Recruitment and Retention of Women in Clinical Studies: Ethical Considerations
Robert J. Levine
INCENTIVES TO ENROLL AND TO CONTINUE PARTICIPATION
The first principle of the Nuremberg Code reads, in part: ". . . the person . . . should be so situated as to be able to exercise free power of choice, without the intervention of any element of force, fraud, deceit, duress, overreaching, or other ulterior form of constraint or coercion. . . ."
Of primary concern is that incentives offered to prospective subjects may be so great as to become an "ulterior form of constraint" which will overcome or overwhelm the individual's capacity to exercise free power of choice; such incentives are commonly called "undue inducements."
In the literature on inducements or incentives, some are usually considered unproblematic. These include (1) altruism; (2) reimbursement for out-of-pocket expenses such as transportation and daycare; and (3) free investigational drugs and other medical goods and services.
There are some who consider the problem of reimbursements more difficult or complicated for women than it is for men. Women are more likely (so it is argued) to require transportation, babysitters, daycare, and other services for which reimbursements are required. In this regard it is worth noticing that men may also require reimbursements for daycare, transportation, and babysitters.
In my view, reimbursements for out-of-pocket expenses are ethically unproblematic. They do not represent constraints in that the recipients do no better than break even; financially, their position is neither improved nor diminished.
Free investigational drugs and other medical goods and services, although
usually considered ethically unproblematic, may be much more powerful inducements than cash payments. The value of free medical goods and services can be overwhelming, particularly for poor people. This is equally true for men and women.
Cash payments are usually considered at least potentially problematic even though their actual value may be substantially less than that of free medical goods and services.
In order to gain some perspective on this issue, let us consider a protocol that was approved by several institutional review boards (IRBs). The protocol consisted of a placebo-controlled, randomized clinical trial of a new "me-too" antihypertensive drug. The subjects were called upon to go through a 2-week "placebo-washout" period followed by 26 weeks of taking either active drug or placebo. Subjects were to be paid $1500 at the conclusion of the trial. In addition, they were to receive free "drug" and free clinic visits each week.
This case is not atypical; many clinical trials offer similar inducements and they are customarily approved by IRBs. I want to make two general points about the inducements in this protocol. Firstly, the inducements offered to participate in this clinical trial make our arguments about the propriety of reimbursing subjects for out-of-pocket expenses seem rather trivial. Secondly, it must be acknowledged that the purpose of this payment is to induce prospective subjects to do what they would not otherwise do. There is nothing necessarily problematic about this. All of us are paid to do things that we would not otherwise do. Payments become problematic only when they might overwhelm a prospective subject's judgment about whether she or he wishes to participate.
In general, IRBs do not seem to worry very much about cash payments or other material inducements in situations in which the risk is acceptably low or in which the risks appear to be in reasonable proportion to the anticipated direct health-related benefits to the subject. In such cases, there is a tendency to let market factors determine the amount of payments. By contrast, where the risk is high and there is no reasonable expectation of proportionately high direct health-related benefits to subjects, IRBs tend to aspire to the ideal of altruism by keeping payments low.
In the category of incentives to enroll or to continue participation, I am not aware of any considerations that apply peculiarly to women that do not apply equally to men.
FREE ABORTION AS AN INCENTIVE
One of the reasons that women are often excluded from clinical trials of new drugs is concern that during the course of the clinical trial they might become pregnant and that the fetus could be deformed. In short, there is great
anxiety about repeating the thalidomide experience.
This raises some questions: Is it permissible for a sponsor to offer a free abortion to women who become pregnant? If so, is it permissible to exclude women who would not freely choose to have an abortion if they became pregnant during the course of a clinical trial?
I believe that our responses to such questions must take into account several features of the clinical trial. Following are attributes of clinical trials listed in descending order of their providing ethical justification for responding ''yes" to one or both of the two questions. (In general, the justification for responding "yes" to the second question must be stronger than that required for a "yes" response to the first question.)
The purpose of the clinical trial is to evaluate a new contraceptive or contragestive agent. That is, the subjects enrolled with the expectation that they would avoid or terminate pregnancy.
The agent being evaluated is known to be teratogenic in humans.
Preclinical tests of the agent being evaluated suggest that it might be teratogenic in humans.
Preclinical tests for mutagenicity and teratogenicity have not been done.
Preclinical tests show no evidence of mutagenicity or teratogenicity.
But what, one might ask, if a drug known to be teratogenic in humans offers the only possible means of arresting or delaying the progress of a lethal or disabling disease and a woman, fully informed of the risks, says she wants the drug, but cannot accept abortion as a matter of conscience. I would make it available to her if she agreed to use a highly effective means of contraception and if she agreed to assume all of the burdens of giving birth to an impaired child, should that occur. In other words, I do not believe coerced abortion can be justified. (Even if a woman agreed to assume all of the burdens of giving birth to an impaired child, I would not hold the child accountable for such a decision. Such a child should be offered financial assistance as necessary for any special needs [e.g., special education] he or she might have as a consequence of the impairment.)
