Legal considerations governing research on human subjects issue both from federal and state laws and from the institutional framework of sponsor agencies and research organizations that interpret and implement these laws. At the federal level, Congress has passed statutes and various federal agencies have promulgated regulations and guidelines to govern research on human subjects, including policies directly pertinent to the participation of women in clinical studies. The extent to which the individuals and organizations involved in the conduct of research must adhere to these policies depends on the type of policy: statutes and regulations must be followed by public and private entities, while guidelines are recommendations. Guidelines lack the force of law, but nevertheless have been quite effective in eliciting the desired behavior. Guidelines are also highly influential in setting standards for appropriate conduct.
The U.S. Constitution, with its protections of individual rights and provisions for equal protection under the law, provides further constraints on the behavior of public organizations and the private individuals and entities who work for them. Those involved in the conduct of human research also are governed by state constitutions, statutes, regulations, and liability decisions. Developed primarily through state appellate court decisions, the record of liability decisions is known as the law of torts. State tort liability rules are relevant to both the inclusion of women in, and the exclusion of women from, clinical studies. The greatest fears about liability for inclusion stem
mainly from the possibility of injury to offspring resulting from women's participation in clinical drug trials.
Health-related research and development in the United States is supported by the federal government (predominantly through the National Institutes of Health [NIH]), the pharmaceutical industry, and private foundations. This institutional structure can affect the conduct of research because it is the source not only of funding, but also of procedures for reviewing the ethics of scientific research-including whether a proposed plan for selecting research participants is just-and of the legal requirements applicable to research.
NIH, located within the Public Health Service (PHS) of the Department of Health and Human Services (DHHS), is the single largest supporter of biomedical and behavioral research and development (health R&D) in the world. NIH underwrites approximately 73 percent of all health R&D supported by the U.S. federal government, and about 30 percent of all health R&D in the United States. In fiscal year 1992 the projected NIH health R&D budget was $8.4 billion (NIH, 1992).
NIH is an extraordinarily complex organization. It includes a federation of 16 institutes, the National Library of Medicine, two divisions, and four centers. All of these components are coordinated by the Office of the Director. NIH policies have a profound effect on other organizations that have health-related missions, including other federal agencies, awardee institutions, and private foundations and corporations that support or conduct health R&D. Many institutions simply adopt NIH policies and procedures. Other organizations, both public and private, adapt NIH policies to suit their own structures and needs.
Most NIH funding components have a twofold structure. Extramural programs support health R&D projects carried out by research institutions throughout the United States and in at least 80 nations worldwide. Intramural programs, operated by federal employees, conduct research on the NIH campus in Bethesda, Maryland, and at a number of other locations throughout the country (Maryland, North Carolina, Colorado, Florida, and other states). Approximately 88 percent of the NIH research budget is disbursed to nonfederal institutions through grants-in-aid, contracts, and cooperative agreements (NIH, 1992). Awards are made by NIH funding components operating with the advice of a large and carefully regulated peer review system. Grant applications submitted to NIH by extramural institutions are typically reviewed by initial review groups (IRGs), commonly called study sections, that conduct scientific merit review and assign a priority score to each application that it recommends for funding. Approximately 20 percent
of applications recommended for funding are actually funded in any given fiscal year.
In addition to their responsibilities for assessing scientific merit, IRGs also are asked to identify ethical concerns associated with proposed research in relation to the rights and welfare of human research subjects, care and use of laboratory animals, and scientific misconduct. If an IRG or an NIH staff person raises an ethics concern, a bar to funding is entered into the computerized grant-tracking system. No award can be made in support of a barred project until and unless the concern has been resolved. Most grant applications also are reviewed by a national advisory council or board that considers program relevance and the public importance of the application. Boards and councils also have authority to raise ethical concerns that place a bar to funding. Ethics concerns raised by IRGs are reviewed and resolved by national advisory boards or councils.
Contract proposals and cooperative agreements for biomedical and behavioral research are reviewed in a manner similar to that utilized to review grant applications. Technical Evaluation Groups (TEGs) carry out assessment of the proposal in a manner analogous to that of IRGs. As is the case for IRGs, TEGs are expected to identify any ethics concerns associated with proposals for contracts or cooperative agreements. No contract or cooperative agreement can be finalized until and unless ethics concerns have been resolved.
In making awards, NIH and other PHS agencies operate under the general authority of the Public Health Service Act, which requires the secretary of DHHS to operate a wide variety of health-related regulatory, research, demonstration, and service programs. Responsibility for these programs is delegated by the secretary, or in some cases directly by the Congress, to the agencies and program directors throughout PHS.
Regulatory responsibility for protecting the rights and welfare of human research subjects has been delegated to the Office for Protection from Research Risks (OPRR). Although that office is located within NIH for organizational purposes, it acts on behalf of the secretary of DHHS. The primary instrument that OPRR uses in meeting its responsibility is the promulgation and implementation of regulations codified in the Code of Federal Regulations (C.F.R.) for the protection of human subjects. Those regulations require that, before awardee institutions are permitted to carry out research involving human subjects, they must provide adequate assurance to OPRR that they will comply with the regulations. The primary requirement of the regulations is that before work is begun and at intervals of no more than one year during the conduct of research involving human subjects, each research project shall be reviewed and approved by an institutional review board (IRB). The U.S. Food and Drug Administration (FDA) has also issued regulations at 21 C.F.R. 50 & 56 that include congruent require-
ments for IRB review of studies involving experimental drugs, devices, and biologics (see below).
IRBs are administrative bodies established to protect the rights and welfare of human research subjects recruited to participate in research activities conducted under the auspices of their affiliated institutions. Each IRB has authority to approve, require modifications in, or disapprove all research activities that fall within its jurisdiction.
The FDA, another PHS agency, functions under the Food, Drug and Cosmetic Act (FDCA). The FDCA provides the commissioner of FDA with authority and responsibility to regulate (among other articles) the testing and marketing of drugs, biologics, and medical devices involved in interstate commerce. The commissioner also is subject to policies and regulations issued by or under the authority of the secretary of DHHS and by many of the provisions of the PHS Act. In those rare instances when the FDA conducts clinical research, that research is subject to DHHS regulations. When NIH conducts or supports clinical studies involving the testing of investigational drugs, biologics, or devices, NIH and NIH awardee institutions are subject to FDA regulations governing such items. Institutions that conduct clinical studies funded by NIH or another PHS agency involving drugs, biologics, or medical devices are subject to all applicable policies and regulations of both NIH and FDA.
Institutions conducting research on drugs, biologics, and medical devices without any public funding, including research conducted by scientists in the employ of pharmaceutical companies and research conducted by academic scientists and others supported by the pharmaceutical industry or private foundations, are subject only to federal policies promulgated by FDA. According to the Pharmaceutical Manufacturers Association (PMA), the pharmaceutical industry contributes just over half of total health research dollars-$10.9 billion in 1992 (PMA, 1992). Industry research is carried out by pharmaceutical companies without federal funding either onsite or at universities. Privately funded pharmaceutical research carried out in a university setting, however, may be subject to DHHS regulations, because individual institutional policies frequently require investigators to conform to DHHS policy, independent of the funding source of a particular study. Private foundations (such as the Pew Foundation) and professional organizations (e.g., the American Lung Association) underwrite a much smaller, but not insignificant, percentage of health research. Such privately funded, nonpharmaceutical research would technically not be subject to any federal policies, but again, if conducted at an institution that receives federal funds, it would likely be subject to DHHS policy.
CURRENT FEDERAL POLICIES
Current federal policies that affect equity in clinical studies take the form of statutes, regulations, and agency guidelines and memoranda. These policies govern research funded, conducted, or otherwise regulated by the federal government, its agencies, and departments. The policies vary: some appear to promote inclusion of both genders, others refer to inclusion of women and minorities, and still others specify conditions applicable to women of childbearing potential and pregnant women. Application of a particular policy may depend on funding origin (e.g., NIH, Department of Veterans Affairs, Department of the Army, private funding, and the like), type of research (e.g., drug development or observational study), nature of condition studied (e.g., life-threatening), or fertility status of the proposed study participant (e.g., postmenopausal women, pregnant women, women of childbearing potential, men). Particularly in the area of drug development, clinical studies that receive federal funding or are performed at institutions supported by federal funding may be subject to a number of policies prior to a drug's entrance into the market. Recently such policies have become more consistent.
These sometimes overlapping policies are perhaps best understood through evaluation on an agency-by-agency basis. We will begin with a discussion of the policies of or affecting DHHS-funded research, focusing particularly on NIH and FDA policies. This will be followed by a discussion of relevant policies of other federal agencies and departments. Because many policies have been revised since 1990, the rationale for revising the earlier policy will be noted where relevant and available. In addition, the type of study, the condition studied, and any provisions to encourage the performance of scientific analyses to identify gender differences will be highlighted. Finally, where relevant, provisions applicable to fertile women will be contrasted with policies applicable to fertile men.
National Institutes of Health
As of this writing, NIH policy on study population composition of intramural and extramural research is in transition. The future policy is reflected in Section 131 of the recent NIH reauthorization legislation, the National Institutes of Health Revitalization Act of 1993 (P.L. 103-43, 107 Stat. 133 to be codified at 42 U.S.C. 289a-2). This provision of the law mandates the inclusion of women and racial and ethnic groups in NIH intramural and extramural research. It is to be implemented in fiscal year 1995 through NIH guidelines scheduled to be published in December 1993; draft versions of these guidelines were not available to the committee.
The NIH policy on study population composition currently in effect is
expected to be revised for implementation during the transition to the legislatively mandated policy (J. LaRosa, deputy director of the NIH Office of Research on Women's Health, personal communication, October 1993). The policy in effect as of this writing was introduced in 1990 and consists of five components:
- the August 1, 1990, NIH policy memorandum on inclusion of women and minorities applicable to intramural research;
- the August 1990 policy notice entitled "NIH/ADAMHA Policy Concerning Inclusion of Women in Study Populations," applicable to extramural research;
- the September 1990 policy notice entitled "NIH/ADAMHA Policy Concerning Inclusion of Minorities in Study Populations," applicable to extramural research;
- an explanatory memorandum entitled "NIH Instruction and Information Memorandum OER 90-5" describing the application of the two policy notices applicable to extramural research;
- the recently instituted requirements in PHS grant applications and continuation applications to identify proposed and recruited study populations by gender and minority composition.
The legislative mandate and its implementing guidelines will likely supersede at least the first four components of the current policy. It is possible that the PHS grant application will be modified as well, to accommodate the policy changes.
The current NIH policy, which became effective in February 1991 (referred to collectively here as the "1991 NIH Policy") will only be briefly described, while the new legislative mandate (referred to here as the "Act") will be discussed in detail below, highlighting areas of known controversy.
The 1991 NIH Policy
Following the issuance of the 1990 GAO report, NIH promulgated a strengthened policy to govern the awarding of federal research grants. The new policy applies to a wide variety of extramural clinical research projects, including:
Human biomedical and behavioral studies of etiology, epidemiology, prevention (and preventive strategies), diagnosis, or treatment of diseases, disorders or conditions, including but not limited to clinical trials.
Standard language articulating the extramural policy now appears in all requests for proposals (RFPs) announced in the NIH Guide for Grants and Contracts:
Applications for grants and cooperative agreements that involve human subjects are required to include minorities and both genders in study populations so that the research findings can be of benefit to all persons at risk of the disease, disorder, or condition under study.
The applicant must describe the proposed study population composition and provide a "compelling" justification for gender or racial and ethnic group exclusion. The investigator must also address gender and racial and ethnic issues "in developing a research design and sample size appropriate for scientific objectives of the study." The NIH Policy Notice further explains that the inclusion of women and racial and ethnic groups in study populations will be considered a matter of scientific and technical merit in peer review.
What constitutes a "compelling" reason for exclusion has generated debate both within the scientific community and within Congress (Wittes and Wittes, 1993). The explanatory memorandum to NIH staff and peer advisory groups directs them to consider sufficient only "strong scientific or practical reasons" for the exclusion of women or racial and ethnic groups from clinical research. Some of the potentially "acceptable" justifications listed include: research on a "predominantly or exclusively a male condition'' (e.g., prostate cancer); research that presents an "unacceptable risk for women of childbearing age"; certain pilot and feasibility studies in which "gender differences may not be germane"; research in an area that "has already been extensively studied in women"; and in certain instances, studies that would be "prohibitively expensive" (NIH, 1990).
