Abortion

Approximately 60% of abortions in the US are obtained by people who are already parents (Guttmacher Institute n.d.). They choose abortion without the epistemic limitation of not knowing what parenthood is like. Their subjective values and preferences as a parent could lead them to choose abortion, in addition to any objective values that contribute to the decision.Footnote ⁴ For the 40% of people who are not already parents and have abortions, they lack subjective experience of parenthood but can rationally decide if they want to transform in this way, at this time. Some individuals could be interested in such a transformation but for objective reasons choose abortion.Footnote ⁵ For others, choosing abortion occurs for medical reasons in the context of a desired pregnancy, and may, in part, be a choice to preserve future procreative potential.Footnote ⁶ Having an abortion is not to decide to never undergo the transformative experience of becoming a parent, only to not do so with a particular pregnancy.

Yet people making abortion decisions routinely face social, cultural, and legal scrutiny. Protecting people from post-abortion regret is one reason given for limiting individual agency to choose abortion. The concept of post-abortion regret dominates in debates over abortion access, legislation, and restrictions, although it is a fiction constructed by anti-abortion interests, not a clinically observed or evidence-based concept (Kelly Reference Kelly2014). Crisis Pregnancy Centers (CPCs) in the US developed the concept of “post abortion regret syndrome” or PAS in the 1970s to further a religiously motivated (Christian) anti-choice agenda. Conservative interest groups strategically disseminated the rhetoric of PAS to influence public policy and legislative action starting in the 1980s through the present.

In a 2007 US Supreme Court decision, Justice Anthony Kennedy wrote that “it seems unexceptionable to conclude some women come to regret their choice to abort the infant life they once created and sustained.” Kennedy noted that, even though “we find no reliable data to measure the phenomenon,” post-abortion regret is plausible and “severe depression and loss of esteem can follow” (Gonzalez v. Carhart, 550 U.S. 124 (2007)). The fiction of PAS made it to a decision from the highest court in the US, despite credible peer-reviewed research published seven years prior showing low rates of regret and adverse psychological outcomes among those who obtain abortions (Major et al. Reference Major, Lynne Cooper, Zubek, Richards, Wilhite and Gramzow2000).

Kate Greasley (Reference Greasley2012) analyzes categories of PAS used to argue abortion is unethical. These include an empirical assertion that PAS is more common than post-natal regret, and a moral stance that PAS is something from which people must be protected for reasons of beneficence. Greasley shows how each type of PAS utilized in abortion debates relies on flawed evidence and logic. Many studies putatively corroborating PAS are funded by, or authored by scholars affiliated with, anti-choice interest groups like the Charlotte Lozier Institute and contribute to the abortion infodemic (Pagoto et al. Reference Pagoto, Palmer and Horwitz-Willis2023). For example, three articles researched and funded by anti-abortion interest groups (though not about PAS) were recently retracted by the publisher for conflicts of interest and problems with data collection and interpretation (Health services research and managerial epidemiology 2024).

Health researchers have responded to misinformation about PAS by conducting research disproving it. The Turnaway Study, which followed people in the US who sought and were either denied or received abortions over a period of five years, found regret was most experienced by those who wanted, but were denied, abortions, and subsequently relinquished the baby (Foster Reference Foster2020). Individuals denied abortions who parented their child initially felt regret over being denied an abortion, which diminished over time. Demonstrably diminished post-natal regret after one becomes a parent affirms the idea of parenthood as a transformative experience that changes one’s overall values and preferences. However, as the growing literature on motherhood regret shows, this transformation is not guaranteed.

Sterilization

If abortion is the choice against a particular pregnancy, tubal ligation and hysterectomy can be the choice against any (future) pregnancy. Tubal ligation permanently prevents pregnancies by cutting, tying, or blocking the fallopian tubes, and is usually irreversible. Hysterectomies involve the removal of the uterus altogether. They are usually performed for medical reasons beyond preventing pregnancies, but infertility is a permanent side effect. Arguably, these choices against future fertility (which they are, regardless of the reasons for the choices) are more complicated from the perspective of transformative experience because they foreclose certain means of transforming in the future, making them a different kind of decision.

