Justice and the Inclusion of Women in Clinical Studies: A Conceptual Framework
Debra A. DeBruin
It often has been charged that women1are not adequately included as subjects in clinical research.2 In particular, it has been charged that women frequently are excluded from studies concerning conditions affecting both women and men, and that when women are included in such studies they tend to be underrepresented.3
This Institute of Medicine study of the legal and ethical issues relating to the inclusion of women in clinical studies focusses on this charge of exclusion and underrepresentation. This is a worthy topic for study; the charge demands attention. However, our moral analysis of our practices concerning the inclusion of women in clinical research will fail to capture all that it should if we restrict our focus to the charge of exclusion and underrepresentation.
We must contemplate a more complicated picture. Imagine a society with the following sorts of research practices: women are excluded from and underrepresented in clinical studies concerning conditions affecting both women and men. Even when women are included in adequate numbers in such studies, researchers often fail to do the analysis necessary to determine whether the gender of the study subject affects the results of the study (e.g., whether the condition in question manifests itself differently in women and men; whether the drug on trial affects women and men in different ways). Sometimes when researchers do perform such gender-specific analysis, they do so without regard for advancing knowledge about women's health. (For example, this hypothetical society has a research program that concentrates on the question of how women can pass along certain sexually transmitted diseases to men, but ignores questions
of how these diseases affect the women themselves and how they should be treated.) Besides having these practices regarding research into conditions affecting both women and men, our hypothetical society pays relatively little attention to the study of health concerns that are specific to women. In general, then, this hypothetical society fails—not completely, but to a significant extent—to study research questions appropriate for women. It fails, to significant extent, to include women in its clinical studies—not just in the sense that it fails to include women, or sufficient numbers of women, as subjects, but also in the sense that it fails to incorporate a concern for women's health into its research agenda.
Why should we concern ourselves with this hypothetical society? For one thing, it would be philosophically instructive to develop a moral analysis of this society's research practices. More importantly, however, we should care about this hypothetical society because I have modeled it on our own society. When we realize the society I've described is not really just hypothetical, we can begin to see how urgent is the need for moral analysis.
It is difficult to quantify the exact extent of women's exclusion from and underrepresentation in clinical studies in our society. (Indeed, I shall here address only the question of exclusion, since the ambiguity of the term "underrepresentation" makes it so difficult to tell when women are included in "sufficient" numbers and when they are not.4) Only since 1992, when the National Institutes of Health revised its policy on the awarding of federal research grants, have researchers been required to include information on gender representation and analysis in their study proposals and renewal requests.5 Prior to that, information on inclusion was incomplete, and what data there are have not yet been systematically gathered.
Nevertheless, the evidence available so far supports the claim that women have been excluded from clinical studies. The charge of exclusion has been made most conclusively with respect to certain areas of research, e.g., the study of cardiovascular disease. Although cardiovascular disease is the leading cause of death for both women and men, women have been excluded from a number of key studies providing the basis for much of our current understanding of this type of disease—e.g., the Multiple Risk Factor Intervention Trial, conducted on 15,000 men; the Physician's Health Study of aspirin's preventive power against heart disease, conducted on 22,071 men; and "all the large trials on cholesterol-lowering drugs, [which] leave researchers with little or no information about how these drugs work in women."6
Some critics of our society's clinical research practices suggest that the exclusion of women is not limited to certain areas of research such as the study of cardiovascular disease, but that it is a systematic problem. These critics note that federal guidelines for research sometimes present barriers to the inclusion of women. For example, until recently, Food and Drug Administration
guidelines for drug research advised that women of childbearing potential should be excluded from early phases of drug testing, except in rare cases (e.g., when the inclusion of the women in the drug trial has lifesaving potential for them).7 Kinney et al. have argued that this policy led, in practice, to the exclusion of women of childbearing potential from later phases of drug testing as well.8 Moreover, the critics cite suggestive anecdotal evidence to support their claim about the systematic nature of exclusion:
[T]he first twenty years of a major federal study on health and aging included only men. Yet two-thirds of the elderly population are women. The recent announcement that aspirin can help prevent migraine headaches is based on data from males only, even though women suffer from migraines up to three times as often as men. . . . Most amazing is the pilot project on the impact of obesity on breast and uterine cancer conducted—you guessed it—solely on men. . . . And in basic research, even female rats are frequently excluded as research subjects!9
We also have evidence that, even when women are included in studies, researchers often fail to perform the analysis that's necessary to determine whether the gender of the research subject affects the results of the study. For example, "there is a premenstrual rise in asthma deaths," but no one has ever studied "'whether there's a connection between the cycle and the bronchi, or differences in how medications work premenstrually. . . . '"10 Also, AIDS studies generally fail to include gynecological exams of women subjects, even though AIDS often manifests itself in women through severe, persistent gynecological problems.11
When researchers do perform such gender-specific analysis, they sometimes do so without regard for advancing knowledge about women's health. The most shocking example of this phenomenon I'm aware of comes from AIDS research. Most studies of women with AIDS are designed to investigate how women transmit AIDS to their male sex partners or to their fetuses. Faden et al. describe the following study as typical of AIDS research involving women:
