Abstract
Roe v. Wade was heralded as an end to unequal access to abortion care in the United States. However, today, despite being common and safe, abortion is performed only selectively in hospitals and private practices. Drawing on 61 interviews with obstetrician-gynecologists in these settings, we examine how they determine which abortions to perform. We find that they distinguish between more and less legitimate abortions, producing a narrative of stratified legitimacy that privileges abortions for intended pregnancies, when the fetus is unhealthy, and when women perform normative gendered sexuality, including distress about the abortion, guilt about failure to contracept, and desire for motherhood. This stratified legitimacy can perpetuate socially-inflected inequality of access and normative gendered sexuality. Additionally, we argue that the practice by physicians of distinguishing among abortions can legitimate legislative practices that regulate and restrict some kinds of abortion, further constraining abortion access.
The current legal framework for abortion rights in the United States was established in the 1973 Roe v. Wade (410 U.S. 113 1973) decision. While often thought of as unilaterally “illegal” prior to Roe, abortion was legally available on a limited basis for women who met the medical justification for a “therapeutic” abortion when there was a threat to the life or health of the pregnant woman (Mohr 1978). Studies of abortion pre-Roe demonstrate that an individual woman’s access to legal abortion care was significantly affected by her social position, including her race, class, and geographic location (Reagan 1997). Women with financial means, for example, had better access to a wanted abortion, whereas women of limited means were stymied (Joffe 1995). White women represented the overwhelming majority of hospital-based abortion recipients; of abortions performed in New York City during 1960 to 1962, only 7% were for nonwhites (Lader 1966). Even in states where abortion was legal, discretionary application of these laws meant that physical location mattered (Packer and Gampell 1959), as in California, where women in the San Francisco Bay Area obtained six times as many abortions as women in the Los Angeles area in 1969 (Jackson, Tashiro, and Cunningham 1971). In practice, this differential system of providing abortion care meant that one key aspect of managing fertility was a privilege available only to some, thereby extending social systems of privilege (further) into reproductive decisions.
Roe was supposed to eliminate this privilege-associated unequal access (Garrow 1998; Petchesky 1990). Superficially, the access problem has been rectified; today, abortion is widely available to most women (Jones and Jerman 2014). Inspection of how and when abortion services are offered, however, demonstrates that inequalities persist; even as women broadly have access to abortion, they do not have equal access to the same kinds of care, with some receiving hospital-based care and most receiving care from designated outpatient facilities.
Of the estimated 1.1 million abortions performed each year, 63% take place in just 329 stand-alone, outpatient clinics that specialize in abortion care (19% of all facilities; Jones and Jerman 2014). An additional 31% take place in 510 other kinds of outpatient clinics (30% of all facilities). In contrast, hospitals constitute 35% of the facilities where abortions are performed but collectively do only 4% of the total number of abortions (Jones and Jerman 2014), suggesting significant selectivity in the patients they serve. Research suggests that women who seek abortion care in nonclinic settings (i.e., in a hospital and/or from their regular doctor) are often turned away (Freedman 2010; Stulberg et al. 2011). Certainly hospital-based care is far more expensive than care that an outpatient clinic provides, but if it were only a question of cost, we would expect the number of hospital-based abortions to be even smaller, with only very rare emergency cases occurring in hospitals. Considered together, these findings demonstrate that differential access to abortion persists, despite Roe.
Freedman (2010) argues that this unequal access can be explained in part by the structural constraints on the provision of abortion. Less research has investigated the way interpersonal interaction influences the decision to selectively provide abortion care in nonclinic settings. Scholarship on doctor-patient relationships points to the discretion physicians wield in providing medical care, often choosing what services to provide based on judgments about whether the patient is “good” or a “problem” (Barr 1983; Bessett 2010; Biener 1983; Lorber 1975; Michaelsen 2012). Other research has illustrated how these evaluations are grounded not only in ideas about normative patients but also in gendered norms (Fisher and Groce 1985; Hoffman and Tarzian 2001; Macintyre 1977) and the perceived social class of the patient (Strong 1979), with consequences for whether the patient receives care and what kind. Moreover, these interactions, in turn, reinforce normative ideals about gender, race, and social class (Bessett 2010; Fisher and Groce 1985).
Here, we illustrate how physician support for a patient’s abortion decision—as evidenced by offering to perform the procedure, coordinating her care, and/or demonstrating empathy—is partially informed by socially significant criteria, such as the patient’s embrace of responsibility to prevent pregnancy, display of vulnerability, and/or desire for motherhood. These criteria are relevant to hegemonic expectations of both gender and sexuality. To capture this intersection, we refer to normative gendered sexuality for women, defined as the set of social expectations about appropriate sexual and reproductive behavior and responsibility to which women, in specific, are held accountable. This concept emphasizes the interconnections between gender and sexuality. We further find that physicians’ support, or lack thereof, has consequences for the kind of abortion women can access. In their capacity as gatekeepers, physicians’ personal judgments about the legitimacy of women’s reasons for abortion or their assessments of what their professional community will accept as legitimate can affect the care they give these patients. This creates a system wherein different patients have access to different kinds of care (e.g., hospital vs. clinic) and those who successfully perform normative gendered sexuality can have more options than those who do not.