Like decisions about some abortions, decisions about tubal ligation or hysterectomy could involve a preference for a childfree life. They can also be about not having more children. Many tubal ligations are performed on people who have recently given birth. In these cases, like in 60% of abortions, the reproductive choice is made by someone who has already transformed into a parent. For a first-time parent opting for a tubal ligation during a planned c-section, the choice to prevent future pregnancies is made prior to becoming a parent, but after the decision to become a parent. Becoming infertile may, in some cases, not be a choice about parenthood at all. It could involve a choice to trade future procreative potential in favor of managing a medical condition. And the choice against carrying a child does not necessarily foreclose the experience of parenthood.

Yet individuals choosing to permanently limit their reproductive capacity to manage medical conditions are sometimes told that they, too, will regret their decision. Data indicate they rarely do, and that regret, when it does occur, diminishes over time (Danvers and Evans Reference Danvers and Andrew Evans2022). The projection of regret by others on behalf of a patient is especially baffling when they are choosing sterilization to prevent harm or suffering from a medical condition, which involves objective and subjective reasons that could quite reasonably outweigh possible future preferences for procreation.

When tubal ligations are sought by non-parents confident that they wish to remain childless, they are sometimes met with skepticism (Bahadur Reference Bahadur2018), especially by medical professionals concerned about irreversibility (American College of Obstetricians and Gynecologists 2020; Hillis et al. Reference Hillis1999). As with abortion, these claims are framed in terms of regret and the possibility they will change their mind about having (more) biological children. While we think it is fair to attribute some of this anticipatory concern about patients’ future regret to essentialist ideas about women and their roles and desires regarding motherhood, some clinicians may understandably worry about the irreversibility of sterilization due to the unjust history of coercive sterilization in the US, particularly of women of color (Roberts Reference Roberts2014) and disabled women (Lombardo Reference Lombardo2022). However, such concerns are different in kind from worries about regret. Addressing those concerns is a matter of enhancing and enacting reproductive justice (Ross Reference Ross2017), whereas invoking potential regret may, as we’ll argue, undermine self-determination and constitute a form of reproductive injustice.

Medical procedures and parenthood

Arguably, invoking potential future regret implies that, because we cannot know our future preferences, forgoing the transformative experience of child-bearing is susceptible to regret. Notably, the inverse, that one might regret child-bearing, is rarely implied in clinical contexts, although an increasing body of literature suggests that motherhood regret (and parenthood regret) occurs more often than reported, and that social norms and stigma prohibit open conversation about it (Donath Reference Donath2015, Reference Donath2017; Heffernan and Stone Reference Heffernan and Stone2021; Moore and Abetz Reference Moore and Abetz2019; O’Reilly Reference O’Reilly2019). As far as we are aware, those considering assisted reproductive technology are not informed about the body of literature on motherhood or parenthood regret, nor are they asked to consider whether they will regret undergoing assisted reproductive procedures to become parents.

We should note a relevant distinction that circumscribes the scope of our argument in this essay. Our primary focus is on the relevance of regret risk as part of the risks, benefits, and alternatives communicated by clinicians during informed-consent processes for medical procedures. If a symmetrical application of regret risk over becoming a parent were to be incorporated into medical informed consent it would only be for treatments and procedures intended to produce parenthood as a result. This would not, for example, include counselling as part of one’s prenatal visits because, once one is pregnant, there is no medical procedure to become pregnant that would necessitate informed consent. That said, in what follows in this paper we will argue against its systematic inclusion in medical informed consent at all.Footnote ⁷

Capacity and voluntariness

In US healthcare contexts, adults are presumed to have medical decision-making capacity unless declared legally incompetent by a court, or they display reasons for questioning their capacity leading to its evaluation. Because capacity comprises the ability to not only make a choice, but to understand, evaluate, and apply relevant information about that choice (Applebaum and Grisso Reference Appelbaum and Grisso1988), information sharing is also a feature of capacity determinations. If, for example, a patient is not incorporating information that a clinician believes is relevant to a decision into their decision-making, the clinician might question their capacity to make the decision. Similarly, having adequate information may be necessary to making a voluntary decision. Insufficient disclosure, or irrelevant disclosure, may impede an individual’s freedom to make an informed decision.

