Lessons from PrEP: A Qualitative Study Investigating How Clinical and Policy Experts Weigh Ethics and Evidence When Evaluating Preventive Medications for Use in Pregnant and Breastfeeding Women

All experts in our study noted that designation of pregnant women as a vulnerable group raises the bar for inclusion in research, particularly clinical trials. They described this as a paradox of needing evidence to gather evidence: more evidence-based interventions in pregnancy are needed, but prior evidence is needed to justify inclusion of pregnant women in research. Many experts described how the overall lack of clinical trials data on effectiveness and safety of medications during pregnancy/breastfeeding presents a significant challenge for clinicians and translational scientists. Experts described being caught in a true dilemma: on the one hand, it is difficult to justify the inclusion of pregnant women in clinical research because an intervention is not yet proven safe. On the other, clinicians are making clinical recommendations for pregnant women every day, often without robust evidence, and in the participants’ views, potentially placing women and pregnancies at risk. Participants with a clinical background noted that avoiding risk in the research setting can result in the delivery of substandard clinical care.

“I really do appreciate the sentiment of wanting to be careful, but I think it ends up being that because of this lag, we end up treating women inadequately and inappropriately for periods of time while we wait for data to appear.” Participant 716; healthcare provider, region of the Americas

Experts reported that strict exclusion of pregnant women from upstream research displaces risks downstream to clinical care. As they explained, restricting research because of concerns for the safety of the unborn child limits availability of safety data, which ultimately constrains clinicians’ ability to provide evidence-based therapies for pregnant women.

“I think too much caution means that people are just doing stuff on their own when it could lead to more potential harm than doing it in at least a study setting when you [are] gathering more data. I think we need more information, and research is the way to get it.” Participant 586; healthcare worker/researcher, region of the Americas

Experts expressed concern that the decision to move to implementation of new interventions in pregnant women is often based on studies with non-pregnant women, leading to untested assumptions about safety during pregnancy or continuation of possibly harmful pre-pregnancy medications.

“I’m on listservs where all these smart people start using whatever great-looking antiretroviral combination is working so well when the woman is not pregnant, and they just continue it during pregnancy. That’s crazy.… They should change to something that they know is safer or safe until there’s enough data to show safety and efficacy in the pregnant woman.” Participant 408; healthcare worker, region of the Americas

Several experts noted that medical practice and research remain haunted by past experiences with under-studied drugs, such as thalidomide, which caused serious fetal harm. The resulting fear of fetal harm was thought to be the reason for imposing a high barrier to investigating therapies for pregnant women, inadvertently creating evidence gaps for preventing or treating illness during pregnancy.

“We are so afraid [of fetal risk] that we don’t make a way to investigate this important time…. Because we have not investigated that time of a woman’s life…we’ll have interventions for babies, for adolescents, for non-pregnant women, but the pregnant women will always remain uninvestigated.” Participant 978; researcher, African region

When asked about the reasons that might justify greater inclusion of pregnant women in research, most participants thought that investigating treatments for severe conditions affecting women during pregnancy, including HIV, should be given priority. When probed, those participants believed it unethical to exclude pregnant women from research studies when the condition was severe and the interventions could provide direct benefit to the pregnant woman.

“[I]f it’s a drug that might be needed to treat a condition in a woman that is serious, not studying it in pregnancy to me seems unethical… [A]nd on the other hand, just saying that you could study any drug in a pregnant woman, that doesn’t seem right either. So again, it depends on, to me, what the condition is, whether one studies it or not, and I would like to see more studies in pregnant women than less…instead [of being] reliant on observational gossip.” Participant 328; researcher, region of the Americas

While it was a minority view among experts, a few maintained that classifying pregnant women as a vulnerable population was warranted and extra precautions should be taken, even if these precautions slow development of new interventions for use during pregnancy.