Together, these interactions aggregate into what we term a stratified legitimacy of abortions. We find that some abortions are deemed legitimate while others less so in a hierarchy informed by broader patterns of stratification. Consequences for this continuation of unequal access post-Roe include the ongoing stigmatization of abortion and the retrenchment of normative gendered sexuality. Moreover, we argue that the practice by physicians of distinguishing among abortions according to social criteria undercuts claims about the preeminence of medical expertise in abortion care. When abortion care provision decisions are made based not on medical facts but on social criteria, this opens the proverbial door for people outside of medicine (e.g., legislators and voters) to regulate abortion, as we see in the recent spate of state-level abortion restrictions.
Background
The bifurcation of abortion care between hospitals and outpatient clinics, with the latter performing more than 90% of the abortion procedures in the United States (Jones and Jerman 2014), owes in part to the history of abortion advocacy following the Roe decision. Seeking to meet the demand for abortion services quickly and efficiently, while also aiming to fundamentally change the physician-patient relationship (Woodward and Armstrong 1979) and challenge the power of medicine and the male physician (Ehrenreich and English 1978), women’s health supporters in the 1970s turned to the freestanding, outpatient abortion clinic model—and away from hospital-based care (Nelson 2015). There are several benefits to the clinic model, including efficiency, physician skill specialization, and the greater certainty for patients that providers and staff they interact with will support their abortion decision (Dixon-Mueller and Dagg 2002).
Contemporary consequences of this arrangement, however, include the marginalization of abortion services (Freedman 2010; Joffe 1995), with physicians who do not specialize in abortion largely avoiding performing the procedure. A recent national survey showed that 97% of obstetrician-gynecologists encountered abortion-seeking patients, even as only 14% reported performing abortions (Stulberg et al. 2011). In effect, abortion provision has been “turfed” to freestanding clinics that provide abortion services to all medically eligible women who seek them. In contrast, hospitals and the physicians who work in them are often selective about which abortions they perform or refer out to other providers.
This setup echoes the provision of abortion before Roe wherein physicians served as gatekeepers to care (Joffe 1995). Pre-Roe, hospital abortion committees, typically composed of colleagues from several different specialties in a given hospital, would meet periodically to hear a physician present the case of a patient who was seeking an abortion and adjudicate whether to approve the request (Solinger 1992). A core feature of these committees was the requirement that physicians tell women’s stories in ways committee members found compelling.
Here, we investigate whether the current abortion provision landscape shares the pre-Roe interpersonal patterns of physicians choosing which abortions to perform or coordinate care for based on social criteria rather than medical ones. The impact of such selectivity is unlikely to manifest in abortion census numbers, as women who are able will go to great lengths to get abortion care, including substantial travel and monetary outlay (Jones, Upadhyay, and Weitz 2013; Roberts et al. 2014); strictly speaking, women are still able to obtain abortions. Yet there are important implications for women’s access as clinics decline and for the cyclical relationship between abortion stigma and the marginalization of services.
Data and Methods
To investigate physician decision making about abortion care, we drew on 61 interviews with obstetrician-gynecologists (hereafter abbreviated as “ob-gyns”) practicing throughout the United States, collected as part of two data sets. Although broadly similar, the two data sets were collected to address different research questions. Described in greater detail below, the first derived from practicing ob-gyns who completed a residency program that integrated abortion training and the second from ob-gyns who are or have been employed by Catholic hospitals, which formally prohibit abortion care. We anticipated that these two samples would represent distinct—perhaps even opposing—viewpoints on and settings for abortion care.
Data Set A, the first sample, consisted of interviews conducted in 2006 to 2007 as part of the Impact of Abortion Training Study. The study interviewed graduates of four ob-gyn residency programs, 5 to 11 years after completion of residency. The programs, each representing one of four regions in the United States (West, Midwest, Northeast, and South), offered abortion training as a routine, integrated part of training. We mailed a letter of introduction to the directors of these four programs, asking them to forward recruitment materials to all graduates of their program from the years 1996 to 2001 (approximate total 150). The materials described the study as a qualitative assessment of the impact of abortion training. Forty physicians returned signed consent forms by mail, and 30 graduates completed the interview. Of the 30, 22 were women and 8 were men (Table 1). Their ages ranged from 34 to 50, with a median of 40. Nine lived in the West, nine in the Midwest, seven in the Northeast, and five in the South. At the time of the interviews, 20 were employed in a group ob-gyn practice, 6 by academic medical centers, 3 by physician-owned health maintenance organizations (HMOs), and 1 by the military. Interviews were semistructured, guided by the responses of the interviewee, and did not always include the same questions. Nonetheless, every interview covered the same set of topics. Relevant to this analysis, interviews included discussion of physicians’ narratives of abortion training, decision making around abortion practice, and the emotional experience of providing abortions. Respondents completed a written consent. This study was approved by the Committee on Human Research at University of California, San Francisco. See Freedman (2010) for further detail on the recruitment and interview methods.
Respondent Characteristics from Abortion Training (2006–2007) and Institutional Influences (2011–2012) Studies.
Note: HMO = Health Maintenance Organization.
