OBJECTIVE: To examine the association between distance traveled for an abortion and site of postabortion care among low-income women.
METHODS: We conducted a retrospective cohort study using claims data from 39,747 abortions covered by California's Medicaid program in 2011-2012. Primary outcomes were the odds of abortion-related visits to an emergency department (ED) and the original abortion site, and the secondary outcome was total abortion care costs. We used mixed-effects logistic regression adjusting for patient and abortion characteristics to examine the associations between distance traveled and subsequent abortion-related care at each location.
RESULTS: Among all abortions (N539,747), 3% (95% CI 2.9-3.3, n51,232) were followed by an ED visit (3% first-trimester aspirations, 2% second trimester or later, and 4% medication abortions) and 25% (95% CI 24.1–24.9, n59,745) were followed by a visit to the original abortion site (4% first-trimester aspirations, 3% second-trimester or later, and 77% medication abortions). Women traveling farther for their abortions had higher odds of visiting an ED (100 or more miles compared with less than 25 miles, first-trimester aspirations: adjusted odds ratio [OR] 2.29, 95% CI 1.50–3.49; medication abortions: adjusted OR 2.30, 95% CI 1.34–3.93) and lower odds of returning to their abortion site for follow-up (100 or more miles compared with less than 25 miles, first-trimester aspirations: adjusted OR 0.36, 95% CI 0.18–0.70; second trimester or later: adjusted OR 0.52, 95% CI 0.31–0.88; and medication abortions: adjusted OR 0.33, 95% CI 0.23–0.50). Costs were consistently higher when subsequent care occurred at an ED rather than the abortion site (median cost $941 compared with $536, P,.001).
CONCLUSION: For most patients, greater distance traveled for abortion was associated with increased likelihood of seeking subsequent care at an ED. Increasing the number of rural Medicaid abortion providers and reimbursing providers for telemedicine and alternatives to routine follow-up would likely improve continuity of care and reduce state costs by shifting the location of follow-up from EDs back to abortion providers.
Across the United States, many women travel long distances for abortion care. Nationally, in 2008, 17% of women traveled more than 50 miles for abortion care. Among all U.S. counties, 90% have no abortion provider because most are concentrated in urban areas. Long travel distances likely have consequences for careseeking after an abortion.
Like with any medical procedure, some women want subsequent care after abortion—either routine follow-up to confirm the abortion is complete or because they are concerned about symptoms. Patients who have in-clinic procedures, defined as first-trimester aspiration or second-trimester or later abortions, may or may not be advised to return for a follow-up visit. Requiring all patients to return for follow-up after first-trimester aspiration is not clinically necessary and for many, the costs may be too high. Patients having a medication abortion are routinely advised to return for follow-up to rule out continuing pregnancies. Although evidence is growing for alternatives to routine follow-up after medication abortion,in-person follow-up continues to be common.
In follow-up care, health care providers counsel patients about symptoms or perform ultra-sonograms to confirm that abortions are complete. For patients who want follow-up care, returning to the abortion provider represents successful continuity and good quality of care. However, some women seek care at a local emergency department (ED), either on the abortion provider’s advice if urgent or self-referring without notifying the abortion provider. Given that women are traveling long distances for abortion care, we sought to understand how that distance affects where women seek care after an abortion.
MATERIALS AND METHODS
We conducted a retrospective observational cohort study of abortions and subsequent health care visits among beneficiaries of the California Medicaid (Medi-Cal) program. Medi-Cal data present a unique opportunity because they can be used to track all subsequent care among a cohort of patients having an abortion longitudinally after their abortions with no loss to follow-up. We obtained data on all abortions covered by the fee-for-service California Medi-Cal program in 2011 and 2012 from the California Department of Health Care Services. The study was approved by the institutional review boards of the University of California, San Francisco and the California Health and Human Services Agency.
California Medicaid is administered on a fee-for-service or managed care arrangement with roughly equal numbers of women enrolled across the two at the time of the study. Women may obtain California Medicaid coverage when pregnant with the Medi-Cal Access Program as well as “presumptive eligibility” and may obtain abortions under these categories of coverage. Although both the Medi-Cal Access Program and presumptive eligibility fee-for-service programs are fee-for-service programs, only the traditional fee-for-service billing records contain complete information for care provided to a particular beneficiary; therefore, we received data only for beneficiaries with traditional fee-for-service coverage. We received claims that represent approximately one fourth of all California Medicaid-covered abortions in 2011–2012.
The primary outcomes were the odds of abortion-related visits to an ED and the original abortion site. The methods used to identify abortions and episodes of follow-up care were based on our previous work. The secondary outcome was total abortion care costs. All reimbursed costs for each billed item on the day(s) of the abortion procedure and at subsequent abortion-related follow-up visits (minus the costs of any contraception) were summed to get the total cost for abortion and subsequent related care.
