Abortion Providers' Experiences with Medicaid Abortion Coverage Policies: A Qualitative Multistate Study

Abstract

Objective

To evaluate the implementation of state Medicaid abortion policies and the impact of these policies on abortion clients and abortion providers.

Data Source

From 2007 to 2010, in-depth interviews were conducted with representatives of 70 abortion-providing facilities in 15 states.

Study Design

In-depth interviews focused on abortion providers' perceptions regarding Medicaid and their experiences working with Medicaid and securing reimbursement in cases that should receive federal funding: rape, incest, and life endangerment.

Data Extraction

Data were transcribed verbatim before being coded.

Principal Findings

In two study states, abortion providers reported that 97 percent of submitted claims for qualifying cases were funded. Success receiving reimbursement was attributed to streamlined electronic billing procedures, timely claims processing, and responsive Medicaid staff. Abortion providers in the other 13 states reported reimbursement for 36 percent of qualifying cases. Providers reported difficulties obtaining reimbursement due to unclear rejections of qualifying claims, complex billing procedures, lack of knowledgeable Medicaid staff with whom billing problems could be discussed, and low and slow reimbursement rates.

Conclusions

Poor state-level implementation of Medicaid coverage of abortion policies creates barriers for women seeking abortion. Efforts to ensure policies are implemented appropriately would improve women's health.

Medicaid, a jointly funded federal-state insurance program, was designed to provide comprehensive coverage of health care for low-income U.S. residents. Women make up three-quarters of the adult Medicaid population, and in 2005, almost 7.5 million reproductive-aged women were enrolled in the program (Guttmacher Institute 2007). Although women insured by Medicaid make up nearly one-third of annual abortion clients (Jones, Finer, and Singh 2010), federal and state policies limit their ability to use their insurance for abortion.

The Hyde Amendment, which prohibits the use of federal funds for abortion, was passed in 1976 and has been re-approved every year since. Currently, federal funds can only be used to cover abortion when the pregnancy is a result of rape or incest, or threatens the life of the woman (hereafter referred to as Hyde-qualifying cases). Thirty-two states and the District of Columbia follow the federal example and restrict the use of Medicaid funds to Hyde-qualifying cases. South Dakota, in direct violation of federal law, only covers abortion when a woman's life is endangered. Seventeen states use their own funds to cover all or most abortions (Guttmacher Institute 2011a).

Little is known about how Medicaid abortion coverage policies are realized in practice or influence service delivery. We compared abortion providers' experiences securing Medicaid reimbursement in states with policies that limit abortion coverage to Hyde-qualifying cases to states with policies indicating broad coverage of abortion. We then investigated how the daily workings of these policies impacted abortion providers and women.

Materials and Methods

We recruited abortion providers in 10 states that restrict Medicaid funding for abortion to Hyde-qualifying cases (Florida, Idaho, Iowa, Kansas, Kentucky, Maine, Pennsylvania, Rhode Island, Wisconsin, and Wyoming) and in five states where policy indicates that Medicaid funding should be broadly available for abortion in all or most cases (Arizona, Illinois, Maryland, New York, and Oregon). We purposively selected states for geographic diversity, and variation in the number of facilities that provide abortion and the number of publicly funded abortions reported in each state. Within states, we purposively selected for diversity in facility size and type (abortion clinic, hospital, nonspecialized clinic, or private physician's office).

We mailed introductory letters to all known facilities in selected study states and followed up with a phone invitation to participate in an in-depth telephone interview. Because of the sample's diversity of facility structures, we asked front-line staff to direct us to the person who could best answer questions about Medicaid coverage of abortion. We then screened that individual for eligibility and interest in participation. We stopped recruitment within individual states once we reached saturation of state-level themes and stopped recruitment overall once this occurred across study states.