Data Set B, the second set of interviews, was conducted in 2011 to 2012 as part of the Institutional Influences on Physician Practices Study. The study focused on ob-gyns currently or with past experience practicing in a Catholic hospital. Catholic hospitals highly restrict the provision of reproductive services, including abortion. We interviewed 25 ob-gyns recruited from participants in a national survey (N = 1,154) of physician beliefs and practices related to ob-gyn care (Stulberg et al. 2010) who agreed to be recontacted for an interview (n = 237), with subsequent snowball sampling to yield six more interviews, for a total of 31 interviewees. Nineteen respondents were women and 12 were men (Table 1). Ten practiced in the West, nine in the Midwest, four in the Northeast, and eight in the South. Their ages ranged from 31 to 61, with a median age of 48. The group was highly diverse religiously. Respondents largely sought out Catholic hospitals for work not due to a desire for a religious workplace but rather because of geographic proximity, clinical reputation, or preexisting arrangements with their practice. At the time of the interviews, 24 were employed by group ob-gyn practices, 3 by academic medical centers, 2 by a physician-owned HMO, and 2 by public health clinics. As with the residency graduate interviews, interviews were structured to allow respondents to identify what was important to them and did not always follow the same format. However, key topic areas were addressed each time. Relevant to this analysis, all interviews included physicians’ narratives of what they liked and disliked about delivering ob-gyn care in their hospital work settings and how the hospital’s religious doctrine may have shaped how they provided care. Although the Catholic hospital physicians interviewed were a bit more conservative regarding abortion (as evidenced by lower rates of abortion training and by expressions of abortion opposition by a few subjects), the majority did support a woman’s right to abortion despite their hospital’s opposition. Respondents provided verbal consent. This study was approved by the institutional review boards at University of California, San Francisco, and the University of Chicago. See Freedman and Stulberg (2013) for more detail on recruitment methods.
Respondents from both samples provided a wide variety of care and procedures for patients of diverse ages and came into regular contact with patients experiencing socially and medically complex pregnancies that they may or may not have wanted to continue. Few physicians in these studies provided abortion care routinely; most physicians reported that they, their group practices, or their institutions restricted abortion practice and referral. Nevertheless, exceptions were made periodically, and in some cases, physicians went out of their way to coordinate care for a patient with a known provider, including thorough follow-up.
Interviews were completed either in person (25%) or over the phone (75%), using interview guides designed with a small number of required questions and numerous optional prompts to allow for fluid conversation. In general, interviews lasted about 60 minutes. All interviews were recorded and transcribed. Initial review by all three authors suggested both data sets could offer insight into physicians’ decision making about abortion provision. The third author analyzed all 61 transcripts in ATLAS.ti 5 for discussion related to how and when physicians decide to provide or refer for abortion and, in collaboration with the second author, wrote memos identifying analytic themes across the data. They identified the recurring narrative of a stratified legitimacy of abortions. Informed by these memos, the first author then iteratively coded the transcript excerpts to flesh out the construction of stratified legitimacy. Throughout this process, we met regularly to refine and confirm the coding scheme. Disagreements were resolved through discussion until consensus. We considered coding to be complete when no new themes emerged.
Our findings are meant to sketch out this emergent narrative of stratified legitimacy, with consideration for what its presence—and perhaps dominance—means for abortion care in the United States. This is surely not the only narrative of abortion among ob-gyns, but tracing it out has important implications for understanding the current landscape of abortion care, over 40 years after Roe. In the analysis below, no persons, institutions, or locations are identified to protect respondent anonymity.
Results
We find physicians make distinctions among abortions and abortion patients, often with concrete consequences for the level of empathy patients receive and where they are able to obtain abortion care. Strikingly, although we anticipated the two data sets would yield different logics of legitimacy, the contours and underpinnings of physicians’ abortion care decisions were instead similar. Despite Roe’s intent to eliminate variation in access to abortion, we find unevenness in the kind of abortion experience a woman may have. We argue this is largely the result of providers’ role as gatekeepers. Although providers do not impede access to abortion in general, they can nonetheless shape a woman’s experience: some women may receive support and care from their physicians, whereas others will not. Below, we illustrate how the discretionary decisions by physicians from both data sets are constructed in relation to—and serve to reproduce—normative gendered sexuality.
Indicated versus Elective Abortions
The majority of the respondents in our samples did not perform abortions after training and some professed moral discomfort or objection to the procedure. Only three of the physicians in Data Set A routinely provided abortions at the patient’s request (for any reason) at the time of the interview. Physicians in Data Set B provided no abortions within Catholic hospitals, although some had access to other surgical sites where abortion was permitted. However, even respondents from both groups who were uncomfortable with abortion, broadly speaking, reported scenarios where they would offer abortion or help a patient obtain the procedure. To explain these exceptions, physicians asserted a distinction between “indicated” and “elective” abortions, both defined based on medical information about the health of the fetus or the pregnant woman. Medically or genetically indicated abortions denoted pregnancies with a poor or fatal fetal prognosis or those that were medically deemed to place the woman’s life in jeopardy. Elective abortions, in contrast, were abortions for any other reason, including poverty, career, mental health, timing, separation/divorce, and sometimes even the woman’s long-term physical health (some considered this last category to be medically indicated; others did not). This distinction between indicated and elective had a loose relationship to gestational age. Although prenatal testing enables ever-earlier diagnosis of fetal anomalies, most fetal anomalies are not recognized until the second trimester, meaning most “indicated” abortions take place after the first trimester. Thus, later abortions often take place for the most highly legitimized “indications.”