“Pregnant women are [a] vulnerable group and just like [a] vulnerable group, they must be protected…. [W]hile we need more information for research, pregnancy is very sensitive and they must be protected, because certain damages done around that period may cause irreversible harm, so it’s a balancing act, whatever drives the research must be a balancing act.” Participant 527; healthcare worker/researcher/ethicist, African region

Because of the lack of clinical trial data establishing efficacy and safety of interventions for use during pregnancy, all experts described the need for resourcefulness when evaluating new treatments for implementation with pregnant women. While they agreed that RCTs remain the gold standard for evaluating new clinical interventions, most recognized that obtaining RCT data for safety and efficacy during pregnancy was often not possible if there was any indication of fetal harm in preclinical trials. Registries were viewed as an essential but imperfect resource, given limitations on generalizability.

“[T]here will never be randomized controlled trial data on most of these things because anything that’s going to show up on an RCT would’ve had such a strong signal in preclinical trials that it would never have gotten there. … For me, the real question is, ‘How can you generate enough registry data to give yourself some confidence that you really believe it?’” Participant 408; healthcare worker, region of the Americas

In the absence of RCT data, all experts described triangulating and extrapolating data from animal studies, clinical case reports, and registry or post-market surveillance data to make decisions.

“[T]he more different sources of data we have, the easier it is to make decisions and to not feel like you are making decisions in the absence of information and that you can make informed decisions….” Participant 131; researcher, African region

However, many experts were quick to caution against several common errors in this approach. First was the need for careful consideration when extrapolating from currently available data to a different setting or population—important biological, genetic, and environmental differences could limit the meaningfulness of data from a particular population. Second, some experts noted that early reports of adverse outcomes, anecdotes, or case studies can be weighted heavily, and can stop interventions from moving forward to implementation research in pregnancy.

“I think many people are really influenced by anecdotal evidence and it usually just messes things up. I think it’s very important to have the real hard evidence in making these decisions, because RCTs would give us good evidence that would help us to make decisions based on some research being done properly.” Participant 978; researcher, African region

Other experts described how stories about single adverse events travel quickly and can breed mistrust among frontline clinicians and in the community, reducing uptake of potentially beneficial interventions, and reducing community members’ willingness to participate in research.

When asked about how best to weigh early data, clinical experts observed that policy makers can sometimes be overly cautious when early data suggests any risks during pregnancy, possibly granting too much weight to early animal data relative to the health impacts to women and their fetuses. One expert argued that even with some evidence of toxicity, this should be weighed against the benefits and use potentially limited to less vulnerable stages of later pregnancy, rather than taking an all-or-nothing approach:

“[T]he use of Efavirenz is another good example, where we had some animal data and some case reports suggesting issues with neural tube defects with use in the first few weeks of the pregnancy. [A]lthough that wasn’t definitive data, based on that, the FDA changed their recommendations. … In use later, we had no indication that there was an issue from the observational data…. So one could’ve said, ‘Well, you can’t use this drug in the first four weeks, but you can use it afterwards,’ but rather the FDA and the European group went way out and said, ‘You should never take this drug during pregnancy,’ which was an over-interpretation of the data.” Participant 328; researcher, region of the Americas

When we asked experts how they consider offering interventions to women during pregnancy, careful evaluation of the maternal/fetal risk–benefit ratio was central to their deliberation. Although this dilemma has often been framed as a dichotomous choice about whether to prioritize maternal or fetal interests, many experts resisted the dichotomy and argued that the woman’s and baby’s well-being are too intimately connected to be considered separately. Participants appealed to a range of ethical positions reflecting a continuum from a woman-centered prioritization to more infant-centered positions (Table 2).

Among those who endorsed a woman-centered view, different reasons were offered. Some made an appeal to a woman’s right to make decisions over her own health. Others appealed to a personhood argument, describing the woman as “the patient” during the pregnancy, while they believed the fetus holds a lower moral status until term or delivery. On either account, these experts argued that we should minimize risks to the fetus, but prioritize the woman’s health over fetal health, especially when the mother’s health was at risk.

