Despite improvements in abortion training since the mid-1990s and evidence that the availability of training increases the likelihood of postresidency provision, only half of residents who intend to include abortion in their practice actually do so.1 While popular media and academic literature commonly report that physicians with abortion training avoid providing abortions for fear of harassment and violence,2–5 little is known about professional obstacles that might affect abortion provision.

In 1996, the American Council of Graduate Medical Education (ACGME) mandated that obstetrics and gynecology residencies provide abortion training, and a national survey conducted in 2002–2003 showed that obstetrician-gynecologists who trained after 1996 were more likely to perform abortions after completing their residency than were those who had trained before the policy change.1 Nonetheless, the overall number of facilities providing abortions declined by 14% between 1992 and 1996, by 11% between 1996 and 2000, and by 2% between 2000 and 2005.6 In 2005, 69% of metropolitan U.S. counties and 97% of nonmetropolitan counties had no abortion provider at all.6 Physicians are not in short supply for most urban abortion clinics, but rural clinics and those in especially conservative areas struggle to find physicians.6–8

Since legalization, abortion services have increasingly become consolidated into the socially insulated settings of specialized abortion clinics. These clinics, which provide 93% of abortions,6 are largely segregated from other medical settings; as a result, they are vulnerable to harassment, violence and targeted legislation, such as laws that impose burdensome and unnecessary requirements on their architecture, landscaping or staffing.9

Many members of the reproductive rights community have advocated for integrating abortion into full-spectrum obstetrics and gynecology and primary care settings, to take the burden off the clinics and to normalize abortion as a standard component of reproductive health care. However, integration remains rare, and little is known about why that is so. Research to understand the factors associated with abortion provision by physicians after residency has thus far been limited to quantitative studies. For example, a 2002–2003 cross-sectional study of 2,149 U.S. practicing obstetrician-gynecologists found that while abortion training was associated with an increased likelihood of provision, postresidency practice restrictions were associated with decreased odds of provision.1 However, the professional barriers that recent graduates of obstetrics and gynecology residency programs face when they wish to provide abortions have not been explored in depth; we conducted a qualitative study as a start to filling that gap.

METHODS

Graduates of four obstetrics and gynecology residency programs that had offered routine, opt-out abortion training since before 1996 were chosen for study. Opt-out programs allocate time for abortion training in residents’ schedules and expect residents to participate in that training unless they have religious or moral objections. The opposite, opt-in programs, arrange for residents to get abortion training during their elective rotations, often at off-site abortion clinics. In 2006, 51% of obstetrics and gynecology residency programs responding to a national survey reported having opt-out training, 39% reported opt-in training and the remaining 10% reported no abortion training options.10 (The ACGME mandate includes an exemption for programs with religious or moral objections to abortion.) We recruited graduates of opt-out programs because we wanted to interview physicians who had been trained in a medical setting that considered abortion training a routine aspect of residency education.

We purposively selected four programs with a strong history of abortion training that predated the ACGME mandate; each represented a different of the United States region (West, Midwest, Northeast and South). In 2006, we mailed a letter of introduction to their directors and asked them to forward our study recruitment materials to all graduates (who numbered approximately 150) from the years 1996–2001. Twenty-seven percent of the physicians returned signed consent forms by mail. Despite the low response rate, we did not ask the programs to conduct a second mailing, because we had reached our goal (aimed at balancing regional representation) of interviewing at least five physicians from each program represented in the study. Thirteen percent of respondents had opted out of abortion training for moral reasons (one of these had opted back in), and therefore the study included diverse perspectives.

The lead author, a sociologist with training in qualitative methodology, conducted in-depth interviews with respondents either in person or over the phone. The interview guide was designed with required questions and optional prompts to allow for fluid conversation, and was modified slightly during the research process to reflect new questions that arose from the early findings. Interviews took 30–60 minutes to complete and focused on participants’ abortion training and subsequent professional experiences. Topics covered in detail included physicians’ abortion training, professional paths since residency and decision making regarding abortion provision. Interviews were transcribed, and analytic themes that emerged were coded by the first author with Atlas.ti 5.0.

The analysis used grounded theory methods, which take an inductive approach to generate theory from the data, rather than test a hypothesis or a preselected theory. In this approach, research questions are initially very broad. After examining the data and noting recurring and meaningful themes, the researcher formulates theories or connects the data to existing social theories that offer explanatory value to the subject. The study was approved by the University of California, San Francisco, institutional review board.

In this article, we use initials based upon pseudonyms to denote physicians’ names. Thus no persons, institutions or locations beyond general region are identified in connection with the data.