The study, begun in early 1990, is designed to assess whether the rate of HIV transmission from mother to infant can be reduced by continuous oral AZT treatment to HIV-infected pregnant women, intravenous AZT during childbirth, and oral AZT treatment to the newborn infant. The study also seeks to evaluate the safety of AZT for both the pregnant woman and newborn infant. Originally, the study included no maternal health component; attention was focused exclusively on the fetus. No gynecological care was provided to the women, there was no requirement that an internist be included on the study team to meet the women's non-obstetrical health needs, and AZT for the women was discontinued immediately after delivery. . . . The Women's Health Core Committee of the [AIDS Clinical Trial Group] revised the protocol in 1991. Under the changed protocol, whatever treatment the women had been receiving during the study
would be continued for six weeks postpartum, while the infants would continue to receive their study treatment for eighteen months.12
Faden et al. also note that ''[g]iven that the first cases of AIDS in women were reported to the CDC in 1981 and the number of AIDS cases has been rising faster in women than in men nearly every year since 1986, it is extraordinary that the NIH's first major study of how HIV disease manifests itself and progresses in women will not begin accruing patients before the end of 1994.''13
Finally, critics commonly charge that health concerns that are specific to women receive relatively little research attention. The long list of oft-cited gaps in women's health research includes study of "the effects of hormone replacement therapy in reducing heart disease and bone loss, the impact of a low-fat diet on preventing breast cancer, and the use of vitamin D and calcium supplements to prevent bone loss and reduce colon cancer."14
In this paper, then, I shall assume that (1) women are excluded from and underrepresented in clinical studies; (2) even when women are included in sufficient numbers, researchers often fail to perform the analysis necessary to determine whether the gender of the research subject affects the results of the study; (3) when researchers do perform gender-specific analysis, they sometimes do so without regard for advancing knowledge about the health of women; and (4) health concerns specific to women receive little research attention. I shall not—I cannot—make any definitive empirical claims about the full extent of these phenomena. However, I shall say this: We must not so focus on any individual phenomenon that we lose sight of the complete picture composed of all four phenomena. Given this complete picture, I shall reiterate that our society fails—not completely, but to a significant extent—to study research questions appropriate for women. It fails, to a significant extent, to include women in its clinical studies—not just in the sense that it fails to include women, or sufficient numbers of women, as subjects, but also in the sense that it fails to incorporate a concern for women's health into its research agenda. Whatever the extent of this failure, I shall argue that it is morally unacceptable wherever it occurs. I shall argue that these research practices result from and further perpetuate the oppression of women in our society. (Toward this end, I shall present an analysis of some aspects of women's oppression.) Given this connection between the oppression of women and these research practices, and given the systematic nature of oppression, it would be quite surprising if this failure to incorporate a concern for women's health into the research agenda were limited to isolated areas of research.
I shall also assume that these practices concerning the inclusion of women in clinical studies result in inequalities in the quality and availability of care, which have a detrimental impact on women's health. After a two-year study, the Public Health Service Task Force on Women's Health Issues concluded in its 1985 report that "[t]he historical lack of research focus on women's health
concerns has compromised the quality of health information available to women as well as the health care they receive."15 It stands to reason that this is true for health concerns specific to women. When these concerns are not studied, but those specific to men are, there will be inequalities in the availability and quality of care. When women cannot get care for their conditions because those conditions haven't been studied, or when they get care based on inadequate information about their conditions, their health will inevitably suffer. Further, there is evidence that this assumption is true for conditions that affect both men and women. For example, there is evidence that, because of the way women have been included in (or to a large extent excluded from) AIDS research, women with AIDS receive less, and lower quality, care than their male counterparts. There is also evidence that some drugs interact with women's menstrual cycles so that drug regimens based on studies that that excluded women or failed to perform gender-specific analysis are ineffective or harmful for women. There is much we don't know about how women's health is affected by our practices concerning their inclusion in studies, precisely because the relevant studies have not been done. What we do know gives us cause for worry.16
My aim in this paper is to provide a framework for thinking about two broad philosophical questions: (1) Are our practices concerning the inclusion of women in clinical studies unjust, and, if so, why? And (2) if they are unjust, what sorts of remedies ought we to adopt?
THE INJUSTICE OF OUR PRACTICES
Are our practices concerning the inclusion of women in clinical studies unjust? The answer, in a word, is yes. I shall devote this section to an analysis of why they are unjust.
To answer this question, we need some understanding of what justice requires. Contemporary philosophers typically embrace what I'll call "the distributive paradigm of justice."17 That is, they typically define justice as the proper distribution of benefits and burdens among individuals in a society. Then, as one would imagine, they disagree about what makes such a distribution "proper."
In her recent book, Justice and the Politics of Difference, Iris Marion Young argues that the distributive paradigm of justice fails to capture all there is to justice. In the first place, she contends, not all of the concerns of justice are matters of the distribution of benefits and burdens. Oppression qualifies as a concern of justice—indeed, justice requires that we eliminate oppression—but some important aspects of oppression are not purely matters of distribution. (I'll talk more about this later.) In the second place, Young notes, the distributive
paradigm considers only how social arrangements affect individuals as such. She argues that such an exclusive focus on individuals fails to capture important aspects of justice. After all, people are oppressed not as individuals, but as members of groups. Therefore, Young concludes, we need a broader characterization of justice than the distributive paradigm provides, one that recognizes both distributive and nondistributive matters as concerns of justice and acknowledges the moral significance of both individuals and social groups.
I agree with Young about these matters, and so I propose that we accept the broader characterization of justice. Of course, this does not provide us with anything approaching a complete account of what justice requires. Still, it gives us enough of a basis for discussion so that we can achieve a reasonable understanding of the issues confronting us.
