Abortion has been legal in the United States since 1973, yet some public policies impede abortion access for many women. The Hyde Amendment, which Congress first passed in 1976 and renews annually as part of the appropriations process, prohibits federal Medicaid funding for abortion except when a pregnancy resulted from rape or incest or endangers the woman’s life.* The “rape, incest and life endangerment” exceptions have been omitted, reinstated, redefined and debated throughout the history of the amendment;1 each modification has simultaneously reflected and reinforced deep cultural ambiguities about the definitions of these circumstances. Much remains unclear (and hence contestable) about how health care providers and insurance companies should best identify, document or confirm these cases.

As of January 2010, some 32 states and the District of Columbia ban the use of state Medicaid funding for abortion except in cases of rape, incest and life endangerment. South Dakota covers abortion in the case of life endangerment only, in violation of federal law. The remaining 17 states provide Medicaid funding for all or most “medically necessary” abortions, primarily as a result of court orders.2

Some evidence suggests that providers do not receive Medicaid reimbursement for abortions that should qualify for funding.3–5 The number of states in which no abortions were publicly funded in the past year increased from 13 in 20013 to 20 in 2006.4 In that same year, 92,455 rapes were reported to police.6 As an estimated 5% of rapes result in pregnancy,7 at least 4,623 rape-related pregnancies may have occurred in 2006. At least some of the women involved likely were eligible for Medicaid reimbursement, assuming that women who have been raped are about as likely to terminate their pregnancies as are women in general.

We investigated three questions: What are providers’ experiences with Medicaid reimbursement for abortion in cases of rape, incest and life endangerment? What is the process for applying for Medicaid reimbursement? What factors facilitate or hinder reimbursement?

METHODS

We attempted to recruit representatives of abortion providers in eight states in which Medicaid funding is limited to cases of rape, incest and life endangerment: Florida, Idaho, Kansas, Kentucky, Mississippi, Pennsylvania, South Dakota and Wyoming. We selected these states because they varied in their numbers of providers and of publicly funded abortions in 2001, the most recent year for which this information was available.3 To protect respondents’ identities in states with one or only a few abortion providers, we present most results in the aggregate; we also highlight limited results from one state in a way that does not compromise anonymity.

We compiled a list of abortion providers in each selected state from listings on the Internet and from the National Abortion Federation (NAF) membership roster as of October 2007 (which NAF provided and is not public). We also asked each respondent to recommend other providers to contact.

We sent a letter inviting participation to all known providers in the selected states. Providers were eligible to participate if they responded yes to two questions: One asked whether they had had any patient in the past five years who had terminated a pregnancy that was the result of rape or incest or that endangered her life, and the other asked if they had sought Medicaid coverage for such patients in the past five years. Our initial sampling frame comprised 128 providers, of which 14 declined to participate, seven were unreachable after eight attempts and 68 were screened out as ineligible (49 because no Medicaid funding had been sought for a qualifying case in the past five years). A network of 15 providers identified one respondent who participated on its behalf but who reported only on the provider where the respondent was based.

The final sample consisted of 25 respondents representing 25 providers in six states. One respondent worked for two providers and reported on both. Two respondents reported on the same provider.

Between October 2007 and February 2008, three trained research staff members conducted 60-minute telephone interviews using a semistructured guide; interviews were digitally recorded. Respondents gave oral informed consent and received $75 for participating. Closed-ended questions asked about respondent and provider characteristics, as well as the numbers of cases eligible for and reimbursed by Medicaid in the past year and in the past five years. Open-ended questions asked about respondents’ experiences with low-income women seeking abortions in cases of rape, incest and life endangerment: Respondents were asked to detail the circumstances of these cases, the services provided, and the process and outcome of seeking Medicaid reimbursement. Each interviewer’s first two interviews were reviewed by another researcher to ensure data quality.

Interviews were transcribed verbatim and thematically coded in Atlas-ti, version 5.2. We used a thematic analysis approach, guided by our research questions. To develop the codebook, we initially drew up a set of codes based on the interview guide. Next, two research staff members coded all of the interviews, reviewed each other’s codes to ensure consistency and refined the codebook iteratively. We also summarized stories of patient cases and indicated whether cases received Medicaid reimbursement. We calculated the total number of rape, incest and life endangerment cases reported, and calculated frequencies and percentages to summarize characteristics of respondents and providers. On the basis of respondents’ self-reported information, we classified them according to their role in the provision of abortion services (executive director, administrator, physician, or clinical support staff) and the type of provider they worked for (abortion clinic, private physician’s office, nonspecialized clinic or hospital). The Western Institutional Review Board reviewed and approved the study protocol.