RESULTS

Overview

Of the 40 physicians who agreed to participate, 30 were available to complete interviews. Nine each had been trained in programs in the West and the Midwest, seven in the Northeast and five in the South, although some had moved to different regions to practice. Seventy-three percent of respondents were female, approximating the proportion of new obstetrics and gynecology residents who were women in 2007–2008, the most recent year for which data are available (76%);11 participants’ ages ranged between 34 and 50 years, and were mainly clustered near 40.

Most respondents practiced general obstetrics and gynecology. Twenty were in private practice, and the rest worked in academic institutions (six), HMOs (three) or the military (one). Eighteen respondents had wanted to offer elective abortions after residency, five had wanted to provide abortions only under specific circumstances, and the remaining seven had not wanted to provide any abortions. However, only three physicians were offering elective abortions at the time of the interview; two of these had done fellowships in family planning and currently taught abortion in obstetrics and gynecology residency programs, and one was providing abortions in her private obstetrics and gynecology practice in a large western city. Three physicians who did not provide abortions cited violence against abortion providers as a primary deterrent. Several others mentioned violence as a concern, but found other barriers more prohibitive.

Physicians’ narratives focused on professional barriers to providing abortions, which fell into three categories: practice prohibitions, strain on peer relationships and institutional restrictions. Respondents described both explicit and implied prohibitions on abortion provision where they work. Some physicians in private practice learned about these prohibitions before they were hired, and others discovered them later. Seven participants described the desire to maintain professional civility or collegiality in small group practices and felt that including abortion in their practice would not be conducive to either. Finally, 12 physicians, in both religiously affiliated and nonsectarian hospitals, reported institutional restrictions.

Practice Prohibitions

• Prohibitions made explicit before hire. Some respondents had asked directly about abortion provision when interviewing to join a practice and had been told that it would not be permitted. Others said that their employers had given them an unsolicited warning at the interview that abortion provision would not be tolerated. While not written into a contract, agreement not to perform abortions had been an explicit condition of employment for some and sometimes was expressed in an intimidating manner. For example, Dr. K recounted her experience in interviewing with a practice in a midsize city in the Midwest:

Agreement not to perform abortions had been an explicit condition of employment for some.

“In this group, you interview with all of the … partners. And the one partner who’s very senior in the group and very prolife, basically his only job is to sit with you and just tell you …‘If you join this group, you will not be performing abortion procedures. And if that’s a problem for you, then you will work elsewhere. Okay?’”

Although Dr. K had wanted to continue performing abortions, she took the job because the practice was one of only two obstetrics and gynecology practices in her area, and both prohibited abortion.

Similarly, Dr. S had been directly threatened by an outgoing senior partner while interviewing for a position in an obstetrics and gynecology private practice in a large midwestern city. Dr. S remembered, “He leaned across the desk and said, ‘If I ever find out you did elective abortion any time in your professional life, you’ll never practice medicine in [this state] again. Do you understand that?’”

In contrast, some groups communicated their abortion prohibitions in a more collegial way. For example, Dr. D, practicing in a small southern town, recalled the interview with his private group practice, in which they discussed his having participated in abortion training during residency. A senior member of the group with strong antiabortion views pressed him to explain why he had participated. The partners told him during the interview, “We’re not going to be doing that.” And Dr. M, practicing in the Northeast, recounted:

“When I finished my residency, I went to [a northeastern state], and I was working in a small hospital. … No one at the hospital would ever perform an abortion. … It wasn’t a religious hospital, but it was a very conservative town, and they just felt like they didn’t want to be associated with doing terminations. And they told me that at the interview.”

•Restrictions discovered after hire. Other physicians learned about restrictions after beginning employment. For example, Dr. G had extensive abortion training; as a result, when she was hired by a large private group practice in a major midwestern city, she was asked to be on its abortion committee, which was designed to screen and approve abortions in the practice. The committee, which includes physicians with different areas of specialization (e.g., family practice and pediatrics) and a chaplain, discusses every case under consideration. However, Dr. G said, “the policy that we have is basically no elective abortions”; the committee approves abortions only for women whose fetus has a fatal anomaly or for whom the pregnancy may cause serious health risks, and refers other women elsewhere.

Dr. P said that abortion provision in her private, nonsectarian hospital in a southern city was similarly restrictive. Nobody had told her directly that the hospital does not allow elective procedures, but she figured it out after learning that several signatures are required to confirm that abortion is a “necessity.” As she recounted:

“Let’s say we have a 16-week anomaly. … We [need] a signature from the chief of staff, the maternal-fetal medicine doctor, the OB chief.”

A few physicians attempted to moonlight while working in private practices where abortion provision was prohibited, and they were surprised to find out that their groups prohibited it outside the practice as well. Dr. K, from the Midwest said, “I brought it back to the group, and they nixed it and said absolutely not, just because they didn’t want my name associated with the [abortion] clinic.” Relatively new and powerless in the practice, none of the physicians who had gotten similar responses to their request to moonlight in abortion clinics pursued it.

