IRB: Ethics & Human Research
Ideal vs. Real: Revisiting Contraceptive Guidelines
We read, with interest, the lead article in the September-October 2010 issue of IRB: Ethics & Human Research.1 We are grateful to Chris Kaposy and Françoise Baylis for keeping the important issue of the ethical use of contraception in research on the forefront of issues in ethics involving human participants.
In many ways, we agree with the central points of the Kaposy/Baylis paper: contraceptive requirements for participants in research must be evidence-based, must apply equally to men and women, and must carefully balance the risks and benefits of participation in research (p. 7). Yet despite these general agreements, we take issue with both the characterization of our proposed policy and with the implications of some of the terminology employed in that counterproposal. In this article, we will clarify our position and its motivation, as well as describe several experiences that have resulted from the implementation of the policy at the University of Nebraska Medical Center (UNMC).
Sources of Evidence
Kaposy and Baylis devote a large portion of their paper to a discussion of mapping contraceptive requirements to the available data on teratogenicity. For the Food and Drug Administration’s (FDA’s) use-in-pregnancy category D and X drugs, where there is known risk to a fetus and/or the drug is contraindicated in pregnancy, our policy suggestions agree with theirs.
For category A drugs, where adequate, well-controlled studies in pregnant women have failed to demonstrate a risk to the fetus, Kaposy and Baylis propose that the researcher may not require abstinence or the use of contraception. The UNMC policy is similar: researchers cannot require contraception unless they provide a justification that the institutional review board (IRB) deems ethically acceptable. In practice, in the six-plus years the policy has been in effect, the IRB has not received a single request to justify the use of contraception in category A drugs. We attribute this to two causes: (a) an acknowledgement on the part of investigators that category A drugs pose no risk to a potential fetus and that, therefore, protections are unwarranted, and (b) the small number of drugs that actually fall into category A.2The few protocols that included category A drugs used them in combination with drugs in another category; thus, any contraceptive requirements for those protocols were related to those drugs with potentially higher teratogenic risk.
It is with category B and C drugs that the UNMC policy and the policy suggested by Kaposy and Baylis diverge in terms of emphasis. In both cases (as with category A drugs), Kaposy and Baylis claim that the UNMC policy “tips the balance clearly in favor of fetal protection” (p. 6). In these categories, as with category A above, they contend that using the FDA use-in-pregnancy categories as an evidence base for the policy requires that the IRB lessen the restrictions on women participants in research. They imply that permitting investigators to require two forms of contraception when there is no positive indication of fetal risk is excessive. Such requirements impede a woman’s autonomy.
They’re right. Insofar as they describe their work as an “ideal policy” based on the FDA categories, their proposal does a much more consistent job of securing women’s autonomy. The problem is that Kaposy and Baylis only consider one form of evidence here: the FDA categories. Yet there are other forms of evidence that are essential for the creation of a good policy; here, we were particularly attuned to the context of research. Our policy development team consisted not just of IRB members and chairs, but also of other important stakeholders in this work—namely, the investigators. We want to make the point that any piece of evidence must be balanced with information gained from actual experience of clinician investigators. Effectiveness rates of contraceptives must be considered in light of actual—not just ideal—use of the intervention.3Additionally, investigators were good sources of information for us about what kind of contraceptive language cooperative group and corporate sponsors of protocols would likely accept for their protocols.
All of this information combines to provide the evidence for the creation of a good institutional policy. We also take issue with the title of Kaposy and Baylis’ article, as it implies that the UNMC policy isnotevidence-based. Rather, the UNMC policy reflects a broader conception of evidence than does the ideal policy, and in doing so better mirrors the field of medicine, where clinical guidelines are routinely altered in light of contextual and other features to provide the best possible care for the patient.
Purposes of Policy
Our original motivation for creating the contraceptive policy at UNMC was to better preserve women’s autonomy by ensuring that they had contraceptive choices available to them that matched the reproductive risk of the trial. Yet this goal can only be achieved if the proposed policy is adopted by the research community. The recognition of the implementation challenges is what led us to include investigators in the task force that created the policy. We needed to know from their perspective what standards and language would likely be able to be integrated with corporate and cooperative group protocols.
Kaposy and Baylis might argue that this is backwards; ethics review committees ought to set the appropriate standards, rather than respond to the demands of others. While we are concerned about the notion that IRBs alone can set the “gold standard” for investigators to follow, we recognize that the claim that establishing a more objective set of ethical standards has at least theoretical merit. But this was not an academic exercise for us; rather, we needed a policy that would both meet our ethical standards but also be workable for the protocols that came through the door. It would be unrealistic to think that we could create a policy that would, in effect, prevent a large portion of research that was presented to university investigators from being conducted on our site because of the contraceptive requirements. It certainly was unlikely that the administration would back such a move, no matter how well-grounded ethically. Further, to the extent that clinical trials benefit participants, blocking the conduct of an otherwise beneficial study because of contraceptive requirements could very well cause much more harm than good. Rather, if we demonstrated a willingness to work with investigators—and, therefore, the university—while still maintaining fidelity to our ethical commitments, we would have a much greater case for defending the few times we did challenge contraceptive language.