RESULTS

Overview

Respondents were predominantly women and had been involved in abortion provision for an average of 14 years (Table 1). Most were executive directors or administrators. The majority of providers were abortion clinics, some were nonspecialized clinics and office-based physicians, and one was a hospital-based clinic. Providers varied in abortion caseload and gestational age limits for performing abortion. Seventeen were registered as Medicaid providers at the time of the interviews; the rest had stopped seeking reimbursement for Medicaid patients within the past five years. The median costs were $450 for a medication abortion, $425 for a first-trimester surgical abortion and $900 for a second-trimester abortion, but costs varied considerably. Respondents reported that the majority of patients pay for abortions out of pocket.

We asked respondents to report the total number of abortions their facility had performed in the previous year that should have qualified for Medicaid reimbursement. Twenty respondents reported the total number, and five reported numbers for only part of the year.

Of the 245 reported abortions that should have qualified for Medicaid reimbursement, 143 were not reimbursed. Of the 102 that were reimbursed, 99 were in one state; within that state, 27 qualifying abortions were not reimbursed. Eighteen respondents reported that no qualifying abortions were reimbursed.

Barriers to Reimbursement

•Administrative burden. Most respondents said that trying to obtain Medicaid reimbursement entailed extensive paperwork, long delays and excessive staff time. Some described spending hours or days trying to reach Medicaid staff who could aid them in submitting their claims. Overall, respondents expressed a sense of futility about the process of seeking reimbursement, as the following comment suggests:

“Essentially … you fill out the form, and you explain the circumstances, and you provide all of the information, and then … submit it and then wait for the rejection. And when they tell you what the reason for rejection is, then you resubmit it and continue to do so until it just is no longer worth pursuing, and then you quit!”—Executive director, abortion clinic

•Nonexistent or poor relationship with Medicaid staff. Respondents generally described their relationship with Medicaid staff as “nonexistent” or “poor,” and said that this contributed to the sense of administrative burden. Respondents reported not being able to reach Medicaid staff or obtain consistent or informed responses to questions. Some encountered antiabortion attitudes from claims processors. An abortion clinic administrator explained, “When you tell them that you want to verify benefits for abortion services, … they go, ‘Oh, my God, I knew one of these days I’d get a call like this.’ ” In another case, when the Medicaid claims reviewer stated that the policy did not cover abortions, a respondent who was on the clinical support staff§ of a hospital replied that in some cases it did, and the reviewer repeated it did not. “Whoever is reading it doesn’t … think this woman deserves an abortion or doesn’t believe in abortion, so they just deny the claim,” the respondent noted.

•Clinicians’ reluctance to sign forms. To file a Medicaid claim for a patient’s abortion, most respondents reported they are required to fill out forms certifying that the woman is seeking an abortion because of rape, incest or life endangerment. Specific requirements and forms varied across states, and often across providers and Medicaid offices within states. Respondents in some states reported that a patient who has been raped must either provide a police report or sign a statement certifying that she was raped but cannot provide a police report (which is typically because she did not file one). In the absence of a police report, some clinicians refuse to sign the form. The executive director of an abortion clinic explained that the doctor’s concern is not about whether women “really were raped. … What he’s concerned about is [being] accused of Medicaid fraud.”

Respondents also reported that clinicians are at times reluctant to sign forms verifying that a pregnancy is life-endangering. Some Medicaid reimbursement forms require a signature from the patient’s primary care provider or the physician providing care for the health condition making the pregnancy life-endangering; clinicians sometimes refuse to sign because they oppose abortion or they fear accusations of coercing patients or of Medicaid fraud. In one case, described by a hospital’s clinical support staff member, a patient’s primary care physician had told her that she could not use contraceptives because of her health condition, but that she could always get an abortion if necessary. When the woman went to the respondent’s clinic for a procedure, her primary care physician would not sign the necessary forms because, in the respondent’s words, he “does not believe in abortion. He told her she’d die, but wouldn’t sign the form.”