For example, in the US people considering abortion are frequently (and surreptitiously) directed to CPCs. The American College of Obstetrics and Gynecology notes that “71% of CPCs use deceptive means such as spreading thoroughly debunked misinformation and 38% do not clearly state on their home page that they do not provide abortion care” (American College of Obstetrics and Gynecology 2022). CPCs provide inadequate information, and sometimes misinformation, about pregnancy and abortion. This can include false information about an individual’s pregnancy, such as when CPCs offer ultrasound services and intentionally give an inaccurate estimated gestational age so that individuals believe they either can no longer obtain an abortion or have more time to deliberate about abortion (Montoya et al. Reference Montoya, Judge-Golden and Swartz2022).

Although capacity and voluntariness relate to information sharing, we presume that in the kinds of cases of abortion and sterilization we are considering, which fall outside of acute medical emergencies, there are no special reasons that capacity is in question, beyond epistemic limitations due to either inaccurate or misinformation sharing, or inaccessible information related to transformative experiences. Voluntariness may also be complicated by incomplete or inaccurate information, as well as social and political circumstances constraining an individual’s options. These could include direct coercion by a partner or family member, or constraints on one’s range of available options, like financial duress, lack of social or familial support to raise children, avoiding being tied to an abusive partner, or concerns about future availability of reproductive choices due to restrictive legislation.

Information sharing: what and why?

There are both ethical and legal components of information sharing and disclosure requirements in medical informed consent. Tom Beauchamp describes how US legal case Salgo v. Leland Stanford, Jr. University Board of Trustees, 154 Cal. App.2d 560 (1957) created the concept of “informed” consent and determined that disclosing an intervention’s risks and benefits was “a logical extension of the already established duty to disclose the treatment’s nature and consequences” (Beauchamp Reference Beauchamp2010, 54, italics in original). But what kind of risk is regret, and is it one that ought to be disclosed for adequate (i.e., ethical) informed consent?

Relatedly, one might wonder whether regret risk is even permissible to disclose without creating undue coercion against a choice or otherwise infringing on self-determination. The ways in which medical literature has responded to false theories of PAS to prove low incidence of abortion regret, for example, raise questions of what kinds of regret warrant empirical investigation related to medical decisions in the first place. Understanding what role regret could play in medical informed consent may help determine whether such empirical research is valuable—or justifiable.

Beauchamp notes that three US federal court decisions from 1972 “had a particularly massive influence in demanding a more patient-oriented standard of disclosure,” according to which the “patient possesses enough information to enable an intelligent choice” (2010, 58). The move in the US from a clinician-centered to patient-centered standard shifted power dynamics and perspective of informed consent. Yet it still leaves open the question of who is a “reasonable person” and what views get attributed to them. This is a shift not entirely unique to the US. In rulings related to obstetric consent, UK courts have similarly shifted to patient-centered standards for disclosure, requiring disclosure of previously undisclosed risks.Footnote ⁸ It may be easier to identify underdisclosed risks than whether all information disclosed is relevant and appropriate disclosure (or could be misleading, nudging, coercive, or biased).

Available alternatives introduced in informed consent processes that detail risks, benefits, and alternatives, are also complex. In cases of sterilization there may be appropriate clinical alternatives to trial prior to sterilization that meet a patient’s needs. But if a patient wants or needs an abortion, non-clinical alternatives to abortion (parenting or adoption) also involve pregnancy and delivery, which are riskier than abortion. For some individuals seeking abortion or sterilization because of other health factors, the risks of pregnancy may be much greater than generalized statistical risks. They may increase even more due to social or geographic context. Therefore, when an alternative is a guaranteed or potential pregnancy, it is not a less medically risky alternative.

Information sharing for informed consent requires at least two forms of trust in the information’s source: relevance and reliance conditions. One is trust that the information is relevant. Another is confidence in a source’s reliability. We’ll suggest that when a clinician introduces the possibility of future regret to a patient as part of informed consent it is neither certain to be relevant nor reliable. In analyzing informed consent, Robert Martin (Reference Martin2023) considers a lay person compared to a physician aiding an injured person they encounter. In each instance, the injured person authorizes the individual to move them by requesting that they help them move to seek medical care. But the physician has reasons related to their medical expertise to believe moving the person is more harmful than not moving them. In this case, Martin argues they have an obligation to share this information with the injured party, and that it would be reasonable for the injured party to then withdraw their consent to being moved.