Monitoring of study populations is accomplished through reporting requirements specified in PHS grant and continuation applications. These applications require awardees to identify annually proposed and actually enrolled study populations according to gender and the five designated racial and ethnic categories (American Indian/Alaskan Native, Asian/Pacific Islander, Black [not of Hispanic origin], Hispanic, and White [not of Hispanic origin]). A summary of study population composition during the first full year of the policy's implementation is expected to be issued by NIH soon.
Most of the NIH intramural research program is subject to a different, less restrictive policy that requires only that gender-based exclusions be indicated and a "clear rationale" be provided. It currently reads, in its entirety:
The inclusion of women must be considered in the study populations for all clinical research efforts. Exceptions would be studies of diseases which exclusively affect men or where involvement of pregnant women may expose the fetus to undue risks. Gender differences should be noted and evaluated. If women are not to be included, a clear rationale should be provided for their exclusion.
In order to provide more precise information to the medical community, it is recommended that publications resulting from NIH- or ADAMHA-conducted research specify, in the abstract or summary, the gender(s) of the research subjects or patients [Rall, 1990].
Large-scale intramural projects that are implemented through contracts are not subject to the foregoing policy, but rather are considered part of the extramural program for policy purposes.
The Act The NIH Revitalization Act of 1993 became law on June 10, 1993. The Act requires that women and ethnic and racial groups be included as subjects in each intramural and extramural clinical research project supported by NIH. It further requires that a clinical trial that includes women and racial and ethnic minorities as participants be designed and carried out to provide for valid analysis of whether the variables being studied affect these subpopulations differently than other participants. Furthermore, NIH is instructed to conduct or support outreach programs for recruitment and retention of women and racial and ethnic group participants in clinical research projects. The law allows for exemptions in cases of research that are inappropriate with respect to the health of the subjects, the purpose of the research, or other circumstances determined by NIH.
NIH is required to promulgate implementing guidelines by December 7, 1993. These guidelines are to specify when the inclusion of women and members of racial and ethnic groups as subjects in clinical research is inappropriate; how clinical trials must be designed to have adequate representation of subpopulations to distinguish whether a treatment affects a subpopulation differently than the other members of a research project; and the operation of outreach programs.
The Act includes express instructions for the NIH guidelines. For example, it specifically prohibits cost considerations as a reason for determining that the inclusion of women and racial and ethnic groups is inappropriate for a clinical trial. In other clinical research projects, however, an exception from the prohibition on cost considerations is allowed when the data that would be obtained by the research project is or will be obtained through other means that provide data of comparable quality. An example of where NIH may allow the exclusion of certain subpopulations on cost grounds is where a body of research on those groups exists and will be analyzed for differential response through other means, such as meta-analysis. Another exception from the prohibition of cost considerations involves cases where there is already a substantial body of scientific data demonstrating that there is no significant difference between subpopulations in the effects of the variable being studied. This prohibition on the consideration of cost differs from the 1991 NIH Policy, which allows cost to be an acceptable rationale for women's exclusion from certain clinical trials.
NIH has been provided with some degree of latitude in the development of their guidelines. For example, the act requires that clinical trials be designed and performed to enable a valid analysis of whether women or racial and ethnic groups respond differently than other subjects in the research. Concern has been expressed about the interpretation and implications of this clause. (IOM, 1993; Wittes and Wittes, 1993). The language of the report suggests the intent of Congress:
Although the Committee has given the Director of NIH the general authority to define [valid analysis]. . . the Committee intends for that definition to include an analysis of not only whether differences among study populations exist statistically, but an analysis of what those differences are as well as their relative importance for various population subgroups in the study.
This would indicate that women and racial and ethnic groups are to be enrolled in clinical research projects in numbers large enough to provide statistical significance, and that the analysis of such research is to include distinctions among and between the various population groups, including women. These concerns have already been discussed in Chapter 4.
The NIH policy does not automatically require that study designs provide statistical power to perform gender analysis except "whenever there are scientific reasons to anticipate differences between men and women." Those who favor the language in the Act over that in the NIH policy contend that the scientific literature on gender-mediated effects is sufficiently sparse that one will not always be able to "anticipate" such differences (S. Wood, scientific director of the Congressional Caucus for Women's Issues, personal communication, May 1993).
Other provisions of the Act require NIH to establish internal and external committees to advise it on issues in women's health research, including gender differences in clinical drug trials and disease etiology, course, and treatment. The Act also requires NIH to determine the extent of women's representation as senior physicians and scientists in the institutes and to carry out activities to increase the extent of such representation. Finally, the Act mandates the creation of a national data system and clearinghouse on research for women's health (see Chapter 2).
The NIH Revitalization Act directs NIH to define the terms "minority group" and "subpopulations." The Act does not clarify the extent of required inclusion of racial and ethnic groups in clinical research projects. The reasons for the unequal legislative specificity between women and minority groups is discussed in the House Committee on Energy and Commerce report accompanying the bill: the committee explained that representative inclusion of minority groups poses a complex problem because not only are there variations between Caucasians and people of color, but there
are also variations among variations (or subpopulations) of the same racial or ethnic group. The report emphasized the importance of identifying such variations in clinical studies, but explains that this statutory language was chosen to prevent quotas or numerical goals for participation in clinical research projects (U.S. Congress, 1993).
Food and Drug Administration Policies
FDA policies concerning the inclusion of women in clinical studies are reflected in guidelines. It is important to note that guidelines do not have the force of law or regulation, and they do not have to go through the same public comment procedures as regulations. Guidelines are not binding on the agency, nor do they create or confer any rights, privileges, or benefits for or on any person. They are provided as an aid to organizations involved in the evaluation of new drugs for FDA approval who wish to market such drugs. FDA guidelines are used to specify the information that data reviewers of such applications will expect in the research that supports the safety and efficacy of a drug.
There are three FDA guidelines that are particularly relevant to inclusion of women in clinical trials. Two of these are discussed in detail; discussion of a 1988 guideline is incorporated into the discussion of the 1993 guideline, where its content has been reiterated.
In 1977 the FDA issued guidelines for drug development: "General Considerations for the Clinical Evaluation of Drugs." The 1977 guidelines specifically stated that pregnant women and "women who are at risk of becoming pregnant" should be excluded from Phase I studies. They further recommended that a "woman of childbearing potential" be excluded from "large-scale clinical trials" (i.e., Phase III studies) until all three segments of the FDA Animal Reproduction Guidelines have been completed. The guidelines further provided that ''if adequate information on efficacy and relative safety has been amassed during Phase II, women of childbearing potential may be included in further studies provided Segment II and the female part of Segment I of the animal reproductive studies have been completed" (FDA, 1977:10). Segment I animal testing covers fertility and reproductive performance; Segment II covers teratogenesis; and Segment III covers perinatal and postnatal effects.
Women of childbearing potential, however, could receive investigational drugs in the absence of adequate animal reproduction studies when: (1) the drug was considered to be a life-saving or life-prolonging measure (e.g., cancer therapy), (2) the drug belonged to a class of compounds for
which teratogenic potential had already been established in animals, or (3) the woman had been "institutionalized" for a period of time adequate to verify that she was not pregnant. When, under this exception, an investigational drug is used to treat a serious disease, the guidelines recommend that the investigator point out the lack of reproduction studies during the informed consent process, test the woman for pregnancy, and advise her of contraceptive measures. As Merton (1994) notes, the guidelines do not mandate the performance of reproduction studies at any time, and, further, do not mention how the results of such studies should affect the inclusion of women.
Lactating women are specifically mentioned in the guidelines, with the recommendation that excretion of the drug or its metabolites should be determined when feasible prior to their usage of the drug. If a woman becomes pregnant during the trial, the guidelines recommend fetal follow-up.
The guidelines broadly defined a woman of childbearing potential as a "premenopausal female capable of becoming pregnant" (FDA, 1977). Included in the definition were women on oral, injectable, or mechanical contraception; women who were "single"; and women whose partners had been vasectomized or were using mechanical contraception. This exclusion was based on a concern that women might become pregnant during the course of a clinical trial and a judgment that the potential risks of exposing a fetus to an experimental drug of unknown fetal toxicity were greater than the potential benefits of the information that would be gathered by including women of childbearing potential in these early trials.
Critics of the 1977 FDA guidelines had questioned for years whether the guidelines reflected gender stereotyping (e.g., female susceptibility and male invulnerability) more than concerns about good science (see Kinney et al., 1981). At a recent conference evaluating the issues concerning the inclusion of women in clinical trials, cosponsored by FDA and the Food and Drug Law Institute, critics claimed that an asymmetry existed in the risk-benefit analyses for research on men of reproductive potential and women of reproductive potential. They noted that according to the guidelines, research involving agents thought to cause reproductive harm in male animals could be conducted in men depending on "the nature of the abnormalities, the dosage at which they occurred, the disease being treated, the importance of the drug, and the duration of drug administration" (FDA, 1977). In practice this meant that even a drug known to have teratogenic effects in animals could be tested in men if they were simply informed of the risks and advised not to conceive while participating in the trial. The 1977 guidelines would have excluded women of reproductive potential from such trials based on the fact that potential offspring might be harmed.
The background paper accompanying recently issued FDA guidelines explains that the 1977 guidelines may have discouraged participation of
women in drug development studies and may have resulted in a "paucity of information about the effects of drugs in women" (FDA, 1993:4). As explained by FDA, the 1977 guidelines had come to be considered by many to be "rigid and paternalistic, leaving virtually no room for the exercise of judgment by responsible female research subjects, physician investigators, and IRBs" (FDA, 1993).
FDA released a new guideline, "Guideline for the Study and Evaluation of Gender Differences in the Clinical Evaluation of Drugs," on July 22, 1993. This guideline modifies and revises the section of the 1977 version concerning inclusion of women of childbearing potential in clinical trials. The introduction to this guideline indicates that the broad principles that are outlined for the inclusion of women in the early phases of clinical trials will also be applied to FDA approval processes for biological products and medical devices.
The new guidelines provide four underlying observations that persuaded the agency of the need for revision of the 1977 and 1988 directives on women in clinical drug trials. These observations are:
- Variations in response to drugs, including gender-related differences, can arise from pharmacokinetic (the effect of the body on the drug) differences or pharmacodynamic (the effect of the drug on the body) differences.
- Gender-related variations in drug effects may arise from a variety of sources (e.g., differences in biology, behavior, etc.) that can affect the pharmacokinetics of some drugs. Identification of these effects enhances the ability to treat both genders appropriately.
- In the evaluation of potential gender-related differences, evaluation of pharmacokinetic differences should precede pharmacodynamic differences because they are more common and because they can be used as an indicator of possible need for later pharmacodynamic studies that further narrow the affected variables for a drug.
- Since there are few documented gender-related pharmacodynamic differences of clinical significance, and pharmacodynamic/effectiveness studies may be difficult to conduct, such studies are not routinely necessary. However, pharmacokinetic studies and the development of blood concentration data to detect important pharmacodynamic and effectiveness differences related to gender are still needed.
Inclusion of both genders in clinical studies The guideline explains that subjects in a given clinical study should reflect the population that will
receive the drug when it is marketed. Although the explicit exclusion of women from Phase I and early Phase II trials is lifted, the new guideline encourages but does not create a strict requirement that women be included in the early phases of drug trials. Rather, FDA now expects analysis of clinical data in drug applications to identify whether there is a gender difference in the response to the drug, and, if so, the basis of the gender difference (if it can be determined).
The guideline suggests that patients of both genders be included in the same trial to permit direct comparison of genders within the studies. The guideline also explicitly discredits routine exclusion of women from bioequivalence trials because changes during the menstrual cycle may cause intrasubject variability (differences or variations in an individual's response to the same amount of drug). Instead, inclusion of women in these studies is expected to indicate if there is a possible need for concern about the variations in response to a drug based on the hormonal fluctuations of the menstrual cycle.
Analysis of effectiveness and adverse effects by gender This section reiterates FDA's 1988 guidelines for clinical and statistical sections of new drug applications, "Guideline for the Format and Content of the Clinical and Statistical Sections of New Drug Applications." The new guideline outlines FDA's expectations for the data analyses of gender differences and other subgroup differences, which are expected to be performed and explained in an application for approval of a new drug. Depending on the findings of these analyses, FDA may require further testing of a drug on specific populations to determine if there are instances in one or more populations where the drug is not as effective as in the overall population, or if there are instances of adverse reactions distinct to a certain population.
