uterus. As evidence of this harm, Clarke pointed out that educated women had a lower birth rate. While these arguments can be dismissed on the basis of faulty logic and false premises, they illustrate a bias toward finding evidence that women are inferior or legitimately kept from some activities. Bleier (1988) offers a more modern example in the field of neuroscience: major scientific journals tend to publish research supporting the prevailing wisdom that there are biological differences in the brains of males and females and to find flaws in research that contradicts this widely held view.
Altekruse and McDermott (1988) provide examples of similar biases in clinical research, arguing that scientific inquiry—were it to fully recognize the differences between men's and women's social roles (and perhaps in their biology)—would differ in several critical areas. For example, the fields of sports medicine, psychiatry, and occupational medicine would focus on different issues if women's recreational and nonrecreational activities and stresses dominated. Indeed, the specialties themselves might have devel-
oped differently, focusing on leisure activities instead of "sports," or on the hazards of domestic activities instead of those of the workplace.
These and other critics suggest that mere recognition of potential gender-related differences by male scientists is not enough: what is needed are female scientists and physicians, because females in these positions are likely to choose different problems and frame clinical problems differently (Healy, 1991). For example, one study on human immunodeficiency virus (HIV) showed that there are more female participants when the principal investigator or study director is female (Cotton et al., 1993). According to this view, the low number of women scientists, particularly at the upper levels of the research hierarchy, is directly related to the level of women's participation in clinical studies and the amount of attention devoted to women's health concerns. Recent analyses of women's advancement in medical and scientific careers (e.g., by the NIH Office of Research on Women's Health) are also evaluating institutional barriers to women's attainment of coveted positions in these fields.
Arguing from a broader perspective, Belenky et al. (1986) suggest that women's ways of knowing differ importantly from men's. Their interviews suggest that women are more likely than men to use interactive modes to collect information and to trust subjective responses. Applying this line of argument to clinical studies and practice, suppose that obstetrical and gynecological needs were handled by female instead of male practitioners (as had been common in the United States until the eighteenth century). The roles of these physicians and their language might have developed very differently, perhaps concentrating more on trusting the birthing mother's actions and "interacting" with the natural processes under distress, rather than "controlling" the birthing process and "fixing" problems with invasive procedures (Bogdan, 1990).
A final example of male bias in the collection of information is the strong tendency to believe "hard" data that are "objective," even when these data have little clinical relevancy. Stephen Jay Gould (1981) illustrates the consequences of this illusion of objectivity through numbers with a complex story of political and social discrimination against classes of people proven by "science" to be "inferior":
This . . . is the story of numbers once regarded as surpassing all others in importance—the data of craniometry or the measurement of the skull and its contents. The leaders of craniometry . . . regarded themselves as servants of their numbers, apostles of objectivity. And [their results] confirmed all the common prejudices of comfortable white males—that blacks, women and poor people occupy their subordinate roles by the harsh dictates of nature.
Male norm refers to the tendency to conceive of men gender neutrally, as persons, rather than to conceive of men in terms of their sex. Thus, men's identity and experience becomes the characterization or standard of what it is to be a human being (DeBruin, 1994). An example is the argument that men make more appropriate research subjects for drug studies because, for example, women have menstrual cycles that produce deviations from the "normal" pattern of drug disposition observable in males. Basic biologic differences between the sexes are valid and relevant for some scientific research, but when male-only studies are regarded as the scientific "ideal" for conducting research (e.g., to ensure homogeneity across the groups being compared), the male physiology becomes the implicit "normal" standard for judging etiology, pathology, and response to intervention. By contrast, female physiology needs to be examined only to the supposed extent that it deviates from the male norm. This can promote the misuse of scientific rationales to justify protectionism and current practices, but it can also compromise subsequent clinical practice by leaving clinicians and patients uncertain about the applicability of research findings to women.
The male norm may play a role in the design of clinical studies that include only men, but that are focused on a disease or condition that affects both genders. These single-gender studies are designed for a number of practical reasons, such as the convenience and availability of a single-gender population for study (e.g., prisons, VA hospitals), or because of an interest in a particular manifestation that occurs more frequently in males (e.g., midlife cardiovascular events). Nevertheless, all-male studies have also occurred when no particular practical or scientific justification is put forward for excluding women, either because the investigators have a conscious preference for studying men or, more likely, because they have been blinded to the need to include women by the male norm gender lens.