•Inconsistent HMO requirements. Respondents identified inconsistent and unclear Medicaid HMO requirements as another layer of administrative burden. For example, in one city, Medicaid operates through a number of HMOs, whose requirements for certifying that an abortion qualifies for reimbursement differ. A respondent in that city stated:

“One [Medicaid HMO] requires a police report … to pay for … pregnancies that are the result of the rape. One … requires a police report and that [the gestational age, as estimated on the basis of an ultrasound] match. One … just requires … forms that the patient and the doctor have to fill out for this state. … One requires the forms plus a letter from me stating that the patient’s been counseled.”—Clinical support staff, hospital

Each HMO in that city has different gestational age requirements for abortion and has contractual relationships with some but not all providers, sometimes creating difficult situations for women. For example, an abortion clinic administrator stated that she sees many women who have been shuffled from provider to provider, seeking a clinic that has a contractual relationship with their HMO. In a recent instance, as a client searched for a provider, her pregnancy progressed into the second trimester. Once she found a clinic that contracted with her HMO, she learned that the HMO required special authorization to cover the abortion, because she was more than 16 weeks pregnant. Although the HMO covered the cost, the delay in the abortion increased the complexity and cost of the procedure.

•Difficulty identifying rape cases. Difficulties in identifying eligible rape cases emerged as another barrier to obtaining Medicaid reimbursement. Some cases are likely missed because patients do not readily disclose rape or because interpretations of the term “rape” vary. For example, some women who were raped by an intimate partner do not immediately describe their situation that way. One respondent commented:

“It’s generally very hard for women to disclose that, but then again, I’ve seen cases where women … all of a sudden say …‘Yes, I’m a victim of domestic violence, and … I’m technically raped.’… They have a hard time saying that … and owning up to [the idea] that rape is rape is rape, you know? That they didn’t give their consent.”—Clinical support staff, abortion clinic

In addition, some providers do not systematically screen for rape histories. Sometimes, providers find out that a woman’s pregnancy resulted from rape only when police contact them, seeking to collect a specimen. Some respondents were conflicted about screening for rape and ambivalent about initiating discussions with patients about eligibility criteria for Medicaid coverage of abortion, as the comments of one respondent reflect:

“I think we as providers also struggle with … do we say to every single patient on the phone …‘Well, no, your Medicaid doesn’t cover unless these are the circumstances’? And I’m not sure why we struggle with that—if it’s just … a difficult situation to have over the phone or if we don’t want to seem like we’re putting ideas in people’s heads.”—Administrator, abortion clinic

A few respondents voiced suspicion of some patients’ reports that they were raped. One remarked:

“The stories are interestingly the same … in each and every incident. They went to a party, they got drunk, they woke up two days later, their clothes were off. … I’m not saying that that did not happen. But these patients can typically not tell you anything except they got drunk, they went to a party, they woke up two days later and their clothes were off.”—Clinical support staff, hospital

Some respondents questioned whether women’s reports of rape were valid, yet acknowledged the desperate circumstances that may lead women to report that their pregnancy resulted from rape. One said:

“There are certainly … a fair number of women who know that this is the only way they can get their abortion covered, and even though they weren’t raped … are in a desperate enough situation that they need to say whatever they need to say to get it covered.”—Administrator, abortion clinic

•Difficulty establishing life endangerment. Efforts to establish that an abortion is necessary to protect a woman’s life are often thwarted by different definitions of life endangerment among providers, Medicaid staff and patients. One respondent described the case of a woman who had jumped out a second-story window of a burning building when she was 13 weeks pregnant. She was hospitalized with a broken back and inhalation burns, and requested an abortion. The hospital did not do elective abortions, so a resident called the respondent’s facility. The respondent related:

“Our physician called the [Medicaid] director and described the patient’s condition, and asked him if Medicaid would cover it in a case of threat to the mother’s life. … Because she was immobile, there was a greater risk for pulmonary embolism [from the pregnancy]. The [Medicaid] director said, ‘She can still continue this pregnancy.’… So, they denied that they would pay. We did the abortion for no payment. We did it for free, because she had nothing.”—Executive director, abortion clinic

The administrator of a nonspecialized clinic described further discrepancies, saying that Medicaid’s interpretation means “the immediate threat of endangerment, not the long-term.” Respondents described cases of women dying of AIDS or suffering from other serious conditions, who could not get Medicaid-covered abortions. In one case, a woman awaiting a kidney transplant was denied coverage by both Medicaid and Medicare, even after the clinic made formal appeals. The clinic administrator recounted that the woman had lost one leg because of diabetes, but the public programs decided that “she could either be bedridden, or she could … use crutches, … but the pregnancy would not cause death. … [It would only] delay the transplant.”