Defining the pharmacokinetics of the drug in both genders FDA emphasizes the importance of pharmacokinetic studies and pharmacokinetic screens to define gender-related differences in drug responses. It recommends the use of "pilot studies" to ascertain the presence of significant pharmacokinetic differences before conducting controlled trials. This enables the early identification of possible variations in dosing regimens that can be built into the larger clinical trials. Emphasis is placed on the heightened importance of early pharmacokinetic studies for drugs with a narrow therapeutic range (drugs that have a small range of concentration between the point of effectiveness and the point of toxicity) where the generally smaller size of women could require modifications in dosing.
In addition, the FDA lists three gender-related facets of pharmacokinetics that "should be considered during drug development": (1) variations in pharmacokinetics caused by the menstrual cycle, including comparisons of
premenopausal and postmenopausal patients as well as changes during the course of the menstrual cycle; (2) the effects on pharmacokinetics of estrogen replacement therapy and systemic contraceptives (oral contraceptives and long-acting progesterone such as Norplant®); and (3) the effect of oral contraceptives on pharmacokinetics.
Gender-specific pharmacodynamic studies Pharmacodynamic and effectiveness studies conducted separately on men and women are not expected by FDA unless the analyses by gender of clinical trials and pharmacokinetic screens indicate a significantly different gender-related response. In such cases, FDA will look for these additional studies. Once again, the importance of such additional studies increases should the early studies indicate a gender-related difference in response to a drug with a narrow therapeutic range.
Precautions in clinical trials including women of childbearing potential FDA will rely on the informed consent document and the investigator's brochure to advise participants of the need to take precautions to prevent the inadvertent exposure of a fetus to potentially toxic drugs. These documents should contain explanations of all available information about the degree of risk of fetal toxicity. It is FDA's belief that large-scale exposure of women of childbearing potential should not take place until after the results of animal toxicity tests are analyzed. It recommends that clinical protocols include provisions for the use of contraception (including counseling in the selection and use of contraception) or abstinence for the entire time a subject will be exposed to the drug (sometimes beyond the completion of the study), use of pregnancy tests before exposure to the drug, and timing studies in accordance with the menstrual cycle.
Potential effects on fertility In the clinical evaluation of drugs that carry the risk of causing abnormalities in reproductive organs or their function (as identified in animals), the risks of exposing individuals of reproductive potential must be weighed against the potential benefits of the drug. In cases where such drugs do proceed into clinical trials, such studies should include monitoring and laboratory studies, as well as long-term follow-up, to enable detection of potentially deleterious effects.
Other DHHS Policies
DHHS Policy on Pregnant Women as Research Subjects
In 1975, DHHS adopted regulations concerning pregnant women as research subjects. These regulations supplement the Model Federal Policy
with respect to DHHS-funded research; the regulations have not been adopted by the other 15 federal agencies that adopted the Model Federal Policy. Selwitz and Wermeling (1992), however, have noted that the regulations have had an impact on IRB decisionmaking at many research institutions, regardless of whether projects are funded by DHHS.
Subpart B of Part 46 of Title 45 of the Code of Federal Regulations is entitled "Additional Protections Pertaining to Research, Development, and Related Activities Involving Fetuses, Pregnant Women, and Human In Vitro Fertilization" and describes protections when the pregnant woman and/or the fetus is the subject of research. Where the pregnant woman is the subject of research, the regulations specify that research cannot be approved except where "appropriate" studies on animals and nonpregnant individuals have been performed (45 CFR section 46.206 (a)(2)) and
(1) the purpose of the activity is to meet the health needs of the mother and the fetus will be placed at risk only to the minimum extent necessary to meet such needs, or (2) the risk to the fetus is minimal [45 CFR section 46.207 (a)]
"Minimal risk" is defined in another part of the regulations to mean that the probability and magnitude of harm or discomfort anticipated in the research are not greater than those ordinarily encountered in daily life, routine physical examinations, or psychological tests (45 CFR 46.102(i)).
In addition to the foregoing limitations, the regulations require that the "mother and father" be legally competent and give their informed consent after being fully informed regarding the possible impact of the research on the fetus. Exceptions are made for the father's informed consent if: the reason for the activity is to meet the health needs of the mother, his identity or whereabouts cannot reasonably be ascertained, he is not reasonably available, or the pregnancy resulted from rape (45 CFR sections 46.207(b)).
Policies of Other Federal Agencies and Departments
The committee contacted the 15 other agencies and departments that have adopted the Model Federal Policy for Protection of Human Subjects (discussed above) to ask whether they had additional policies related to the inclusion of women in clinical studies. Two of those surveyed had such policies: the Department of the Army has a specific policy on pregnancy testing (Department of the Army, 1992), and the Department of Veterans Affairs has a policy generally promoting the inclusion of women in clinical studies (Department of Veterans Affairs, 1992). The Department of Energy explicitly recommends that the guidance provided in Subpart B of the DHHS regulations be followed if the proposed research involves pregnant women (Department of Energy, 1992).
All laws and policies of state and federal government are expected to conform to the ultimate source of legal authority in this country-the U.S. Constitution. The Fourteenth Amendment's express protections against laws that deny "equal protection" and "life, liberty, and property" are particularly relevant to questions of participation in clinical research. The Fourteenth Amendment, and the Bill of Rights generally, have been interpreted by the Supreme Court to require equal access to government health benefits, along with a high degree of personal liberty in matters affecting health care. This has led some legal experts and advocates to conclude that exclusion of women from government-sponsored or government-regulated research violates constitutional standards of liberty and equality.
Officials of federal agencies understand that their policies and activities may be subject to constitutional challenge and review. For example, a lawsuit seeking to invalidate a federal biomedical research policy would likely name as defendants the Department of Health and Human Services and its secretary. Private firms, such as pharmaceutical manufacturers, generally are not considered arms of the government. Nevertheless, a citizen's petition filed with the FDA in December 1992 argued that the policies and practices of the industry could count as government action, for purposes of construing the Fourteenth Amendment, to the extent that the FDA or other federal agencies encourage the firms that they closely regulate to exclude women from studies (NOW Legal Defense and Education Fund, 1992). Were industry action held to be government action, Fourteenth Amendment principles would apply to industry policies of exclusion.
Liberty, Privacy, and Bodily Self-Determination
The Fourteenth Amendment provides that "No state shall... deprive any person of life, liberty, or property." This protection of bodily self-determination and private decisionmaking about matters closely affecting health is often labeled the "right to privacy." Women and their advocates have sometimes appealed to this right to protest exclusion from, for example, acquired immune deficiency syndrome (AIDS) research: women with AIDS or human immunodeficiency virus (HIV) infection have a right to choose for themselves whether to take on the health risks of drug research. Advocates say that the principles of private choice and good science should determine the extent of female participation in research, not the principle of government paternalism.
The Supreme Court has repeatedly held that decisional privacy about matters affecting health care is among the liberties protected by the Fourteenth Amendment. The Court has upheld this right with regard to the ter-
mination of artificial nutrition and hydration (Cruzan v. Director, Missouri Department of Health) and abortion (Roe v. Wade). The Court reaffirmed the right to private abortion decisionmaking in Planned Parenthood of Southeastern Pennsylvania v. Casey, in which it recognized the right of a woman "to choose to have an abortion before viability and to obtain it without undue interference from the State." While acknowledging a state interest in the life of at least some fetuses, the Court also stressed the importance of reproductive privacy for women's liberty, linking a woman's "unique" reproductive liberty to her ability "to participate equally in the economic and social life of the nation."
Unresolved for now is the question of the standard of review that the Court should apply in Fourteenth Amendment privacy cases relating to women's health. In Roe v. Wade, the Court applied a standard it calls "strict scrutiny." Under this standard, the Court strikes down laws or policies that constrain "fundamental" rights and liberties unless the government can cite a "compelling state interest" that cannot be furthered by "less restrictive means." In the Cruzan case, however, the Court applied a weaker standard of review known as "rational basis,'' under which the Court upholds laws or policies that are rationally related to a "legitimate state interest." A third standard of review appears to have been utilized in the Casey decision, where the Court considered whether the state laws in question "unduly burdened" the woman's Fourteenth Amendment right of private abortion decisionmaking. Casey did not expressly appeal to the idea of a general, "fundamental" right to privacy requiring "strict scrutiny" as Roe v. Wade had two decades earlier. As a consequence, there is uncertainty among lawyers and jurists about precisely what standard of review would apply in future Fourteenth Amendment privacy cases affecting women's health, including cases involving exclusion from research.
The uncertain status of privacy jurisprudence makes it difficult to predict the outcome of challenges to policies that exclude women from health research. Nevertheless, scholars have argued that Cruzan's recognition of a right to refuse artificial nutrition and hydration, premised on freedom from intervention in private decisionmaking related to health care, also implies a right to take part in risky clinical studies, i.e., if one can terminate one's own life, one should be able to assume the risk of taking an untested drug. This argument may be particularly strong when made on behalf of women suffering from terminal illnesses who seek access to experimental drug therapies.
Privacy vs. Fetal Protection
Objections to women's participation in clinical studies are often premised on an interest in doing what is best for women's own health and well-being; but in some instances they are also based on an interest in fetal and
child welfare. The former objection assumes that policy makers ought to protect women (but need not protect men to the same degree) from risky activities. The "right to refuse treatment" analogy rejects such intervention as unwarranted paternalism, suggesting instead that decisions affecting one's own physical person must be treated as personal and private. Objections based on fetal and child welfare assume that policymakers are entitled to limit women's choices where those choices impose risks on the unborn. If women have a constitutional right to abortion that is virtually unfettered in the early weeks of pregnancy, however, the case for fetal protection seems weak. Some have argued that the case for fetal protection is stronger if it is the children those fetuses will become (in pregnancies intended to go to term) that the policy seeks to protect (see, for example, Robertson, 1994; Steinbock, 1994).
Tribe (1991) and other constitutional experts assert that the unborn are not persons within the meaning of the Constitution, and that the case for fetal protection is without significant support. Yet the question of the legal status of the unborn is complex. The unborn are ascribed numerous legal interests, including those provided by the laws of inheritance and personal injury. Casey expressly recognizes a governmental interest in the welfare of an unborn fetus that "may become a child." Nevertheless, the Supreme Court has held to the principle that government may not deny women and their physicians the liberty to abort previable fetuses. Thus, both Roe and Casey are inconsistent with any notion that the previable fetus is a constitutional person with a right to life equivalent to that of newborns and children. Casey also invalidated provisions of a Pennsylvania statute requiring that pregnant women attest to having notified their spouses of plans to abort.2 The Court saw no legally enforceable role for third-party notification and consent in the procreative decisionmaking of adult pregnant women. This is the strongest constitutional basis for an argument that husbands and fathers also have no legally enforceable role in deciding whether a pregnant woman may expose herself and a fetus to research-related health risks, even in the therapeutic context.
Some states have attempted to control the behavior of pregnant women in order to protect fetuses. Between 1982 and 1987, institutions in 18 states noted 36 different attempts to override maternal refusals of medical treatment, such as forced caesarean sections and intrauterine transfusions; judges issued court orders for an overwhelming majority (86 percent) of these interventions (Green, 1993).3 Many states have also tried to prosecute women for engaging in behavior during pregnancy that is suspected or known to be risky for the fetus. For the most part, efforts to charge a pregnant woman with crimes ranging from delivering drugs to a minor and criminal abuse and neglect to disobeying doctor's orders have been unsuccessful. Courts have shown a reluctance to include a fetus in the definition of child in
existing child abuse laws (Green, 1993). But where courts have been hesitant, state legislatures have not: 12 states have made it a crime for a pregnant woman to use drugs and alcohol, and 19 states have passed laws allowing prosecutors to charge women with child abuse if they give birth to children with illegal drugs in their blood (Green, 1993).
Since the Supreme Court has not ruled on the constitutionality of these state statutes and court decisions, the committee cannot draw any firm conclusions about whether fetal interests limit women's rights to decisional privacy. Until these fetal protectionist policies are challenged, they remain enforceable as law. As a result, equal protection rather than privacy principles may ultimately be a stronger basis for a constitutional test of policies that, on their face or in effect, exclude women from full participation in clinical research.