Regardless of their motivations, a primary concern raised about male-only clinical studies relates to their external validity and the generalizability of their results to women, who will likely be diagnosed and treated based on the findings of such research. As a case in point, three out of four randomized clinical trials on cholesterol-lowering drugs were conducted using only middle-age men (Muldoon et al., 1990), yet half of the prescriptions for these drugs were written for women over 60 years of age (Wysowski et al., 1990).
Uncertainty of treatment can also lead to wide variation in treatment patterns (Wennberg, 1991). As a result of greater ambiguity in applying the treatment to women, some clinicians may adopt a conservative approach (withholding treatment for women until further evidence is collected), while others may apply a liberal approach (assuming that the study findings can
be generalized to all patients). These arguments have even greater cogency when considering children and pregnant women. Ironically, if researchers deliberately avoid these groups because of concerns about harm, they may be placing them at greater risk by withholding a potentially useful treatment.
This tendency to view men's identity and experience as universal—to subscribe to ''false universalism," as defined by philosopher Debra DeBruin—may operate in a subconscious fashion. The subtlety and lens-like transparency of false universalism makes it difficult to attribute malice or deliberate intent, but it nevertheless may result in making women "invisible" in a general sense and "deviant" with regard to clinical studies.
QUESTIONS OF RACE AND ETHNICITY
The Diversity of Women
Gender is not the only significant difference to which science may be insensitive. There are other important differences among groups that can have a significant impact on health and illness, and have historically received scant or inadequate attention by clinical researchers. A growing body of evidence (Journal of the American Medical Association Special Issue on Hispanic Health, 1991; Amaro, 1993; Annals of Epidemiology Special Issue on African American Health, 1993; Amaro and Vega, in press), including substantial contributions from scholars of color some of whose work is cited here, indicates that variables such as race, ethnicity, rural and urban background, socioeconomic status, and sexual orientation should be examined in clinical studies.
Recent federal legislation on the inclusion of women and racial and ethnic groups in clinical research has major implications for the investigation of race and ethnicity as explanatory variables in clinical studies. One provision of the recently enacted NIH Revitalization Act requires that clinical research be designed and carried out in a manner that provides valid analyses of whether the variables under study affect women or members of racial and ethnic groups differently than other participants. In order to comply with the intent of the legislation, researchers will have to develop effective outreach, recruitment, and retention strategies and to employ scientific methods that are appropriate and valid for ethnically diverse populations.
This promises to contribute to scientific knowledge on the impact of gender, race, and ethnicity on disease and treatment. It also presents researchers with complex conceptual and methodological challenges, requiring them to clearly define and operationalize race and ethnicity and to show specifically how these variables relate to the research questions under study. In addition, the scientific community must adapt current methods and de-
velop new, more appropriate methods for specific racial and ethnic groups. As a result, methodology becomes a critical factor in determining the risk-benefit ratio for participation and assessing the ethics of conducting such research. To aid researchers in this complex task of inclusion, it may be advisable to draw on the expertise of those researchers who routinely examine the impact of gender, race and ethnicity in their respective fields of research (e.g., social anthropologists, sociologists, social epidemiologists, psychologists, etc.) in order to avail themselves of the most accurate and current conceptualizations and measurements of these variables relevant to their own investigations. The complexity of proper conceptualization and measurement of such variables is treated in the section which follows.
The inclusion of women from diverse racial and ethnic groups therefore confronts each researcher with a set of complex issues that must be addressed and used to inform the research design and approach, including:
- the constructions of race and ethnicity in each research study;
- the specific context and meaning of research within participants' racial and ethnic communities.
Related issues relevant to the recruitment, retention, and the protection of research participants through the informed consent process will be treated in the following section.
Constructions of Race and Ethnicity
Biology or Sociology?
Health research, including clinical research, has given little systematic consideration to race and almost none to ethnicity. Even when race is included as a variable under study, investigators generally fail to establish whether race is considered to have a biological or a sociological connection (Gamble and Blustein, 1994; Williams, in press).
Biological constructions of race highlight the impact of genetic characteristics or genotypic differences along racial lines. The genetic model of racial differences has been based on the following assumptions: "(1) race is a valid biological category; (2) the genes which determine race are linked to those that determine health; and (3) the health of a population is largely determined by the biological constitution of the population" (Krieger and Bassett, 1986). This concept of race has been criticized by those who believe that "race is a societally constructed taxonomy that reflects the intersection of biological, cultural, socioeconomic, political, and legal factors, as well as racism" (Williams et al., 1993).