As a result, we engaged in a consensus-building process by which we could effectively preserve women’s autonomy and yet respect other values on these issues. All things being equal, we would not allow contraceptive requirements for women unless there was good reason to believe there was a potential for fetal harm. Yet there are numerous people who extend greater protections to the fetus for a number of reasons: the notion of the fetus as a being of full moral standing,4the desire to provide as much of an “open future” for the resulting child as possible,5the contention that the fetus is an “innocent” deserving of protection and advocacy, etc. Although we disagree with that position, not all of the reasons listed are without merit.
A second important result of the consensus-building process was the removal of abstinence from the policy. When the investigators indicated that there would likely be significant backlash if they were required to include abstinence as a contraceptive method, we were forced to consider the motivation behind its inclusion in the first place. For one thing, we wanted to ensure that individuals who were not engaging in procreative sexual activity on a routine basis (e.g., nuns and lesbians) would not be required to use contraception. But the other reason we wanted it included was to respect a woman’s choice to refrain from sexual activity, for whatever reason and for whatever duration. The investigator task force members readily accepted barrier contraceptive methods as one acceptable form of birth control. Since barrier contraception is only utilized when someone is actually engaging in sex, the option of barrier birth control effectively includes the concept of abstinence within it: individuals who are not sexually active will not use them—or any—birth control. Kaposy and Baylis acknowledge this feature in their discussion of the way the UNMC policy does not require individuals who are not sexually active or who cannot get pregnant to use contraception (p. 5). Thus, we served our main goal of protecting women from risks of unnecessary contraception or exclusion from research while still creating a policy that served investigator needs. Given that our goal was the practical implementation of the policy, not the statement of an ideal standard that we may never reach, the consensus was that the inclusion of barrier contraceptive methods as an acceptable form of birth control effectively included the same amount of autonomy preservation that abstinence affords.
In fact, it may be the case that leaving abstinence off our list does more to preserve women’s autonomy than its inclusion. The word “abstinence” itself has several connotations that we would do well to avoid. One possible interpretation—the one the investigators on the task force were familiar with and concerned about—is that someone who is abstinent (now) need not worry about contraception. Period. The data do not bear out this claim: because of a failure to recognize the mutability of relationships, birth control through abstinence is ineffective, and abstinence-only education increases the likelihood that contraceptive choice will be inadequate.6It is also not clear that potential subjects will share a common understanding of the definition of “abstinence,” which may reflect differences in norms among social groups or general confusion about the way the term is used in different venues.7Once again, barrier contraception preserves the notion that those who are not engaging in procreative sexual activity need not use any contraception, but should that change in the future (however short- or long-term that might be), options are available.
Further, we are concerned with the use of the term “abstinence” itself. Because of the use of the term abstinence in a larger political context, there may be considerable normative baggage associated with the term, at least for some women. Faced with the frequent conception that women who embrace their sexuality, or take precautions to ensure their safety during sexual encounters, are “loose” and impure8and deserve whatever consequences come to them, we are worried that women may elect to use abstinence as their preferred method of birth control not because of a commitment to abstain from sex, but in order to be seen as virtuous. This side effect of the many uses of the word “abstinence” undermines a woman’s right to autonomy in issues of sexual expression, as well as increasing the chance of unplanned pregnancy.
Using Evidence to Shape Policy, from Conception to Implementation
Since implementing its contraceptive policy over six years ago, the UNMC IRB has received very little resistance from investigators or sponsors. We credit this outcome to the inclusion of a variety of stakeholders in the policy development process. To further the development of the policy and its underlying ethical foundation, we have begun an empirical investigation into women’s willingness to participate in research as a function of reproductive risk;9work continues on that project.
Additionally, we continue to evaluate and augment the policy as needed. For example, we are considering altering our definition of women without childbearing potential from those exhibiting 12 months amenorrhea to women exhibiting 24 months amenorrhea as standards of practice change. This is another example of how we continue to ground our approach to this policy in evidence.