The wholesale exclusion of women from research would raise obvious equal protection concerns. So, too, would what constitutional lawyers term the "underinclusion" and "overinclusion" of women relative to men in clinical research. "Underinclusion" in this context could signify, for example, the inclusion of women in clinical research in numbers lower than good science requires; by contrast, ''overinclusion" in this context could refer to the preferential inclusion of women in numbers great enough to diminish the scientific merit of the research.
The equal protection clause of the Fourteenth Amendment provides that: "No State shall . . . deny to any person within its jurisdiction the equal protection of the laws." Although this clause makes express reference only to "State" government, the Supreme Court has long held that it applies to both the state and federal government. Courts have recognized that the principle of equal protection is embodied also in other key provisions of the Constitution, including Article I and the due process clauses of the Fifth and Fourteenth Amendments. Significantly, each state has a constitution, and many state constitutions contain provisions similar to those found in the federal constitution that are held to require equal protection of the law.
The equal protection clause prohibits the federal government-including its regulatory agencies and research units-from engaging in certain forms of discriminatory disparate treatment. The Supreme Court has yet to decide a challenge to a demographic restriction on access to clinical studies, but legal experts maintain that research policies that result in the exclusion of women as a class, whether on their face (with explicit exclusionary language) or in effect (because they result in disproportionate participation of men and women), may be found to contradict the equal protection clause (Charo, 1994; Merton, 1992). Policies that explicitly favor participation of women in research also raise equal protection concerns.
The most controversial research policies have excluded women or categories of women through explicit policy language. For example, the 1977 FDA guidelines (now superseded) recommended the exclusion of women of childbearing potential from early phase drug trials; the regulations governing research on human subjects require that IRBs approve research on pregnant women only if it involves treatment for the mother or fetus and minimal risk to the fetus (45 C.F.R. § 46.207). Other policies, however, may exclude women because their aggregate effect is the "underinclusion" or exclusion of women from full participation in research. For example, if evidence is found that the benefits of participating in research accrue disproportionately to men, it could be argued that the federal research policies effectively result in disparate treatment of men and women. At the same time, it is conceivable that a man could challenge the constitutionality of the NIH Revitalization Act of 1993 if its language was legally determined to be an express preference for the participation of women in clinical studies and sufficient evidence was available to prove that the benefits of participation accrued disproportionately to men.
The Supreme Court has concluded that the equal protection clause restricts the right of the government to treat similarly situated persons and groups differently. Nevertheless, the Constitution does permit differential treatment that is, in the language of the federal courts, "reasonable"-that is, if government can cite a "rational basis" for discrimination that is tied to ''legitimate state interests." A long-standing exception to this principle is the rule that merely reasonable disparate treatment premised on racial categories is presumptively unconstitutional. Legal restrictions that curtail the rights of any racial group are immediately suspect, and courts are required to subject such restrictions to its most rigid review standard: "strict scrutiny." Only in rare instances, when government can identify "compelling" state interests that can be promoted only by racial discrimination, will the courts permit a racial restriction to stand.
Today, legal restrictions that impair the equal treatment of women are also inherently suspect. As explained by Tribe (1987), "until the early 1970s, the Supreme Court routinely upheld sexually discriminatory laws whenever they could be rationally related to government purposes reflecting traditional views of the 'proper' relationship between men and women in American society." In the late 1970s, the Supreme Court held in Craig v. Boren that policies and practices involving gender classifications are constitutionally suspect. Establishing an "intermediate" level of scrutiny for gender discrimination cases, the Court held that gender classifications must have an "exceedingly persuasive justification" and be substantially related to "important governmental objectives."
Under this standard, women's treatment in the context of government research may be unconstitutional under the principle of equal protection if it results in disparate public benefits, unless there is an exceedingly persua-
sive justification. If a policy does not exclude women on its face, but its effect (alone or in combination with other policies) is a disproportionate accrual of benefits to men over women, a claim can theoretically be made that such a policy is unconstitutional under the equal protection clause. To successfully claim that a facially neutral policy results in discriminatory treatment in effect, however, the Court must find that the government agency or state actor implementing such a policy had the actual intent to discriminate (Personnel Administrator of Massachusetts v. Feeney). Unless intent can be shown, facially neutral policies that result in disparate impact need only show a rational relationship to a legitimate government objective. Under this level of scrutiny, the Court usually gives great deference to the government agency and upholds the policy.
Policies That Exclude Pregnant Women
The federal regulation barring the use of pregnant women in research except in limited circumstances (45 C.F.R. § 46.207) is an example of a facially neutral policy (it can be classified as gender-neutral because it doesn't explicitly exclude all women as a class) that arguably results in disparate treatment of women (because only women can be pregnant). In a constitutional challenge to the regulation, the government would likely argue that the policy is linked to an important government objective-that of protecting potential life—and that there was no intent to discriminate against women in creating the pregnancy classification.
The question of whether the differential treatment of pregnancy represents a suspect gender classification was first dealt with in Geduldig v. Aiello. The Supreme Court held that the equal protection clause was not violated by California when it excluded pregnancy-related disability from coverage under the state's disability insurance program. The Court found that the classification in California's program was not between men and women but between pregnant and nonpregnant persons; despite the argument that only women can be pregnant, the Court found that this policy was not on its face treating women differently. Consequently, the Court construed the arguments of the plaintiffs to be an in effect challenge and inquired as to whether California intended to discriminate against women in designing the program. Finding no intent to discriminate, the Court then cited the state's "legitimate interest in maintaining a self-supporting... insurance program" and upheld the constitutionality of the program. This decision, permitting disparate treatment of pregnancy, appears to be directly relevant to any inquiry about the constitutionality of federal regulations governing research on pregnant women.
A more recent Supreme Court decision-International Union, UAW v. Johnson Controls-has sent a different signal about disparate treatment of
pregnant women. Johnson Controls excluded pregnant and fertile women employees from selected jobs in its battery manufacturing plant out of concern about the effects of exposure to certain chemicals on the unborn and about its liability for injury to the fetus. Employees filed a lawsuit under Title VII of the Civil Rights Act of 1964, as modified by the Pregnancy Discrimination Act. Title VII prohibits sex-based employment discrimination, and the Pregnancy Discrimination Act makes it clear that for the purposes of Title VII, discrimination based on pregnancy is discrimination based on sex (Newport News Shipping & Dry Dock Co. v. EEOC). The Court held in favor of the employees, stating that women could give their informed consent to potentially risky employment and that Title VII did not permit the company to single out women in its concern about offspring.
Johnson Controls was brought under a federal civil rights statute governing private sector discrimination in the terms and conditions of employment, but the notion that it is wrongfully discriminatory to exclude women from a benefit on grounds of fetal protection could apply by analogy to the context of clinical studies. The employment-related finding in Johnson Controls may apply more directly to the research study context only in what Merton (1992) describes as the "rare circumstance that subjects are paid and their participation in research is fairly characterized as employment."
Exclusion of Women of Childbearing Potential
Although the 1977 FDA guidelines that recommended the exclusion of women of childbearing potential have been superseded, there may be some value in analyzing the constitutionality of such a broad restriction under the equal protection clause. Categorical exclusion of all women of childbearing potential-without regard to an individual woman's actual childbearing capacities and intentions, or to the effect of the substance under study on the unborn-would arguably be unconstitutionally broad under the equal protection clause. Similarly, it can be argued that such policies are unconstitutionally underinclusive where they exclude only women of childbearing potential, ignoring fertile men whose reproductive capacities and offspring may be affected by the substance under study (see Chapter 7). Again, any policy that restricts the research participation of women of childbearing potential will also be questionable under the decision in Johnson Controls, in which the Court struck down an exclusion policy that applied to fertile women as well as to pregnant women.
Policies That Favor Inclusion of Women
Equal protection challenges also apply to benign classifications, which favor treatment of one group over another, usually to remedy past discrimi-
nation. Affirmative action policies are examples of benign classifications. It could thus be argued that the NIH Revitalization Act, which calls for affirmative inclusion of women in clinical studies, could arguably be considered a benign classification. If the constitutionality of this statute is challenged by a man on grounds of disparate treatment, the intermediate scrutiny test developed in Craig v. Boren for application to exclusion would also be applicable to favored inclusion. The success of such a challenge would likely hinge on whether the government could present evidence of past discrimination against women in research to justify the disparate treatment.4 The committee declines to speculate on the outcome of such a challenge, especially in light of currently available data on the participation of women (see Chapter 2).
Both individuals and organizations involved in the conduct of research must deal with another set of legal considerations-liability. Fear of potential legal liability has been cited as one of the chief reasons that some women have been excluded from clinical research (NAS, 1991; Flannery and Greenberg, 1994; Merton, 1992). The possibility of injuries to the women themselves, however, has not been the basis for this concern. Increasing the number of women in clinical studies will simply add them to the class of all clinical research participants, and concern about participant injuries has never been great enough to halt research involving human subjects. Rather, the focus of liability concerns is on possible injury to offspring when pregnant women and women of childbearing potential are included in clinical drug trials. Although recent evidence may indicate that exposure of the father to some chemicals may cause harm to a developing fetus (see Chapter 7), the focus has overwhelmingly been on the potential for harm to offspring resulting from the mother's exposure either before or after conception.
More recently, pharmaceutical companies have begun to recognize that they could also be liable for not including women in clinical research (Bush, 1993). For example, a pharmaceutical company may be liable if a drug that has never been tested in women is nevertheless marketed for use by both genders and prescribed for a woman who then suffers an adverse reaction. Similar approaches to liability could be used as well where men, or subpopulations of women or men, were not included in a study population but suffered an injury. This creates a paradox for pharmaceutical manufacturers whose efforts to exclude women in order to protect themselves from liability may actually risk liability for exclusion.
Potential Liability for Inclusion
The threat of liability exists for injury to any subject of clinical research. Although fear of liability has never operated to exclude men as a class from participating in research, a general discussion of liability for research injuries is instructive in understanding how this liability operates in the context of research involving pregnant women and women of childbearing potential. The following sections discuss the incidence of research injuries, the legal theories that could serve as the basis for legal action, and the individuals and entities that might be found liable.
Incidence of Research Injuries and Subsequent Legal Actions
The reported incidence of research injuries generally appears to be quite low. NIH and FDA do not require investigators or sponsors to report research injuries, and there is no registry of injuries for publicly or privately sponsored research. A 1975 survey of principal investigators found that only 3.7 percent of research participants had been injured, with less than I percent of the injured participants suffering permanently disabling or fatal injuries (Cardon et al., 1976). The incidence of injury in this survey was not separated by gender or by age; thus the incidence of injury to women, or to offspring as a result of women's participation in clinical studies, has not been quantified.
Legal recourse is sought in only a small percentage of research injuries, and an even smaller number ever reach court. The NIH Office of the General Counsel is only aware of three legal actions for clinical research injuries where NIH was involved in the past 20 years5 According to the FDA Office of the General Counsel, that agency has never been the subject of a legal action resulting from a clinical trial injury (R. Blumberg, personal communication, August 1993).
There have been approximately two dozen reported legal cases concerning research injuries-that is, cases in which a written opinion was officially published and thus available to courts, lawyers, and the public. Opinions are normally published in case reporters; occasionally, an opinion may be available only through on-line legal databases such as LEXIS and WESTLAW. For the most part, reported opinions are rendered by appellate courts and federal district courts. Cases that are decided at the state trial level usually are not reported, nor are injury claims that are settled out of court (case records are often sealed as part of the settlement agreement). Thus, the small number of reported research injury cases do not reflect all actions initiated as a result of research injury.
At a recent Institute of Medicine (IOM) workshop on AIDS vaccine clinical trials, the general counsel for a small U.S. pharmaceutical company
stated that nine legal actions had been filed against his firm seeking damages for adverse reactions resulting from clinical trials of one drug. As of 1991 (the date of the workshop), the company had not lost a court judgment for any of these cases, but several cases had been settled at substantial cost to the company (IOM, 1991 b).