The ways in which race is defined are highly variable and reflect strong
sociological influences (Williams et al., 1993; Gamble and Blustein, 1994; Williams, in press). While genetic and biological factors may contribute to racial differences in health and response to certain treatments, these differences may also reflect "habitual behaviors of a social group in response to the constraints of its environment" (Williams et al., 1993). When researchers do not give sufficient attention to understanding when and how biological differences between races are the consequences of differential living conditions, they may draw erroneous conclusions about the role of biology in racial differences (Williams, 1993).
This conceptual confusion has led to varying definitions of race across disciplines and to inconsistencies in the ways national health data systems and health research have defined race and ethnicity (Yu and Liu, 1992; Amaro, 1993; Williams, 1993, in press; Gamble and Blustein, 1994). The most common ways of assessing race have been self-identification and interviewer or provider assessment. The latter relies primarily on skin color, a phenotypic attribute that does not always correspond well to genotypic differences or to self-identification (Gamble and Blustein, 1994; Williams, 1993, in press). Methods of classification have also changed over time and across geographic locations, reflecting arbitrary and inconsistent definitions of race (Hahn, 1992; Gamble and Blustein, 1994; Williams, 1993, in press).
This lack of attention to the conceptualization of race and its assessment leads to inappropriate conclusions in health research. After a thorough review of the use of race in the health services research literature from 1966 to 1990, Williams (1993) concluded that "using race only as an afterthought and/or in a mechanical and a theoretical manner does not shed any light on the ways in which racial differences are built into the institutions of society, and serves only to perpetuate the distortion of social reality." Thus, clinical researchers must pay careful attention to the conceptualization and measurement of race and must provide an explicit rationale for the omission or inclusion of analyses to investigate racial and ethnic differences.
Faulty Constructions of Ethnicity
The study of ethnic group and subgroup differences in clinical research is embryonic. Researchers rarely consider differences within ethnic subgroups, often inappropriately grouping individuals from different ethnic backgrounds (Amaro, 1993; Lex and Norris, 1994; Yu, 1994; Zambrana, 1994; Williams, 1993, in press). National health data also combines information for Hispanics, Asians and Pacific Islanders, and American Indians and Alaskan Natives, ignoring the vast differences between their subgroups (Yu and Liu, 1992; Amaro, 1993; Williams, 1993, in press; Yu, 1994).
As with race, definitional issues arise in the measurement of ethnicity.
Yu (1994) highlights the complexity of this task with the following questions:
What should one use as the criteria for defining whether someone is an Asian/Pacific Islander or a member of subgroups of its populations, such as Chinese, Japanese, Filipinos, Hawaiians, Samoans, Guamanians or others? Should the definition be based on ancestry? Self definition of ethnicity? Or interviewer observation of "race," as has occurred in several government surveys prior to 1976?
Despite the conceptual and measurement issues, most scientists who have written on this topic believe that the relationships between race, ethnicity, and health should be investigated. And just as researchers should articulate how race and ethnicity relate to the questions under study, they must also consider how differences in results by race or ethnicity could be explained-not only by biological factors but also by socioeconomic status and other social status characteristics, such as being the direct target of racism (Williams, 1993, in press; Gamble and Blustein, 1994).
Subgroup information must also be disaggregated, however, because there are substantial differences in socioeconomic status, access and utilization of health services, migration, and political history within each racial and ethnic group (Yu and Liu, 1992; Amaro, 1993; Williams, in press; Lex and Norris, 1994; Zambrana, 1994; Gamble and Blustein, 1994).
One Model of Health Does Not Fit All Groups
Historically, the absence of comprehensive epidemiologic information on the health problems of Asians and Pacific Islanders, American Indians and Alaskan Natives, and Hispanics resulted in placing all these groups into a larger social category of "minorities" (Amaro and Vega, in press). Presumably, to the extent that these groups were exposed to the underclass social conditions faced by African Americans, generalizations derived from that group could be extended to cover other groups as well. Instead, the few dedicated studies of other racial and ethnic groups have demonstrated important differences in mortality and morbidity across and within groups. Consequently, clinical investigators should not employ gross conceptualizations or assessment categories of race and ethnicity based on a "minority" model of health. Categorizations of study samples as "white and nonwhite" or ''white and minorities" are scientifically unacceptable. Generalizations from research should be limited to the racial and ethnic groups represented in the study sample, and conclusions regarding racial or ethnic differences (or the lack of such differences) should be limited to groups examined through detailed analyses of differences in each study.