The primary independent variable was travel distance, calculated using methods described elsewhere,10 based on distance categories in existing literature. Of all abortions, 89% had a full and valid address available for both the beneficiary and health care provider, thus allowing us to calculate distance traveled. Of those missing distance traveled (n54,316), the vast majority (98%) were younger than 21 years of age and covered by a specific minor consent program; the California Medicaid program suppresses data for these participants.
We examined the characteristics of the sample and estimated the proportion of abortions with follow-up visits by distance traveled, age, race, procedure type (medication abortion, first-trimester aspiration, or second-trimester or later abortion), and source of care. For the subset who had subsequent care, we calculated the number of days between the abortion and first abortion-related subsequent care at EDs and the original abortion site by abortion type. We did not calculate the number of days to the first subsequent care visit at the original abortion site for medication abortion because we could not ascertain the date of routine follow-up visits when bundled codes were used. We also calculated median costs reimbursed for abortion and all follow-up care by source of follow-up care, procedure type, and distance traveled.
We then built three multivariable mixed-effects logistic regression models for each procedure type to examine the factors associated with the following outcomes: abortion-related follow-up care at EDs, abortion-related follow-up care at the original abortion site, and abortion-related follow-up care at any location. The models controlled for sociodemographic and other potential confounders. We accounted for clustering afforded by multiple abortions by the same woman using random-effects specifications. The abortion was the unit of analysis. Statistical significance was set at P,.05 for all comparisons and adjusted odds ratios (ORs).
RESULTS
The data set contained 39,747 abortions obtained by 36,720 beneficiaries. (Table 1) Among all abortions, 3% (95% CI 2.9–3.3, n51,232) were followed by an ED visit, many of which involved no treatment—3%(n5630) after first-trimester aspirations, 2% (n5136) after second-trimester or later procedures, and 4%(n5466) after medication abortions. Subsequent visits to EDs were significantly higher after medication abortions than both first-trimester aspirations and second-trimester or later abortions (both P,.001, not shown). Subsequent visits to EDs were also significantly higher after first-trimester aspirations than second-trimester or later procedures (P5.003). Among all abortions, 25% (95% CI 24.1–24.9, n59,745) were followed by a visit to the original abortion site—4% (n5783) after first-trimester aspirations, 3% (n5204) after second-trimester or later procedures, and 77% (n58,758) after medication abortions. In total, 29% (n511,542) of abortions were followed by an abortion-related follow-up visit at any location within 6 weeks of the abortion—9% (n5569) among first-trimester aspirations, 9% (n51,959) among second-trimester or later procedures, and 80%(n59,014) among medication abortions (not shown).
Rates of subsequent care and abortion-related costs differed by abortion type as well as distance traveled (Fig. 1; Appendix 2 detailing median costs by abortion type and distance traveled). Median costs were generally lowest for women traveling less than 25 miles and for women missing information on distance. Median costs were higher when subsequent care occurred at an ED compared with care occurring at the abortion site ($941 compared with $536, P,.001). This was consistent for all distances and abortion types. Median costs were highest for women obtaining second-trimester or later abortions.
Among all abortion types, women obtaining follow-up care in an ED first sought that care an average of 10.4 days after the abortion with a median of 5 days after the abortion (Fig. 2). Among women who sought care at EDs (n51,229), 12% (n5150) of women visited the ED on the day of the abortion, 10%(n5128) visited the day after the abortion, and the remaining 78% (n5951) sought ED care 2 days or more after.
Women who sought subsequent abortion-related care after first-trimester aspiration and second-trimester or later procedures did so on average 12.9 days after abortion with a median of 11 days after the abortion (Fig. 3).
In multivariable models for women obtaining first-trimester aspiration abortions, greater distance was associated with ED visits (Table 2) with those who traveled 25–49 miles (adjusted OR 1.52, 95%CI 1.16–1.98), 50–99 miles (adjusted OR 1.73, 95%CI 1.26–2.39), or 100 or more miles (adjusted OR 2.29, 95% CI 1.50–3.49) significantly more likely to have an ED visit after the abortion than women traveling less than 25 miles. Conversely, women who traveled 25–49 miles (adjusted OR 0.75, 95% CI 0.57–0.99), 50–99 miles (adjusted OR 0.60, 95% CI 0.41–0.88), or 100 or more miles (adjusted OR 0.36, 95% CI 0.18–0.70) were significantly less likely to have a follow-up visit at the original abortion site than women traveling less than 25 miles. Among women obtaining first-trimester aspiration abortions, Hispanic or Latina and Asian ethnicities were associated with reduced likelihood of any follow-up visit at an ED, the original abortion site, or other source of care. Women obtaining first-trimester aspiration abortions at a hospital had higher odds of having an ED visit, a follow-up at the abortion provider, or any other source of care compared with women who had their abortions at an outpatient clinic. Women who had their first-trimester aspiration abortion at a physician’s office had lower odds of having a follow-up visit at the original abortion site compared with women who had their abortions at an outpatient clinic.