Ever-changing diagnostic and procedural codes required on the claims forms also play a role in the different understandings of life endangerment. One respondent pointed to the dehumanizing application process and described the detrimental impact of using codes to communicate the patients’ life-endangering circumstances:

“When you bill … sometimes they may not see that the woman’s life was in danger. They may not see it that way, even though the woman’s life is in danger. You know it’s based on numbers, and codes and digits, not a human being, so that makes it difficult.”—Administrator, abortion clinic

Some respondents reported conflicts with Medicaid staff about what paperwork is necessary to establish life endangerment. An abortion clinic administrator described a case in which a Medicaid HMO insisted that the woman fill out the paperwork required when a pregnancy resulted from a rape even though the patient was seeking to terminate a life-endangering pregnancy. The administrator, recounting her discussions with Medicaid staff, related:

“We were like, ‘Are we speaking the same language?’… And they’re like, ‘But that’s part of the form.’ And we’re like, ‘You’re not listening to us.’”

•Financial issues. Respondents reported varying reimbursement rates, and some stated that rates were so low, they would not be able to stay in business if they continued to bill Medicaid. Some providers received a flat reimbursement regardless of the gestational age at abortion, even though the cost of abortion increases as the pregnancy progresses. For the most part, respondents reported that if they were reimbursed at all, the level was low relative to the cost of providing the abortion, and that once they accounted for the staff time required, the value of the reimbursement was substantially reduced. As one respondent remarked:

“It is disheartening, working hard, maybe getting paid. When I do, it’s … less than half of what we would normally charge cash. And I still have to pay my billing company 10% of the net, or 10% of whatever we get.”—Administrator, abortion clinic

This respondent reported that Medicaid reimbursed $212 for an abortion that cost $420. Another, the executive director of an abortion clinic, calculated that given the administrative cost of pursuing reimbursement, “It took me 42 days to get $1.22.”

Respondents reported that because of low reimbursement rates and the expense of staff time to pursue Medicaid dollars, many providers avoid working with the system by offering discounts, providing loans or absorbing the costs of abortions themselves. Some respondents said that they absorb between $1,000 and $60,000 annually in free or reduced-cost services. A physician at an abortion clinic said, “We would find the money for anybody no matter who it was … because that’s just our philosophy.” On the other hand, a private-practice physician said, “If it’s not an emergency, we don’t have to treat them. We can send them to other places.”

All but two respondents reported relying on abortion funds—nonprofit groups that provide grants or loans to help women pay for an abortion.8 Most respondents described working with the funds as positive and essential. “I think the fund does a lot of work for Medicaid patients. They really help out a lot of women,” said an abortion clinic administrator. “I don’t know what [these patients] would do without the fund, to be honest with you.”

Although many respondents regarded abortion funds as a dependable source of coverage, they were aware that those funds cannot cover all women. They commented that if Medicaid reimbursed all qualifying cases, abortion funds would be able to cover more women who are not eligible for Medicaid—for example, women whose pregnancies are approaching maximum gestations for having an abortion. As the administrator of a nonspecialized clinic remarked, “They do try to help. But … they’re limited.”

Giving Up on Medicaid

Some respondents no longer contracted with Medicaid, or worked with Medicaid as little as possible, because of the numerous administrative and systematic barriers they experienced. One related:

“We cannot get a Medicaid referral because we are not a Medicaid provider, but we are not a Medicaid provider because they seldom ever pay for abortions.”—Administrator, abortion clinic

[Some] states … make it so difficult for providers to obtain reimbursement that a de facto ban is in place.

Some saw the administrative burdens as related to a generally “conservative” or “antichoice” climate in their state, and suggested that states purposely make it so difficult for providers to obtain reimbursement that a de facto ban is in place. The executive director of an abortion clinic said the Medicaid office had created enough of a “hassle” for abortion providers in the state that the clinic had stopped working with the office, and said that the office no longer has any worries “because nobody’s applying for [Medicaid reimbursement of abortion].”

One respondent reported that even after Medicaid’s refusal to pay was successfully challenged in court, obtaining reimbursement in his state remained a struggle. In the year following the court order, Medicaid began to reimburse claims, but increased the “hassle level” with longer delays and more complex requirements. Several providers reported experiences like the following:

“There was only about a period of a year that we could get any money out of Medicaid at all, and it was a major struggle. … They just made it so difficult and so impossible that you get the things submitted, and three or four months would go by, and they’d say, ‘Oh we’re still working on it.’… And we finally said, ‘Ah, forget it.’ ”—Executive director, abortion clinic

Other concerns led some providers to withdraw from the Medicaid program. A respondent from the state whose providers had the greatest success obtaining reimbursements reported that the clinic no longer accepted Medicaid because “the state started conducting raids on abortion files.” The respondent said:

“We had no idea what was going to ever happen with those charts [that Medicaid took]. We had no idea about patient confidentiality, … and they just kept telling us, ‘There’s nothing you can do. If you agree to participate in state programs, then we can do this at any time we feel like.’”—Executive director, abortion clinic

When the System Sometimes Worked

Some respondents reported that with persistence, they could obtain Medicaid reimbursement. Respondents from the state in which almost all of the successful reimbursements occurred (where obtaining reimbursement reportedly had grown easier over time) attributed that success to an intervention begun in 2001. The intervention simplified the forms required to submit a Medicaid claim. It also educated providers and Medicaid HMO staff about women’s rights to obtain Medicaid reimbursement for abortion in certain circumstances, and about providers’ and insurers’ obligations to seek reimbursement. And it educated women about their rights to seek Medicaid reimbursement and about the circumstances under which they would be eligible to do so. One respondent reported that the intervention led to clearer and more streamlined protocols and forms, and an improved understanding of patients’ rights. Some reported increased collaboration among providers in the state following the intervention.

“I remember getting one payment from them and feeling like I should frame it, ’cause it was just such a miracle.”

Another respondent in that state, an administrator from a nonspecialized clinic, described staff members’ longevity and increasing assertiveness as key to improvements: Many staff members had been with the provider for 8–25 years, had learned the Medicaid reimbursement process, and had become stronger and more confident. As the administrator expressed it, “We’ve gotten better also at saying ‘No, you have to approve this!’ and ‘Yes, you have to pay us!’… and ‘No, it’s not for you to question, because the law doesn’t say you have to question it.’”

Success also depended on knowledgeable staff at Medicaid. One respondent described the importance of a helpful encounter in the past and the cost and impact of Medicaid staff turnover, recalling a time when a woman in the Medicaid office “talked me through the entire … form … and what I’d need to put in each box in order to get the claim paid.” Unfortunately, the respondent continued:

“She’s no longer there, and when we try and get someone like that to help us, it doesn’t work. So, that was back in the late 90s, and I remember getting one payment from them and feeling like I should frame it, ’cause it was just such a miracle.”—Executive director, abortion clinic

Experiences of successful reimbursement were not uniform within this state. Although some respondents reported a fairly systematic process for obtaining reimbursement, others reported that no eligible cases in the previous year were reimbursed. One who felt that the system was still not working well explained that many Medicaid HMO personnel did not know how to handle these cases:

“We called the state. We called [the city], we called the capital, and said, ‘Help us.’ And nobody—nobody—even knew anything. It wasn’t like they were refusing to help, but they just had no idea how this … would work.”—Administrator, abortion clinic

The few respondents outside of this state who reported having received reimbursement had taken unusual measures to get it. In one state, Medicaid owed a provider $10,000 and had not paid for a year. The respondent, the executive director of an abortion clinic, worked with a state legislator to put pressure on the state Medicaid office and was later reimbursed. The respondent commented, “It literally took almost an act of Congress to get that to happen.” Others who reported success had a larger volume of potentially eligible cases, which helped to establish a relationship with the Medicaid staff handling their claims. Similarly, longtime providers frequently reported that in 1973–1976, when federal Medicaid covered all abortions for all program participants, they had more contact with the Medicaid office and were able to develop a relationship with the office and a streamlined procedure. Now, with eligible abortions strictly limited, most providers have minimal interaction with Medicaid offices.

Respondents’ Overall Assessments

Twenty respondents reported that the system does not meet women’s needs, three had mixed assessments, one believed that the system is satisfactory and one did not provide this information. Respondents who reported that the current system is inadequate emphasized that Medicaid should provide reimbursement for low-income women’s abortions in all circumstances and that “a total change of attitude” among Medicaid administrators is needed. One put it this way:

“Medicaid should be expanded to cover any woman who wants to have an abortion in the state. … I don’t think it should be limited to rape, incest, life of the mother, and I don’t think that providers should have to jump through hoops in order to get reimbursement, and the reimbursement needs to be equal to what the doctors are doing.”—Physician, abortion clinic

Two respondents emphasized the need to simplify the system by designating a staff member or office specializing in abortion within the Medicaid office to process claims, and one highlighted that providers’ working with a nonprofit abortion fund would better meet women’s needs.