Initially, participants were eligible for interview if they reported they had sought Medicaid coverage of a Hyde-qualifying abortion in the last 5 years. After recruiting the first third of study participants and finding that many were ineligible because they did not work with Medicaid, we removed the requirement to collect information on why facilities no longer contracted with Medicaid. Participants were then eligible if they reported they had experience working with at least one woman seeking a Hyde-qualifying abortion. In all, the study team attempted to contact 225 potential participants by phone. Of these, 16 percent were unreachable, 18 percent declined to participate, 36 percent were ineligible,1 and 30 percent were eligible and interested in participating.

Five interviewers, trained in qualitative data collection techniques, conducted all interviews. Interviews were semi-structured and primarily consisted of open-ended questions about participants' experiences seeking Medicaid coverage of abortion in Hyde-qualifying cases. Respondents were asked to describe the circumstances of these cases, the services provided, and the process and outcome of seeking Medicaid coverage. We also asked participants to estimate the number of claims their facility submitted in the previous year for Hyde-qualifying cases and the number of those cases ultimately reimbursed by Medicaid.

Interviews were conducted between October 2007 and March 2010. During this time, there were no changes to the Hyde Amendment or to study states' policies regarding Medicaid coverage of abortion.

Data were approached using framework analysis, a method well suited for applied qualitative research (Ritchie and Spencer 2002). All interviews were recorded, transcribed, and coded in ATLAS.ti version 5.2 or 5.5 (Scientific Software Development, Berlin, Germany). Codes were initially generated from research questions. Revisions to the codebook were made as new themes emerged. Research team members reviewed each other's coding to ensure inter-coder reliability. We then summarized individual and combinations of codes and identified patterns within and across codes, extracting illustrative quotes pertaining to identified themes. Microsoft Excel 2007 was used to develop basic descriptive statistics regarding demographics, participant and facility characteristics, and responses to close-ended questions.

The Western Institutional Review Board approved all procedures. Participants gave oral consent prior to participation and received $75 remuneration. To protect the identities of abortion providers and facilities, we present quotes and results without identifying the states in which they are located, although we do provide the type of facility in which the participant worked and his or her self-identified role at the facility (administrator, clinical support staff, counselor, financial manager, multiple roles, or physician).

Results

Facility and Participant Characteristics

We interviewed 68 participants representing 70 facilities. Two participants worked for two facilities and reported on both. One participant worked for three facilities and reported on all of them. In three cases, two participants worked for one facility and were interviewed separately because the participants felt they could only answer a portion of the interview questions.

Participants working in 15 states were interviewed; an average of five interviews was conducted in each state (range 1–9). On average, the facilities represented 32 percent (range 2–66 percent) of all known abortion facilities in each state and provided an average of 51 percent (range 4–98 percent) of annual abortions in each state (Table 1). Of the 70 facilities, most (74 percent) primarily or exclusively provided abortion services. The majority (72 percent) of facilities provided between 400 and 3,000 abortions annually. Participants had an average of 14 years of experience in abortion provision and most held administrative roles.

Table 1.

Facility and Participant Characteristics

Facility characteristics
Percentage of state abortion-providing facilities represented in the sample, mean (range)	32 (2–66)
Type of facility, n (%)
Abortion clinic	52 (74)
Private physician's office	4 (6)
Nonspecialized clinic	9 (13)
Hospital	5 (7)
Number of abortions provided, past year, n (%)
<400	8 (11)
400–999	17 (24)
1,000–2,000	18 (26)
2,001–3,000	15 (22)
>3,000	7 (10)
Refuse or do not know	5 (7)
Percentage of states' annual abortions that participants reported they provide, mean (range)	51 (4–98)
Participant characteristics
Age in years, mean (range)	46 (22–82)
Position,* n (%)
Administration	49 (71)
Clinical support staff	13 (19)
Counselor	14 (20)
Financial manager	10 (14)
Physician	7 (10)
Years in abortion care, mean (range)	14 (0.5–35)

^*Participants could select more than one position.

Participants' Estimates of Success Securing Medicaid Reimbursement for Hyde-Qualifying Cases

Participants estimated that in the year prior to interview, the facilities they worked in attempted to secure Medicaid reimbursement for 1,884 abortions provided in Hyde-qualifying cases (Table 2). Participants reported receiving Medicaid reimbursement in 58 percent (n = 1,102) of qualifying cases.