We might think something similar happens when a patient seeks certain procedures, including abortion or sterilization. The patient authorizes the pregnancy termination or sterilization by requesting it, triggering the informed consent process in which they receive relevant information about risks, benefits, and alternatives. The individual can then revise their consent, including electing an alternative or retracting authorization entirely. Moreover, Martin argues that it may be reasonable for the physician in the injured person example to override the person’s consent to be moved for reasons of non-maleficence because they have information the injured person does not: that the risks significantly outweigh benefits and what the person is requesting will in fact be counter to their objective of receiving beneficial medical care.

Is the risk of regret the kind of information to share with or disclose to a patient that a clinician might be obligated to share? If a clinician believes the risk of regret is a significant risk, like in the injured person case, would they be permitted to override their patient’s authorization for reasons of non-maleficence? We argue below that the answer in both of these scenarios should be no. Per Martin, trustworthiness comes from the “goodwill” we judge to be present when an agent “has our health-related interests at heart and is seeking to promote those interests and not their own” (2023, 8). Introducing the specter of regret potentially undermines relevance conditions of informed consent, and likely undermines reliance conditions, as does overriding a patient’s decision. Introducing regret is not promoting another’s interests and may impose one’s own values.

As Martin (Reference Martin2023) says, “an agent must have a particular epistemic position to validly consent. Not surprisingly, there is significant discussion and disagreement over what this position is.” Potentially transformative choices further complicate any view of a sufficient epistemic position. Even on a patient-centered account of relevant disclosure, uncertainty about what a patient will value in the future generates uncertainty about what information will be relevant to their deliberation in the present. This returns to questions regarding the relevance of regret, as well as questions of capacity and voluntariness.

Others have introduced various puzzles that transformative experiences raise for healthcare, particularly how many of these experiences, when related to illness or disease, are unchosen (Carel and Kidd Reference Carel, James Kidd, Bain, Brody and Corns2019, Reference Carel and James Kidd2020). Hofmann (Reference Hofmann2024) suggests that the unchosen nature may resolve the apparent puzzle because one does not decide to transform and only makes choices about their course of treatment once transformed. Yet some transformative illness experiences, while not chosen, are progressive and there is no clear moment or completion of transformation, while treatment decisions must be made.

Hofmann (Reference Hofmann2024) suggests that cases in which a person’s choices will result in an avoidable transformative experience (i.e., a preventable disease) are a genuine puzzle for consent to their current choices. In these cases, Hofmann argues that the potentially transformative nature of the individual’s choices may justify nudging as “paternalistic prevention.” Yet Hofmann focuses on an individual’s “lifestyle” choices about which a clinician may counsel them, not clinical decisions. We’ll grant that in the kinds of cases Hofmann considers, while the clinician is counseling about personal, not clinical, choices, nudging might be justified if the clinician had information from their clinical expertise to which the patient would not otherwise have access.

Introducing regret risk into a clinical discussion could be a form of nudge as well. However, a clinician raising the risk of future regret over abortion or sterilization is not introducing information specific to their clinical expertise nor to which the patient does not otherwise have access.Footnote ⁹ Therefore, introducing regret risk does not meet the justificatory threshold that other cases of advising or nudging about personal choices might.

Second-personal anticipatory regret

One form of anticipatory regret we’ll call second-personal anticipatory regret, which occurs when one agent introduces the possibility of future regret on the part of another agent. Whether intended by the speaker or not, a potentially pernicious function of second-personal anticipatory regret regarding abortion and sterilization is the imposition of anticipatory regret on an agent deciding about and for their future self. When others raise the prospect of future regret to someone considering abortion or sterilization, the primary function of “regret” in these cases might be the expression of the other person’s values, or the assertion of social norms and expectations, especially those imposed on women-identifying people that they should want, have, and mother, children. It highlights how regrets “are formed in the context of the society, and they can demarcate the cultural ethos, norms of behaviour, and gendered ideologies” (Donath Reference Donath2015, 345–46).

Concerns that people seeking abortions will regret their decision enforce norms of “good” or “loving” mothers. Concerns that people preferring childfree lives, and taking active steps to guarantee these preferences, will regret this choice assume women necessarily should have maternal instincts, needs, or desires. It tells women that they ought to be unfulfilled by not experiencing motherhood, even if they do not recognize this desire or need within their own preferences and values.