Because of the extensive disclosure involved in the informed consent process, those injured in research seldom have a basis to pursue legal action (Flannery and Greenberg, 1994). When participants take part in the informed consent process, they may feel they have assumed the risks involved in research and therefore be less likely to initiate legal action (Clayton, 1994). Participants may also be less likely to start legal proceedings because they usually receive excellent medical care if they have an adverse reaction. In addition, corporations generally take action to avoid risk, and plaintiffs have difficulty in proving that their injuries were caused by the research. All of these considerations may contribute to the lack of legal activity in the area. With regard to injuries to offspring, until very recently FDA guidelines discouraged the use of women of childbearing potential in early phases of research; hence the full potential for liability for injuries to offspring may be unrealized.
Even though there are few reported cases of research-related injury, fear of liability has played a key role in the exclusion of women in their childbearing years and pregnant women from many clinical studies. Factors that contribute to the intimidating legal landscape include the following:
- Legal actions can be extremely costly to defend, even if the plaintiff's case is weak and the question of liability is uncertain, and companies are inclined to take any action that appears likely to eliminate or reduce the risk of becoming involved in litigation (Flannery and Greenberg, 1994)6;
- variation in liability rules among the states;
- uncertainty introduced by having highly technical and complex scientific issues evaluated by a judge and jury untrained in the sciences;7
- a manufacturer's inability to rely on FDA approval as protection against liability;
- perception that the number of legal actions filed and the size of awards have increased dramatically in recent years;
- fear of loss of public confidence because of adverse publicity (IOM, 1990).
Theories of Liability for Inclusion
Because of the paucity of reported decisions related to research-related injury, prediction of the risk of liability from the inclusion of women in clinical studies is difficult. Because of this difficulty, the committee found
it helpful to examine the theories on which such an action might be brought. Legal actions for research injuries have generally been based in tort law, the branch of the law that allows persons injured by certain conduct to seek monetary compensation. Tort law is the province of state law, created through state legislation and the state "common law" developed through judicial decisions. Although variations exist among the states, state tort rules are similar enough that generalized statements of liability can be made. For example, many states, through their courts and state legislatures, have adopted portions of the Restatement (Second) of Torts, the American Law Institute's synthesis of the major principles of contemporary tort law (ALI, 1977). The three legal bases for a legal action for research injury are battery, negligence, and strict liability. The most common application of negligence in the area of research injury is lack of informed consent.
Battery Battery generally is defined as unlawful and intentional bodily contact, directed at another person without that person's consent (Keeton et al., 1984). Battery is an intentional tort; thus an intentional contact that is unpermitted, harmful, or offensive is a necessary element. In the context of research, if someone is used as a research subject without his or her knowledge or consent, all of the potential defendants may face a legal action for battery. Damages awarded in a battery action will generally include compensatory damages, which are designed to compensate the plaintiff for any harm resulting from the contact and will include payment for medical expenses, loss of wages, and pain and suffering. Punitive damages, which are intended to punish the defendant for willful, injurious behavior, may also be awarded in a battery action (Keeton et al., 1984).
Negligence Negligence is deviation from acceptable standards of conduct. To establish negligence, the plaintiff must prove that: (1) the defendant had a legal duty to the plaintiff; (2) the defendant breached that duty; (3) the plaintiff suffered an injury; and (4) this injury was caused by the defendant's breach of its duty (Keeton et al., 1984). The duty owed may be the general obligation that persons have to exercise reasonable care toward other people and their property. A duty may be mandated or implied by a statute, regulation, or guideline, as is the case in the context of research on human subjects. The duty also may be shaped by whether the defendant reasonably could have foreseen the injury to the plaintiff.
The existence of a duty will in turn create responsibilities on the part of the defendant to exercise due care. The level of care that a defendant must exercise to avoid liability, called the standard of care, is determined by what is reasonable under the circumstances. The defendant will be found to have breached a duty if his or her behavior fell below the standard of care (Keeton et al., 1984). In the research context, the standard of care is shaped
in part by the conduct deemed reasonable when one is entrusted with human lives. Federal policies for research on human subjects also may help set the standard of care.
To recover damages for negligence, the plaintiff must also prove that the defendant's actions caused the injury. Causation can be a difficult concept: the plaintiff must show that, "but for" the actions of the defendant, the injury would not have occurred. This is "cause in fact." But the plaintiff must also establish legal cause, or proximate cause. The defendant is only responsible for the damages that are aforeseeable consequence of his or her behavior. For example, courts have differed as to whether a party may be liable for injuries to offspring that occur as a result of an injury to the woman before the child's conception (see below for more on preconception liability).
In most negligence cases, only compensatory damages are awarded. In cases where the defendant's behavior was reckless or an extreme departure from the standard of care, commonly known as gross negligence, however, some states allow a plaintiff to recover punitive damages as well (Keeton et al., 1984). Third parties may also recover money damages for injury to a research participant caused by negligence. For example, a spouse or child may claim loss of consortium, the legal term for the loss of "the society and affection" of the injured person.
Finally, while it is necessary that there be some injury in order to recover damages, the injury need not necessarily be tangible. In some cases, for example, plaintiffs exposed to toxic substances recovered damages for anxiety caused by fear of getting cancer (Reisman, 1992).
Strict liability Pharmaceutical manufacturers, because they are in the business of selling a product, may also be held to the legal principles governing product liability. Under strict liability, a person injured by a product can recover damages without having to show that the manufacturer was negligent (ALI, 1977). Strict liability proponents believe that manufacturers should bear this cost because they can distribute the costs of injuries to all persons who benefit from the product through their power to set the market price of the product (Campbell, 1969).
A manufacturer may be strictly liable if it sells a product "in a defective condition unreasonably dangerous to the user or consumer" (ALI, 1977). A manufacturer may keep its product from being considered "unreasonably dangerous" by giving appropriate warnings; the adequacy of these warnings is often the legal issue in contention. In a research context, the manufacturer is not selling the drug directly to the participant; nevertheless, one state court rejected the notion that strict liability would not apply to drugs in the experimental phase because they were not sold (Merton, 1992, citing Gaston v. Hunter).
Courts in some states have exempted drug manufacturers from strict liability by adopting an explanatory comment from the Restatement of Torts. Comment k states that a drug is not unreasonably dangerous, and thus its manufacturer is not subject to strict liability, if the drug is "properly prepared and marketed and a proper warning is given" (ALI, 1977). Commentators have argued that drugs in the experimental phase are particularly strong candidates for "comment k protection," because the language of comment k specifically refers to experimental drugs: "new or experimental drugs as to which, because of lack of time and opportunity for sufficient medical experience, there can be no assurance of safety" (Flannery and Greenberg, 1994). States differ in the extent of their adoption of comment k for drugs; some states have chosen to apply its protection only on a case-by-case basis.
Some courts have found that manufacturers have a duty to warn consumers directly-for example, for prescription contraceptives (Flannery and Greenberg, 1994). Under the "learned intermediary doctrine," however, manufacturers can usually satisfy their duty to warn for prescription and investigational drugs by informing physicians and investigators of any risks of harm. The physicians are then responsible for prescribing drugs only for appropriate indications and for monitoring their use.
Legal actions for research injury based on a negligence theory frequently involve the doctrine of informed consent. Because nearly all research is conducted after securing consent from participants, most legal actions by participants for research injuries will be based on whether the information given to the participant before securing consent adequately warned of the potential risks (Reisman, 1992). The doctrine of informed consent requires that physicians secure consent from the patient before medical treatment is administered, and this consent must be based on information given to the patient about the risks and benefits of the proposed procedure and the alternatives (Wadlington, 1984).
There is a distinction between the nature of the consent necessary to avoid a legal action for battery and that necessary to avoid an action for negligence. Consent to avoid a charge of battery is a form of first-order assent to a bodily intervention, sometimes referred to as simple consent. By contrast, consent to avoid an action for negligence generally requires what has come to be called informed consent, a consent based on the disclosure of all facts "which are necessary to form the basis of an intelligent consent by the patient to the proposed treatment" (Salgo v. Leland Stanford Jr. University). In the context of research, if persons are subjected to a study without their knowledge or consent, any of the potential defendants may be
sued for battery (see Appendix D). If the initial consent to participate has been secured but it was obtained without adequate disclosure of risks and alternatives, the legal action will be based on lack of informed consent, which is considered to be negligence.
This distinction is important. Damage awards for negligence will be smaller than those for battery, because they only compensate the plaintiff for his or her injuries; for an intentional tort such as battery, punitive damages may also be awarded. The statute of limitations-which limits the number of years during which a legal action can be initiated-is usually longer for a negligence action. For a battery, the plaintiff needs to prove there was an intentional contact with her body without her consent. In a negligence action for lack of informed consent, the patient will usually present the testimony of experts on how much information would be considered to be a reasonable disclosure under the circumstances. This is part of demonstrating to the court the standard of care for informed consent, which is necessary in order to prove that the defendant breached this standard. The plaintiff also must then prove that she would not have chosen to be a participant in the research had she been given more complete information (Shack v. Holland).
The federal regulations on informed consent for research will influence the standard against which a defendant will be judged; that is, whether the degree of disclosure was reasonable given the circumstances. Meeting these regulations, however, does not shield potential defendants from liability. The mechanisms for securing informed consent in any research protocol will also be subjected to scrutiny under the particular state's criteria for the standard of care for negligence actions. Some states allow physicians to set the standard by inquiring into the extent of disclosure that is customary for physicians practicing in the community. Other states apply a more objective standard: what a prudent person in the patient's position would want to know about the possible risks (Wadlington, 1984). Many states have adopted a more subjective standard based on how much information is needed in order to allow a particular patient (the plaintiff) to make a decision (Keeton et al., 1984).
A recent decision from a North Carolina court indicates that courts may require a higher standard for informed consent in nontherapeutic research injury cases. In Whitlock v. Duke University, the plaintiff, a participant in a nontherapeutic study looking at high-pressure nervous syndrome in underwater diving, signed a consent form advising him of known physical risks and the possibility of unknown risks. The plaintiff claimed he suffered organic brain damage from the dive and sued the researcher and the university, citing negligent failure to warn about the danger of brain damage. The court cited the DHHS regulations on informed consent for nontherapeutic research and held that the researcher and the university had a duty to inform
the plaintiff of all ''reasonably foreseeable" risks. Although the court found that in this particular case the risk of brain damage was not reasonably foreseeable, it nevertheless articulated a strict standard for informed consent to nontherapeutic research (Kobasic, 1988, citing Whitlocki
Liability of Potential Defendants
Targets of legal action-potential defendants-in the research context may include public and private entities who sponsor or oversee research (such as NIH, FDA, and pharmaceutical manufacturing firms), as well as those who approve and conduct research (such as IRBs, investigators, physicians, and research institutions). If an offspring is injured as a result of a parent's participation in a drug trial, the parent could also be a defendant, together with the research sponsor (see below).
Government agencies The liability of government agencies, because they conduct research in the name of the federal government, is a special one and is spelled out in federal law. Historically, under the doctrine of sovereign immunity, the United States could not be sued, nor could its agencies and departments. The Federal Tort Claims Act (FTCA), passed in 1964, sharply restricted federal sovereign immunity; government agencies that conduct, sponsor, or oversee research, such as NIH and FDA, may now be held liable if a research participant is injured as a result of negligence (28 U.S.C. §§ 1346(b), 2674 ). Plaintiffs must sue the government agency in order to recover; the Federal Employees Liability Reform and Tort Compensation Act of 1988 (P.L. 15-11) protects federal employees from personal liability for tortious actions committed within the scope of their employment. The federal government cannot be sued for injuries resulting from the negligence of persons not directly employed by the federal government but who are working with full or partial federal financial support. For example, NIH could not be sued for the negligence of an investigator in an extramural research protocol (United States v. Orleans).
FTCA prohibits legal actions for battery and those arising out of the exercise of a "discretionary function," a provision that has been applied to exempt some activities where the government employee was acting pursuant to a federal statute, regulation, or guideline (Wion, 1989). It is possible that this "discretionary function exception" to FTCA may operate to bar a legal action for research injury against a federal agency such as NIH or FDA. In deciding whether a particular employee's actions qualify under the discretionary function exception, the court will examine whether the employee's actions involved an element of choice or judgment and whether this choice was a permissible exercise of policy judgment (Wion, 1989). If the federal policy is a general one, allowing federal employees some discretion in how
it is implemented, the federal agency may not be sued in tort if the employees' actions result in injury. If the policy is written with detailed requirements, leaving little discretion for federal employees, however, the agency may be subject to liability if an employee's implementation of that policy causes injury (Wion, 1989).
