Finer, Lawrence B. PhD; Wei, Junhow
Mifepristone, also known as RU-486, is a synthetic steroid approved by the U.S. Food and Drug Administration (FDA) in September 2000 for early medication abortion. Misoprostol, a synthetic prostaglandin analogue, often is used with mifepristone to produce uterine contractions that expel the products of conception.1 The FDA-approved regimen for mifepristone administration entails a 600-mg dose, but ongoing clinical studies performed in multiple countries suggest that alternative mifepristone–misoprostol regimens are as or more effective than the FDA-approved regimen.2–9 As early as 2001, an estimated 83% of providers were using 200 mg of mifepristone.10
The mifepristone and misoprostol sequential regimen has been shown to be an effective, safe, and acceptable method for pregnancy termination in the first trimester.6,11,12 Trials in the United States and the United Kingdom resulted in efficacy rates of 92–95%.13–18
In the period just before its approval, mifepristone was described as “the little white bombshell” owing to its potential to change the nature of abortion provision in the United States.19 Approval by the FDA raised expectations that abortion would become more widely accessible and that women would make use of an option that was perceived to be more private, more agreeable to some women, and, because it could be delivered more privately and without surgical facilities, offered by a wider range of providers, such as private obstetrician-gynecologists and family practitioners.20,21
Recently published data demonstrate that mifepristone use did grow between its introduction and 2005.22 However, more recent and detailed data allow us to examine provision patterns by provider and physician type and in greater detail. In this article, we attempt to answer three main questions. First, how rapidly has mifepristone provision increased over time? Second, how does mifepristone provision vary by provider type? And third, has mifepristone increased access to abortion, particularly for women in areas that lack a surgical abortion provider?
MATERIALS AND METHODS
This study was found to be exempt by the Guttmacher Institute’s institutional review board. We relied on two main data sources. The first was a 2005 national census of all known abortion providers in the United States,22 the 14th in a series of censuses conducted periodically since 1974 by the Guttmacher Institute. These data generally are recognized as the most complete source on U.S. abortion incidence and provision and more complete than state reports compiled by the Centers for Disease Control and Prevention.23 We used these data to assess the number of abortions performed nationwide, the proportion of abortions performed using mifepristone for many providers, and the location of surgical providers.
The second main data source consisted of deidentified sales data from the sole U.S. distributor of mifepristone, Danco Laboratories (New York, NY). This included monthly sales data, in units purchased, from the first shipments in November 2000 through May 2007 and limited information about the purchasers, including provider type (clinic, hospital, physician), affiliation with a provider member organization, physician specialty (eg, ob-gyn, family practice), and aggregated geographic location (city, state, zip code, county). We also obtained summary data for the year 2007 as a whole. To protect provider privacy, we were not given access to provider names and street addresses. Finally, to calculate the proportion of eligible (ie, early) abortions that were performed using mifepristone, data on abortions by gestational age were obtained from the Centers for Disease Control and Prevention’s abortion surveillance system, adjusted for year-to-year changes in completeness of those reports.24
The abortion-provider census data define a “provider” as a physical site where services are offered. Several physicians offering abortions at one site are considered a single provider, whereas an agency with several sites constitutes multiple providers. We assumed that each unique mifepristone purchaser was a provider and that mifepristone was provided in the same city where it was shipped, although this may not be true in some cases (eg, where an agency purchases mifepristone for multiple clinic sites).
In the data, each sale unit consisted of three 200-mg tablets. The number of units purchased by providers was used to estimate the number of mifepristone abortions performed, using conversion factors provided by the distributor of mifepristone. Specifically, the conversion factor accounts for the fact that major associations of providers, which provide the large majority of mifepristone abortions, have established protocols that shift to a 200-mg dose. Many or most independent providers follow these protocols as well. The conversion factor also accounts for the fact that returns are quite minimal and that most providers do not keep additional stock of the drug (personal communication, A. Long, 2009).
We tabulated the estimated numbers of mifepristone abortions and providers by year, provider type, and physician specialty. We also calculated the proportion of all abortions and of eligible abortions that used mifepristone. Mifepristone–misoprostol medication abortion is approved by the FDA for use up to 7 weeks after the last menstrual period, but clinical research indicates that the method can be effective through 63 days of gestation and perhaps later.13,25,26 For this reason, we used the total number of abortions at less than 9 weeks of gestation as the denominator for our estimate of the proportion of eligible abortions performed by mifepristone.