In some ways, invoking potential regret might most look like a “nudge” that frames information in a way as to encourage, or some would say manipulate, a particular decision by the patient. The ethics of the “nudge” in healthcare are debated (Blumenthal-Barby and Burroughs Reference Blumenthal-Barby and Burroughs2012; Prainsack Reference Prainsack2020; Schmidt and Engelen Reference Schmidt and Engelen2020). The nudge of second-personal anticipatory regret in these cases promotes misogyny while parading as beneficence, or perhaps paternalism not in the sense of promoting beneficence, but in the sense of patriarchal values. It diminishes rather than enhances personal autonomy and empowered decision-making. Healthcare providers and systems often value women based on their “reproductive capacity and provision of care for children” (Hay et al. Reference Hay2019). These providers and systems may frame second-personal anticipatory regret about the loss of reproductive capacity as a way of supporting an individual’s future self, while imposing gender norms.

But what about second-personal conveyance of data about regret? We’ll call this claim that clinicians should disclose available regret data the statistical objection. It may look like a value-neutral disclosure of statistical evidence about incidences of regret, presumed to be relevant to an informed consent process in which a patient weighs risks, benefits, and alternatives. Are data about regret risks the kinds of risks that clinicians should share for adequate consent?

There are at least two concerns with the content of the data, and a third concern related to its existence. First, there are questions as to whether it is within the scope of a clinician’s professional obligations to introduce regret risk as a relevant clinical risk in an informed consent process. As already discussed, a patient’s potential for regretting an abortion or sterilization procedure is information accessible to her outside of (and prior to) a clinical encounter in which she is seeking this care. Such a deliberative process neither relies on clinical information nor reflects a clinical risk. Second, there are questions as to whether the available evidence suggests that regret rates are high enough to be considered a relevant risk. Existing studies suggest show very low regret rates for abortions and sterilizations. Much higher rates of regret occur in less contested procedures (Wilson et al. Reference Wilson, Ronnekleiv-Kelly and Pawlik2017) without regret risk being routinely introduced into informed consent. Thus, even if it were the kind of risk relevant for a clinician to report, it is not a sufficiently likely risk to report it as a risk. Alternatively, clinicians could affirm that most people are satisfied with their choice without introducing the language of regret at all and the imposition of gendered norms and expectations around parenting and motherhood that come with it.

The third concern regarding the statistical objection relates to our earlier discussion about how abortion regret data is largely produced by either bad actors advocating against abortion, or as correctives to these interests and literature. Although the abortion regret example is an extreme one, it illustrates that the very assumptions in place that motivate which kinds of regret are and are not studied and for what procedures are value-laden, as are the conclusions drawn from such research. Recent studies on regret rates in vasectomy patients (Charles et al. Reference Charles2023) and tubal ligation patients (Danvers and Evans Reference Danvers and Andrew Evans2022) found similar rates in both groups. Yet the paper on vasectomy concluded that regret is “very rare” (Charles et al. Reference Charles2023, 111) and therefore included no recommendations on regret-related counselling, while the paper on tubal ligation concluded that “counseling about sterilization should reveal the unpredictability of future desire” (Danvers and Evans Reference Danvers and Andrew Evans2022, 433).

Providing statistical data about what other people who have undergone similar decisions end up thinking or feeling is also not predictive for how a particular individual will think or feel. If the theory of transformative experiences is correct, then we can reflect on what our future self might think or feel, but it will only be a conjecture since we don’t have the information available to us about what we will be like post-transformation. Statistical data about the experiences of others can’t tell us where along the distribution we will land. Instead, it may cause us to use post-transformation experiences of others as objective reasons, when they are actually backward-looking explanations for decisions those individuals made pre-transformation. Additionally, the norms imposed by second-personal anticipatory regret about abortion and sterilization overlook that the limited available data suggest one may also regret choosing to become a parent or to have more children.