DEALING WITH DROPOUTS
The passage in Department of Health and Human Services regulations that speaks to this issue is Section 46.116(a)(8) of Title 45 of the Code of Federal Regulations: ". . . in seeking informed consent the following information shall be provided to each subject. . . . A statement that participation is voluntary . . . and the subject may discontinue participation at any time without penalty or loss of benefits to which the subject is otherwise entitled."
The philosopher Lisa Newton argues that the completely unfettered freedom
to withdraw which is required by ethical codes and regulations "is an anomaly in ethics, since it appears to be . . . in direct conflict with [one's] ordinary duty to keep [one's] promises" (Newton, 1984). She recommends that future formulations of regulations should recognize that the relationship between researcher and subject is "binding on both sides, hence to be taken very seriously on both sides, and go on to specify the circumstances that shall be taken to negate or cancel that mutual commitment."
IRBs seem to be acting as if they partially agree with Newton as they routinely approve plans to withhold part or all of cash payments from subjects who withdraw without having some plausible excuse such as an adverse reaction to a drug.
What about "quiet dropouts"—those who simply miss appointments without ever asserting their right to withdraw without penalty? Is it permissible to contact them (e.g., by repeated telephone calls) to try to persuade them to return?
The answer, I believe, is "yes." I believe that it is important to anticipate such behaviors in the informed consent process. Prospective subjects should be made aware of plans to contact them because some may consider such attempts unwelcome intrusions or threats to their privacy.
In some clinical trials such plans have been rather elaborate. For example, in the Multiple Risk Factor Intervention Trial protocol, the investigators enlisted the aid of an investigative services firm in locating subjects with whom they had lost contact for 12 months or more. After attempts to locate the subjects through "local resources and Social Security" failed, they provided the investigative services firm with the subject's name, last known address and telephone number, place of employment and work phone, and names and addresses of persons indicated by the subject as those likely to know his whereabouts. These procedures were publicized widely without adverse criticism.
Some commentators have expressed concern that women may be "less empowered" by their nature to assert their rights to withdraw from research, particularly given their greater (than men's) proclivities to value and maintain relationships. This could be problematic in some types of clinical research. This is a potential problem that should be called to the attention of IRBs and investigators.
I would like to call attention to another likely problem—that in some randomized clinical trials women might be inclined to either drop out or engage in covert noncompliance. Let us consider, for example, the Breast Cancer Prevention Trial (BCPT), a placebo-controlled trial of tamoxifen for the prevention of breast cancer. Newspapers around the country have popularized the debate over the so-called "serious" adverse effects of tamoxifen such as induction of cancer or life-threatening liver damage. But they have reported little or nothing about the so-called "minor" side effects which, in my view, are much more likely to undermine the validity of BCPT.
Tamoxifen, in the dose prescribed in BCPT, produces a relatively high rate
of unpleasant side effects such as vaginal dryness and hot flashes. Many women may reconsider their commitment to endure such discomforts a full 5 years, particularly if they recall and believe the statements made to them during the informed consent process about the initial null hypothesis or state of "clinical equipoise" that justified beginning the randomized clinical trial. One such woman might ask, "Does it make sense for me to put up with hot flashes and vaginal dryness for 5 years when the doctors say that tamoxifen is not known to be more or less beneficial than placebo which, in turn, is the same as nothing?" If she decides that it does not make sense, she may either drop out of the trial or engage in covert noncompliance, skipping some or most of her doses of tamoxifen.
But, one might ask, is that not what the process of informed consent is designed to prevent? Can we not anticipate that women who are unwilling to accept a chance of these symptoms will refuse to enroll in the study? Some will, of course, but there is a great difference between the acceptance of a statistical chance of the occurrence of symptoms by a women who has never experienced them and the subsequent lived reality of hot flashes and vaginal dryness.
If there is a high rate of dropouts or covert noncompliance, this will seriously undermine the validity of this clinical trial. Perhaps this should alert us to the need to involve in the design of such clinical trials people who are sensitive to such possibilities.
COSTS OF INVOLVING WOMEN AS RESEARCH SUBJECTS
Some sponsors of research, both industrial and governmental, have claimed that inclusion of women as research subjects raises the costs of doing research to the extent that their exclusion is justified on ethical grounds. Such sponsors hold that there is an unacceptably high cost-benefit ratio.
One cannot deny categorically the validity of cost-benefit considerations in determining whether certain research projects should be done. There are some interesting bits of knowledge that are simply not worth the expense of their acquisition. In general, however, I would deny the validity of such considerations as a justification for excluding women from most research projects.
Women have been excluded from studies of new drugs owing to concern that there might be damage to their fetuses should they become pregnant. One cannot deny the validity of this concern. As a partial response to this concern, I favor the development of no-fault compensation systems for injured research subjects. This does not mitigate the grief of bearing or being a handicapped child. It would, however, relieve the at times overwhelming financial consequences for both sponsors and subjects.