To assess geographic variation in mifepristone provision, we tabulated the estimated number of mifepristone providers and abortions as well as the total number of abortion providers and abortions (both medication and surgical) in 2005 by state, county, and metropolitan and micropolitan statistical area. (Metropolitan areas contain at least one urbanized area of 50,000 or more inhabitants, and micropolitan areas have one or more urbanized areas of at least 10,000 but less than 50,000 inhabitants.) We also compared them with the populations of girls and women aged 15–44 years in the areas they served. In addition, we used geographic information systems software to plot the locations of mifepristone providers (by zip code) and compared them with the locations of facilities that provide surgical abortions to examine the geographic distribution of areas where only mifepristone was offered. Specifically, we calculated the number of mifepristone-only providers who were more than 50 miles away from a known surgical provider. Because data on mifepristone providers were deidentified, we were unable to distinguish between mifepristone and surgical providers at a smaller level than zip code, so the exact location of providers was unknown. Distances between providers were approximated as follows: a mifepristone provider was considered mifepristone-only if it was not in the same zip code as a known surgical provider. A mifepristone-only provider was considered to be more than 50 miles away from a surgical provider if no part of the zip code in which it was located was 50 or fewer miles away from the edge of a zip code in which a surgical provider was located.
Some 208 providers purchased mifepristone in November and December 2000, the first 2 months of its availability. The estimated number of mifepristone providers increased rapidly through 2004 to 877 providers in that year and then more slowly through 2007 to 902 in that year (Table 1). The 887 mifepristone purchasers in 2005 represented about half of the 1,787 total abortion providers reported in the Guttmacher provider census from the same year.
The estimated number of mifepristone abortions increased from about 55,000 in 2001 (the first full year of availability) to about 158,000 by 2007 (Table 1). Again, the increase was steeper through 2004 and more gradual afterward. Although the estimated numbers of mifepristone providers and abortions increased, the overall number of abortion providers decreased slightly from 1,819 in 2000 to 1,787 in 2005, the most recent year in which data were collected, and the overall number of abortions decreased from 1.31 million in 2000 to 1.21 million in 2005.
The total number of abortions in 2007 is not available, but if the downward trend in abortion since 2000 continued through 2007, mifepristone would represent 14% of all abortions performed in that year (Table 1). Although the overall number of abortions decreased, the proportion of all abortions performed at less than 9 weeks of gestation increased between 2000, when it was 58%, and 2004, the most recent year for which data are available, when it was 63%. Assuming this gestation trend also continued through 2007, we estimate that about 21% of eligible abortions were performed using mifepristone in that year.
In 2007, clinics represented about 46% of mifepristone providers but performed 88% of mifepristone abortions. Physicians represented 51% of providers but accounted for just 11% of abortions, and hospitals accounted for 4% of providers and 1% of abortions. These distributions are similar to the percentages of all (including surgical) abortions provided by the different types of facilities in 2005.22
Among physicians who ever provided mifepristone between November 2000 and May 2007, two thirds were obstetrician-gynecologists and 13% were family practice physicians (Table 2). Obstetrician-gynecologists accounted for an even higher proportion of mifepristone abortions provided by physicians during this period (79%) compared with family practice physicians (11%).
In 2005 (the most recent year for which data on all abortions were available), the states with the highest proportion of abortions performed using mifepristone were Wyoming, Vermont, Iowa, Arizona, and Utah (Table 3). In Wyoming, the total number of abortions was extremely small (70 in 2005), although it is worth noting that most Wyoming residents seeking abortions obtain them out of state.23 In Vermont, Iowa, and Arizona, a high proportion of abortions are performed by clinics affiliated with member organizations, and these facilities are much more likely than other providers to offer mifepristone.
The proportion of providers in each state in 2005 who provided mifepristone varied from 0% to 100%. States with the lowest proportion of mifepristone providers were South Dakota (0%), Hawaii, Idaho, Oklahoma, the District of Columbia, and West Virginia. The highest proportions were in Wyoming, South Carolina, Kentucky, Louisiana, North Dakota, and Kansas.
In four states, data from Danco Laboratories (New York, NY) indicated a greater number of mifepristone providers in 2005 than the total number of providers in the Guttmacher census. Although Guttmacher sent questionnaires to all providers believed to have provided surgical abortion or purchased mifepristone, some did not complete and return the questionnaire, so the Guttmacher study may have omitted some mifepristone providers recognized in the other data source or some may have purchased the medication but not provided it in 2005 or both.