Distance was not significantly associated with ED visits among women having second-trimester or later procedures (Table 3). However, women traveling 25–49 miles (adjusted OR 0.54, 95% CI 0.33–0.89), 50–99 miles (adjusted OR 0.42, 95% CI 0.24–0.74), or 100 or more miles (adjusted OR 0.52, 95% CI 0.31–0.88) for a second-trimester or later abortion were significantly less likely to return to the site of their abortion for follow-up care than women traveling less than 25 miles.
Among women obtaining medication abortions, greater distance traveled for an abortion was significantly associated with ED visits (Table 4). Patients having a medication abortion traveling 100 miles or more were significantly more likely to have an ED visit after the abortion than women traveling less than 25 miles (adjusted OR 2.30, 95% CI 1.34–3.93). Conversely, patients having a medication abortion travel ing 25–49 miles (adjusted OR 0.79, 95% CI 0.65–0.96), 50–99 miles (adjusted OR 0.42, 95% CI 0.32–0.55), or 100 miles or more (adjusted OR 0.33, 95%CI 0.23–0.50) had lower odds of returning to the original abortion site than women traveling less than 25 miles. Additionally, Asian women had significantly lower odds of an ED visit after a medication abortion and significantly higher odds of returning to the abortion site compared with white women.
DISCUSSION
This study finds that traveling greater distances for an abortion is associated with an increased likelihood of seeking subsequent care at an ED and a reduced likelihood of seeking subsequent care at the original abortion site. This finding is intuitive in that when symptoms arise, traveling back to the abortion provider may be too burdensome. These findings suggest that women living in areas with no abortion provider may be compelled to use EDs for follow-up care.
Previous research among a similar population found that 1 in 38 women visit an ED within 6 weeks after an abortion for abortion-related care. However, in the majority of these visits (67%), the patient was observed and sent home without a diagnosis or treatment. Patients in rural areas and patients using Medicaid or Medicare have been found to have more ED visits both generally and for nonemergency care. Together, these studies suggest that distance to an abortion provider is an important determinant of ED use. Women may visit EDs for follow-up care, but they would more appropriately be seen and counseled at the original abortion site or, perhaps, by phone.
When a patient is concerned about a problem, she would ideally be seen by a health care provider who is trained in abortion care and is most familiar with the expected postabortion symptoms. Compared with ED staff, abortion providers are typically better equipped to evaluate abortion patients’ symptoms and avoid unnecessary use of additional interventions such as antibiotics or reaspiration. Returning to the abortion provider may also avoid poor or delayed treatment from hospitals with policies that restrict or prohibit abortion,which may exacerbate abortion stigma.18 Additionally, as was found here, ED costs are higher than other sources of care.
Strategies to reduce ED visits include improving health care provider–patient communication and materials about nonurgent postabortion side effects. Health care providers may also emphasize the safety of abortion and differentiate signs and symptoms of common side effects from serious complications.
As of March 2016, the label for medication abortion no longer recommends a follow-up visit. New technologies such as low-sensitivity or semiquantitative urine pregnancy tests, phone calls, or internet communication with women and other alternatives to routine follow-up are being tested. These technologies most benefit patients who cannot return for follow-up. This study supports fully reimbursing abortion providers for all forms of follow-up care after medication abortion. Encouraging these alternative forms of follow-up care will likely result in cost-savings because women may be less likely to seek such care at EDs.
A strength of the study is that it used data from a complete system with virtually no loss to follow-up, and thus we can reasonably expect that all follow-up visits, regardless of source of care, were captured. However, this study has limitations. First, generalizability may be limited because it relies on data from only one state, from a population of Medicaid enrollees, on a specific fee-for-service plan. Women in western states travel shorter distances to reach an abortion provider than women in the Midwest and South; in more rural states visits to EDs after an abortion may be even more common. Nevertheless, we expect that the association between long distances and ED visits applies to other states and to women outside Medicaid programs, because distance makes it harder to return to the original abortion site regardless of insurance type. Second, data on distance were more likely to be missing for younger women, and this may have affected the precision of the effect sizes of distance among the nonmissing groups. Third, we did not assess adverse events, and thus we do not know the extent that women obtaining subsequent care were experiencing adverse events compared with seeking routine care, advice, or reassurance. Having detailed clinical data would have led to more precise estimates.
Increasing the number of abortion providers accepting Medicaid, particularly in underserved and rural areas,could help reduce state costs by shifting the location of follow-up from EDs back to abortion providers. Strategies to mitigate the urban concentration of Medicaid-accepting abortion providers include reimbursing the use of telemedicine, ensuring that abortion training is part of medical education, integrating abortion provision and postabortion care into primary care settings and residency training, expanding the types of health care providers who can offer abortion (such as nurse practitioners), and increasing reimbursement rates to incentivize additional existing providers to accept California Medicaid.
These results fill a gap in the literature, helping to explain why some EDs may see women after an abortion. Visits to EDs after abortion do not necessarily represent emergencies and may represent routine follow-up care.