It is very unlikely that we will ever see another "thalidomide." Tests that
would have detected the problem caused by thalidomide are now done routinely in the preclinical phase of drug development. But let us consider a hypothetical case. Consider a new drug that, like thalidomide, caused a severe congenital anomaly by a mechanism that could not be detected by preclinical testing. Let us further assume that, like thalidomide, the anomaly it caused also occurred "naturally" in a small percentage of babies whose mothers did not take the new drug.
If the first administration of this drug to women who could and did become pregnant were conducted in the context of a carefully monitored clinical trial, the association between the drug and the anomaly could be detected after the occurrence of only a small number of anomalies. If instead, the first administration of this drug to women who could and did become pregnant took place in the relatively unmonitored context of clinical practice, then, as with thalidomide, the association between the drug and the anomaly would not be established until hundreds or thousands of babies were born with the anomaly. As with thalidomide, each obstetrician would believe the anomaly she or he observed was one of the unusual cases that occurred "naturally."
Women also are excluded from "basic research" designed not to develop drugs but rather to understand basic physiological processes. They are excluded because phasic variations in their physiological processes make such research difficult and more costly. It is more costly because larger numbers of subjects must be studied. Such exclusion is also unjust. Women are entitled to have basic information about their physiological processes studied and understood. The "norms" established by doing research on men should no longer be considered the normal . Data derived from studies of healthy women also reflect a normality, albeit, in some cases, a more complicated normality. Moreover, such studies often become the bases for subsequent development of diagnostic, prophylactic, and therapeutic products.
TWO SEPARATE AGENDAS ON THE INCLUSION OF WOMEN
Considerations of distributive justice provide the ethical grounding for arguments that women ought to be included as subjects in research. I want to point out that there are two very distinct claims, each grounded in distributive justice, and that the two claims should give rise to two very different responses.
As a matter of fairness, women as individuals ought not to be excluded from research and its associated direct benefits.
As a matter of fairness, women as a class of persons ought not to be deprived of the benefits of research, generalizable knowledge, and the development of new therapies.
The first claim is related to a shift in public opinion about research. In the 1960s and 1970s the prevailing image of research was that it was a burden from which persons would want to be protected. Research was seen as dangerous and exploitative. In the 1980s, largely in response to the highly articulate and effective AIDS activists, research has come to be seen as beneficial. It is a context in which subjects can receive ''promising new therapies" and medical care that is both excellent and free of charge. Hence, the argument that nobody should be deprived arbitrarily of access to such benefits.
A response to the first claim would be to open the doors to all persons, including women, to all research projects. The result would be the inclusion of varying numbers of men and women in studies. While this would satisfy the claims of equal access to putative benefits, it would not necessarily result in the development of either knowledge or products that were especially relevant to either men or women. At least theoretically, such an open door policy might make it much more difficult to develop knowledge or products especially relevant to either gender. For example, a drug that was more effective than placebo in women but not in men could, in a clinical trial involving both men and women, be found no better than placebo when the results obtained from both men and women were averaged together.
Now let us consider the second argument: that women as a class of persons ought not to be deprived of the benefits of research, generalizable knowledge, and the development of new therapies. A response to this claim does not entail establishing an open door policy. Rather, it involves the conduct of studies involving as subjects only men and only women and, in some cases, men and women in a design intended to compare them in a formal way. For example, one might develop a placebo-controlled, randomized clinical trial of a new drug in which an equal number of men and women would be enrolled and subjects would be stratified according to gender.
I think it is very important to be attentive to the differences in these two justice-based claims so that we can proceed intelligently to develop our research policies. The current National Institutes of Health policy calling for involvement of women and minorities in research unless their exclusion can be justified seems to be responsive only to the first of these two claims.
WOMEN IN (SOME) DEVELOPING COUNTRIES
Finally, I want to comment on a problem that we in the United States will find inescapable as it becomes necessary to conduct some studies involving subjects in developing countries. Time does not permit my providing an account of the intense arguments surrounding this issue. For the purpose of starting a discussion, I shall just read one paragraph from the recently published Council for International Organizations of Medical Sciences (CIOMS) International
Ethical Guidelines for Biomedical Research Involving Human Subjects (CIOMS, 1993):
Obtaining the informed consent of women, including those who are pregnant or nursing, usually presents no special problems. In some cultures, however, women's rights to exercise self-determination and thus give valid informed consent are not acknowledged. In such cases, women should not normally be involved in research for which societies that recognize these rights require informed consent. Nevertheless, women who have serious illnesses or who are at risk of developing such illnesses should not be deprived of opportunities to receive investigational therapies when there are no better alternatives, even though they may not consent for themselves. Efforts must be made to let such women know of these opportunities and to invite them to decide whether they wish to accept the investigational therapy, even though the formal consent must be obtained from another person, usually a man. Such invitations may best be extended by women who understand the culture sufficiently well to discern whether prospective recipients of investigational therapies genuinely wish to accept or reject the therapy.
Council for International Organizations of Medical Sciences in collaboration with the World Health Organization. 1993. International Ethical Guidelines for Biomedical Research Involving Human Subjects. Geneva: WHO.
Newton, L. 1984 Agreement to participate in research: Is that a promise? IRB: A Review of Human Subjects Research 6(2):7–99.