Context and Meaning of Research in Different Communities
The meaning of participation in clinical research studies for different racial and ethnic groups, including women, has important ethical implications that also affect the scientific validity and merit of the research. For all the major "minority" groups in the United States, participation in research cannot be separated from experiences with a health care system that, according to a former secretary of Health and Human Services, has clearly demonstrated undeniable evidence of discrimination and racism (Sullivan, 1991). Distinctions between health research and health services, which are clearly demarcated for health professionals, are not necessarily clear to the general population. In addition, the history of unethical research that has directly or indirectly led to exploitation, deception, disease, and at times death among diverse ethnic populations adds to the distrust of research in these communities (Brandt, 1985; Jones, 1993; Gamble and Blustein, 1994; Lex and Norris, 1994; Yu, 1994). This history of abuse requires that extraordinary measures be taken to guarantee the ethical treatment of participants from diverse racial and ethnic groups in clinical research (see below).
Each racial and ethnic community has a distinctly different history with regard to research (Gamble and Blustein, 1994; Lex and Norris, 1994; Yu, 1994). Clinical researchers must be familiar with the experience of unethical research in each of these groups in order to ascertain the best methods of addressing distrust and fear and for ensuring appropriate conditions for participation. Among American Indian and Alaskan native peoples, for example, a long-standing history of distrust of government personnel and policies is rooted in an even longer history of disease, coercion, broken contracts, and denial of basic rights (Lex and Norris, 1994). In an effort to protect native peoples from potential abuse, the Indian Health Service has developed detailed guidelines to obtain informed consent for certain invasive procedures, especially those pertinent to reproduction (Lex and Norris, 1994).
The infamous case of the Tuskegee syphilis study conducted by the U.S. Public Health Service, in which 400 poor African American sharecroppers from Macon County, Alabama, infected with syphilis, were denied treatment when treatment options became available, has come to be symbolic of the racism within U.S. medicine and research in particular (Brandt, 1985; Jones, 1993; Gamble and Blustein, 1994). Accounts of unethical and cruel treatment of African Americans as research subjects continues to instill distrust of the medical profession and to affect participation in research and health promotion and intervention efforts (Gamble and Blustein, 1994).
Among Hispanics, women have often been the target of the most blatant unethical research and medical practices. In the early 1960s, poor women in Puerto Rico served as some of the first subjects of experiments with birth control pills, in part because of a lack of legal protection regarding drug
testing (Yu, 1994). A study conducted in 1969 in Austin, Texas, told poor Mexican American women that they had been provided with contraceptives, when in fact many of the subjects received placebos; those who became pregnant and requested abortions were denied such services (Adams and Cown, 1972). Such research practices are indicative of a broader climate of unethical medical practices and violations of the reproductive rights of poor women and women from diverse racial and ethnic groups (Rodriguez-Trias, 1976; Gamble and Blustein, 1994). For example, public health fertility control campaigns targeted to increase the use of sterilization among women in Puerto Rico have resulted in inappropriate and uninformed use of this procedure (Presser, 1965; Scrimshaw and Pasquariella, 1970; Vasquez-Calzada, 1973; Rodriguez-Trias, 1976; Schensul et al., 1982).
The legacy of mistrust toward research among Asian and Pacific Islander populations is just as profound but distinctly different from that of the groups discussed above. The history of unethical research with this population includes U.S.-funded research conducted in the countries of origin, as well as the overall oppressive political conditions to which some Asian and Pacific Islander populations have been subjected. An example provided by Yu (1994) is the involuntary radiation exposure of the inhabitants of the Pacific atoll of Eniwetok (now part of the Marshall Islands) as a result of U.S. nuclear testing. These experiences have contributed to distrust of researchers and other individuals perceived, accurately or not, to be "government" representatives.
RECRUITMENT AND RETENTION ISSUES
Feasibility, Logistics, and Cost
Under some circumstances, investigators who wish to recruit and retain women in their studies may encounter problems related to feasibility, logistics, and cost. The NIH Revitalization Act specifically discounts the issue of cost in its requirements for investigators to increase the representation of these groups in their studies, yet investigators will need practical guidance in order to comply with the mandate. Some of these problems may apply to women in general; others are relevant primarily to poor women or to women in racial or ethnic groups. Taken together, however, these practical issues may limit the extent to which equity in research can be achieved, and they may prompt investigators to request exemptions from requirements for inclusiveness.
The feasibility of recruiting and retaining participants for clinical studies depends on the number of women in the reference population and their characteristics in relation to the study. If eligible women are rare relative to men, or if eligible women in an ethnic, socioeconomic, or age group of
interest are relatively rare in comparison with affluent young white women, a greater effort will be needed to identify and attract them.