Table 2.

Participants' Estimates of Abortions Provided and Covered by Medicaid in Cases of Rape, Incest, or Life Endangerment of the Woman in Year Prior to Interview

			Subsample of Five Nonrestricted States
	Total Sample	Subsample of 10 Restricted States	Policy Implemented Correctly	Policy Implemented Incorrectly
Participants' estimates of abortions provided to Medicaid clients in Hyde-qualifying cases, n	1,884	742	695	447
Participants' estimates of abortions provided to Medicaid clients in Hyde-qualifying cases covered by Medicaid, n (%)	1,102 (58)	398 (54)	671 (97)	33 (7)

Participants' success securing Medicaid reimbursement differed in states with restrictive versus broad Medicaid abortion policies. In the 10 sampled states with restricted coverage, participants reported that 54 percent (n = 398) of Hyde-qualifying abortions were reimbursed by Medicaid, compared with 62 percent (n = 704) in the five sampled states where state Medicaid coverage of abortion should be available in most cases. Further, participants in restricted states almost universally expressed considerable challenges obtaining reimbursement from Medicaid. In the five states with nominally broad coverage of abortion, two divergent patterns emerged as follows: participants in two states reported consistent success obtaining coverage (97 percent, n = 671), but participants in the other three states reported securing reimbursement in only 7 percent (n = 33) of Hyde-qualifying cases (Table 2).

As participants experienced similar challenges in the 10 states with limited abortion coverage and in the 3 states where broad coverage should be available but largely was not, we grouped the states by participants' quantitative and qualitative assessments (explored below) of how likely they were to secure Medicaid coverage in Hyde-qualifying cases. In the two states where participants reported consistent success obtaining Medicaid reimbursement, 97 percent (n = 671) of cases were reimbursed, compared with only 36 percent (n = 431) in the 13 states where participants commonly experienced barriers obtaining Medicaid reimbursement (data not shown).

Participants' Reports of Success Securing Medicaid Reimbursement for Hyde-Qualifying Cases

In two study states, participants indicated that they consistently receive Medicaid reimbursement for abortion care. These participants described themselves as “lucky” that Medicaid functioned well in their state and recognized the difficulties other states face securing Medicaid coverage of abortion.

In both states, participants had recently moved from submitting paper claims to electronic billing systems, which led to streamlined, consistent, and relatively simple claims procedures. Although switching to the electronic process proved initially challenging, participants reported that over time it improved their experience and success obtaining reimbursement. The electronic system, often described as “user friendly,” helped to eliminate billing errors that occurred when paper-based claims were submitted. One participant with multiple roles at an abortion clinic said, “When we used to do paper claims…often times there would be errors, and I'd have to resubmit a lot.… It's not usually an issue [now] since we have electronic billing.” The electronic billing system also made it easier to confirm women's eligibility and enrollment in Medicaid, regardless of whether a client had his or her insurance card; this system facilitated provision of care and helped participants feel confident that they would receive reimbursement. As one administrator at an abortion clinic said, “Most of the time we are very certain that the patient has that coverage and that they will cover the visit…. We rely on that database very heavily.” In addition, the majority of participants found that electronic billing helped ensure that they would receive timely reimbursement.

Another component of success in these two states was participants' relationships with responsive Medicaid staff that were able to provide billing support. Participants in both states said that they were able to access a Medicaid staff person or department who was experienced in abortion billing and had held that position for a number of years, providing continuity of support.

Despite the positive aspects of working with Medicaid in these two states, participants reported that Medicaid managed care organizations (MCOs) complicated the process of applying for reimbursement because of the different claims procedures for each MCO. One hospital-based clinical support staff explained, “We have all these different sub-types [MCOs] of Medicaid…that makes things very complicated for our financial people because they all have different contact people; they have different eligibility criteria.”