DISCUSSION

The rape, incest, and life endangerment exceptions to the Hyde Amendment prohibitions are intended to accommodate a limited set of situations in which women are eligible for federal Medicaid coverage of abortion. Our six-state study suggests that with a few exceptions, this policy is not being applied or is implemented arbitrarily. Applying for reimbursement takes excessive staff time and is hampered by administrative burdens and antagonistic or ill-informed Medicaid staff, and reimbursement levels are low. Respondents overwhelmingly reported that in their view, the current Medicaid system does not meet the needs of poor women in these circumstances.

Medicaid’s failure to reimburse for qualifying abortions pushes financial responsibility for these procedures onto others. Providers themselves shoulder much of the cost through reduced fees, sometimes absorbing costs entirely. Most providers in our study also turn for financial help to abortion funds, which emerged as critical in the struggle to compensate for Medicaid’s failure to meet its obligation to pay for qualifying abortions.

These burdens have led many abortion providers to stop seeking Medicaid reimbursement for qualifying cases or to stop accepting Medicaid patients altogether. Notably, 49 of the 128 providers we invited to participate in our study were ineligible for this reason. Nationwide, the proportion of physicians not accepting new Medicaid patients increased from 19% in 1996–1997 to 21% in 2004–2005, largely because of long delays in reimbursement.9,10 Our findings are consistent with this national trend, although abortion providers may face heavier bureaucratic requirements, stronger political opposition and greater stigma than other health service providers.

Some respondents, mostly in one state, reported successful reimbursement experiences. Key elements of success included low staff turnover, a statewide intervention and strong personal relationships with staff in the Medicaid office. On the other hand, working with a dedicated Medicaid staff member can be a hindrance if that person does not support abortion and introduces obstacles to reimbursement. Moreover, even within this state, not all providers are successful at obtaining reimbursement, and many reported significant obstacles.

The policy implications of this study are twofold. In the short term, we call for interventions to ensure that Medicaid reimburses qualifying cases in accordance with federal law. At a minimum, such interventions would include streamlining forms and administrative processes, reducing structural inefficiencies in the Medicaid system, educating providers and Medicaid staff about cases that qualify for reimbursement, screening regularly for rape, reducing the burdensome and complex requirements for proving instances of rape, and holding Medicaid legally accountable for funding qualifying cases.

In the longer term, we argue that abortion should be a broadly covered service under Medicaid. Policies that fund abortions only in certain circumstances impose a heavy administrative burden on providers and women, to the detriment of women’s health. An estimated 18–37% of Medicaid-eligible women who would have an abortion if funding were available instead carry their pregnancies to term.11–17 Low-income women may experience delays of up to three weeks in obtaining an abortion12,18 as they seek to amass the money needed for the procedure; sometimes they become stuck in a cycle of gathering money only to find that the cost of an abortion has increased because the pregnancy has progressed.5 Abortion is a safe and common component of women’s health care, and the earlier abortions are, the less expensive, less complicated and more widely available they are.19–21 Public policies that facilitate access to early abortion best serve the needs of women who have decided to terminate a pregnancy.1,22

Our findings should be interpreted in light of the following limitations. First, the results are not generalizable beyond the six states in which our respondents worked. While we approached all abortion providers we could find in each state, we did not reach all providers. Second, the effects of nonparticipation bias are not known, and the experiences of hospital-based providers and private physicians’ offices are likely underrepresented. However, these providers account for a small proportion of abortion services in the United States,23 and characteristics of practices participating in our sample mirrored those in a 2005 national study.22 Third, although some providers in our sample had stopped filing Medicaid claims, we screened out those who had not filed a Medicaid claim in the previous five years. Though some of our respondents reported minimal contact with Medicaid, the results do not reflect the experiences of providers who had stopped filing Medicaid claims more than five years ago or who had never accepted Medicaid. Finally, estimates of the number of rape, incest and life endangerment cases were based on respondents’ reports and may not accurately reflect the number of cases qualifying for reimbursement. In addition, the number of eligible cases likely was underestimated because women may have underreported rapes, providers inconsistently screened for sexual assault and some providers did not screen for Medicaid eligibility or did not work with Medicaid. However, because of the extreme nature of qualifying cases, their low incidence and the rarity with which Medicaid reimbursed for them, we believe that respondents are likely to remember such events clearly.

Further research is needed to document Medicaid reimbursement experiences in other states, to estimate the number of eligible claims that are denied, to evaluate state-level interventions intended to expand Medicaid coverage for abortion and to gauge the number of providers who have stopped trying to obtain Medicaid reimbursement for abortion. Perhaps most important would be operations research with abortion providers and Medicaid offices to design interventions to ensure public funding of qualifying cases.