The feasibility of studying women is also affected by social conditions that may limit the personal autonomy of women. For example, women may have less flexibility than men regarding lifestyle or keeping appointments during work hours. Personal safety may also be a greater concern with respect to evening appointments, which would require provision of transportation, security guards, and the like. Arranging child care during clinic appointments may be more of a problem when studying women than when dealing with men.
For situations in which menstrual cycle variation or other reproductive variables are related to the issues under study, the feasibility of the study will depend on the ability to schedule or plan around this variability. This may increase recruitment costs (to find enough women with the right reproductive or hormonal status) or increase sample sizes (to balance reproductive or hormonal status within the study population). The most extreme version of this problem arises when all women in their reproductive years are considered to be inappropriate study participants because of potential harm to future offspring (see Chapter 7). If gender is thought to be a modifier of the effects or pathways under study, leading to an interest in studying men and women or subgroups of men and women separately, then an inclusive study will have a larger sample size (i.e., enough in each subgroup) than a study in which one or more subgroup is excluded to achieve homogeneity.
Outreach and Access
The factors that motivate women to participate in clinical studies particularly women from diverse racial and ethnic groups, and the elderly and the poor—may be different from those that apply to men. Similarly, the channels of influence whereby women are persuaded to enter and remain in clinical studies may also be culture- or gender-specific. Achieving the goal of a diverse study population may require investigators to develop and implement gender-specific recruitment programs. This could require special or different staffing (e.g., elderly women may be the best people to recruit elderly women into studies) and costly duplication of certain aspects of recruitment for each special population.
Community Attitudes Towards Research
Decisions regarding appropriateness of individual participation in clinical studies can be influenced by a communal perception of risk and benefit, leading to an overlay of motivation or resistance based on group or commu-
nity membership. For example, peer pressure or solidarity within a group may enhance an individual's desire to participate in a project that would benefit the group, or to refrain from participating even though personally willing to participate. An example of community attitudes favoring participation of women in a study might be a grassroots movement in favor of participation in a study of breast cancer, a high-profile condition affecting a high proportion of women. An example of attitudes fostering resistance might be fears of African American women about participating in research based on stories of covert sterilizations in the past or the exploitation of African American men in the Tuskegee study.
Appropriateness of Participation
Any discussion of the participation of women in clinical studies assumes that this is a good thing—that is, that participation in research is beneficial, to society if not to the individual, and that it is appropriate for researchers to solicit participation from individuals in the population at large. The informed consent process (see below) is intended to allow anyone who so wishes to decline or drop out of participation, but there may be some situations in which it is inappropriate, in principle, to request their participation.
For example, is there a threshold standard of living that should be met before it is ethical to solicit a woman's participation in a research study? Does the answer to this question differ according to the level of direct benefit the individual could derive from study participation? That is, is it appropriate to solicit the participation of economically disadvantaged or socially marginal individuals when the study provides a treatment that would otherwise be unavailable to them? This would mean either that there is no placebo condition or that even the placebo condition is an improvement over the care she would otherwise receive. Conversely, is it inappropriate to recruit poor, inner-city women into a study that will benefit society but that will cost them money and time, possibly jeopardize their work performance, and not offer a reasonable assurance of health benefit? Is the obligation to serve society through participation in studies one that accrues only to those for whom society is meeting the minimum need for food, clothing, and shelter? By this reasoning, is research participation one way that persons who owe a debt to society can pay this debt? The committee raises these questions in order to highlight the potential difference between participant and investigator perceptions of justice in clinical studies.
PROTECTION OF RESEARCH PARTICIPANTS
Potential for Exploitation
The history of unethical research and broken social contracts with racial and ethnic groups in the United States provides the context for understanding the importance of ensuring ethical practices in clinical research with women in these groups. Because of the lower status ascribed to ethnic populations, especially women in these groups, they are especially at risk for abusive and exploitative research practices. Clinical research with these populations must clearly articulate safeguards for the ethical treatment of participants.
There has always been a tension between, on the one hand, the investigator's need to recruit and retain study participants and, on the other hand, the study participants' right to uncoerced involvement in and free withdrawal from a study at any time. The issue of diversity may further strain the already delicate balance between these two ends: in light of the NIH mandate, the investigator now stands to gain or lose directly in accordance with his or her success in recruiting a diverse study sample. The gain, if in compliance, may be a higher priority for funding or the ability to receive continuation funds. The penalty for noncompliance may be losing research funds. It will be the investigator's and the IRB's responsibility to see that the needs and rights of the participants are balanced against the desirability of including them in the study. One of the most important safeguards for protection of human subjects is informed consent.
