The Washington Stand reported in detail on Monday on a massive new study of U.S. women who have consumed the abortion pill. The study is the first in a planned series from the Ethics and Public Policy Center (EPPC), a Washington, D.C.-based nonprofit that describes itself as “the premier institute working to apply the riches of the Jewish and Christian traditions to contemporary questions of law, culture, and politics, in pursuit of America’s continued civic and cultural renewal.”

The study of nearly 866,000 abortions procured by use of the prescription drug mifepristone is the largest of its kind to date and full of findings that, its authors say, should prompt action by the U.S. Food and Drug Administration (FDA) to review and address the high rate of serious complications the drug inflicts on women.

The study is significant in many ways in addition to the size of the study population. First, it covers a sample of women who have taken the drug in recent years, whereas prior safety and efficacy studies date back a decade or more. The EPPC study examined the impact of the drug under the loose conditions that have actually prevailed in the past decade. These conditions, as prior investigators have warned, include deletion of multiple medical visits originally recommend by the FDA, the use of “no test” protocols designed to confirm pregnancy and to assure it is not ectopic, online distribution of the pills, and use for at least three and sometimes more weeks later in pregnancy than when the drug was originally approved in the year 2000. Mifepristone is known to become less effective and more hazardous for women as the pregnancy proceeds.

Mifepristone now dominates abortion provision in the United States, constituting nearly two-thirds of the procedures carried out in the country. It is tailor-made for a laissez-faire national policy on abortion as our already weak system of abortion incidence and injury reporting makes tracking safety especially difficult when women experience harms and are treated not by the medical personnel who have supplied the drug but by third parties who may be unaware of or misinformed about their medical history. In its first 16 years, mifepristone harms, including injuries and deaths, were subject to mandatory reporting to the FDA. In 2016, that mandate was lifted, and the FDA required the reporting of harm only when use of the pill led, as it sometime does, to a maternal fatality.

The EPPC study, like a few smaller predecessors, gets around this challenge by accessing deidentified claims data for individual women, where a medical history pre- and post-use of mifepristone could be statistically tracked and tallied. The 865,727 abortions studied involved 692,873 women because of repeat mifepristone abortions on some of the patients. EPPC paper authors Jamie Bryan Hall and Ryan T. Anderson report that the women were broadly representative of the U.S. female population and were not prescreened for inclusion in the observational study. Nearly one-fifth of the study population had multiple mifepristone abortions.

The results are striking and of such magnitude as to warrant prompt and thorough FDA review. The study found that 10.93% of the women who took mifepristone experienced a serious adverse event. The sequelae examined included sepsis, infection, hemorrhage, need for blood transfusion, hospitalization, emergency room (ER) visits related to the abortion, need for a surgical abortion due to incomplete abortion, and other abortion-specific complications. The highest specific count of injuries and percentage of cases with serious adverse events were ER visits (40,960, 4.73%), hemorrhage (28,658, 3.31%), need for surgical abortion (24,563, 2.84%), infection (11,707, 1.34%), and ectopic pregnancy (3,062, 0.35%). The authors note that some women suffered more than one of these conditions.

The EPPC study is the largest of its kind to date and is also notable for its inclusion of claims data from a variety of sources, and not just a single insurer or agency. The claims data was obtained from a commercially available source and is presumably available to other researchers for further studies. The database included “information on hospital and office visits, diagnoses, procedures, and prescriptions processed through private health insurance, Medicaid, Medicare, TRICARE, and the Department of Veterans Affairs.” It did not include transactions where the provider — like an HMO — is also the insurer, or cases where the abortion transaction was in cash. This is a common analytical framework according to the authors and allowed for longitudinal tracking of the information in the study.

In broad terms, the EPPC study tracks with prior attempts, primarily by Charlotte Lozier Institute (CLI), to analyze Medicaid claims data in search of adverse events. CLI (Studnicki et al.) reviewed Medicaid claims data for mifepristone use from 1999 to 2015 (therefore not overlapping with any of the EPPC study population). The study found a sharp increase after 2002 in the rate of ER visits following ingestion of mifepristone. The study was published in a peer-reviewed journal, reached a wide audience, and was praised by a publication editor as “fine” work. After the study’s findings were cited by attorneys in a case that ultimately reached the Supreme Court, politically-motivated critics influenced the journal to retract the study and two others, unrelated topically, by the same authors.

One criticism worthy of discussion regarding any study of adverse reactions to a medication or procedure is whether the injuries involved are truly serious. A critic of the CLI study suggested ER visits not be considered serious if the woman’s reasons for going to the ER were precautionary and not substantive. This triggered another study by Studnicki et al., published in the International Journal of Epidemiology and Public Health Research in September of last year. The study focused on ER visits within 30 days of a pregnancy outcome — whether a chemical abortion, surgical abortion, or live birth — and compared them to ER visits by women who were not pregnant. Among other findings, using standard codes to identify the acuity of a condition leading to the ER visit, the Studnicki study found, “The likelihood of an [Emergency Department] visit with a severe or critical acuity code following the use of abortion drugs was more than twice as high as that of women who were not pregnant.” 

In November 2024, Jonathan Abbamonte of The Heritage Foundation Center for Data Analysis published a study using another data source regarding emergency room use. The report drew upon public access data available for various time periods from South Carolina, New Jersey, and Arkansas. Abbamonte points out that only a few states make free public access data available for their emergency departments (a policy that health advocates might do well to seek installing in states that have limited data from this source). His data cover South Carolina from 2016-2023, New Jersey (a generally abortion information-limiting state) from 2016-2022, and Arkansas from 2019-2021. Abbamonte concluded that ER visitation rates have increased in these states due to complications from induced abortions since the FDA removed the in-person medical visit requirement in 2020-2021. The Heritage study also found evidence that reported miscarriage rates in those states also increased, an indication that misreporting of abortion drug use as miscarriage, as some pill providers recommend, may be occurring.

Overall, then, evidence continues to accumulate that abortion drugs, administered without medical evaluation, distributed by mail, and consumed under conditions (e.g., 11 weeks of pregnancy) outside even today’s lax FDA requirements, are causing disproportionate harm, both in absolute terms and relative to surgical abortion. The recent statement from FDA Commissioner Dr. Marty Makary that FDA intends to take no action on the harms of mifepristone is premature and worrisome. The harms of this drug have now been identified in states that fund abortions with their own Medicaid funds (Studnicki), in a cross-section of insurance contexts including public and private programs (Hall, Anderson), and emergency room data in states with diverse abortion policies (Abbamonte).

Congressional committees and the FDA should take an exhaustive look at these findings. The data at issue are available for second and third looks by other scholars as well. The well-being of both women and the unborn merit renewed efforts to ascertain the truth about a drug with such a questionable risk profile.