Finally, we aimed to assess the extent to which mifepristone-only providers were located in areas not served by surgical abortion providers, particularly surgical providers likely to be publicly known. In 2005, the large majority of surgical providers (95%) were located in metropolitan areas. This was true for mifepristone providers as well: 96% were located in metropolitan areas, 3% were located in micropolitan areas, and only 1% were located elsewhere.
Only 14 mifepristone-only providers were located more than 50 miles away from any surgical provider (Table 4). Only five mifepristone-only providers of 10 or more abortions were located farther than 50 miles from any surgical provider of 400 or more abortions.
In 2005, 403 counties (13% of the 3,141 total U.S. counties) had one or more abortion providers and 307 counties (10% of all counties) had one or more mifepristone providers. However, 84% of mifepristone abortions were performed in just 93 counties (which comprise 3% of all counties in the United States). Of the 62 million girls and women aged 15–44 years, an estimated 58%, or 36 million women, lived in a county with a mifepristone provider compared with 66%, or 41 million, who lived in a county with any abortion provider.
Our study is subject to some limitations. Fifty-three percent of the providers who reported any abortions in 2005 in the Guttmacher provider census did not indicate whether they performed any mifepristone abortions. In cases in which the census did not have information on the medication/surgical breakdown, we assumed that the provider performed some surgical abortions (and therefore was not a mifepristone-only facility). If some of these providers did in fact perform only medication abortions, we would have underestimated the extent to which mifepristone had extended to new areas. However, the vast majority (93%) of the 351,000 abortions performed by these providers were surgical abortions. Moreover, among the providers whose status was known, 88% performed surgical abortions; the percentage among those who did not report abortion types was likely similar or higher.
Because the data we received from the distributor of mifepristone were deidentified, we could not definitely match each provider to one in Guttmacher’s census (or establish that a provider was a new provider not counted in the census). For this reason, our only option was to perform aggregated analyses at the level of geographic units such as counties and zip codes. We identified providers as mifepristone-only if they did not share a zip code with a surgical provider from Guttmacher’s census. We assumed that mifepristone providers and surgical providers sharing a zip code were the same provider. Because they could have been two providers (one mifepristone-only, one surgical-only), we may have underestimated the number of mifepristone-only providers. However, we would have correctly identified the zip code as not being mifepristone-only.
Furthermore, data aggregation at the zip-code level may have led us to underestimate the number of mifepristone providers more than 50 miles away from a surgical provider. We defined such mifepristone providers as those located in zip codes that did not in any part touch a point 50 or fewer miles from the edges of any surgical provider’s zip code. Thus, although the closest point of a mifepristone provider’s zip code may have been less than 50 miles away, other parts of the same zip code may have been more than 50 miles away.
We assumed that each unique mifepristone purchaser was a provider and that their mailing cities were their service-delivery cities. It is possible that some larger agencies may have distributed mifepristone from a central site (which may or may not have performed abortions) to clinics located in less accessible areas. Therefore, we may have underestimated the number of mifepristone providers and the number located farther away from known surgical providers or attributed some abortions to locations where none were performed. However, in at least one large provider network, 75% of all facilities order individually rather than centrally (personal communication, E. Talmont, 2009).
Because the estimated number of mifepristone abortions performed in a particular year was based on the number of units purchased that year, the estimate may be greater or less than the actual number performed. Year of purchase may not have corresponded precisely to the year mifepristone actually was used. However, the number of facilities that purchased mifepristone in a particular year but did not provide any in that year is likely to be negligible (personal communication, A. Long, 2008).
According to the 2005 Guttmacher abortion-provider census, an estimated 1,086 providers offered mifepristone. The difference between this estimate and the 887 reported above is likely the result of three factors. First, there may be some providers who purchased mifepristone in a previous year but used it in 2005, although the extent to which this happened is likely minimal. Second, because the mifepristone data consist of purchasers whereas the Guttmacher data consist of physical sites, it is likely that some purchasers distributed mifepristone at multiple sites. Finally, the 2005 provider census data were weighted to account for providers who did not respond to the items about early medication abortion. If providers of early medication abortion responded to the survey at a higher rate than nonproviders, the weights may have overestimated, or inflated, the total number of abortion providers in the census who offered this service.