Informed Consent Process
Central to the informed consent process are the requirements that a competent person be given the information necessary for a full understanding of the risks and benefits involved in participating in a research study, and that no coercion or manipulation is used to obtain the potential participant's permission. The informed consent process may become more complex in light of the issues discussed in the preceding sections, such as community attitudes toward research and the legacy of historical exploitation. Monetary compensation for participation in clinical research might be considered an undue influence in any potential subject population; the poorer the population, the more likely it is that such compensation might render poor women vulnerable to disproportionate experimentation (Yu, 1994). Similarly, the real or perceived increased access to health services provided by participation in clinical research may offer relatively more incentive for poor women to participate in experimental treatments than would be the case for other groups.
Although access to participation is desirable, therefore, it is also important to ensure that all potential participants understand the risks and limitations of participating in clinical research. This of course is a concern with all human subjects; the process of informed consent was developed precisely to ensure that participants receive and understand such information. In some of the populations of concern here, however, the process of informed consent is more complex than in other groups.
Insofar as the ability to comprehend complicated information depends on literacy, education, and proficiency in English, for example, it is questionable whether current methods for obtaining informed consent are adequate for many participants from racial and ethnic groups, especially the poor and recent immigrants (Hurh and Kim, 1982). Valid consent also depends on the investigator's willingness, efforts, and commitment to convey information in language the participant can understand. Current guidelines for informed consent are not uniform from institution to institution; some do not require that written informed consent forms be provided in the participant's native or dominant language. The more usual practice is to supply a literal verbal translation of the form or a translation of its major points, by a translator who may or may not be a member of the research team. This type of translation can introduce bias and inconsistency into the informed consent process.
Another problem of informed consent that affects immigrant populations and those who have lived under oppressive political regimes is the lack of understanding of basic human rights as defined in this country (Yu, 1994). Those who have lived in societies where torture and government coercion are common may fear retaliation for participating or not participating. In addition, social status may unknowingly exert undue influence on an individual's decision to take part in a clinical study. Yu (1994) reports that some Asian and Pacific Islander populations are influenced to sign consent forms because the person who asked for their participation is considered credible and not because they understand the risks involved.
CONCLUSIONS AND RECOMMENDATIONS
Within the scientific community, there is no consensus concerning whether scientific objectivity can be achieved. Some scientists believe that the research process cannot easily be disentangled from the social world within which it is conducted. Societies stratified by gender, race, ethnicity, and socioeconomic status provide different "lenses" through which to see and understand social and scientific reality. These unconscious biases may permeate the entire scientific research process, influencing the research topics selected, the definition and operationalization of concepts examined, the study design, the method of data collection employed, and the research
participants chosen for inclusion. Furthermore, such unconscious assumptions contribute to the view that men's physical makeup and experiences are the standard by which to measure and compare women's; to the extent that women's experiences differ from the established male norm, they may be categorized as deviant. These biases impede the progress of the scientific enterprise and produce findings that are not valid for large segments of the population.
The committee recommends that NIH and IRBs engage in educational efforts that will ensure that investigators are aware of such gender biases and that studies are equitably conceived and designed with respect to gender.
One way to reduce the influence of such gender biases may be to have a greater number of women scientists active in the research enterprise, through, for example, identification and removal of any institutional barriers to their increased participation. The perspectives they bring to bear may differ markedly from those of their male colleagues, thus aiding in the dissolution of unwarranted and inaccurate assumptions about women in the research enterprise.
The committee recommends that NIH continue its efforts to encourage women of all racial and ethnic groups to become scientific researchers and to assume positions of authority within the scientific hierarchy.
Variables such as race, ethnicity, and socioeconomic status can affect health outcomes. The lack of attention to or inadequate conceptualization and measurement of these variables in clinical studies has resulted in findings that are inapplicable to particular racial, ethnic, and socioeconomic groups. For example, in order to accurately determine the effects of race on health and treatment outcomes, it is important to clearly distinguish the biological and sociological components of race. Standard methods of data collection may be inappropriate to certain cultural groups and may need to be modified to ensure that the information obtained is valid and for the risk-benefit ratio to be acceptable. Thus, studies must be planned, designed, and executed to produce valid and generalizable results to the populations under investigation. Investigators and IRBs should utilize the expertise of scholars with experience in studying these populations to avoid the weaknesses evidenced in earlier research.