I suggest that we focus on particular considerations that might be offered in arguing for or against our practices concerning the inclusion of women in clinical studies and see why we must conclude, all things considered, that they are unjust. Let's begin by discussing considerations weighing against our practices concerning the inclusion of women. First, we must take into account the detrimental effects we're assuming these practices will have on women's health: the deaths, the disabilities, the illnesses, the suffering.18 Second, we must note how these negative health effects can result in further disadvantages to women, such as time lost from work and impaired ability to function in personal relationships. Third, we must see that these harms to women can negatively affect those who have relationships with the women who suffer—their employers or employees, their friends and lovers, their children. Even purely distributive models of justice would have to take all these concerns into account. And these concerns alone weigh heavily against our practices concerning the inclusion of women in clinical studies.
But the distributive considerations are compounded by nondistributive ones. We must recognize that it's no accident that women suffer the harms I've just listed. Rather, these harms are one result of the oppression of women in our society, and, accordingly, they become a part of that oppression. How are these harms the result of oppression? There's a lot one could say about this, but I would like to focus on how two particular aspects of oppression can give rise to these harms.
One central aspect of oppression concerns how groups, and individuals in those groups, are conceived of in society. Let me explain. Oppressive societies take the dominant group's identity and experience to be, not the particular identity or experience of one group in society, but the universal identity and experience—that is, the norm.19 I shall call this aspect of oppression "false universalism." The upshot for our discussion here is this: Our society does not conceive of men in terms of their gender; it conceives of them gender-neutrally, as persons. Thus, men's identity and experience serves, in effect, as the
characterization or standard of what it is to be a person. This is true only of men's identity and experience; it is not true of women's.
According to sociologists Kimmel and Messner, there is a sociological explanation for this conceptual phenomenon. "[T]he mechanisms that afford us privilege are very often invisible to us. . . . Thus, white people rarely think of themselves as 'raced' people, rarely think of race as a central element in their experience."20 (Hence we use the phrase "people of color" to refer to everyone but white people.) According to Kimmel and Messner, the same point holds for gender. That is, men rarely think of themselves as gendered persons, rarely think of gender as a central element in their experience. Indeed, Kimmel and Messner suggest that white men, when asked what kind of being they see when they look at themselves in the mirror, tend to respond ''a person" or "a human" (race- and gender-neutral), but white women tend to respond "a woman" (race-neutral but gender-specific), and black women tend to respond ''a black woman" (race- and gender-specific).21
Our use of language further bears out the point that our society takes men's identity and experience as the characterization or standard of what it is to be a person. After all, the way we talk reflects the way we think. In our language, masculine terms such as "he" and "man" serve as gender-neutral terms; feminine terms such as "she" and "woman" cannot.22 Also, we qualify gender-neutral occupation names when women, but not when men, hold the occupation in question. For example, we speak of "women doctors," "women professors," and "lady cops" but not "male doctors," "men professors," or "gentleman cops." Her gender is relevant; his is not. She is conceived of in gendered terms; he is not. His identity and experience serves as the characterization or standard of what it is to be a doctor, professor, police officer—or even a person.
False universalism has two consequences. First, it makes women invisible in the following sense. When we think "person," we tend to think not of women but of our paradigm persons—men. Women tend to disappear from the conceptual scene when we're thinking in gender-neutral terms, because we tend to think of women in terms of their gender. On the other hand, men tend not to disappear, because we tend not to conceive of men in gender-specific terms but as the standard of what it is to be a person. The second consequence of false universalism is that, insofar as they are different from men, women are conceived of as inferior, deficient, or deviant. If they were not, if their differences were viewed as mere differences, this would challenge men's claim to universality, their ability to serve as the standard of what it is to be a person.
We should note that while false universalism qualifies as a concern of justice, it is not purely a matter of the distribution of benefits and burdens among individuals in society. Rather, this aspect of oppression is primarily a matter of our conceiving of men in gender-neutral terms, as the norm of personhood, and of women in gendered terms, as deviant from this norm. More generally, then,
false universalism is primarily a matter of how we conceive of women and men in our society. How we conceive of men and women does have some distributive consequences, e.g., the allocation of resources for studies of men's or women's health concerns, but how we conceive of men and women cannot be satisfactorily analyzed as a matter of the distribution of benefits and burdens. Thus, a purely distributive model of justice could not fully account for this concern of justice; to do so, we need the sort of broader model Young provides.
A second important aspect of oppression is that society subordinates members of oppressed groups to members of dominant groups.23 Domination and subordination are typically analyzed as imbalances of power. However, this is too simplistic an account of dominance and subordination.
In the first place, such an account fails to recognize the source of whatever power imbalances are involved in dominance and subordination. These power imbalances arise from socially prescribed norms. Let me explain. Society establishes norms that apply to individuals on the basis of their membership in certain groups. What these social norms do is prescribe what members of particular groups should be like. Thus, these norms fix (parts of) our social conceptions of members of particular groups. For example, gender norms prescribe what roles and responsibilities women should assume, how women should look, walk, talk, sit, interact with others, and much more. In general, they provide us with a picture of what a "real" or a "good" woman is like. That is, gender norms specify how we are to conceive of men and women in our society.24 Moreover, the social reality of who gets how much and what kinds of power is built into these social conceptions of groups. Perhaps the most obvious example of this connection between power and norms is this: In many societies, "good" women are not aggressive or ambitious, though ''good" men are. These gender norms condemn women, but praise men, for possessing traits that in many ways give individuals power. Therefore, any analysis of the social reality about power relations must be given in terms of the social conception of groups.