The outcome of a legal action for research injury against a federal agency, such as NIH or FDA, is uncertain. Much of the conduct of research sponsors and investigators is spelled out in federal policies, but there are also areas of considerable discretion on the part of federal employees. For example, in FDA regulations governing new drug applications (NDAs), there are no product-specific regulations; the regulations contain only general information about the criteria for approval. Consequently, legal actions against FDA in connection with NDAs have been dismissed by federal courts under the discretionary function exception (Wion, 1989). But in one case where an injured plaintiff claimed FDA made a decision to release a polio vaccine based on inadequate animal test data, the court held that FDA's assessment of the animal data was a scientific determination, not the exercise of a discretionary function, and thus subject to liability (Griffin v. United States). It is not clear whether the particular actions of a researcher/investigator in implementing federal policy regarding research on human subjects would be considered a scientific determination or a discretionary policy judgment.
IRBs and research institutions The research institution and its IRB may also be held liable for approving a negligent protocol or for not closely monitoring or supervising the ongoing research. Under the tort doctrine of "respondeat superior," hospitals may be held liable for the negligent activities of their employees (Campbell, 1969). In Friter v. IOLAB Corporation, a Pennsylvania court recently found that a hospital, as a participant in a clinical investigation for the FDA, had assumed a duty under federal regulations to ensure that informed consent was obtained from research participants.
There have been no reported cases in which IRB members were successfully sued for breaching their duty to protect research subjects, male or female (Flannery and Greenberg, 1994). Nevertheless, the potential exists for liability of both IRBs and their individual members. An IRB could be found negligent, for example, if it approves an informed consent process later found to be inadequate. Federal policies on informed consent are detailed but leave some discretion to the IRB-for example, whether the research warrants inclusion of a statement about possible unforeseeable risks to an embryo or fetus. In addition, the regulations permit an IRB to waive certain elements of the consent form if it determines that only minimal risk is involved (Bordas, 1984).
Federal regulations require that an IRB ensure that investigators obtain
the consent of the participants, but most IRBs trust the investigator to obtain proper consent. Regulations authorize IRB members to observe the consent process, but a 1982 poll found that only 2 out of 100 IRB members had ever required anyone other than the investigator to be involved in the informed consent process (Robertson, 1982).
Given the potential conflict of interest that an investigator brings to the consent process (it is in an investigator's best interest to acquire participants), a court could find that an IRB negligently relied on the investigator if it finds the participant never gave informed consent or was coerced into consenting.
Courts may also find IRBs negligent if they fail to conduct a continuing review of approved research, as required by statute. A 1978 study by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research revealed that only half of the IRBs had a policy requiring investigators to report participant injuries; just over a third designated someone to observe a research project; and only half of this group routinely did so (Bordas, 1984). The courts might find that the negligent actions of the investigator were an intervening cause of the injury, absolving the IRB from liability, but if the IRB should have anticipated the actions of the investigator, it can still be found negligent for exposing participants to the risk (Bordas, 1984, citing Prosser, 1971).
If the IRB is part of a state-owned institution, it may be considered an agent of the state. The doctrine of sovereign immunity might bar legal actions against the state, thus protecting the IRB and its members from liability. Some states retain their own sovereign immunity, but over the past few decades the doctrine has been weakened by exceptions. There is also the possibility that members of hospital-based IRBs might be protected by state statutes that provide immunity to members of hospital review committees (Bordas, 1984). Thus, while there is potential for IRBs to be the targets of legal action, both as individuals and as an entity, no IRB has been successfully sued for a research injury, and they may be granted immunity under various state-created immunity doctrines.
Liability for Injuries to Offspring
As mentioned above, the greatest source of concern about liability is the possibility of injury to offspring when women of childbearing potential are included in clinical drug trials. Fear of liability to offspring may be based on a number of factors. First, the fear is inspired by the experiences of diethylstilbestrol (DES) and thalidomide cases, where offspring were seriously injured by the mother's ingestion of the drug during pregnancy (see Appendix C for DES Case Study). As a result, there is a high level of concern about the inclusion of women of childbearing potential in drug
trials, particularly in early phase studies when effects on both adult and offspring are largely unknown. Second, although it may be difficult for a plaintiff to prove causation, the magnitude of harm that could be alleged to result from in utero exposure to a drug is great. Third, the statute of limitations usually is longer for cases of injury to children, and damage that occurs in utero may not show up until years later, leaving potential defendants liable for an indeterminate amount of time. Finally, securing a parent's consent to the research is unlikely to preclude recovery by a child for injuries in utero, because strong arguments can be made that a parent cannot waive a child's rights to sue. The rationale behind such a policy is to avoid conflicts of interest between parents and children (Clayton, 1994; Apicella v. Valley Forge Military Academy and Junior College).
Tort law and prenatal injuries Initially, recovery for prenatal injuries was denied because there was no authority for such an action; an unborn child was considered to be part of its mother, without a separate legal existence. This view persisted until 1946, when a federal court in Bonbrest v. Kotz held that a child may recover after birth for injuries suffered as a viable fetus. Many states adopted this rule, with the requirements that the child be born alive and that the injury occur when the fetus was viable (Shack v. Holland). The viability rule was subsequently abandoned in most jurisdictions, and prenatal injury claims in general now are seen as compensable, sometimes with large damage awards (Clayton, 1994).
Courts have shown some reluctance to recognize a cause of action for preconception injury. It may be difficult for the plaintiff to establish the necessary causal link between the defendant's behavior and the child's injuries. Courts are also uneasy with the notion of extending liability back to before the point of conception, because this greatly expands the boundaries of foreseeable injuries for which a defendant can be held responsible (Merton, 1992). Some states have allowed recovery for preconception injuries, while others have denied recovery.8 The best-known cases for preconception injury are those of "DES granddaughters," women whose mothers were exposed to DES in utero; nearly all of these claims have been rejected by the courts (Sherman, 1990; Enright v. Eli Lilly & Co.). (See Appendix C for more information on the DES cases.)
Potential plaintiffs in a legal action for injuries to offspring include the child, as long as he or she is born alive, and the parents. The tort principle will vary depending on whether the child was born alive and what the parents would have done if they had been adequately informed about risks to offspring. When the plaintiff is the child, the likelihood of a successful legal action depends in part on the ability of the parents to show they would not have participated in the research had they known about the risks. Damage awards to these children can include pain and suffering and damages to
cover their special needs over the course of a lifetime (Clayton, 1994; Keeton et al., 1984). If the parents claim, however, that they would not have had the child if they had been informed of the risks, the child's action is one of wrongful life. Very few states have recognized such an action (Clayton, 1994).
Parents of a liveborn child who allege that they would not have participated in the research if they had known about the risks may recover the additional medical expenses that are not covered in the damage award from the child's legal action. Most of the time they may not recover for their own pain and suffering, because they are considered to be "bystanders" to their child's injury. If the parents claim that they would not have had a child had they been fully informed of the risks of the research, their action is one for wrongful birth. If the child is stillborn as a result of a research injury, the parents may bring an action for wrongful death; in states that permit such a cause of action, the damage award is usually relatively small (Clayton, 1994).
Potential defendants In addition to all of the defendants discussed above, it may be possible for the child to sue a parent for participating in a clinical drug trial, depending on the law of parent-child immunity in the state. Under parent-child immunity, children are not permitted to sue their parents. Some states have begun to limit parent-child immunity; others have eliminated it; still others have declared that only some parental decisions will give rise to liability; and a few states retain complete immunity (Clayton, 1994).
Existing case law in this area focuses on the potential of a woman's behavior to result in a bad pregnancy outcome. At least three states have addressed children's claims that they were injured by their mothers' behavior during pregnancy, although none were in the context of research. Two states, Michigan and New Hampshire, permitted the child to recover damages; in one of these cases, the child alleged injury resulting from the mother's ingestion of tetracycline while she was pregnant (Grodin v. Grodin; Bonte v. Bonte). A third state, Illinois, denied the cause of action, stating that to allow such a claim would intrude too deeply into the lives of pregnant women (Stallman v. Youngquist).
For states that allow such recovery, the child must show that the woman was negligent in her behavior-that her choice was one that other reasonable people would not have made. In the research context, this would require an inquiry into whether it was reasonable for the woman to choose to participate in the protocol. According to one commentator, other factors that should be considered include the general deference given to patients to choose their own treatment, the seriousness of the woman's medical problem, and the availability of nonfetotoxic alternatives. The commentator also
notes that wrongful life actions-"I should have never been born"-are extremely unlikely to succeed against parents (Clayton, 1994).
In cases where alleged injury to offspring stems from the mother's participation in a drug trial, pharmaceutical manufacturers and the clinical investigators may also be sued. One commentator has suggested that, when the informed consent of the woman was obtained, the third-party defendant may join the woman as a defendant in the action, claiming that she was an intervening cause of the injury or that her choice to participate in the trial was contributory negligence (Merton, 1992). In tort law generally, the presence of an intervening cause may operate to extinguish liability for the other party, but public policy may dictate that a parent not be recognized as an intervening cause in this situation. Another commentator has argued that allowing third-party defendants to escape liability by bringing the woman into the action would contradict the rationale for not allowing parents to release the legal claims of their children (Clayton, 1994). Finally, the court also may seek to have the defendant with greater financial resources be responsible for paying the damages; in most cases this will be a third party rather than a parent.
Sufficiency of informed consent Under the limited case precedent available, the likelihood of a successful legal action for offspring injury has depended on whether obtaining informed consent from the woman is sufficient, in all circumstances, to avoid liability. The committee is aware of emerging data concerning the possibility of male-mediated developmental effects (see Chapter 7); nevertheless, this possibility has not operated to exclude fertile men as a class from clinical drug trials. Hence the committee frames the issue in terms of the woman's participation in clinical trials.
Of all of the reported cases of research injury, only two concerned injuries to offspring. Both cases came out of the University of Chicago experiments with DES in pregnant women, where there was a failure to secure the woman's consent to participate in the research (see Appendix C). There is no case precedent to set the boundaries for liability for injuries to offspring when the woman's valid consent to the research has been secured.
At least one court has held that the informed consent of the woman is sufficient to avoid liability for injury to offspring if the research is a therapeutic intervention for the fetus (Roberts v. Patel). The rationale of this one case seems to imply that if the woman has a serious illness and is participating in a clinical trial of a treatment that could be beneficial to her, there will likely be no recovery in tort for injuries to her or to her offspring, as long as there was no negligence involved and the informed consent of the woman was properly secured. The risk of liability may be higher, however, for both the woman and the pharmaceutical company, if the intervention is a benefit to the mother but is not for a serious illness, or if there are safer alternative
interventions available and the risk to the fetus is known or suspected to be significant.
For trials of drugs that are completely nontherapeutic to either the woman or fetus (for example, early phase trials of drugs where testing is done in healthy participants), the outcome of an offspring injury case is even more unclear. DHHS regulations on research in children may be instructive here. Under the current regulations, in some cases parents may consent for their minor children to be research subjects in protocols that do not directly benefit the children. DHHS may fund such research only where the IRB finds that:
- The risk represents a minor risk over minimal risk.
- The intervention or procedure presents experiences to subjects that are reasonably commensurate with those inherent in their actual or expected medical, dental, psychological, social, or educational situations.
- The intervention or procedure is likely to yield generalizable knowledge about the subjects' disorder or condition that is of vital importance for the understanding or amelioration of the subjects' disorder or condition.
- Adequate provisions are made for soliciting assent of the children and permission of their parents or guardians (45 C.F.R. 46.406). Permission of both parents is required unless one parent is deceased, unknown, incompetent, or not reasonably available, or when only one parent has legal custody of the child (45 C.F.R. 46.408(b)).
For research involving greater than minimal risk, there is no category for research of no benefit to the child; however, the regulations do specify that, for research that would not otherwise be approved but "would present an opportunity to understand, prevent, or alleviate a serious problem affecting the health or welfare of children," the IRB must submit such research to DHHS for approval (45 C.F.R. 46.407). A court may find, however, that these regulations governing research on minors do not apply to offspring injury resulting from a woman's participation in a clinical study.