Interviewees articulated distinctions between the two types of abortion that revealed the social evaluations embedded in each category. Typical of several respondents, one physician (Data Set A) explained, Most of the second-trimester ones are genetic [i.e., indicated] . . . and, you know, those people are a little different than people coming in for elective [abortions] . . . and so they had a very different flavor to them.
Elective and indicated abortions were not distinct simply in a medical sense but also in a more ethereal sense—they had a different meaning or “flavor” that the physician could not more precisely articulate. Nonetheless, this flavor was central to the empathy she felt toward abortion patients in cases where there was a fetal anomaly. Similarly, for several respondents, pregnancy circumstances, like fetal anomalies, compelled them to put aside their personal feelings about abortion and provide care to patients terminating their pregnancies.
Beyond underlying personal decisions about participating in abortion care, the indicated-versus-elective binary was used more formally to delineate which abortions physicians were institutionally permitted to perform. One physician (Data Set B) explained that despite her Catholic institution’s prohibition on performing abortions, “if it’s medically indicated and [the] patient’s life is in jeopardy, we can do it in our hospital.” Respondents reported that their hospitals had either policies about when terminations were permitted on site or generally understood parameters that specified which abortions could be performed, all of which related to the health of the fetus or woman. In some cases, a committee of physicians, administrators, and/or clergy was authorized to approve or deny a physician’s request to schedule an abortion procedure. One physician (Data Set B), who had served on a Baptist hospital’s “Termination Committee,” described the hospital’s policies positively, highlighting the way in which they codified indicated abortions as legitimate. He explained, It’s actually a fairly liberal policy that essentially just says, you know, “Look, if the mother’s life is in danger carrying a pregnancy or if the fetal condition is essentially terminal, we are not going to make that mother go through the pregnancy. We can terminate those pregnancies.” So I actually think that’s pretty good for a Baptist-based hospital.
By the physician’s assessment, through recourse to the health status of the fetus and the woman, this policy enabled physicians to perform at least some abortions—and that was a good thing.
Several respondents, as with the physician above, described their hospital’s or practice group’s policy in positive terms, seeing the ability to provide even just a few abortions as an enormous advance over complete prohibition. The framing of an abortion as indicated offered them a medical justification for the abortions they performed. One physician (Data Set A) who received pushback from nurses and administrators at her hospital over her plan to perform a second-trimester abortion for a woman carrying a fatal fetal anomaly, for example, related her indignant response: “[I said,] ‘First of all, this is none of their business. Secondly, it’s medically indicated.’” In so doing, she leveraged her medical authority to insist on the validity of this abortion. Using the label of “medically indicated” served to justify this abortion without endorsing abortion more broadly. For this physician, it was a victory in an abortion-hostile environment.
While expedient in this instance, the physician’s deployment of medical authority in abortion decision making also implicitly casts abortions that do not bear the imprimatur of medical authority (i.e., elective abortions) as less legitimate. As Waggoner (2013) argues of the strategy behind preconception care initiatives, leveraging a normative discourse, in this case, of physician authority, may have unintended consequences. Even as hospital policies enabled physicians to perform a select number of abortions—and impact the lives of individual patients—they had the additional effect of exacerbating the socially illegitimate status of other abortions, banishing them from many hospitals. As one physician (Data Set B) said simply, with no further comment, “In terms of elective terminations, we don’t do them.” The distinction between indicated and elective along with the informal and formal hospital policies enabled physicians to cite outside benchmarks or consensus about appropriate abortions in justifying their own provision or refusal. It also reified the idea that the same medical procedure (i.e., an abortion) could have qualitatively different meanings that physicians are qualified to identify.
Scholars have critiqued the construction of “medical necessity.” Skinner (2012:17), for instance, argues that, rather than understand justifications of medical necessity as a set of scientific facts, one should see them as “discursive representations” (see Farquhar 1996) in which arguments of medical necessity cater and respond to cultural norms that make the absence of choice their primary source of legitimacy. In the case of abortion, where women are the ones whose choice is displaced for a medical justification, there are additional layers of gender at play. Women’s medical needs as they identify them are systematically undervalued in medicine, with women less likely to receive adequate pain management than men, for example (Hoffman and Tarzian 2001). A physician’s endorsement of an abortion for reasons outside of the woman’s stated desires for her pregnancy may allow him or her to provide some patients with abortion care, but it may also inadvertently reinforce a devaluation of women’s own medical decision making.
Moreover, despite the medicalized language, the distinction between healthy and unhealthy fetuses is not clear-cut. Indeed, across our interviews, there was variation in which anomalies respondents cited as indications, underscoring the discursive construction of “medically indicated.” The clinical language of indicated itself obscures the ambiguity that often attends fetal anomalies: whether a fetus would have a good quality of life is not always known. Although many fetal anomalies are incompatible with life, others are associated with shorter or even very short life, physical disabilities, and/or mental disabilities. When those disabilities make quality of life just too low or duration of life just too short is not necessarily a medical question but, rather, the subjective judgment of the pregnant woman, often in coordination with her partner. In this vein, we found variability across sites in whether a given indication resulted in termination, consistent with research finding that some populations are more likely to terminate for a given anomaly than others (Rapp 1999).