Further, it is a mistake to think that all dominance and subordination is a matter of imbalances of power. Sometimes—indeed, probably far more often than not—women's subordination to men is a matter of women's interests being taken less seriously than are men's, or of the arenas in which women have power (indeed, more power than men have) being valued less than those in which men have power. For example, in our society, women have primary responsibility for domestic concerns; thus, in day-to-day matters, women often have more power in the domestic sphere than men do. (Indeed, women often almost single-handedly run the home.) Men, on the other hand, (still) have more power than women in the world of paid employment and in politics. But work and power in the public sphere (paid employment, politics) are highly valued in our society, whereas work and power in the domestic sphere are not.
We should note that, on the standard account of dominance and
subordination, it appears that we can analyze this aspect of oppression on a purely distributive model of justice. Assuming we can speak meaningfully of the distribution of abstract goods like power,25 the standard account says simply that society allocates power (or certain kinds or amounts of power) to members of some groups but not to others. I have just argued that this account of dominance and subordination is overly simplistic. On the more realistic account I have just sketched, dominance and subordination cannot be analyzed in purely distributive terms. There are two reasons for this. First, on my account, dominance and subordination is sometimes a matter of power imbalances. However, I have argued that the social reality of who gets how much and what kinds of power is built into the social conceptions of groups provided by our social norms. Since, as we have seen, we cannot give a purely—even a primarily—distributive account of our social conceptions of groups, and since an analysis of domination and subordination must be given in terms of an account of our social conception of groups, we cannot give a purely—even a primarily—distributive account of dominance and subordination. Second, on my account, sometimes dominance and subordination is not a matter of imbalances of power, but instead a matter of women's interests being taken less seriously than men's are, or of the arenas in which women have power (indeed, more power than men have) being valued less than those in which men have power. These aspects of dominance and subordination cannot satisfactorily be analyzed in terms of the distribution of benefits and burdens to individuals in society. For these two reasons, purely distributive models of justice could not fully account for domination and subordination, even though it is a concern of justice. To fully account for dominance and subordination, we need the sort of broader model Young provides.
Now what does all this discussion of oppression have to do with the harms we've decided women suffer as a result of our practices concerning their inclusion in clinical studies? Lots. Remember that my claim is that women suffer these harms because given our practices, we fail to incorporate a concern for women's health into our research agenda, and that we fail to do so because of women's oppressed place in our society. I'd like to suggest three links between women's oppression and our practices concerning their inclusion in clinical studies. First, we've seen that oppression makes women invisible. Acknowledging this allows us to come to an important realization: we can abhor oppression without necessarily vilifying men. In the case at hand, we need not assume that members of the medical establishment sit around consciously thinking, "We're men, and we have the power to determine what will be studied and what will not. We will promote research on men's health issues, and, for the most part, ignore women's health issues." Instead, we can see how the oppression of women creates in men a conceptual blindness to the special needs of women.
Second, we've seen that oppression makes women appear deviant or
problematic. We know that one of the reasons that women have traditionally been excluded from clinical studies of conditions that affect both men and women is that, as The Washington Post put it, "their hormonal fluctuations [have been] said to 'confound' or confuse research results."26 Women's cycles appear not only to be different from men's physiology, but also to be problematic. Because men are conceived of in gender-neutral terms as paradigmatic persons, researchers too often feel they can simply avoid the "problems" caused by women's cycles by studying only men. (I'll have more to say about this below.)
Third, we've seen that oppression imposes gender norms on women which subordinate women to men. These gender norms can dictate what research will be done and what will not.27 For example, consider research on birth control methods. By far most of this research focuses on birth control methods for women. In our society, social norms place primary responsibility for birth control on women; this norm is usually understood as developing out of the norms that give women primary responsibility for children. These norms contribute to women's subordinate status in society by demanding that women devote themselves to child care—undervalued labor in our society—while men are freed to advance their status in more highly valued activities, and to pursue leisure activities. The birth control research agenda based on these norms also subordinates women's interests to men's. Women must bear the costs of birth control: the inconvenience of acquiring it, which often requires a visit to her doctor for a prescription; the monetary cost of purchasing it; the psychological costs of bearing responsibility for its proper use, which are not inconsiderable especially since our gender norms frown on women who plan for sex; the physical costs of using it, from discomfort to serious health risks. Men needn't bear these costs, but they do reap the benefits of women's sacrifices.
Consider another example of how gender norms can influence research agendas. In our society women's primary sex roles are those of wife and mother. These roles define women in terms of their relations to men and children, they require that women provide service to their men and children, and they take priority over the woman's other roles-e.g., of career person, or of caretaker for herself. Thus, these roles subordinate women to others. Society's conception of women in terms of their relations to men and children influences research agendas. For example, we've seen that as far as AIDS research goes, women have been primarily studied as "vectors" of the disease. That is, research on women concentrates on how women transmit AIDS to their male sex partners and their children. In contrast, little research has been done on how AIDS affects women themselves.28 Here women's interests are subordinated to those of their men and children.
Thus we can see that our practices concerning the inclusion of women in studies and the harms attendant upon these practices are one result of the oppression of women in our society, and, accordingly, they become a part of that
oppression. Since justice requires the elimination of oppression, this serves as an especially powerful consideration weighing against our practices concerning the inclusion of women in clinical studies. Notice that this consideration would weigh against our practices even if they did not have a detrimental impact on women's health. Insofar as these practices result from and are a part of women's oppression, they must be remedied out of a moral concern for the elimination of oppression, no matter what additional harms are related to the practices.