If a woman consents to participate in a trial where the treatment is not for a serious illness and presents a known or possible significant risk of harm to the fetus, what would be the result? Because there is no legal precedent for such a scenario, the committee can only speculate about the outcome. If the research has been pursued negligently, or the informed consent process was not legally sound, it is possible that the offspring would recover under the same legal principles applicable to medical malpractice, as discussed above.
If there is no negligence, and informed consent of the woman has been properly secured, it may still be possible for the offspring to recover. Because it is the IRB's responsibility to ascertain whether the benefits of a
particular research protocol outweigh the risks for any class of potential participants, the IRB could be liable for approving a protocol that permits participation of pregnant women where there is no benefit to the woman and some risk to the fetus. At least one commentator has noted that there may be settings where third parties are not entitled to rely on even fully informed consent of the woman to guard against liability, particularly if the protocol poses serious risks to the fetus while offering little benefit to the woman (Clayton, 1994).
Yet another commentator noted that there is no precedent for imposing liability on a researcher who has properly obtained informed consent for harm to a participant's offspring. Because the harm is done by the woman's choice to participate in a research protocol, with full knowledge that it might have damaging consequences, the woman's negligent choice is an intervening cause in the injury (Merton, 1992). There is considerable disagreement over whether a court would find a mother legally liable for injury to offspring resulting from her participation in a clinical trial, but some agreement about the remote likelihood of successful legal action by an injured offspring against a pharmaceutical manufacturer who has obtained valid informed consent from the mother (Reisman, 1992; R. Blumberg, FDA Office of the General Counsel, personal communication, August 1993; Flannery and Greenberg, 1994).
Consent of the father9 As mentioned above, there may also be risks to offspring when fertile men participate in trials of drugs that may cause damage to germ cells. The issue here, however, is whether the father's consent is necessary to avoid liability for injury to the offspring when the mother participates in a clinical trial of little or no benefit to her and with risk of harm to the fetus. There is no explicit policy or case law, but analogies from related areas may apply. For medical treatment of minors, one parent's informed consent is sufficient (Holder, 1985); DHHS regulations for research on children also require the consent of only one parent. Because the concern is with harm to a fetus rather than a child, federal regulations with respect to fetal tissue research may also be relevant; where such research is permitted, it requires informed consent of both father and mother, unless the father's identity and location cannot reasonably be ascertained, the father is not reasonably available, or the pregnancy resulted from rape (45 C.F.R. § 46.209-210). This committee is concerned with the participation of women in clinical research, however, and thus the case law on abortion, where a woman chooses for her own benefit to undergo a medical procedure, may be more applicable. An earlier section of this chapter discusses how the Supreme Court has invalidated state laws requiring a woman to obtain spousal consent or give spousal notification before obtaining an abortion. Based on these examples, it is unlikely the father's consent could
be required in order for a pregnant woman to participate in a clinical trial, or that the addition of the father's consent would shield a manufacturer from liability.
Is Johnson Controls Relevant?
Some commentators have suggested that the language of the recent decision by the Supreme Court in Johnson Controls, discussed in the foregoing section on constitutional issues, might be influential in future case decisions in other contexts. While certainly not decisive, it may be pertinent in decisions relating to liability for injuries to offspring from clinical research. The company argued that fear of tort liability justified its policy of excluding women from jobs with potentially high lead exposure. The Court rejected this argument noting that "if, under general tort principles, Title VII bans sex-specific fetal protection policies, the employer fully informs the woman of the risk, and the employer has not acted negligently, the basis for holding an employer liable seems remote at best." Since Johnson Controls was a case of employment discrimination, this comment on the unlikelihood of liability is not to be interpreted as definitive, but it may provide support for persons who assert that adequate informed consent from the woman would be sufficient protection against liability.
Liability for Exclusion
Liability for excluding women from clinical trials may be a serious risk, particularly for pharmaceutical manufacturers and, indirectly, for physicians. Manufacturers' liability results when, after a drug is on the market, evidence emerges that the drug is more dangerous or less effective in women (Flannery and Greenberg, 1994). For example, a woman may have an adverse reaction from one of her prescriptions and discover that the drug was never tested in women. Her injury is not a research injury as this term has been used thus far. In this example it is the woman's exclusion from clinical research, not her inclusion, that caused the injury. Increased awareness on the part of women that they are not always represented in the populations tested may contribute to an increasing number of legal actions for exclusion.
Under strict liability principles, manufacturers may be held liable for the defective design of a product, and a drug that has not been adequately tested may be found to be defectively designed (Flannery and Greenberg, 1994). In addition, manufacturers must warn about not only the known risks, but also foreseeable risks that should have been known if "reasonable, developed human skill and foresight" had been applied (Flannery and Greenberg, 1994, quoting ALI, 1977).
The duty to warn about foreseeable risks requires that pharmaceutical manufacturers apply state-of-the-art testing methods to their products. With all of the recent publicity about physiological differences between men and women with regard to drug efficacy, dosing, and adverse reactions, it would be difficult to argue that all-male studies of drugs that may be used by women represent state-of-the-art testing methods (Flannery and Greenberg, 1994). In states with a case-by-case determination of the application of comment k (which protects pharmaceutical manufacturers against strict liability-see above), manufacturers will have a difficult time arguing that a drug is "unavoidably" unsafe if they fail to test the drug in a population that might foreseeably use it (Bowles, 1992). Also, if the courts find that the manufacturer deliberately avoided learning about whether a risk was associated with its drug, the manufacturer could be liable for punitive damages as well (Flannery and Greenberg, 1994).
In contrast to the lack of reported legal cases of injury from inclusion in research, there are a number of cases in which damages were awarded to plaintiffs, in part because of inadequate testing of the drug before it was released into the market. Courts have no qualms about scrutinizing the research design of a clinical trial and criticizing research sponsors, not only for unreliable technique and sloppy data handling but also for a lack of response to actual market conditions (Merton, 1992, citing Tinnerholm v. Parke, Davis & Co). In one such case, West v. Johnson & Johnson Products, Inc., a woman who claimed that she had suffered from toxic shock syndrome resulting from use of a tampon marketed by Johnson & Johnson Products was awarded damages. The court found that the company had failed to study the basic microbiology of the human vagina, to test for vaginal infections, and to include women with a history of vaginitis in their human studies (Merton, 1992). In Taylor v. Wyeth Laboratories, the court found that a prudent manufacturer, once aware that women with type A blood experience a disproportionate number of pulmonary embolisms, would have looked at the relationship between blood type and blood-clotting risk in women taking oral contraceptives (Merton, 1992). Courts have also awarded damages to injured plaintiffs on the grounds of inadequate testing for adverse effects, even in cases where the FDA did not require further testing (Merton, 1992, citing Barson v. E.R. Squibb & Sons, Inc.).
In legal actions where the plaintiff claims inadequacy of premarket testing, however, plaintiffs may face some difficulty in proving that their injuries were caused by a failure to test for foreseeable risks. In Jones v. Ortho Pharmaceutical Corp., the plaintiff sought to hold Ortho liable for her carcinoma in situ, claiming that they failed to conduct clinical trials that would have established whether the oral contraceptive she took caused cancer. Because a tort action must prove that the defendant's behavior caused the injury, the plaintiff attempted to persuade the court that because Ortho
did not conduct the proper clinical trials, which would have enabled her to show the causal link between the contraceptive and her cancer, the company was presumptively liable. The California intermediate appellate court would not allow the plaintiff to presume causation from a lack of premarket testing (Merton, 1992).
Thus, although there is a general lack of case law on liability for injuries to research participants, there is some precedent for liability for exclusion from research. The case law suggests that if a drug was found to cause injuries to women, and yet women had been excluded from clinical trials of the drug, the pharmaceutical manufacturer might be held liable for failing to test the drug in women. For some drugs, however, the potential for teratogenic or mutagenic effects is low or the negative effects are manifested after a long latent period. For these drugs, even adequate testing in all relevant populations may not reveal their potential to cause harm. At least one commentator has noted that establishing surveillance systems, or requiring companies to keep track of and report adverse drug reactions, may provide plaintiffs with a source of evidence that a company knew or should have known that a particular drug or dose level was potentially dangerous and required further testing or a more adequate warning label (Gibbs and Mackler, 1987).
For physicians, liability resulting from exclusion of women from drug trials arises in the form of negligent drug prescription. For example, the physician could be liable for prescribing a drug to a woman: (1) for a different purpose than that for which it was initially designed and tested or (2) in disregard of the drug's label that it has not been tested in women. With regard to liability for offspring injury resulting from exclusion of women (or men) from drug trials, information about testing for reproductive and developmental effects is often not available for all drugs for which such testing would be appropriate.
CONCLUSIONS AND RECOMMENDATIONS
The many federal regulations governing research on human subjects do not provide investigators and IRBs with clear answers on issues concerning the inclusion of women and racial and ethnic groups in clinical studies. Recent changes in policies, however, have made them more consistent. All of the recent changes have been implemented to promote the inclusion rather than exclusion of women. Consistency and, where possible, congruence among these policies is important to promote compliance and prevent confusion.
Policies and regulations issued by the FDA or other PHS agencies must be harmonized with those of NIH. At the very least, policies and regulations issued by NIH and those issued by FDA must not be contradictory. The
closer the congruence between FDA and NIH regulations and policies, the more likely that a regulated institution will understand precisely what is required and be motivated to comply. When NIH updated its policy concerning inclusion of women in 1991, and added the sanction of possible reduction in the project's priority score for noncompliance, many in the research community came to believe that FDA and NIH policies were contradictory. Now that the FDA has updated its 1977 policy, and NIH is in the process of updating its 1991 policy, there is an opportunity to achieve congruence between the positions of the two agencies.
The committee recommends that NIH work closely with the FDA and with other PHS agencies to make regulations and policies on inclusion of women and racial and ethnic groups consistent with one another and, wherever possible, to make them congruent.
If the policies of the two agencies are harmonized, there will still remain the task of educating the research community concerning what is required, and motivating that community to comply. Enunciation of sound and congruent policies, in conjunction with a comprehensive educational program, will ensure that policies and the rationales for the policies are properly understood by the research community.
The committee recommends that NIH, in cooperation with FDA, should institute a comprehensive education program directed at investigators, institutions and IRBs on policies concerning the inclusion of women and racial and ethnic groups in clinical studies.
It is impossible to quantify the risk of tort liability from the inclusion of women in clinical studies at this time, because: (1) there is no complete compendium of unreported cases involving settlements and (2) women have not been included in some major studies in the past. Difficulties of prediction are compounded because tort law is governed by the individual states, with many variations on issues such as whether a woman's informed consent will serve to bar an independent action by a child injured as a fetus during such research. Analysis of existing legal rules and principles seems to indicate that the likelihood of successful damage actions is limited. Nevertheless, broadening the research population to include those groups previously excluded may also generate additional legal actions that will test existing legal doctrine.
A special set of concerns in the research area stems from the differing bases for liability according to which party is a defendant. A pharmaceutical company, for example, might be sued on the basis of strict liability, while a researcher ordinarily would be sued only on the basis of negligence.
With regard to the latter, the new federal policies calling for inclusion of women in clinical studies will help establish new standards that will be relevant to legal actions.
Many of the concerns voiced about liability in the context of research including women are the same as those with regard to the tort system in general. For example, expert scientific testimony is necessary to establish that a particular drug caused an injury. There are inherent difficulties in assuring the unbiased nature of such testimony in what are often highly technical cases.
The committee recommends that current and future initiatives toward general tort reform include attention to issues of research-related injury, including issues of proof of causation.
The question of whether there should be a special compensation scheme for injuries sustained by children as a result of a parent's participation in a clinical study is similar to that raised in the context of research subjects in general. The committee does not recommend adoption-at this time-of a special compensation scheme limited to coverage of children injured prenatally or preconceptually. Any new compensation scheme focusing only on such injuries poses especially difficult problems with regard to establishing causation and averting large numbers of questionable recoveries. Appendix D discusses several existing compensation schemes dealing with children and illustrates these and other difficulties.
The committee recommends that NIH thoroughly review the area of compensation for research injury in general and that consideration of implementation of any compensation scheme include attention to prenatal and preconceptual injuries to children resulting from a parent's participation in a clinical study.
Our current health care reimbursement system does not include coverage for medical care resulting from injuries sustained during research. This could be accomplished through a system of universal access with adequate coverage.