The numbers of mifepristone abortions and providers both have increased dramatically since FDA approval in 2000, even as the total number of abortions performed in the United States has declined consistently over the same period. Although in recent years these increases have begun to slow, mifepristone clearly has become an integral part of abortion provision in the United States—by 2005, about one fifth of eligible abortions were performed using mifepristone. In some states, the percentage was 40% or more.
The steep increases in mifepristone provision through 2004, and the slower increases thereafter, may be the result of the method meeting unfilled demand in the first few years of its availability. This increase paralleled the increase in the abortion rate after legalization in 1973—the U.S. abortion rate rose sharply for several years until plateauing in 1980. However, further increases in mifepristone use seem possible. In France, for example, mifepristone use increased from about 40% of eligible abortions in the early 1990s to more than 80% in 2002. In England and Wales in 2002, 10 years after introduction, the percentage of eligible abortions that used mifepristone was just 18%27,29; by 2007, however, the percentage had risen to 43%.29 Moreover, preliminary data for the United States indicate that the number of units of mifepristone distributed in 2008 increased by 15–20% over 2007, compared with annual increases of 2% to 6% in the previous 3 years (personal communication, A. Long, 2009). It is possible that an increase in the use of misoprostol alone for early abortion also has affected the uptake of mifepristone,30 although such a conclusion would be purely speculative because there are no good data currently available to document this trend.
Mifepristone also may have contributed to a trend toward very early abortions. The proportion of abortions at 8 weeks of gestation or less has been increasing since the early 1990s, but the increase was steepest between 2000 and 2002, right after the introduction of mifepristone. At the same time, however, the early 2000s saw an increase in the number of providers performing early surgical abortions22; these two developments both likely have contributed to this trend.
The large geographic overlap between facilities that provide surgical abortion and those that offer mifepristone means that, in many cases, women are able to choose the type of early abortion procedure they prefer. The differences in mifepristone use by state often can be driven by the decisions of an individual agency or a small number of providers to provide mifepristone. The likelihood of encountering mifepristone availability is even greater at clinics affiliated with member organizations, for whom a higher percentage of early abortions are performed medically; in some states, these organizations have been very proactive in introducing mifepristone.
Compared with other physician types, obstetrician-gynecologists have most commonly incorporated mifepristone into their practices. However, early expectations that approval of mifepristone would result in a wider range of providers offering it have not yet been met. In particular, family practice physicians continue to provide only a small percentage of the total number of mifepristone abortions performed. One limiting factor may be liability coverage, which has been identified as a barrier to provision of abortion services generally, and mifepristone specifically, in family medicine.31 On the other hand, it is possible that the increasing role of advanced practice clinicians may have an effect on the distribution of mifepristone providers going forward; at one large provider network, more than half of mifepristone abortions are performed by advanced practice clinicians.32
Our results suggest that mifepristone has not brought a major improvement in the geographic availability of abortion. This is not to say that there are no mifepristone-only abortion providers, but rather that most mifepristone provision occurs among existing surgical providers and that the majority of providers who offer only mifepristone are located near surgical providers. Advocates aiming to facilitate access to abortion may want to consider whether and how the method could be made available in a broader geographic area.
1. Beal MW, Simmonds K. Clinical uses of mifepristone: an update for women’s health practitioners. J Midwifery Womens Health 2002;47:451–60.
2. Clark WH, Hassoun D, Gemzell-Danielsson K, Fiala C, Winikoff B. Home use of two doses of misoprostol after mifepristone for medical abortion: a pilot study in Sweden and France. Eur J Contracept Reprod Health Care 2005;10:184–91.
3. El-Refaey H, Rajasekar D, Abdalla M, Calder L, Templeton A. Induction of abortion with mifepristone (RU 486) and oral or vaginal misoprostol. N Engl J Med 1995;332:983–7.
4. Hamoda H, Ashok PW, Flett GM, Templeton A. A randomised controlled trial of mifepristone in combination with misoprostol administered sublingually or vaginally for medical abortion up to 13 weeks of gestation. BJOG 2005;112:1102–8.
5. Hamoda H, Ashok PW, Flett GM, Templeton A. A randomized trial of mifepristone in combination with misoprostol administered sublingually or vaginally for medical abortion at 13-20 weeks gestation. Hum Reprod 2005;20:2348–54.
6. Kulier R, Gülmezoglu AM, Hofmeyr GJ, Cheng L, Campana A. Medical methods for first trimester abortion. The Cochrane Database of Systematic Reviews 2004, Issue 1. Art. No.: CD002855. DOI: 10.1002/14651858.CD002855.pub3.