What considerations purport to weigh in favor of our practices concerning the inclusion of women in clinical studies? I shall focus on three such types of considerations. First, we must take into account the role played by considerations of cost in attempts to justify these practices. There's no doubt that including women in sufficient numbers and doing gender-specific analysis would increase the costs of studies. There's no doubt that limits in our resources force us to face difficult decisions about how to allocate those resources. However, it is morally unacceptable to allocate resources on the basis of sex (or, for that matter, other group membership, such as race, sexual preference, etc.) when there is so much at stake for the excluded parties. It would be outrageously immoral to manage the costs of education by excluding, say, black children from our educational system. Likewise, we must conclude that considerations of cost cannot justify failing to include women in clinical studies in sufficient numbers or failing to perform gender-specific analysis.
Second, as we've seen, researchers often try to justify excluding women from studies of conditions that affect both men and women by appealing to the claim that women's hormonal fluctuations "'confound' or confuse research results." However, as we've also seen, this purported justification takes women's cycles not only to be different from men's physiology, but also to be problematic. Hence this purported justification must be rejected because of its connection to the first aspect of oppression I discussed above—false universalism. Moreover, it would be appropriate to strive for clean, uniform data which can be analyzed using simple, elegant models only if such research best served the health care needs of all persons. But we know that it does not, since we know that some diseases present themselves differently in women than in men, and that some treatments affect women differently than men. Therefore, we have ample reason to reject this purported justification.
Third, it is sometimes argued that we should exclude women of childbearing age from clinical studies to protect the well-being of possible or actual fetuses. A full discussion of this issue would require a paper—or a book—of its own. Here I shall briefly make five points: (1) It is not morally acceptable to place a higher value on fetal life and well-being than on women's lives and well-being. If we refuse to study women of childbearing age because we are concerned for the well-being of (possible or actual) fetuses, regardless of the health benefits to women such research would yield, then we place a higher value on fetal life and well-being than on women's lives and well-being. Current
regulations require researchers to consider the impact of research only on individual women actually involved in the research and on (actual or possible) fetuses affected by the research. In the cases of both actually pregnant women and women of childbearing potential who are not pregnant, regulations deny (or, in the case of nonpregnant women, have until recently denied) women participation in studies unless the purpose of the research is to meet the significant health needs of the individual women in the study.29 The impact research would have on the health of women as a group (not just those involved in the study) is ignored. If we omit from our risk-benefit analysis, or refuse to weigh, the possible benefits the research would have for women's health generally, then we fail to place sufficient value on the health of women in general30—we value fetal life and well-being more highly than women's lives and well-being. We cannot be morally justified in doing so. (2) Assuming that we have a legitimate moral interest (at least in some cases) in protecting fetal well-being, we should strive to find a way to safeguard (possible or actual) fetuses while conducting research on women. We should not simply settle for excluding women from research. (3) No matter what view we take about the value of fetal life and well-being, this consideration cannot justify excluding from clinical studies women who cannot or will not have children.31 (4) If we are concerned with fetal life and well-being, then we must take care not only with how we conduct research on women, but also with how we conduct research on men. We have reason to believe that fetal health and well-being is affected not only by agents to which women are exposed, but also by those to which men are exposed. To exclude women but not men from research out a concern for fetal health (as current regulations do) is to discriminate unjustly against women.32(5) Comments (1)–(4) discuss moral considerations involved in the exclusion of women from research out of a concern with fetal health. Some researchers wish to exclude women not so much out of a moral concern for the health of the fetus, but out of a concern for avoiding legal liability for the possible harms fetuses might suffer as a result of research. I cannot discuss liability laws at any length here. I shall say only that our liability laws should be consistent with our moral views on this matter.
It should be clear by now why our practices concerning the inclusion of women in clinical studies are unjust. We now understand the considerations arguing for and against these practices, and there can be no doubt that the considerations against these practices vastly outweigh those supporting them.
Now that we have a reasonable understanding of why our practices concerning the inclusion of women are unjust, we can turn our attention to what we should do to remedy this injustice. It should be clear by now that we must do something; justice cannot tolerate the preservation of the status quo. Where these injustices occur, they must be remedied. The question is: What kind of new policy should we adopt? Of course, any policy we adopt must require more than mere head counting. That is, simply adding more women to studies will not solve the injustices in question here. We must explore whether the sex of the subject is relevant to the condition or treatment being studied. We must not treat women's bodies as deviant or problematic compared to men's when we do so, nor should we treat our study of women as a mere means to gaining knowledge of the health of men and children. However, it is not enough for us to realize how we should proceed once we've included women in our studies. We must also determine how we should allocate the resources available to support research. It is this question that I shall focus on in this section of this paper.
Should we demand gender-neutral allocation of resources for research? In general, gender-neutral policies require us to ignore gender when we reason from the point of view of justice, and prohibit us from treating persons differently on the basis of their gender. Thus, such a policy would require nondiscrimination in the allocation of resources for research. With respect to studies of conditions affecting both men and women, it would require that women be included in studies in sufficient numbers. To exclude or underrepresent women would be to extend special treatment to men—i.e., desirable levels of inclusion in studies—which we would deny to women. With respect to studies of health concerns specific to women, a gender-neutral policy would require that resources for research be allocated on the basis of the same criteria we use to allocate resources for studies of men's health—e.g., how many people are affected by it, how much of a health threat it is.
Or should we insist upon something more than gender-neutrality? Should we adopt (at least temporarily) programs of preferential treatment which allocate a larger share of available resources to studies of women's health than would be allocated by a gender-neutral policy? I shall argue that we should, indeed, adopt a policy of preferential treatment as a remedy for the injustices involved in our practices concerning the inclusion of women in clinical studies. I shall begin by dispensing with the main argument against preferential treatment.