In other instances, despite the absence of overt restrictions, participants found that the culture of their group practice or institution was to discourage abortion provision and refer women elsewhere for abortion services. For example, Dr. F, from a large southern city, said abortions are never done in her practice. She learned this shortly after being hired, when she noticed that abortion providers were listed in the referral book in the office. She casually asked a colleague about whether practice members do abortions, and the colleague explained that because of one senior partner’s opposition, patients were always referred elsewhere for abortions.

Another physician, Dr. R, working in a suburb of a large western city, explained that she does not perform abortions because some staff at the public hospital where she performs surgery are opposed to abortion and refuse to assist in procedures. In Dr. R’s view, the policies of her group practice are not prohibitive, but the culture of the practice makes it so:

“It’s a big deal. I don’t know if the nurses don’t want to be part of it or they all just like to band together … because if you’re the one that says you don’t mind doing it, everyone else is going to look at you. So if there’s an abortion procedure that needs to be done, I send [the woman] to Planned Parenthood. It’s not worth my time and effort to jump through the hoops of the hospital to make that happen. … Actually, in my first couple months in practice, the people that are in my office here told me, ‘Don’t even bother.’”

Dr. W, working in a large northeastern suburb, had encountered abortion restrictions in two work settings. In the first, she remembered, “I asked when I was interviewing, ‘Do you guys do [abortions]?’ They said, ‘No, we usually don’t.’” While abortion was not directly prohibited, she observed that it was not acceptable, as senior members of her practice would cover up the occasional abortion they performed. She explained, “Once in a while there, somebody would sneak one in and call it [an] incomplete miscarriage, [because] bleeding they can use as an excuse. But it really was termination. … I maybe saw one or two cases the whole five years I was in that practice.” In her new practice, Dr. W said she was told that abortions are performed only when medically indicated and that women are otherwise referred elsewhere for the procedure.

Peer Relationships

Some doctors experienced conflict regarding abortion within their private group practices. These participants reported that they tended to avoid performing abortions, or minimized the number they provided, so that they would not provoke conflict. For example, Dr. G recounted a case that made her aware of how she would have to adjust her practice to minimize conflict. One of her patients was carrying a fetus with a fatal anomaly and wanted to have an abortion by inducing labor. Dr. G scheduled the procedure, but two of her partners who were scheduled to be on night call for the practice became upset when they learned they might need to care for a patient undergoing an induction. Dr. G had been unaware that these particular colleagues would not participate in abortion care for any reason. She explained that the procedure “ended up getting changed to another day, but it brought to my attention there are definitely partners who won’t do it at all. You really need to communicate if you’re going to set up something on somebody else’s call day.”

Dr. D, practicing in a small southern town, recalled a time when he and his colleague strongly disagreed about making an abortion referral for a hospitalized patient. Dr. D’s senior colleague confronted him somewhat angrily for making the referral and told him that he had gone to the patient directly to try to talk her out of having the abortion. Dr. D recalled the interaction:

“I said, ‘Listen, we clearly are going to disagree on this topic and not change our minds about it. … I appreciate your opinion, and … I expect you to appreciate mine.’”

Dr. D is prochoice, but except in that one instance, he has tried to avoid any discussion of abortion with practice members, “You still have to work and get along with these people. … There’s only so much upstarting that you can do, especially without having a lot of your own patients yet. So I have remained fairly quiescent about it.”

Dr. H, practicing in a small town in the West, noted that conflict over abortion has been known to divide practices. “I have heard about practices where … those that were not [performing abortions] had the issues with those who were performing them, and the practice ended up splitting over it.” Dr. H implies that in some cases, physicians end up practicing with like-minded colleagues as a result of such conflict.

Conflict over abortion has been known to divide practices.

Institutional Restrictions

Physicians working for large HMOs or health networks, both religiously affiliated and nonsectarian, can find themselves without the autonomy to decide whether to provide abortions. Catholic health networks, which account for one-sixth of hospital beds and yearly hospital admissions in the United States,12 pose extensive restrictions on reproductive health care services provided within their properties and by their employees. One physician, who was on the faculty in her residency program at the time of the interview, remarked:

“The majority of our residents stay in town, and we have a very strong [Catholic] health care system that has a lot of tentacles through the community. … Even though you have an independent practice, they own the building, and they refuse to allow you to do abortions—even if it’s in your own [private] practice. … There’re several private groups associated with that facility, and so it makes it really tough.”