In addition, participants stated that reimbursement rates were consistently lower from MCOs than “straight” state Medicaid. Participants in these states said that state Medicaid programs reimbursed them an average of $403 (range $230–650) for an abortion regardless of the gestation of the abortion, but that MCO reimbursement rates were consistently lower. Few participants knew the exact reimbursement rate from the various MCOs; one participant with multiple roles at an abortion clinic reported specific rates, “We get $174 for an abortion from a Medicaid managed care program even though [state] gives us $230.”

Participants' Reports of Challenges Securing Medicaid Reimbursement for Hyde-Qualifying Cases

Participants in all 10 states with restrictions on Medicaid coverage of abortion, and in the 3 states where state funding should be broadly available but largely is not, reported difficulties obtaining reimbursement in Hyde-qualifying cases.

The biggest challenge that participants in these states faced was consistent rejection of claims that they believed qualified for coverage. Applying for Medicaid reimbursement for abortion was described as “futile,” “a big runaround,” “a huge rigmarole,” or “a big fat circle of confusion.” Most participants in these states said that they had never received reimbursement for Hyde-qualifying cases and did not know of any abortion providers who ever had. One participant with multiple roles at an abortion clinic described her experience being denied reimbursement, “We jump through every hoop they've asked us to jump through; I don't remember ever receiving a payment from them [Medicaid] for these procedures.” Likewise, an administrator at an abortion clinic explained that obtaining reimbursement from Medicaid was not part of her institution's history: “The woman who trained me has been in the business for 28 years and they have never been able to get assistance from public aid for any part of an abortion.”

Participants whose claims were rejected were generally unsure of Medicaid's reason for denying reimbursement. The lack of information about rejected claims prevented many participants from pursuing previously denied claims or correcting future claims. One abortion clinic administrator described trying to find out why a claim was rejected three times:

We took copious notes like you would in a science lab. This was the one factor that was different. It was done on this date. It was turned in on this date. It was rejected on this date. Next try. [We did this] to see if we could…experiment to find the magic pill. We did not get reimbursed. We stopped trying.

Some participants speculated that they received rejections because Medicaid disagreed with participants' assessment of the circumstances of women's abortions. Participants reported that it is not clear how Medicaid defines rape, and that women, providers, and Medicaid often have varying definitions. Likewise, participants reported that there is no clear definition of what characterizes a threat to a woman's life or whose certification of life endangerment is needed to secure Medicaid coverage. One abortion clinic administrator described the issue this way: “We…may believe an abortion is necessary to save the life of a pregnant woman. Oftentimes, when it goes to Medicaid, they don't agree with that assessment.”

Given these challenges, many participants gave up on filing Medicaid claims. The few participants who reported continuing to apply for reimbursement described a complex, paperwork-heavy, and time-consuming billing process in which they repeatedly submit claims only to have them rejected for seemingly arbitrary or insignificant reasons.

Participants reported that they rarely seek help from Medicaid staff to resolve questions about reimbursement for a number of reasons. First, many participants reported difficulty contacting an appropriate staff person. One abortion clinic administrator described her efforts: “You always leave a phone message. You never get a real person, and then of course you never have the right department…. It's grueling.” In addition, some participants said that they have received misinformation about the availability of abortion coverage from Medicaid staff and therefore no longer reach out to Medicaid for billing support.

In the rare circumstance that participants were able to secure reimbursement for abortion after “fighting tooth and nail” for it, participants said that reimbursement rates were inadequate and turnaround time was slow. The few participants who received reimbursement for abortion reported that they received an average of $235 (range $60–498) from Medicaid regardless of the gestation of the abortion, meaning reimbursement rates were often lower than the cost of providing services. One hospital-based physician expressed the need to increase reimbursement rates: “Certainly, the procedure needs to be reimbursed in a very reasonable way…. Although they can check it off on the books, like, ‘Oh, yes, public aid pays,’ but, it's not nearly enough to make it worthwhile to do those procedures.”

Similar to the two states where participants reported successfully receiving reimbursement, the complexity of working with Medicaid was increased by the presence of MCOs. Challenges included difficulties understanding which MCOs covered abortion, under what circumstances they offered coverage, and the different processes for securing reimbursement among MCOs. Participants in these states also reported that the reimbursement rate from MCOs was consistently lower than the rate paid when directly reimbursed by the state, although the exact rate of reimbursements from MCOs was not known.