One interesting application of the contraceptive policy related to men’s use of contraception, rather than women’s. Kaposy and Baylis argue that contraceptive policies disproportionately affect women and their contribution to fetal outcomes (p. 6). As they rightly acknowledge, the UNMC policy does require men who can get their partners pregnant to use effective contraception. Yet we have not focused on the male contribution to reproductive outcomes because of the lack of scientific evidence concerning it. That is, most damaged sperm simply fail to fertilize an ovum successfully, whether in vivo or in vitro. There are no confirmed data to suggest that damaged sperm that result in a successful fertilization serve as a teratogen, and certainly no data to suggest that damaged sperm can result in problems for an established pregnancy.10
This became a particular issue for us on a protocol that required both men and women to use birth control and for female subjects to be withdrawn from the study should they become pregnant. This was justified by the potential teratogenic effects of the study drug. However, the study also required the withdrawal of male subjects should their female partner become pregnant. The investigator’s original justification was that the pregnancy indicated that the subject was not complying with the policy. When asked about other ways in which a partner might become pregnant, as well as about what seemed to be a purely punitive subject withdrawal, investigators responded that they wished to protect the fetus from contact with sperm that might contain traces of the drug and could damage a developing fetus—with no data available about this eventuality. On the basis of lack of evidence, we denied their request: a move with which Kaposy and Baylis ought to sympathize. Although the withdrawal criteria for males and females were asymmetrical, the difference is the result of evidence-based differences in risk. To withdraw men merely to create equal consequences would be to create a gender bias.
Finally, we began to observe that many protocols both required contraception and excluded pregnant women from the study, yet specified no form of pregnancy testing. Other protocols required serum-only pregnancy testing at intervals so close together that the data gained from such tests would be meaningless. By using the same logic of evidence-based restrictions that created our contraceptive policy, we were able to implement a pregnancy testing policy that balanced the likelihood that the test itself constituted a barrier to participation with the actual degree of risk associated with the protocol. Such a policy was much easier to create because of the inclusion of all stakeholders in the creation of the contraceptive policy. We are inclined to believe that, had we originally chosen to pursue amodelpolicy for contraceptive requirements, we would have neither that policy nor our pregnancy testing policy in place. Detailed information about the pregnancy testing policy is outside the scope of this paper, but we look forward to distributing this policy in the near future.
Acknowledgments
We are grateful to Michael D. Arenson for research assistance.
Toby Schonfeld, PhD,is Director, Master of Arts of Bioethics, Center for Ethics, Emory University, and Visiting Associate Professor, Emory School of Medicine, Atlanta, GA;Joseph S. Brown, PhD,is Professor of Psychology, University of Nebraska, Omaha, NE;N. Jean Amoura, MD, MsC,is Associate Professor, Obstetrics and Gynecology, University of Nebraska Medical Center, Omaha, NE; andBruce Gordon, MD,is Chair, IRB and Pediatric IRB, and Professor, Hematology and Oncology, University of Nebraska Medical Center, Omaha, NE.
References
1. Kaposy C, Baylis F.Ethical, evidence-based guidelines for contraceptive use in research.IRB: Ethics & Human Research2010;32(5):1-9. All further references to this article are given in the text as page numbers only.
2. Briggs GG, Freeman RK, Yaffe SJ.Drugs in Pregnancy and Lactation: A Reference Guide to Fetal and Neonatal Risk,third ed. Philadelphia, PA: Lippincott Williams and Wilkins, 2008.
3. Mansour D, Inki P, Gemzell-Danielsson K. Efficacy of contraceptive methods: A review of the literature.European Journal of Contraception and Reproductive Health Care2010;15(1):4-16; Guttmacher Institute. Facts on contraceptive use in the United States. 2010,https://www.guttmacher.org/pubs/fb_contr_use.html.
4. Strong C. The moral status of preembryos, embryos, fetuses, and infants.Journal of Medicine and Philosophy1997;22(5):457.
5. Blustein J.Parents and Children.New York: Oxford University Press, 1982.
6. Isley MM, Edelman A, Kaneshiro B, et al. Sex education and contraceptive use at coital debut in the United States: Results from Cycle 6 of the National Survey of Family Growth.Contraception2010;82(3):236-242; Yang Z, Gaydos LM. Reasons for and challenges of recent increases in teen birth rates: A study of family planning service policies and demographic changes at the state level.Journal of Adolescent Health2010;46(6):517-524; Bersamin MM, Walker S, Waiters ED, et al. Promising to wait: Virginity pledges and adolescent sexual behavior.Journal of Adolescent Health2005;36(5):428-436; Blinn-Pike L, Berger TJ, Hewett J, Oleson J. Sexually abstinent adolescents: An 18-month follow-up.Journal of Adolescent Research2004;19(5):495-511.
7. Byers ES, Henderson J, Hobson KM. University students’ definitions of sexual abstinence and having sex.Archives of Sexual Behavior2009;38(5):665-674.
8. Frankel A, Curtis DA. What’s in a purse? Maybe a woman’s reputation.Sex Roles2008;59:615-622.
9. Schonfeld TL, Amoura NJ, Stoner JA, Gordon BG. Women and contraception in research: A pilot study.Journal of Women’s Health (Larchmont)2009;18(4):507-512.
10. Cordier S. Evidence for a role of paternal exposures in developmental toxicity.Basic & Clinical Pharmacology & Toxicology2008;102(2):176-181.
Toby Schonfeld, Joseph A. Brown, N. Jean Amoura, and Bruce Gordon, “Ideal vs. Real: Revisiting Contraceptive Guidelines,”IRB: Ethics & Human Research32, no. 6 (2010): 13-16.