The committee recommends that health care reform efforts include considerations of medical care for research-related injury.
The committee recognizes that, regardless of their basis or justification, fears about liability are real. On balance, however, the committee concludes that liability concerns should not represent an impediment to implementation of public policies that favor the broader inclusion of women in clinical studies.
See the paper by Professor Debra DeBruin in Volume 2 of this report for further analysis of whether women are owed affirmative action measures as a result of past discrimination in research.
v. Merrell Dow Pharmaceuticals, the Court explained that Frye's general acceptance standard was superseded by the adoption of the Federal Rules of Evidence. Under the Federal Rules, the trial judge makes a flexible determination of whether the evidence rests on a reliable foundation and is relevant to the task at hand. The effect this change in evidentiary rules will have on the liability climate for research injuries is unknown.
"Father" is the term chosen here to include the biological father of the baby, the child's legal guardian, or anyone the law would recognize as having legal responsibility for the welfare of a child.
ALI (American Law Institute). 1977. Restatement (Second) of Torts Section 402A.
Bordas, L.M. 1984. Note: Tort liability of institutional review boards. West Virginia Law Review 87:137-164.
Bowles, L.E. 1992. The disenfranchisement of fertile women in clinical trials: The legal ramifications of and solutions for rectifying the knowledge gap. Vanderbilt Law Review 45(4):877920.
Bush, J.K. 1993. Women's participation in clinical research: The industry perspective. Presented at Conference on the Inclusion of Women and Minorities in Clinical Research, Bethesda, Md., June 28, 1993.
Caban, C. 1993. Personal communication with T. Johnson. May 7. (Dr. Carlos Caban advises the NIH Office of Research on Women's Health with respect to its policy on the inclusion of women in clinical research.)
Campbell, J.E. 1969. Civil Liability for Investigational Drugs: Part II. Temple Law Quarterly 42(4):289-354.
Cardon, P.V., Dommel, F.W., Jr., and Trumble, R.R. 1976. Injuries to research subjects: A survey of investigators. New England Journal of Medicine 295:650-654.
Charo, R.A. 1994. Constitutional issues raised by the exclusion of women from research trials. 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.
Clayton, E.W. 1994. Liability exposure when offspring are injured because of their parents' participation in clinical trials. 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.
Department of the Army. 1992. Memorandum from Gregory P. Berezuk, Chief, Human Use Review and Regulatory Affairs Office, Office of the Surgeon General, re: Human subjects research review board policy regarding pregnancy testing of research subjects (January 21).
Department of Energy. 1992. Protecting Human Research Subjects at the Department of
Energy: Human Subjects Handbook. Washington, D.C.: Office of Health and Environmental Research, Office of Energy Research, Department of Energy.
Department of Veterans Affairs. 1992. Inclusion of women and minorities in clinical research. Veterans Health Administration Manual M-3, Part 1, Chapter 14 (July 24).
Flannery, E., and Greenberg, S.N. 1994. Liability exposure for exclusion and inclusion of women as subjects in clinical studies. 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.
FDA (Food and Drug Administration). 1993. Guideline for the study and evaluation of gender differences in the clinical evaluation of drugs. Notice. Federal Register 58(139):3940639416.
FDA. 1977. General Consideration for the Clinical Evaluation of Drugs. Rockville, Md.: FDA.
Gibbs, J.N., and Mackler, B.F. 1987. Food and Drug Administration regulation and products liability: Strong sword, weak shield. Tort and Insurance Law Journal 22(Winter):194243.
Green, V. 1993. Doped Up, Knocked Up, and Locked Up?: The Criminal Prosecution of Women Who Use Drugs During Pregnancy. New York: Garland.
Harvard Medical Practice Study. 1990. Patients, doctors, and lawyers: Studies of medical injury, malpractice litigation, and patient compensation in New York . Boston: Harvard Medical Practice Study.
Holder, A.R. 1985. Legal Issues in Pediatrics and Adolescent Medicine. New Haven: Yale University Press.
IOM (Institute of Medicine). 1989. Medical Professional Liability and the Delivery of Obstetrical Care. Washington, D.C.: National Academy Press.
IOM. 1991a. Issues in the Inclusion of Women in Clinical Trials. Report of Planning Panel of the Institute of Medicine's Division of Health Sciences Policy. March 1-2. Washington, D.C.: Institute of Medicine.
IOM. 1991b. Expanding Access to Investigational Therapies for HIV Infection and AIDS. Summary of a Conference, March 12-13, 1990. Washington, D.C.: National Academy Press.
IOM. 1993. Inclusion of Women in Clinical Trials: Policies for Population Subgroups. Washington, D.C.: National Academy Press.
Keeton, W.P., Dobbs, D.B., Keeton, R.E., and Owen, D.G., eds. 1984. Prosser and Keeton on the Law of Torts. St. Paul, Minn.: West.
Kinney, E.L., Trautmann, J., Gold, J., Vesell, E.S., and Zelis, R. 1981. Underrepresentation of women in new drug trials. Annals of Internal Medicine 95(4):495-499.
Kobasic, D.M. 1988. Institutional review boards in the university setting: Review of pharmaceutical testing protocols, informed consent and ethical concerns. Journal of College and University Law 15(2):185-216.
Mariner, W.K. 1994. Compensation for research injury. 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.
Merton, V. 1992. The Exclusion of Pregnant, Pregnable, and Once-Pregnable People a.k.a. women) from Biomedical Research. Unpublished Paper (December 12).
Merton, V. 1994. The impact of current relevant federal regulations on the inclusion of female subjects in clinical studies. 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.
NIH (National Institutes of Health). 1990. NIH Instruction and Information Memorandum OER 90-5. December II.
NIH, Office of Research on Women's Health. 1991. National Institutes of Health: Opportuni-
ties for Research on Women's Health. Summary report of a conference held at Hunt Valley, Md., September 4-6.
NIH. 1992. NIH Data Book 1992: Basic Data Relating to the National Institutes of Health. Bethesda, Md.: NIH.
NIH and ADAMHA (Alcohol, Drug Abuse, and Mental Health Administration). 1990. NIH/ADAMHA Policy Concerning Inclusion of Women in Study Populations. NIH Guide 19(31):18-19.
NOW Legal Defense and Education Fund. 1992. Citizens' Petition to the FDA, signed by AIDS Service Center HIV Law Project, NOW Legal Defense and Education Fund, and American Civil Liberties Union, AIDS Project. December 15.
NRC (National Research Council). 1989. Biologic Markers in Reproductive Toxicology. Washington, D.C.: National Academy Press.
NRC (National Research Council)/IOM. 1990. Developing New Contraceptives: Obstacles and Opportunities. L. Mastroianni, Jr., P.J. Donaldson, and T.P. Kane, eds. Washington, D.C.: National Academy Press.
PMA (Pharmaceutical Manufacturers Association). 1992. 1992 survey report: 91 medicines in testing; 3 approved this past year. P. 8 in Development: AIDS Medicines, Drugs and Vaccines. Washington, D.C.: PMA.
Rall, J.E. 1990. Memorandum to scientific directors, clinical directors, chiefs, and ICD clinical branches/laboratories from NIH deputy director for intramural research. Policy on the Inclusion of Women in Study Populations: Intramural and Epidemiological (August 1). Bethesda, Md.: National Institutes of Health.
Reisman, E.K. 1992. Products liability: What is the current situation and will it change (and how) when more women are included in studies? Presented at the Women in Clinical Trials of FDA-Regulated Products Workshop, Food and Drug Law Institute, Washington, D.C., October 5.
Robertson, J.A. 1982. The law of institutional review boards. UCLA Law Review 25:484-549.
Robertson, J. 1994. Ethical issues related to the inclusion of pregnant women in clinical trials (1). 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.
Selwitz, A.S., and Wermeling, D.P. 1992. IRB Policies and Practices: Review of Subject Population. Paper commissioned by the Office of Protection from Research Risks, National Institutes of Health.
Sherman, R. 1990. New DES front. The National Law Journal 12(March 12):1, 26-27.
Steinbock, B. 1994. Ethical issues related to the inclusion of pregnant women in clinical trials (II). 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.
Tribe, L.H. 1987. American Constitutional Law. Westbury, N.Y.: Foundation Press.
Tribe, L.H. 1991. Abortion: The Clash of Absolutes. New York: W.W. Norton.
U.S. Congress. 1993. House Report 103-28:84.
Wadlington, W. 1984. Breaking the silence of doctor and patient (review of J. Katz's The Silent World of Doctor and Patient). The Yale Law Journal 93(8):1640-1651.
Wion, A.H. 1989. Potential implications of the Supreme Court's decision in Berkovitz v. United States. Food, Drug & Cosmetic Law Journal 44:145-156.
Wittes, B., and Wittes, J. 1993. Group therapy. The New Republic April 5:15-17.
STATUTES AND REGULATIONS
45 C.F.R. 46
21 C.F.R. 50 & 56
28 U.S.C. § 1346(b), 2674 (19764)
Federal Employees Liability Reform and Tort Compensation Act of 1988, P.L. 15-11
Albala v. City of New York, 445 N.Y.S. 2d 108 (1981)
Apicella v. Valley Forge Military Academy and Junior College, 630 F. Supp. 20 (E.D. Pa. 1985)
Barson v. E.R. Squibb & Sons, Inc., 682 P.2d 832 (Utah 1984)
Bergstrasser v. Mitchell. 448 F. Supp. 10 (D.C. Mo. 1977), aff'd, 577 F.2d 22 (8th Cir. 1978)
Bonbrest v. Kotz, 65 F. Supp. 138 (1946)
Bonte v. Bonte, 616 A.2d 464 (N.H. 1992)
Craig v. Boren, 429 U.S. 190 (1976)
Cruzan v. Director, Missouri Department of Health, 497 U.S. 261 (1990)
Daubert v. Merrell Dow Pharmaceuticals, 113 S.Ct. 2786 (1993)
Enright v. Eli Lilly & Co., 570 N.E.2d 198 (N.Y. 1988)
Friter v. IOLAB Corporation, 607 A.2d 1111 (Pa. Super. Ct. 1992)
Gaston v. Hunter, 588 P.2d 326 (Ariz. Ct. App. 1978)
Geduldig v. Aiello, 417 U.S. 484 (1974)
Griffin v. United States, 500 F.2d 1949 (3d Cir. 1974)
Grodin v. Grodin, 301 N.W.2d 869 (Mich. Ct. App. 1981)
In re A.C., 573 A.2d 1235 (D.C. Ct. App. 1990)
International Union, UAW v. Johnson Controls, 111 S. Ct. 1196 (1991)
Jones v. Ortho Pharmaceutical Corp., 209 Cal. Rptr. 456 (1985)
Monusko v. Postle. 175 Mich. App. 269, 437 N.W. 2d 367 (1989)
Newport News Shipping & Dry Dock Co. v. EEOC, 462 U.S. 669 (1983)
Personnel Administrator of Massachusetts v. Feeney, 442 U.S. 256 (1979)
Planned Parenthood of Central Missouri v. Danforth, 428 U.S. 52 (1976)
Planned Parenthood of Southeastern Pennsylvania v. Casey, 112 S.Ct. 2791 (1992)
Renslow v. Mennonite Hospital, 67 111.2d 348, 367 N.E. 2d 1250 (1977)
Roberts v. Patel, 620 F. Supp. 323 (N.D. Ill. 1985)
Roe v. Wade, 410 U.S. 113 (1973)
Salgo v. Leland Stanford Jr. University, 317 P.2d 170 (Cal. Ct. App. 1957)
Shack v. Holland, 389 N.Y.S.2d 988 (1976)
Stallman v. Youngquist, 531 N.E.2d 355 (Ill. 1988)
Taylor v. Wyeth Laboratories, 362 N.W.2d 293 (Mich. App. 1984)
Tinnerholm v. Parke, Davis & Co., 285 F. Supp. 432 (S.D.N.Y. 1968), modified on other grounds, 411 F.2d 48 (2d Cir. 1969)
United States v. Orleans, 425 U.S. 807 (1976)
West v. Johnson & Johnson Products, Inc., 220 Cal.Rptr. 437 (1985)
Whitlock v. Duke University 637 F. Supp. 1463 (M.D.N.C. 1986), aff'd, 829 F.2d 1340 (4th Cir. 1987)