The history of government-sponsored health research and health care efforts in racial, ethnic, and socioeconomic subgroups has not been unblemished—past unethical treatment has led individuals from these groups to be
wary of participation in current studies. Because of the requirements of the NIH Revitalization Act of 1993, researchers now stand to gain or lose support in accordance with their success in recruiting and retaining participants from these same groups, the federal mandate has the potential effect of exacerbating past problems of exploitation. Knowledge of the history of health research in relevant racial or ethnic groups and an awareness of the cultural and political frames of reference employed by the members of these groups will enable researchers to avoid perpetuating the problems.
Informed consent is the primary mechanism for protecting subjects from unethical treatment. NIH, IRBs, and investigators must work together to tailor the consent process so that it will be effective for every group that participates in clinical studies. This entails, for example, both understanding and avoiding what might constitute excessive inducement (monetary or otherwise) for members of a group. If the benefits of research are to accrue to all groups equally, then proper study design and fully informed consent are critical elements to the achievement of that end.
The committee recommends that NIH commission a study of attitudinal and institutional barriers to participation in research among women, racial and ethnic groups, and the poor.
The committee recommends that NIH train IRGs and TEGs and investigators in recruitment and retention issues; part of this training should emphasize methodological and ethical issues in conducting research with women of diverse racial and ethnic groups and poor women.
The committee recommends that investigators tailor study designs and recruitment and retention efforts to the specific populations to be included in the study. Investigators must consider the relevance of race, ethnicity, socioeconomic status, and other subgroup variables to their study and develop appropriate definitions, methods, and measurements, to ensure the validity of their research efforts among these groups.
The committee recommends that in designing recruitment and consent procedures, investigators be cognizant of concerns and needs of communities that have a history of exploitation or abuse in previous clinical studies. Investigators also must ensure that such information be presented and carefully explained, orally and/or in writing, in the potential participant's preferred language.
Adams and Cown. 1972. The human guinea pig: How we test new drugs, World (5 December):21.
Altekruse, J.M., and McDermott, S.W. 1988. Contemporary concerns of women in medicine. Pp. 65-88 in Feminism Within the Science and Health Care Professions: Overcoming Resistance, S.V. Rosser, ed. New York: Pergamon.
Amaro, H. 1993. Women don't get AIDS. They just die from it. Presented at the American Psychological Association Annual Meeting, August 20-24, Toronto, Canada.
Amaro, H., and Vega, W. In press. Health status of Latino population. Annual Review of Public Health.
Annals of Epidemiology. 1993. Special Issue on African American Health. 3(2).
Belenky, M.F., Clinchy, B.M., Goldberger, N.R., and Tarule, J.M. 1986. Women's Ways of Knowing: The Development of Self, Voice, and Mind. New York: Basic Books.
Bem S.L. 1993. The Lenses of Gender: Transforming the Debate on Sexual Inequality. New Haven: Yale University Press.
Bleier, R. 1988. Science and the construction of meaning in the neurosciences. Pp. 91-104 in Feminism Within the Science & Health Care Professions: Overcoming Resistance, S.V. Rosser, ed. New York: Pergamon.
Bogdan, J.C. 1990. Childbirth in America: 1650 to 1990. Pp. 101-120 in Women, Health, and Medicine: A Historical Handbook, R.D. Apple, ed. New York: Garland.
Brandt, A.M. 1985. Racism and research: The case of the Tuskegee Syphilis Study. In: Sickness and Health in America, 2nd ed., J.W. Leavitt and R.L. Number, eds. Madison: University of Wisconsin Press.
Cotton, D.J., He, W.L., Feinberg, J., and Finkelstein, D.M. 1993. Determinants of accrual of women to a large, multicenter HIV/AIDS clinical trials program in the United States. Journal of Acquired Immune Deficiency 6:1322-1328.
DeBruin, D.A. 1994. Justice and the inclusion of women in clinical studies: A conceptual framework. In: Women and Health Research: Ethical and Legal Issues of Including Women in Clinical Studies, Volume 2, A. Mastroianni, R. Faden, and D. Federman, eds. Washington, D.C.: National Academy Press.
Gamble, V.N., and Blustein, B.E. 1994. Racial differentials in medical care: Implications for research on women. In: Women and Health Research: Ethical and Legal Issues of Including Women in Clinical Studies, Volume 2, A. Mastroianni, R. Faden, and D. Federman, eds. Washington, D.C.: National Academy Press.
Gould, S.J. 1981. The Mismeasure of Man. New York: W.W. Norton & Company.
Hahn, R.A. 1992. The state of federal health statistics on racial and ethnic groups. Journal of the American Medical Association 267:268-271.