If second-personal anticipatory regret is to be useful, it would have to be limited in scope. Schaubroeck and Hens suggest four conditions “for the prospect of regret to have any real deliberative value,” in which they understand it to be a question (“will you regret it?”) asked by someone else to the deliberating agent. They require that deliberation over prospective regret should be done “(1) via reflection on one’s values, (2) non-coercively, (3) when well-informed about factual circumstances and (4) in a process of self-commitment” (2017, 602). The challenge for second-personal anticipatory regret is whether these conditions can be sincerely met in a clinical context. The inconsistency with which regret is invoked—around certain decisions and not others—reflects the asker’s values that this is a matter one might regret, rather than the deliberator’s, and risks at least mild coercion.

Moreover, there is an outstanding question as to what it would mean to be “well informed about factual circumstances” regarding regret. We could imagine something like the statistical objection becoming routine in all clinical decisions in recognition of how one might regret choosing for a transformative experience just as easily as one might regret choosing to avoid it. Risks of regret would have to be more systematically studied. In the cases we are considering, this might include risks of regretting parenthood, although social stigma can make obtaining accurate data challenging. Such risks would then need to be disclosed to those seeking pregnancy termination or sterilization and assisted reproductive technology. But if there is a role for second-personal anticipatory regret in medical decision-making then it should have implications for nearly all medical decisions, or at least those that may involve major, and potentially transformative, experiences.

Yet data about regret rates don’t help the deliberating individual determine their own risks for regret. For that assessment, knowing that other people do or do not regret their decisions doesn’t illuminate why they do or do not regret them, and how those reasons applied to their values and preferences. This is, fundamentally, the project at hand when one is considering their potential for future regret. So rather than collect more data about regret rates for all procedures, individuals can evaluate their own prospect for future regret with the same information about risks, benefits, and alternatives to, and the nature of, the proposed intervention, that is already embedded into standard informed consent.

First-personal anticipatory regret

The more positive—and potentially productive—function of anticipatory regret is as a tool for identifying one’s own present values and preferences, and what information they need to reflect on what their future values and preference might be. This anticipatory regret is applied prospectively by the agent who might experience it. Though regret is experienced as a backward-looking assessment by one’s present self, anticipatory regret functions as a forward-looking assessment of one’s potential future preferences. By asking the question of whether I will regret my choice, regret functions to help me take an inventory of my known present values, and to shape the values and character I intend to develop for myself.

This may be a process that an agent engages in with others who help them think through their own values and preferences and in that sense the process may be second-personal. Identifying important pieces of information one needs from their medical providers may be a way that anticipatory regret positively enters into medical decision-making and informed consent. But we are introducing a distinction between first-personal and second-personal anticipatory regret that expands on Schaubroeck and Hens, who permit the second-personal form of anticipatory regret insofar as someone else asks (if it is in good faith and non-coercively) “will you regret it?’” Unlike Schaubroeck and Hens, our account focuses on the deciding agent driving their own reflective process regarding their potential anticipatory regret, which therefore remains a first-personal process (though another person can take steps to hold or make space for this kind of open-ended reflection, including medical providers in how they interact with patients and design their clinical encounters).

As Elizabeth Barnes suggests in her response to Paul, she could “rationally predict that I don’t want [kids] given who I am now,” and make a choice based on her “current preferences” and “sense of self” and the values and preferences she desires to cultivate in herself (Barnes Reference Barnes2015). For Barnes, imagining her future self not having children does not induce a pang of regret. Presumably for others it might, and this would inform that individual about their current values and preferences, maybe some that they had not previously identified. At its best, anticipatory regret is a kind of personalized thought experiment that helps us pump our intuitions about our own future self, values, and preference. Yet undergoing this thought experiment does not guarantee one won’t regret their future choice.Footnote ¹¹ As Paul makes clear, some epistemic and personal data about our future selves are not available to us from which to reason (or intuit).

Thus, this first-personal anticipatory regret is potentially productive—but not necessarily predictive—for future regret avoidance when looking back on one’s choice. If anticipatory regret is to have a productive function in these cases, it is for the agent to decide what regrets they can and cannot live with, and if this is even a useful question for their own deliberation, recognizing that imagined counterfactuals are just that: an imagining of what might—or might not—be important to your future self.

Dana Howard argues that anticipating future regret from a first-personal perspective can be a way to authentically access your future self who “reacts to and must live with the choices [you are] trying to make right now” (Howard Reference Howard2022, 90). We can’t know if this future self will regret your current choices, but we can try to connect with this person as an extension of our current selves rather than a completely distinct person hidden behind the transformative experience wall.