When asked why they thought it was difficult to secure adequate Medicaid coverage in Hyde-qualifying cases, participants almost unanimously reported that they felt subtle antiabortion politics pervaded Medicaid. One provider said, “About, Medicaid—I think that there's a lot of politics going on…. They're all anti-abortion.” Alternatively, a small number of participants, usually hospital-based providers, speculated that the difficulties they experienced securing reimbursement were “just the way the Medicaid system is set up,” and that their difficulties were not related to the services they provide.

Participants' Reports of the Consequences of Challenges Working with Medicaid

Participants who worked in the two states where Medicaid practices appeared consistent with state law and the majority of Hyde-qualifying cases were covered said that they experienced few, if any, service delivery challenges working with Medicaid. These participants also reported that their Medicaid-eligible clients were able to access abortion in a timely manner. However, participants working with Medicaid in the other 13 states experienced a number of challenges that had severe consequences for participants as well as their clients.

Participants who said that they experienced problems with Medicaid chose to either struggle to obtain some level of reimbursement or refused to work with Medicaid due to the “hassle level” of doing so. Some participants described feeling “extremely unenthused” about continuing to work with Medicaid, as it was becoming financially untenable to do so. In one extreme case, a participant reported that Medicaid owed the facility $90,000 for past services and that because of Medicaid's slow reimbursement process the facility had to cut staff salaries by 20 percent. Other participants, who reported “giving up” on contracting with Medicaid, said that it was easier and less time-consuming to provide services at a discounted rate, rather than work with Medicaid. These participants had to “eat the cost” of providing care that should have been covered by Medicaid. We received mixed feedback about which strategy appeared to be most cost-effective.

Participants also reported that these challenges made it difficult to ensure low-income women could obtain timely abortion care. When Medicaid coverage of abortion was inaccessible or denied, participants reported that women scrambled to find other resources to cover the cost of the procedure, which led to delays in obtaining a desired abortion or the continuation of an unwanted pregnancy. One abortion clinic counselor explained as follows:

There are certainly women who have an unwanted pregnancy, and wish to terminate, and don't have the funds to. They may, out of necessity, continue the pregnancy because they don't even have $340 dollars to do the termination at that early stage. I've certainly seen people that are as much as 20 weeks [gestation], and when we get to that point, our services are jumpin' to roughly $2,000, and if they don't have $340, they may not have the $2,000…. That might be financially impossible for the patient to get in a timely manner.

Although many participants said that women often rely on family, friends, or partners to help raise money, and use their savings or credit cards, participants also reported that some women are forced to take more drastic measures. One financial manager at an abortion clinic noted, “Women sometimes take money out of their rent, selling their food stamps for cash, and we have even had cases where a woman admitted that she had sex for cash to raise money for her abortion.”

A small number of participants reported working with women with life-endangering conditions. After being told that Medicaid would not provide abortion coverage, women were forced to delay treating their condition while they raised money for the procedure. In these few cases, most participants attempted to mitigate the challenges experienced by women by discounting the procedure or working with abortion funds.2 One participant shared the story of a woman who needed an abortion to undergo life-saving cancer treatment:

It was a first pregnancy and she had a reoccurrence of throat cancer, and had to undergo chemo, and they had to withhold the chemo because they found out she was pregnant, so she had to terminate the pregnancy in order to have chemo, in order to treat the reoccurring throat cancer…. She was only 26 years old.

The participant sought Medicaid reimbursement for this case, but she was denied because Medicaid determined the woman's life was not sufficiently endangered.

Discussion

Findings suggest that policies stipulating Medicaid coverage of abortion do not always translate into coverage of care. Abortion providers working in 13 of 15 sampled states reported experiencing considerable administrative burden when submitting Medicaid claims for abortion. Of concern, providers also reported that women in the majority of sampled states (including three states where broad abortion coverage should be available) have no, little, or extremely hard-won access to Medicaid coverage of abortion even in cases of rape, incest, and life endangerment.