Healy, B. 1991. The Yentl syndrome. New England Journal of Medicine 324(4):274-276.
Hurh. W.M., and Kim, K.C. 1982. Methodological problems in the study of Korean immigrants: Conceptual, interactional, sampling and interviewer training difficulties. Pp. 8193 in Methodological Problems in Minority Research, Occasional Paper No. 7, W.T. Liu, ed. Chicago: Pacific/Asian American Mental Health Research Center.
Jones, J.H. 1993. Bad Blood: The Tuskegee Syphilis Experiment, 2nd ed. New York: Free Press.
Journal of the American Medical Association. 1991. Special issue on Hispanic health. 265(2).
Krieger, N., and Bassett, M. 1986. The health of black folk: Disease, class, and ideology in science. Monthly Review 38:74-85.
Kuhn, T.S. 1970. The Structure of Scientific Revolutions, 2nd ed. Chicago: University of Chicago Press.
Lex, B.W., and Norris, J.R. 1994. Health Status of American Indian and Alaska native women.
Women and Health Research: Ethical and Legal Issues of Including Women in Clinical Studies, Volume 2, A. Mastroianni, R. Faden, and D. Federman, eds. Washington, D.C.: National Academy Press.
Mannheim, K. 1936. Ideology and Utopia: An Introduction to the Sociology of Knowledge. New York: Harcourt Brace Jovanovich.
Muldoon, M.F., Manuck, S.B., and Matthews, K.A. 1990. Lowering cholesterol concentrations and mortality: A quantitative review of primary prevention trials. British Medical Journal 301:309-314.
Nagel, E. 1961. The Structure of Science. New York: Harcourt, Brace and World.
NAS (National Academy of Sciences). 1989. On Being A Scientist. Washington, D.C.: National Academy Press.
Popper, K.R. 1981. The Logic of Scientific Discovery, 3rd ed. Boston: Routledge & Kegan Paul.
Presser, H. 1965. Sterilization and Fertility Decline in Puerto Rico, Population Monograph AE13. Berkeley: University of California Press.
Rodriguez-Trias, H. 1976. Women, health and law: Guidelines on sterilization under attack. Women and Health 1:30-31.
Schensul, S.L., Borrero, MG., Barrera, V., Backstrand, J., and Guarnaccia, P. 1982. A Model of fertility control in a Puerto Rican community. Urban Anthropology 1(1).
Scrimshaw, S.C., and Pasquariella, B. 1970. Obstacles to sterilization in one community . Family Planning Perspectives 2(4):40-42.
Sherwin, S. 1994. Women in clinical studies: A feminist view. Women and Health Research: Ethical and Legal Issues of Including Women in Clinical Studies, Volume 2, A. Mastroianni, R. Faden, and D. Federman, eds. Washington, D.C.: National Academy Press.
Sullivan, L.W. 1991. Effect of discrimination and racism on access to health care. Journal of the American Medical Association 266:2674.
Vasquez-Calzada, J.L. 1973. La esterilizacion femenina en Puerto Rico. Revista de Ciencias Sociales 17(3):281-308.
Wennberg, J.E. 1991. Unwanted variations in the rules of practice. Journal of the American Medical Association 265(10):1306-1307.
Williams, D.R. In press. Race in health services research: Past, present, and future. Health Services Research.
Williams, D.R., Lavizzo-Mourey, R., and Warren, R.C. 1993. Race in the health of America: Problems, issues, and directions. Submitted to Public Health Reports.
Wysowski, D.K., Kennedy, D.L., and Gross, T.P. 1990. Prescribed use of cholesterol-lowering drugs in the United States, 1978 through 1988. Journal of the American Medical Association 263(16):2185-2188.
Yu, E.S.H. 1994. Legal and ethical issues relating to the inclusion of Asian/Pacific Islanders in clinical studies. Women and Health Research: Ethical and Legal Issues of Including Women in Clinical Studies, Volume 2, A. Mastroianni, R. Faden, and D. Federman, eds. Washington, D.C.: National Academy Press.
Yu, E.S.H., and Liu, W.T. 1992. U.S. national health data on Asian Americans and Pacific Islanders: A research agenda for the 1990s. American Journal of Public Health 82(12):1645-1652.
Zambrana, R.E. 1994. The inclusion of Latino women in clinical and research studies: Scientific suggestions for assuring legal and ethical integrity. Women and Health Research: Ethical and Legal Issues of Including Women in Clinical Studies, Volume 2, A. Mastroianni, R. Faden, and D. Federman, eds. Washington, D.C.: National Academy Press.