“I’m a maternalist, so I don’t see the life of the mother and the life of the fetus as equal. I think we need to put appropriate safeguards in place and do as little harm as possible to the fetus, but always keeping in mind that the mom’s life comes first. [G]enerally the way that I think about it is, if a mom really needs a drug for her health, then we need to treat her like she is not pregnant.” Participant 980; healthcare worker, region of the Americas

Other experts holding more woman-centered views along the spectrum strongly defended prioritizing the woman’s health on consequentialist grounds. These reasons included: (1) she can go on to have other children if she loses this pregnancy; (2) she has responsibilities to other children, her family, and her community; and (3) the infant’s survival and future well-being depend on the woman’s health.

However, within each of these ethical appeals, experts’ prioritization of maternal–fetal interests shifted, depending on the context. For example, clinical experts from WHO regions with high maternal and child morbidity/mortality were most inclined to collapse the dilemma of fetal vs. maternal interests, noting that the health and life prospects of the infant, and any other children the woman may have, depend on the health of the mother during and after childbirth. Others recognized that in many circumstances, including the case of HIV prevention and PrEP, the interests of both the woman and future baby are inextricably linked, requiring a more holistic consideration of potential benefits and risks, including that the woman will bear the care burden for any disabilities caused to the baby.

“[A]s a clinician, both are your patients and the infant’s health is tied to the mother’s health…..The mother’s health is critically important for ensuring that her infant reaches term and is delivered safely, so I don’t think her health should be excluded from the equation. But similarly, there should be a priority based on delivering a healthy infant and ensuring that there are no harms to that child because any harms to that child are likely to be life long and likely to be the responsibility of that woman and her family.” Participant 131; researcher, African region

When asked how the interests of the woman and her unborn baby ought to be balanced in the context of HIV prevention, and more generally in the provision of any intervention during pregnancy, experts considered the severity of the illness in the woman against the likelihood of fetal harm as the main consideration. When the illness was severe and the mother’s life was in jeopardy, these experts prioritized intervention regardless of the impact to fetal health.

“The maternal health piece, in almost all cases, would be the primary consideration; you would want to be sure that you are treating a woman for a serious condition adequately, and unless you knew that the agent being used had serious consequences for the fetus, the maternal benefit would outweigh it, and potentially, if you had a life-threatening illness in the mother, even if you knew that there might be a problem in the fetus, one might consider moving forward, if you were saving that woman’s life.” Participant 328; researcher, region of the Americas

When evaluating outcomes, the majority of experts placed more weight on known harms to the mother or fetus when compared to potential fetal harms. In certain situations, known fetal harms trumped potential benefits to the woman.

“[T]he sicker the woman and the more provenly beneficial the treatments are for the woman’s sickness, the more likely you are to accept the potential for known or unknown risk to the fetus.” Participant 408; healthcare worker, region of the Americas

Some experts thought it always ethically required to minimize the use of treatments with potential or unknown risks to the fetus, except when there is no alternative for the mother, and this position was often accompanied by reflection on past controversies. One participant appealed to the future child’s right to as open and healthy future as possible.

“Well, you know there’s the old Thalidomide story, right? That the potential for harm and the lives of those people born with the harmful effects, I think they too have the right to the potential of a healthy future life, so the issue of protecting the unborn fetus so that when they’re born, they’re born as healthy as is possible, is important.” Participant 660; researcher, African region

Experts shared many insights on the contextual factors influencing decisions to move forward with new interventions for use in pregnancy (Fig. 1).

“I think the social and economic context of the clinical care and research is always part of the equation, too. [W]omen in some places are desperate for good medical care, and some places they’re not. And I think when we talk about how we’re going to take care of women when they’re pregnant… we can’t divorce ourselves from that economic and social context, which affects every decision that they make, basically.” Participant 782; researcher/ethicist, region of the Americas

Within resource-constrained settings, medication costs play a large role in treatment decisions during pregnancy. Cost and limited financial resources can impact the availability and acceptability of treatment options when women or in-country policy makers are asked to balance medication costs with treatment effectiveness and the availability of alternative treatment options.