Participants' reports about how lack of access to Medicaid coverage of abortion affects women are consistent with previous studies. A 2009 literature review of 38 studies examining the impact of Medicaid restrictions on abortion found that as many as 25 percent of low-income women are forced to carry pregnancies to term that they would have terminated if Medicaid coverage of abortion was available and that many other women delay their abortions by days or weeks while trying to raise money to cover the procedure. The review also found that Medicaid restrictions on abortion increase public costs for prenatal care, delivery services, and welfare, and have a potentially (although not well documented) adverse impact on child health (Henshaw et al. 2009). There is a dearth of research about the psychological or social impact of being forced to continue a pregnancy or raise money for an abortion when Medicaid coverage is delayed or denied; more research is needed to investigate women's perspectives on the impact of inaccessible Medicaid coverage.

One potential solution to identified Medicaid challenges is to involve state-level Medicaid officials in providing Medicaid staff guidance about abortion coverage policies and appropriate implementation of such policies. In previous research, we found that one state-level coalition consisting of Medicaid officials, abortion providers, legal professionals, and women's health advocates was able to identify and implement solutions to state-level Medicaid challenges. As a result of the coalition's work, abortion providers and Medicaid staff were educated about qualifying circumstances for coverage, claims procedures were simplified, and the rate of qualifying claims reimbursed increased (Dennis, Blanchard, and Córdova 2011). State-level intervention may not be successful in some states for many reasons, including lack of political will or coalition partners, among other factors. In these circumstances, federal oversight of abortion claims may be needed.

However, at a time when more restrictions on abortion access have been passed than ever before (Guttmacher Institute 2011b), abortion providers may be wary of working with state or federal officials. Moreover, previous research has documented that some abortion providers are hesitant to advocate for state- or federal-level interventions due to negative experiences working with policy makers and fear that such interventions would ultimately be short won or inspire backlash against abortion providers (Kacanek et al. 2010; Dennis, Blanchard, and Córdova 2011). Therefore, interventions not focused on state or federal oversight of claims must also be considered to ensure that current Medicaid coverage of abortion policies meets their stated goals. The experiences of participants in the two states with well-functioning Medicaid systems provide evidence of other ways that state administration of abortion coverage can be improved.

One of the most readily replicable aspects of the two successful states' systems was the use of electronic billing, which streamlined billing procedures. Indeed, many health care facilities are converting to electronic records and claims and there are considerable benefits to doing so; research has shown that compared with paper-based claims, electronic claims reduce administrative burden, are more accurate, less expensive to file, and more quickly processed (Yoo and Harner 2006; Blanchfield et al. 2010). Moving to electronic claims could potentially benefit abortion providers working with Medicaid for Hyde-qualifying abortion claims, as well as other reproductive health services. However, abortion providers would have to weigh these potential benefits against the complexity of transitioning staff to utilizing electronic claims, as well as the costs associated with converting to electronic claims; these are challenges that data suggest are more difficult for smaller or independent facilities (Resnick et al. 2009).

Access to responsive and educated Medicaid staff providing billing support facilitated reimbursement of claims in two states. Future state-level interventions should include training Medicaid staff about the availability of abortion coverage and the procedures for filing abortion claims. Given the relative rarity of Hyde-qualifying cases, trainings must incorporate regular ongoing reminders about state and federal policies regarding abortion coverage. In addition, we found in previous research that Medicaid staff do not always have access to up-to-date information about what should be covered by Medicaid (Dennis and Blanchard 2011). We therefore suggest that Medicaid officers work diligently to ensure that all materials provided to staff are current and accurately reflect state and federal policies. This will enable Medicaid staff to provide appropriate support to women seeking abortion care and abortion providers filing claims.