Empathizing with Some Patients
While respondents expressed comfort with what they considered to be medically indicated abortions (even as what fell under that purview varied) through recourse to the medical facts of the pregnancies, a subset also differentiated among abortions using criteria that were more explicitly social, including aspects of patients’ behavior, such as their emotional or maternal performance. 1 For some respondents, this differentiation did not impact patients’ access to care—the physicians performed abortions or made referrals as requested—but their level of empathy for patients was not uniform, and some women seeking abortions were considered more worthy of empathy than others. This led them to feel that those abortions, in particular, were appropriate and legitimate. For other respondents, their social evaluation dovetailed with the medical facts of the pregnancy, often in a tautology. For all respondents, these findings point to the pervasiveness of normative gendered sexuality and its influence on the patient–provider relationship.
Although the literature has found that highly emotional patients are not always considered “good” patients (Lorber 1975), our data suggest that the opposite is true in physicians’ empathetic response to abortion patients. Typical of respondents, one physician (Data Set A) expressed disapproval at what she perceived as patients’ emotional apathy toward abortion while simultaneously linking a patient’s lack of affect about having an abortion to an overall failure to contracept: “There’s definitely some people [who sought abortions] that didn’t seem to care and [abortion] was their form of birth control and that was a little bothersome.” In essence, she asserts that these abortions were less appropriate or moral, justifying her statement through recourse to patients’ emotional expression, even as she was willing to perform the procedures while training in abortion during residency.
Another (Data Set A) explained, “There’s always bad apples in every bunch that were irresponsible, and this was their fix.” The “bad apples” she refers to were women who used abortion to “fix” their contraceptive mistakes. By assigning culpability for unintended pregnancy (i.e., irresponsible use of contraception) and positioning abortion as women’s intentional backup plan, these physicians invoked conceptions of normative gendered sexuality that construct women as primarily responsible for preventing pregnancy (Fennell 2011). This construction further asserts that abortion should be rare (Weitz 2010), with perhaps one “mistake” allowed (Reich and Brindis 2006). Contemporary ideals of a liberal meritocracy hold that women can and should control their fertility (Layne 2003; Rothman 1989). Women who seek abortions have thus failed to control their fertility as socially expected, setting up the typical abortion patient as a woman who is selfish, irresponsible, and uncaring (Cockrill and Nack 2013). Each aspect becomes evidence of the others such that when women failed to perform distress about an abortion, their lack of emotionality was read as evidence that they were irresponsible contraceptors. Their subsequent need for abortion, in turn, did not inspire empathy in physicians—nor would we expect it to given the pervasiveness of normative gendered sexuality.
In parallel, physicians spoke approvingly of women who performed distress or regret over an unplanned pregnancy and the need for abortion. This framing identified women’s emotionality as evidence that they were assuming responsibility for their previous role failure, conveying guilt over their inability to prevent pregnancy. As one respondent (Data Set A) explained, the evident distress of most patients seeking abortions helped her feel better about performing terminations. She said, [Abortion training] made me more [sympathetic] for people who needed the abortion. . . . We do the pre-ops, we talk to the patients, and a lot of time they’re there crying and all that because they really don’t want to do it, but they don’t see any other way of it. So it makes you feel like you’re actually doing something not that negative.
To the extent that the physician felt uncomfortable with abortion provision—internalizing abortion stigma or personally questioning the morality of abortion, perhaps—seeing women who were vulnerable and sad rather than cavalier made her feel empathy toward her patients. In highlighting behavior in ways consistent with normative notions of women as remorseful that they had to seek an abortion (i.e., had failed to prevent pregnancy), this narrative marked some abortion patients as moral, consistent with overarching social ideals about appropriate gendered sexuality.
Attention to patients’ gendered performance was not limited to elective abortions. One physician (Data Set A) asserted that medically indicated terminations are more distressing to women than abortions for other reasons because of the presumed circumstances of the pregnancy. She said, “[Genetically abnormal pregnancies] are desired—that they found out had a problem—and so [abortions are] very kind of traumatic for them.” Her statement links intended pregnancy to indicated abortion and to visible and implied emotional distress. All three signaled that these were legitimate abortions that compelled empathy; simultaneously, the presence of one could imply the other two. Thus, in these accounts, women who had medically indicated abortions had wanted pregnancies. They had, in our culture of reproductive responsibility, done everything right: they had sex responsibly, eschewing contraception not because they were lazy but because they wanted to become pregnant. This inferred behavior led respondents to express greater empathy for these patients. Overall, we find that abortions for pregnancies achieved through normative performances of gendered sexuality—through sex for procreative purposes, that is, intended pregnancies—were more likely to be framed as legitimate.