The most common argument against preferential treatment goes like this: "Programs involving preferential treatment discriminate in favor of members of certain groups. But all discrimination is unjust. Therefore, programs involving preferential treatment are unjust." We must reject this argument, since it depends upon an ambiguity in the meaning of the term "discrimination."
"Discrimination" can mean simply "the drawing of a distinction" or it can mean "the unjust drawing of a distinction.'' When the opponent of preferential treatment claims, in the first step of the argument, that such treatment discriminates, she cannot mean that it discriminates unjustly. That would be to assume what the argument was supposed to demonstrate. So, at this point in the argument, to say that preferential treatment discriminates is just to say that it draws distinctions. That, of course, is true. But in that case the next step in the argument—the claim that all discrimination is unjust—is false. Not all distinction drawing is unjust. The claim that all discrimination is unjust turns out to be true only if, at this point in the argument, "discrimination" means ''the unjust drawing of a distinction." But logic prohibits changing the meaning of key terms from one step in an argument to another. So either the argument assumes what it is supposed to establish, or it contains false assumptions, or it contains a prohibited shift in the meaning of a key term. Thus, we must reject the argument.
So we have no reason to believe that programs of preferential treatment are inherently unjust. But what reason do we have to believe that we would be justified in adopting such a program as a remedy for the injustices involved in our practices concerning the inclusion of women in clinical studies? I shall argue that we have plenty of reason to believe that we would be justified in adopting such a program. Indeed, I shall argue that justice not only permits but requires that we adopt such a program. Let's begin by considering what justice requires of a remedy for the injustices involved in the way we've included women in clinical studies.
Most generally, of course, a remedy must correct the injustices in question. In this case, this means it must address both the distributive and the nondistributive injustices. Because we traditionally have conceived of justice on the distributive paradigm, we are accustomed to focusing only on distributive injustices. But we must not lose sight of the nondistributive ones. Even if it were to turn out, contrary to all evidence and reasonable expectations, that our practices concerning the inclusion of women in clinical studies have had no detrimental impact on women's health whatsoever, we would still have to remedy serious, albeit nondistributive, injustices. For example, we would still have to counter the injustices involved in the way we conceive of women in our society, as reflected in our medical research. Even if all the data that we gather from men apply perfectly well to women, it still says something disturbing about how we conceive of men and women in this society if we continue to study only men: it says that men are the norm for persons, and that, when we're thinking in gender-neutral terms of persons, we needn't think of (or include) women.
In addition, a remedy must respond to a cross-temporal perspective on the moral problems in question.33 That is, the remedy must address the important realization that our practices concerning the inclusion of women in studies have
created justice problems in the past, continue to create such problems in the present (both because their use in the past has deprived us of knowledge crucial to the treatment of women now, and because the practices continue in the present), and will continue to create justice problems in the future unless something is done to solve all these problems.
I contend that a gender-neutral policy cannot correct the injustices involved in our practices concerning the inclusion of women in clinical studies. I shall outline three reasons why such a policy cannot correct those injustices.
The first two reasons concern the inability of the gender-neutral policy to respond to a cross-temporal perspective on the injustices in question. First, such a gender-neutral policy cannot correct for past or present injustices; it is completely future-oriented. When we confront the problem of our failure to incorporate a concern for women's health into our research agenda, we face a history of injustice (involving, as we have seen, both distributive and nondistributive injustices). A gender-neutral policy of nondiscrimination says, in effect, "We'll try to do better in the future, by resolving to incorporate a concern for women then." But simply resolving to do better in the future does nothing to address past or present injustices. True, we cannot resolve to do better in the future without acknowledging past or present injustices. But to acknowledge past or present injustices is not to make amends for them. And we must make amends; justice requires it. Suppose, for example, that there is a society that has a history of the following practice: white people (the dominant race by this society's standards) routinely steal the property of black people. Suppose further that this society sanctions, or at least does not condemn, this practice. Suppose finally that members of this society come to be persuaded that this practice is unjust, and that they decide to remedy the injustice by resolving not to allow whites to steal the property of blacks in the future. This example makes it clear that, when we are faced with a history of injustice, simply stopping the unjust practice, while important, is not sufficient from the point of view of justice. Justice requires that amends be made for past and present injustices. In our hypothetical example, justice requires, for example, that restitution be made to those who have had their property taken from them unjustly. A gender-neutral policy of nondiscrimination simply stops our problematic practices concerning the inclusion of women in clinical studies. Thus, it makes no provision for making amends for past or present injustices. Therefore, a gender-neutral policy of nondiscrimination does not correct for past or present injustices, and so it does not meet the demands of justice.
Opponents of preferential treatment sometimes argue that, while it makes sense to think that we're obligated to compensate individuals we have wronged, it makes no sense to think we're obligated to compensate certain individuals now for wrongs done by others to others in the past. While this objection may seem initially plausible, it cannot succeed in undermining my claim that justice
requires us to make amends for past and present injustices, for two reasons:
This objection cannot undercut my claim that we must make amends for present injustices; it addresses only past injustices (and, indeed, only those injustices done so far in the past that neither those who perpetrated them or benefitted from them nor those who suffered them are alive any longer).