Despite these two states' overall success working with Medicaid, participants reported receiving low reimbursement rates for abortion, as did participants from all study states. Almost all of the participants recommended increasing the overall reimbursement rate for abortion and prorating reimbursement for termination services based on the gestation of pregnancy, as the complexity and provision cost of the procedure increase with gestation. These findings are consistent with the struggles that many health care providers face; nationwide, physicians report reluctance to work with Medicaid largely because of low reimbursement rates (Borchgrevink et al. 2008). Other scholars have noted that increasing Medicaid reimbursement rates is critical to improving access to a variety of health care services (Cohen and Spector 1996; Grabowski 2001; Intrator and Mor 2004; Yoo et al. 2010). State-specific strategies for increasing reimbursement must be developed as states establish their own Medicaid reimbursement rates, and there are no uniform procedures for ensuring the rates are adequate (Centers for Medicare & Medicaid Services 2011). Our findings suggest that this should be a priority for stakeholders working on improving abortion access, as increasing reimbursement levels may be an important incentive for abortion providers to participate in the Medicaid program.

Although some of the challenges abortion providers face when working with Medicaid are not specific to abortion care, we hypothesize that because abortion providers and the services they provide are heavily stigmatized and regulated, abortion providers face heavier bureaucratic requirements, stronger opposition, and greater scrutiny when working with Medicaid. However, more research is needed to test this hypothesis. Regardless of whether the Medicaid challenges abortion providers experience are unique to the services they provide or generally related to working with Medicaid, solutions to these challenges must be identified, tested, and shared to protect women's health.

Limitations

This qualitative study was conducted with a purposive sample of respondents; therefore, results may not be generalizable to other facilities that provide abortion in the states included or to other states where we did not conduct interviews. The limited generalizability of our findings is likely particularly true within states where a small number of abortion-providing facilities are represented. However, this article focuses on patterns that occurred across states, and not on individual state analyses.

In addition, the effects of nonparticipation bias are not known and the experiences of hospital-based providers and private physicians are underrepresented. However, a 2011 study found that hospitals and private physicians' offices provided only 5 percent of abortions provided between 2007 and 2008, whereas specialized abortion clinics provided 70 percent of procedures in those years and 24 percent of abortions occurred in nonspecialized clinics (Jones and Kooistra 2011). The same study also found that most abortion providers have annual caseloads between 1,000 and 4,999. Therefore, the types of facilities represented in our sample and the number of abortions provided annually reflect national abortion provision trends.

Another limitation of this study is that participants' estimates of the number of submitted or reimbursed Hyde-qualifying abortions may be imprecise due to recall challenges. We did not confirm the number of cases submitted or reimbursed with Medicaid claims data or with abortion providers' client files and therefore cannot verify the accuracy of participants' self-reports. We are confident that participants' reports are reasonably accurate because many participants reported that they reviewed their client files and Medicaid records prior to the interview and a small number did so in real time during the interview. In addition, because of the extreme nature of qualifying cases and the rarity of receiving Medicaid reimbursement, participants were more likely to recall those events. Next, estimates of percent of cases reimbursed and reported relationships with Medicaid were remarkably consistent across providers within individual states, suggesting that reimbursement experiences were similar among providers within a state. We did not interview Medicaid officials about their perspectives on the submission of abortion claims, or review Medicaid claims data, and believe future research in this area is necessary.

Despite these limitations, this study provides in-depth data about abortion providers' experiences under policies regarding Medicaid coverage of abortion and how the implementation of such policies affects provision of and access to care, an area lacking rigorous empirical research.

Conclusion

The Affordable Care Act will expand Medicaid coverage to nonelderly individuals with incomes up to 133 percent of the federal poverty level. With this expansion, more women will become eligible for Medicaid and more women will be affected by Medicaid abortion policies. Our findings highlight that state-level variances in how, or if, state-level policies are implemented play a critical role in access to and the provision of Medicaid-covered abortion care. Our findings also suggest that restrictions on the circumstances under which Medicaid covers abortion effectively lead to prohibitions on coverage in all cases, even those “exempted” for coverage by the Hyde Amendment. Given that current policies are not meeting their stated goals, interventions are needed to ensure that state Medicaid programs meet their obligations to cover abortion as outlined by state and federal policy.

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of this article:

Appendix SA1: Author Matrix.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.