“PrEP has great potential [for] preventing infection. But we have so many people who need treatment and I think sometimes I understand the dilemma of the policy maker because with the resources they have, who do they give [PrEP to], do they give to the sick or do they use the resources to prevent new people from getting [sick]…? We don’t have the unlimited resources.” Participant 202; healthcare worker/researcher/ethicist, African region

In addition to cost, logistical factors, such as dosing, availability of alternative options, and pick-up location were important practical factors experts considered when making treatment decisions. Medications requiring too much time, follow-up, or intensive care were recognized as not likely to be taken up by women and therefore less likely to be prescribed.

“[W]ithin the environments where we work you’ve got overloaded services already. We battle to get women into four antenatal visits. We battle to get HIV-positive women into three monthly visits for testing, and negative [women] for testing…Anything that’s just too demanding is going to be highly challenging, unless it’s just a very small number of women that require that level of intensity to treat them, like diabetics…And I know it’s a consideration for the woman because if she’s having to take time off work, pay for fare, spend a day sitting in the clinic, you know she’s not going to be keen to do that either.” Participant 539; researcher, African region

All experts described the challenges of navigating social and cultural attitudes towards the role of women in decision-making, including attitudes of deference from women themselves. In situations of uncertainty, when information or data are imperfect, experts still believed it is important to share what is known with women. This was especially true in situations where the decision is complex and both the mother’s health and the health of the fetus are in jeopardy. Clinicians, in particular, felt it was important to make such a decision with a woman as a shared decision, to support women in weighing the risks and benefits as known, together. Overall, the more uncertainty present in the risk/benefit equation, the more experts wanted the woman to be informed and make her own choice, but with support. Clinical experts reported that women differ in their desire for information and involvement in decisions and found it challenging when women defer to their expertise, but they do not have sufficient evidence on which to base a recommendation.

“Some women want to be very involved with every decision-making, and there are some women who simply want a clinician to tell them what the clinician thinks is best. And that’s tricky when it’s a grey area and the clinician can’t say definitely what’s best.” Participant 221; healthcare worker, region of the Americas

Clinical experts also found it difficult to navigate conversations with pregnant women who prioritized the safety of the fetus at the expense of their own health, with some noting that they would feel obligated to advocate for the woman’s health in such a case.

“[M]ost of them in my experience, whether it’s in Europe or in Africa, most women will have a bias towards the baby’s health…but I think it’s up to us to help make sure that we balance that a little bit. If something is really dangerous for the health of the woman…we should really advise them towards their own health.” Participant 136; healthcare worker/researcher/ethicist, European region

Similarly, clinical experts described the cultural challenges surrounding women’s deference to male partners and expectations in some cultural contexts that male partners must be involved in decisions potentially affecting a woman’s pregnancy. Several experts took a pragmatic approach in considering male partners’ views in contexts where they knew women were likely to defer to husbands.

“The male partner becomes an important player and stakeholder, and quite often even after you have discussed issues with the women, they defer, they don’t make a decision, they want to go and consult first, so that becomes important because whatever he decides is probably what is going to carry the day.” Participant 202; healthcare worker/researcher/ethicist, African region

Finally, several clinical experts made an important observation about the role of women in making decisions around new interventions for use in pregnancy, noting that particularly in developing countries, women are seldom brought into the decision-making processes about when and what new drugs and interventions are made available to pregnant women in the first place.

“I’ve heard this in these conversations where people are talking about rolling out PrEP… particularly in the developing world… where there’s this very patriarchal way of thinking: there are bodies of people who are going to make decisions about what’s best for women, and that kind of rubs me the wrong way. I still think that there has to be this theme through what we do about giving them the information and then helping them in making the choice that works best for them. Participant 221; healthcare worker, region of the Americas