Nonsectarian HMOs can effectively block abortion provision as well, but they do so usually for economic reasons rather than ideological ones. Dr. F, who worked for a large HMO in a western city, recalled, “I wasn’t part of this, but at some point they decided to contract [abortions] out.” Because of the low cost of abortion care in specialized clinics, such contracting out is common. It is further justified by concerns about conflict avoidance such as those discussed above. Dr. N, who worked for the same HMO but in a different city in the West, explained that the HMO sends women elsewhere for abortions and that the administration is “happy to not have to deal with” the issue:

“The chief of my department told me, ‘I think everybody’s just very relieved that we don’t have to worry about this ourselves.’… And she’s somebody who’s actually a supporter, but she was relieved as the chief not to have to deal with … who was going to do [abortions], who wasn’t going to do them, and whether the department had to be all in agreement about providing the service.”

In yet another western city, Dr. V felt especially frustrated when her HMO employer decided to contract out abortion care. She had done a fellowship in abortion and family planning, and had hoped to offer those skills to her patients. She said, “I’m really dismayed about it. And I really love my job here, but it makes me kind of sad. Because I feel like I have this skill, I should be spreading it around.” She looked into moonlighting, but her employer does not permit physicians to work outside the system.

DISCUSSION

Our findings do not confirm the common assumption that physicians avoid abortion provision out of fear of violence or harassment.2–5 Rather, the physicians in our study emphasized that professional obstacles—explicit and subtle practice restrictions and fear of repercussions from colleagues—hinder the integration of abortion into medical practice. Ultimately, the stigma of abortion and ideological disagreement are at the root of the policy restrictions and collegial strain surrounding abortion. However, our study illustrates that medical professionals face myriad challenges. First, physician autonomy has decreased substantially in the past few decades as a result of major structural shifts in health care financing under managed care.13,14 Since the 1980s, physicians have rarely gone into practice alone, but have tended to join private group practices, large HMOs and academic institutions,15–17 whose policies may dictate specifics of their practice. Second, those in charge of these larger organizational bodies and their practice policies must contend with market forces and the political environment within which they function. Therefore, even if they are not ideologically opposed to abortion, they may fear a loss of business due to the stigma and controversy that may surround abortion provision, and thus may contract out or refer out abortion care.

Our findings bring up two related questions: How can residency programs help physicians overcome the professional obstacles to integrating abortion into their practice? And how can programs equip physicians who want to provide abortions to do so outside of restrictive settings—ideally without being forced to forgo other parts of their medical practice? One answer may be to explore how to prepare and support physicians as they make the transition from residency programs to practice. What kinds of information and resources might help them to continue performing abortions if they so wish? The explicit prohibitions described by our study participants suggest that perhaps graduating residents might benefit from instruction on skills needed for contract negotiation or on leadership skills related to conflict management and change of practice. Linking new physicians with colleagues and community members who support abortion provision may also be beneficial.

Finally, residency programs may want to adopt values clarification curricula. Such curricula would give residents the opportunity to identify their moral boundaries around the care they deliver, to articulate their beliefs regarding abortion and to learn about others’ beliefs. In doing so, they would help physicians develop tolerance for practice diversity, as well as a nuanced understanding of distinctions between personal beliefs and professional obligations. Use of values clarification curricula could slowly change the culture of obstetrics and gynecology practice.18,19

Limitations

Our findings are based on an in-depth study of some 30 obstetrician-gynecologists from diverse backgrounds; they cannot be used to draw broad conclusions about a larger population, and they may have been influenced by a number of factors. Although the threat of violence was not a major deterrent to provision among our study participants, the May 2009 murder of an abortion provider may affect the experiences and decisions of physicians in practice in ways that are not reflected here. The study title, Assessment of the Impact of Abortion Training on the Careers of Obstetricians and Gynecologists, may have attracted physicians who are amenable to integrating abortion into their practice.

We do not believe that we have selected for physicians who have experienced disproportionately few or many obstacles to providing abortions. However, even if physicians experiencing obstacles were overrepresented in our sample, this would not necessarily undermine the study’s findings. The research question was formulated in response to findings from a large national survey of obstetrician-gynecologists, which showed that only half of residents who intend to continue providing abortion after residency ultimately do so.1 The aim of our study was to explore why and to describe some of the obstacles that physicians experience; we sought to contextualize these obstacles, rather than to quantify their occurrence. The strength of our qualitative approach is that it allowed deep and personal explorations of a wide range of experiences and distinctions within the group.

Conclusion

The integration of abortion into medical practice can be prohibited or restricted in multiple ways, both overt and subtle. Even where abortion provision is not explicitly prohibited, new physicians may lack power to include it in their practices. The desire to avoid conflict is highly pervasive, but unsurprising, given the persistent controversy that profoundly affects physician behavior and patient care surrounding abortion.