Some physicians described carving out exceptions for certain patients that drew more narrowly on normative femininity, without reference to sexuality. For example, one physician (Data Set A), recounting a case where a woman needed a termination because of a fetal anomaly, explained that she was asked to perform the abortion by her practice partner because otherwise the patient would have to travel to the abortion facility 30 minutes away. The partner was not trained in abortion procedures and, on principle, did not perform them. But, in this case, he deviated from their standard practice of referring abortion patients out and asked his colleague to take care of the patient. Justification for this exception to standard referral practices was based on the patient’s anxiety about going to an unknown provider in an unknown city. She detailed, [This patient] wasn’t comfortable having to go down there [to a designated abortion provider] for a procedure, and she didn’t know anybody there and it just seemed overwhelming. She just wanted to stay in the practice that she was comfortable with. And the physician who was taking care of her didn’t perform those procedures, so he asked me.
In the respondent’s retelling, the exception was rationalized largely because of the emotional expressions of the patient, which generated an empathetic and caretaking response from the referring physician. This patient was hardly alone in wanting to receive abortion care from a doctor she knew, in a location she was familiar with (Weitz and Cockrill 2010). But, in this instance, although her desire for a local abortion from a familiar doctor did not make her a special case, her emotional expression and the understanding that her pregnancy was wanted and the abortion “indicated” resonated with her primary physician and compelled him to pursue an unusual referral for an abortion in their own town. He became her advocate and, in effect, protector.
Other accounts touched on normative expectations about women as naturally maternal who experience abortion as an injury to their fundamental desire to nurture (Siegel 2008). For example, one respondent (Data Set B) shared an account of a newly pregnant patient who had unexplained renal failure: [She] had to go to dialysis three times a week and her dialysis, finally, you know, they’ve gotten better at that and they finally got so good that she became fertile and got pregnant. And she was absolutely torn. She was like a 36-year-old prima gravitas [first-time pregnancy] and she knew if she stayed pregnant she was going to have to go to dialysis five times a week [and] she wouldn’t have a job anymore. Likelihood was she would make it to 24 to 28 weeks [gestation] and then end up raising a child with disabilities while she’s on dialysis. And she finally did decide to have an abortion and thank goodness it still is legal here but, you know, that’s just not the kind of thing that you can legislate. That wasn’t incest, she wasn’t going to die, you know. But I thought that was a pretty darned good reason. And she was pretty torn up about it but she did go have an abortion.
In describing the case, the physician enumerated several medical reasons why abortion was appropriate, including risk to the patient’s long-term health and the likelihood that she would deliver prematurely. Simultaneously, the physician repeatedly called attention to the difficulty the patient experienced in deciding to abort, marking the patient’s emotional experience of deciding to have an abortion as appropriate. She also flagged that the patient took some time to make the decision, indicating that she did not choose abortion without extensive consideration. In this way, the physician’s endorsement of the legitimacy of this patient’s abortion was premised not only on her assessment of the patient’s health but also on her tacit approval of the patient’s expression of emotion. The physician constructed the patient’s emotion as a proxy for her proper embrace of a maternal role: she anguishes over the decision to terminate. Further, the physician asserted her own evaluation of the patient’s reason as “pretty darned good,” both applying her medical authority to the decision and implicitly reifying a hierarchy of reasons for abortion.
This is not to say that physicians did not reflect critically on their emotional responses to patients. For example, one physician (Data Set A) recalled a case of a woman during his residency who knew early in her pregnancy that she, as he recounted, “didn’t want to keep it.” However, she did not seek an abortion until late in the second trimester, thereby increasing the difficulty of the procedure and the risk to the patient. The physician had a negative reaction to the patient’s delay in seeking care, but one he quickly identified as problematic: I remember at the time thinking, I was like, “God dang it, you know, she did this and this is a crappy thing that she’s, you know, doing to herself, putting herself at risk, and me because now it’s a much more difficult procedure, and the fetus because it didn’t need to go through this.” And I, and I remember being inwardly conflicted, you know, wanting to do a good job, but in my heart thinking, “You know, if this doesn’t go right, she almost deserves it.” And as soon as that thought popped into my head, it’s like, “Whoa! This is really bad. This is not a thought that should be coming into my head.”
This physician distinguished himself from other respondents in acknowledging that his resentment toward the patient was grounded in her social behavior—which he perceived as irresponsible and uncaring—and not a clinical assessment. Nonetheless, his story is consistent with those of other respondents in demonstrating the pervasiveness of normative judgments of abortion patients, specifically regarding what is considered appropriate behavior in such decision making.
Protecting Patients from the Clinic
Finally, in addition to differentiating among abortions and abortion patients based on the medical facts of the pregnancy and the woman’s emotional expression, physicians asserted qualitative differences between hospital-based abortions (or medical practice-based abortions taking place in the office or outpatient surgery centers) and ones obtained in dedicated abortion clinics. Specifically, the latter were marked as lower value than the former. This differentiation is not grounded in medical evidence; hospital-based abortions are not medically safer than those in a clinic (Upadhyay et al. 2015). There is also no evidence that patients experience clinic-based care negatively; research regularly finds very high patient satisfaction (McLemore et al. 2014; Taylor et al. 2013). Nonetheless, hospital- or practice-based care means that the patient is much more likely to have an ongoing relationship with her care providers. Respondents’ explanations of what made them empathize with some patients and what set some patients apart and earned them a procedure by their physician or a recommended colleague further illuminate how physicians (re)produce a narrative of stratified legitimacy of abortions.