7. Marions L. Mifepristone dose in the regimen with misoprostol for medical abortion. Contraception 2006;74:21–5.
8. Schaff E. Evidence for shortening the time interval of prostaglandin after mifepristone for medical abortion. Contraception 2006;74:42–4.
9. Schaff EA, Fielding SL, Westhoff C. Randomized trial of oral versus vaginal misoprostol 2 days after mifepristone 200 mg for abortion up to 63 days of pregnancy. Contraception 2002;66:247–50.
10. Henshaw SK, Finer LB. The accessibility of abortion services in the United States, 2001. Perspect Sex Reprod Health 2003;35:16–24.
11. Winikoff B, Ellertson C, Elul B, Sivin I. Acceptability and feasibility of early pregnancy termination by mifepristone-misoprostol. Results of a large multicenter trial in the United States. Mifepristone Clinical Trials Group. Arch Fam Med 1998;7:360–6.
12. Jensen JT, Harvey SM, Beckman LJ. Acceptability of suction curettage and mifepristone abortion in the United States: a prospective comparison study. Am J Obstet Gynecol 2000;182:1292–9.
13. Spitz IM, Bardin CW, Benton L, Robbins A. Early pregnancy termination with mifepristone and misoprostol in the United States. N Engl J Med 1998;338:1241–7.
14. Urquhart DR, Templeton AA, Shinewi F, Chapman M, Hawkins K, McGarry J, et al. The efficacy and tolerance of mifepristone and prostaglandin in termination of pregnancy of less than 63 days gestation; UK Multicentre Study–final results. Contraception 1997;55:1–5.
15. Beal MW. Update on medication abortion. J Midwifery Womens Health 2007;52:23–30.
16. Henderson JT, Hwang AC, Harper CC, Stewart FH. Safety of mifepristone abortions in clinical use. Contraception 2005;72:175–8.
17. Shannon C, Brothers LP, Philip NM, Winikoff B. Infection after medical abortion: a review of the literature. Contraception 2004;70:183–90.
18. Sitruk-Ware R. Mifepristone and misoprostol sequential regimen side effects, complications and safety. Contraception 2006;74:48–55.
19. Talbot M. The little white bombshell. The New York Times July 11, 1999;sect 6:39.
20. Joffe C. Medical abortion in social context. Am J Obstet Gynecol 2000;183:S10–5.
21. Leeman L, Asaria S, Espey E, Ogburn J, Gopman S, Barnett S. Can mifepristone medication abortion be successfully integrated into medical practices that do not offer surgical abortion? Contraception 2007;76:96–100.
22. Jones RK, Zolna MR, Henshaw SK, Finer LB. Abortion in the United States: incidence and access to services, 2005. Perspect Sex Reprod Health 2008;40:6–16.
23. Strauss LT, Gamble SB, Parker WY, Cook DA, Zane SB, Hamdan S. Abortion surveillance–United States, 2004. MMWR Surveill Summ 2007;56:1–33.
24. Henshaw SK, Kost K. Trends in the characteristics of women obtaining abortions, 1974 to 2004. New York (NY): Guttmacher Institute; 2008.
25. Child TJ, Thomas J, Rees M, MacKenzie IZ. A comparative study of surgical and medical procedures: 932 pregnancy terminations up to 63 days gestation. Hum Reprod 2001;16:67–71.
26. Hamoda H, Ashok PW, Flett GM, Templeton A. Medical abortion at 9-13 weeks’ gestation: a review of 1076 consecutive cases. Contraception 2005;71:327–32.
27. Jones RK, Henshaw SK. Mifepristone for early medical abortion: experiences in France, Great Britain and Sweden. Perspect Sex Reprod Health 2002;34:154–61.
30. Rosing MA, Archbald CD. The knowledge, acceptability, and use of misoprostol for self-induced medical abortion in an urban US population. J Am Med Womens Assoc 2000;55:183–5.
31. Dehlendorf CE, Grumbach K. Medical liability insurance as a barrier to the provision of abortion services in family medicine. Am J Public Health 2008;98:1770–4.
32. National Abortion Federation. Timeline of work to enhance the role of certified nurse-midwives (CNMs), nurse practitioners (NPs), and physician assistants (PAs) in abortion care. Washington (DC): National Abortion Federation; 2006.