This objection cannot subvert my claim that we must make amends for past injustices, for three reasons: (a) It fails to recognize that those past injustices, done by others against others (if committed far enough in the past) have present unjust consequences—for example, women now suffer from lack of appropriate health care as a result of our practices concerning the inclusion of women in studies in the past. (b) The objection equates making amends for past injustices with paying compensation for past injustices. But paying compensation is merely one way of making amends. This is a point that never seems to be recognized in discussions of affirmative action and preferential treatment. Consider the following case: Dr. Jones commits malpractice, and Samantha Smith dies as a result. Jones can justly be required to compensate the Smith family for the loss of Samantha's earnings. He cannot, however, justly be required to compensate the Smith family for the loss of Samantha. It is morally inappropriate to speak in terms of paying compensation here, because one cannot put a price on human life which one can then pay in exchange for taking the life. It is not at all inappropriate, however—indeed, quite the contrary—to talk of Dr. Jones making amends for his actions which led to Samantha's death. So even if, as the objection alleges, we ought not require certain individuals to pay compensation to other individuals now for injustices performed by others against others in the past (and even this is false, I think), it does not follow that we ought not require certain individuals to make amends to other individuals now for injustices performed by others against others in the past. (c) The objection loses sight of the moral importance of groups; it focuses solely on compensation (or, if we revise it in light of my previous point, making amends) to specific individuals who have suffered injustices. However, the broader conception of justice I have adopted in this paper opens up the possibility of thinking in terms of making amends to groups, not just to individuals. After all, individual women suffer the injustices involved in our research practices not because they are the individuals they are but because they are women. Furthermore, all women (just because they are women) risk the harms to health and other related harms I listed early on in my analysis of why these practices are unjust. And all women, just because they are women, suffer in some way from the oppression of women in our society (especially the aspects of oppression concerning how we conceive of women). Thus it makes perfect sense to speak of society making amends to women (as a group) for the injustices it has perpetrated against women (as a group).
The second reason why a gender-neutral policy cannot correct the injustices involved in our research practices is that such a policy, although it is completely future-oriented, cannot fully correct for future injustices. As I argued above, such a policy does serve as a resolution to incorporate a concern for the health of women into our research agenda in the future. And that is, admittedly, an important part of correcting for the future injustices in question. But it is not all there is to it. Even if we were to incorporate a concern for women's health into our research agenda immediately, our history of failing to do so would continue to affect people in this society for a long, long time to come. For example, men would continue to receive better health care than they would have received in a truly just society, because a disproportionate share of our resources have for so long gone to the study of men's health. On the other hand, women would continue to be the victims of these inequalities in the availability and quality of care. Adopting a gender-neutral policy of nondiscrimination might well stop those inequalities from broadening, but it would do nothing to narrow them. Justice requires that we close those gaps; since these inequalities are unjust, allowing them to continue is unjust. Since a gender-neutral policy of nondiscrimination would allow them to continue, it cannot correct for all the future injustices involved in our history of failing to incorporate a concern for women's health into our research agenda.
The third reason why a gender-neutral policy cannot correct for the injustices involved in our practices concerning the inclusion of women in clinical studies is that gender-neutrality provides an unacceptable model of what constitutes justice. Gender-neutral policies are premised on the idea that a person's gender is irrelevant from the point of view of justice. That is, they are based on the view that all persons should be treated equally, regardless of their gender. Thus, they require us to ignore gender when we reason from the point of view of justice, and they prohibit us from treating persons differently on the basis of their gender. Such a model of justice fails to incorporate a sufficient and appropriate moral sensitivity to our social context. We live in a society with a history of gender-based oppression. Gender-neutral models do condemn oppression as they understand it; they contend that it is unjust for women to be treated differently from men on the basis of their gender. However, such models of justice do not recognize that our history of gender-based oppression makes gender relevant from the point of view of justice.34
A concern with justice demands that we strive to recognize, to understand, and to overcome oppression. So, as we have seen, we must concern ourselves with (among other things) the intricacies of our oppressive social conception of women, and with the multifarious ways this social conception has affected, and continues to affect, women. We must resist and revise our oppressive conception of women. We must make amends for our history of oppression. We must eliminate gender-based inequalities that oppression has created. To accomplish
all of this, we must attend to gender when we reason from the point of view of justice. None of this is compatible with a model of justice that requires us to ignore gender in our justice reasoning and prohibits us from treating persons differently on the basis of their gender. For example, we could not make amends to women for our history of oppressing them (with all the complex varieties of injustices that involves) if we did not attend to gender in our justice reasoning and extend special treatment to women that we denied to men; quite simply, in our society, women are in a position that calls for the making of amends, and men, as such, are not.35 Also, we could not eliminate existing gender-based inequalities if we did not attend to gender in our justice reasoning and extend special treatment to women that we denied to men; again, in our society, women are in a social position that calls for their being advanced to a point of equity with men, and men, as such, are not in such a position vis-a-vis women. In general, since we live in a society with a history of gender-based oppression, adopting a stance of gender-neutrality blinds us to—forces us to ignore—issues that are relevant from the point of view of justice, and prohibits us from correcting the injustices of oppression since doing so necessitates that we extend special treatment to women which we deny to men.36
We have seen that there are three reasons why a gender-neutral policy of nondiscrimination cannot meet the demands of justice: (1) it cannot correct for past and present injustices; (2) it cannot correct for future injustices; and (3) given the context of our oppressive society, gender-neutrality provides an unacceptable model of what constitutes justice. For all these reasons, we must reject such gender-neutral policies as remedies for the injustices involved in the exclusion of women from clinical studies.