When respondents went out of their way to secure care for patients in their hospitals, they often emphasized that these patients did not deserve the “punishment” of being sent elsewhere for their care. This application of value to the location where a patient could receive her abortion was read not only forward, with respondents concerned about where a woman would receive her abortion, but also backward, with an implied evaluation of the legitimacy of an abortion based on where it was obtained.
In some accounts where physicians lobbied authority figures for permission to provide abortion care on site, they explicitly disparaged abortion clinics. Speaking of a cancer patient whose treatment necessitated her terminating a wanted pregnancy, one physician (Data Set B) tacitly marked going to a clinic as a punishment, insisting she deserved an in-house procedure: I remember speaking to [the ethics committee], saying . . . “This poor woman is suffering. She’s got a malignancy, she doesn’t want to terminate, she realizes she has to and you’re going to make her go to a clinic?”
Such framings of the clinic were about more than the inconvenience of having to go to another location. They were tied to the constructed distinctions among abortions and abortion patients. This physician sought to protect this patient, who did not want to have an abortion, from having to go to a clinic where she would be grouped with women who presumably, in contrast, wanted to have abortions. That her procedure was medically indicated and she was “suffering” rendered her abortion legitimately hospital-worthy to her physician advocate, although the Catholic hospital authorities still insisted that her procedure could not be performed in house.
Physicians were not the only ones to recognize the social meaning for patients of staying with their current caregivers for abortion care. One physician (Data Set B) related the story of being lobbied to perform a termination on site by a priest, to whom we give the pseudonym Father Gregory. The patient in question, a devout Catholic, was carrying a fetus with anencephaly, a fatal anomaly wherein the fetus does not have a fully formed brain. After deliberation and consultation with Father Gregory, who was affiliated with the Catholic hospital, the patient decided to terminate the pregnancy. Catholic hospitals formally allow abortion procedures only when the fetus is already dead (for a more nuanced analysis of these policies and their consequences, see Freedman and Stulberg 2013). In this case, Father Gregory was asking for an exception. He justified his request by invoking the sinister specter of the abortion clinic. The physician related, So apparently, she was one of Father Gregory’s parishioners and she went to Father Gregory and she said she really didn’t want to carry this pregnancy, but she really didn’t want to go to an abortion clinic and why does she have to go to an abortion clinic? Why can’t she stay at [our Catholic hospital]? And this baby’s not gonna make it anyway. And she did this whole thing with Father Gregory. So then Father Gregory calls me and he said, “Well, you know, it’s really okay for you to do this procedure at [the Catholic hospital].”
The physician recalled being taken aback by this request. Although she was supportive of abortion rights and trained to perform the procedure, she felt the priest’s plea failed to recognize the logistical difficulties of securing the appropriate space and staffing for a termination in a hospital with a policy against abortion provision. The Father’s request was more substantial than he realized.
As the physician related the story, she also recognized that the priest’s request was not just about making an exception but also about formalizing the distinction between legitimate and illegitimate abortions. If this physician relented and provided a hospital-based abortion, she believed, this patient and the priest could construct this procedure as not an abortion. This offended her and cemented her decision not to provide the termination. She said, And, you know what? The fact of the matter is, it’s an abortion. I mean, part of it, too, is, I feel like it’s an abortion. They do this in abortion clinics. If you are having this, this is an abortion . . . so, a little bit, I didn’t really feel a great need to sort of bolster her [the patient’s] fantasy that this is not an abortion.
This physician’s reaction to the request for an in-house abortion highlights how the physical location of the procedure can change its social definition. In asking for a hospital-based procedure, the priest was, by the physician’s analysis, trying to protect the patient from feeling like she was having an abortion. The physician believed that the in-house setting could enable them to construct the procedure as so legitimate it ceased to be an abortion altogether.
Discussion
Although Roe aimed to eliminate differential access to abortion care, in practice, this has not occurred. The majority of abortions—over 90%—take place in clinics (Jones and Jerman 2014), with abortion services generally absent from mainstream medical care (Freedman et al. 2010; Joffe 1995). While the lower cost of clinic-based care may explain some of the predominance of clinics in abortion provision, the small but notable number of hospital-based abortions suggests that more than economics is driving this pattern of provision. Building upon literature examining the structural reasons for abortion’s continued marginalization (Freedman 2010), we find that physicians articulate a narrative of a stratified legitimacy of abortions to explain their selective provision and coordination of abortion care. Medical decisions are often fundamentally social, and they can and do replicate inequality. Like the pre-Roe days of therapeutic abortion committees, physicians and institutions make determinations of which women are more or less worthy of receiving the procedure—and of when they will advocate for a patient, given their perception of what colleagues and administrators will support. The consensus around these determinations demonstrates the construction of a broader narrative about abortion that is grounded in the idea that abortion is a morally dubious procedure and not all women are equally justified in seeking abortion care.
Our analysis identifies several ways physicians produced this narrative of stratified legitimacy. First, physicians distinguished among the reasons women sought abortion, privileging “medical” reasons for termination that were, in effect, out of a woman’s control. This can serve, likely inadvertently, to devalue women as medical decision makers. Abortion for reasons of fetal anomaly was treated as the least problematic reason for abortion, both for providing care and for advocating on behalf of a patient to others. Future research should take up the question of why particular fetal indications find more sympathy among physicians and hospital administrations, as the unproblematic construction of the validity of abortions for fetal indications contributes to disability stigma (Asch 1999; Jesudason and Epstein 2011; Saxton 2006).
Second, respondents articulated greater empathy and approval for women who evidenced wanting their pregnancy, remorse over “their” contraceptive failure, or general distress about seeking abortion—that is, their sexuality-related behavior was considered appropriate. These expressions are consistent with normative gendered sexuality wherein women are constructed as maternal, selfless, and responsible for pregnancy prevention. By successfully performing normative gendered sexuality, women were rewarded with the imprimatur of a legitimate abortion and, often, physicians’ willingness to offer or facilitate abortion care. In contrast, the less well a woman performed normative gendered sexuality, that is, the less attached or saddened by the loss of the pregnancy she appeared, the less legitimate her abortion was considered.
Finally, physicians assigned value to the location where the abortion would take place, stigmatizing dedicated abortion clinics and advocating for nonclinic procedures for “legitimate” abortion patients. Although the “punishment” of being sent to a dedicated abortion clinic does not inherently mean a poorer quality of care, physician efforts to help favored patients avoid abortion clinics reinforce social constructs of the abortion clinic as a scary place (Condit 1990; Kimport, Cockrill, and Weitz 2012).
In its grounding in patients’ gendered performances, the stratified legitimacy of abortions has the—likely unintended—effect of serving to perpetuate normative gendered sexuality, tying the receipt of particular kinds of socially valued medical care to idealized and expressive motherhood and distress over pregnancy loss. Future research should investigate the implications of bifurcating legitimate from illegitimate abortions for race and class inequality. To the extent that pregnancy unwantedness overlaps with unintendedness, legitimizing only abortions related to wanted pregnancies is a raced and classed assessment. Women of color, particularly, black women, have higher rates of unintended pregnancy than white women, and lower-income women have higher rates of unintended pregnancy and abortion than middle- and upper-income women (Dehlendorf, Harris, and Weitz 2013). Similarly, low-income women are more likely to be without health insurance and/or a regular healthcare provider (Dehlendorf et al. 2013). When in need of abortion care—for any reason—clinics are typically their only option. The devaluation of the stand-alone clinic can extend to the clients it serves, further marginalizing not only “elective” abortion care but also the women for whom these are the only sources of care. The result is that white and middle- and upper-income women can be more often at the center of the empathetic abortion narrative, offering a strong echo of pre-Roe practices of providing abortions to some—usually privileged—women.
In addition to extending systems of privilege, the stratified legitimacy of abortions may have implications for policy making and advocacy. By inserting social criteria into medical decision making, the narrative of stratified legitimacy makes abortion policy vulnerable to seemingly commonsense policies and regulations. It weakens the claim that abortion care decision making requires medical expertise. This opens the door for legislators and voters who lack medical knowledge to nonetheless claim authority to regulate and restrict abortion through recourse to social knowledge. Our findings on the importance of interpersonal interaction also add depth to advocacy efforts that have focused on social change at the institutional or policy level. While advocates may be tempted to push for the reintegration of abortion into general ob-gyn care—especially in light of contemporary attacks on public funding for Planned Parenthood, a major abortion care provider in the United States, and the rise in state-level restrictions of abortion care at freestanding abortion facilities (Guttmacher Institute 2015)—our findings point to some of the overlooked and substantial challenges such efforts would face in terms of medical culture. Organizational policies certainly constrain physicians’ provision of abortion (Freedman 2010), but our data suggest that interpersonal factors and social norms are at work as well.
These findings must be understood in parallel with studies of physicians with practices dedicated to abortion care (Joffe 1995), who pointedly do not differentiate among patients in providing services (although they may still provide specialized care to some patients). The narrative of stratified legitimacy is not the only narrative of abortion available to—or produced by—ob-gyns. Future research should evaluate the dominance of the stratified legitimacy narrative, although we suspect that the marginalization of dedicated abortion providers (Joffe 1995) combined with their comparatively small numbers means that their discursive contributions carry less weight than those produced by the respondents we analyze here.
Our findings illuminate social questions about the role and meaning of a stigmatized service in contemporary society, illustrating how, in the case of abortion, particular emotional displays, backstories, and health conditions are privileged. We find that interpersonal interactions between physicians and patients, contextualized in ongoing systems of gender, produce a stratified legitimacy of abortions that reflects and reifies social expectations of normative gendered sexuality. Although abortion is in some ways uniquely marked as morally dubious and politically fraught, other kinds of medical care may similarly be characterized by stratified legitimacy. For instance, the concept of stratified legitimacy may be of value in understanding access to care for other stigmatized conditions, such as obesity and HIV/AIDS, or for people in stigmatized positions, such as incarceration. Such examinations can contribute to broader understandings of how provider-patient interactions produce and perpetuate social inequalities.
Footnotes
Acknowledgements
An earlier version of this paper was presented at the 2013 meetings of the American Sociological Association.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data collection was supported by a University of California-Davis Block Grant, the Mary and Lloyd Schwall Foundation, and an anonymous foundation, and partial support for data analysis was provided by the Society of Family Planning.