Programs of preferential treatment, on the other hand, can meet the demands of justice. Recall that, most generally, justice demands that a remedy correct the injustices in question. More specifically, justice first requires that a remedy address both the distributive and the nondistributive injustices involved in our practices concerning the inclusion of women in clinical studies. I have argued that, even if these practices had no detrimental impact on women's health, justice would still require us to adopt a remedy that could counter the nondistributive injustices involved in our practices: the way we conceive of women, and the subordinate weight we give to their interests or value we attach to the arenas in which they have power, as these things are reflected in our medical research. A policy of preferential treatment would counter these injustices in two ways. First, by requiring us to incorporate a concern for women's health into our research agenda, it would help undermine the subordination of women and the false universalist view that men are the norm for persons and that, when we're thinking in gender-neutral terms of persons, we needn't think of (or include) women. Second, by requiring us to extend special treatment to women, it would not merely reject these aspects of women's oppression, it would also make amends for our history of oppression (something
a gender-neutral policy could not do).
Second, justice requires that a remedy respond to a cross-temporal perspective on the moral problems in question. I have argued that gender-neutral policies fail to meet this demand of justice. Unlike such policies, policies of preferential treatment can meet this demand. Unlike gender-neutral policies, policies of preferential treatment can make amends for past and present injustices by offering special treatment to women—that is, by not only resolving to incorporate a concern for women's health into our research agenda in the future, but also going the extra distance to make up for the various (distributive and nondistributive) injustices of the past and present. Also, unlike gender-neutral policies, policies of preferential treatment can fully correct for future injustices. That is, not only can it halt our problematic practices concerning the inclusion of women in studies, but by allocating extra resources to the study of women's health, it can work to close the gaps between men's and women's health care, to eliminate the inequalities in quality and availability of care that our history of these problematic practices has created.
Finally, I have argued that gender-neutral policies fail to meet the demands of justice because gender-neutrality provides an unacceptable model of what constitutes justice. Unlike such policies, programs of preferential treatment recognize that our history of gender-based oppression makes gender relevant from the point of view of justice. Unlike gender-neutral policies, programs of preferential treatment allow—even require—us to take gender into account when we reason from the point of view of justice, and permit—even require—us to extend special treatment to women that we deny to men. As I have just shown, this insures that such programs can make amends for past and present injustices and can eliminate existing inequalities between men's and women's health care, as justice requires a remedy to do.
In summary, then, gender-neutral policies fail to meet the demands of justice, while policies of preferential treatment succeed in doing so. Therefore, justice requires that we adopt a policy of preferential treatment in the allocation of resources for research, to remedy the injustices involved in the our practices concerning the inclusion of women in clinical studies.
Doubtless, critics of preferential treatment will respond to my conclusion here by complaining, as they often do, that it seems impossible for us to know for sure when we have done enough—that is, when programs of preferential treatment have been in place sufficiently long to complete their tasks of making amends for past and present injustices37 and correcting for existing inequalities so they don't continue into the future.38 These critics have a point; it will, indeed, be difficult to determine when programs of preferential treatment have achieved what they were designed to accomplish. However, we should not think that this difficulty undermines the acceptability of programs of preferential treatment. On the contrary, it is inappropriate for us to demand precise answers to questions about justice. As Aristotle says,
Our discussion will be adequate if it achieves clarity within the limits of the subject matter. For precision cannot be expected in the treatment of all subjects alike. . . . Problems of what is noble and just, which politics examines, present so much variety and irregularity that . . . we must be satisfied to indicate the truth with a rough and general sketch.39
Furthermore, to object to the adoption of programs of preferential treatment on the grounds that we won't know exactly when those programs have completed their mission is to send the wrong message to society, especially to members of the group that has suffered the injustices in question.40 It is to say, "I know we cannot determine when we have doled out enough preferential treatment to correct for the injustices in question. So if we do anything, we risk doing too much. I'd rather refuse to remedy the injustices under discussion than risk depriving the dominant group in society of any more goods than is absolutely necessary." Justice simply cannot tolerate this kind of attitude. We must simply accept that precise answers cannot be given, and do the best we can to monitor the moral progress our remedies allow us to achieve.
Finally, we must address the question of who should be held responsible for making the changes necessary for securing justice for women. Of course, we cannot make these changes without the cooperation of everyone involved. For example, researchers must take the steps necessary to include women in their studies, to investigate whether the sex of the research subject is relevant to the condition or treatment being studied, and to avoid treating women's bodies as deviant or problematic compared to men's. Those who train researchers must instruct them about women's health needs, and should encourage more women to become researchers, since women researchers are likely to be more sensitive than men to women's health needs.
Ultimately, however, we must do more than call upon the cooperation of those involved in the pursuit of knowledge about women's health. The appropriate federal agencies must adopt regulations that implement the dictates of justice as I have outlined them in this paper. I cannot give an exhaustive summary of these dictates and their correlative regulations here, but they include the following: Our regulations must require researchers not only to include women in studies but to investigate whether the sex of the research subject is relevant to the condition or treatment being studied. They should encourage the study of women's health issues. They must require the presumption of the inclusion, not the exclusion, of women of childbearing age and pregnant women in studies. They must take the benefits research is likely to have for the health of women as a group (not just the individuals involved in the studies) into account in the risk-benefit assessments done to assess the merits of research proposals. They must insure that policies concerned with the protection of fetal health treat women and men consistently, and do not discriminate unjustly
against women. And they must require the preferential treatment of women's health issues in the allocation of resources for research. I recognize that it will be politically difficult to secure the passage of such regulations. Nevertheless, justice demands that we do.41