As justification for a Food and Drug Administration review of a drug used in medication abortion, two Trump administration health officials have referenced an April report from an anti-abortion group that claims to show a far higher rate of serious side effects from the drug than has been found previously. However, the report, which is not peer-reviewed research, didn’t disclose where it got its data and has substantial methodological issues, reproductive health experts say.
The drug in question, mifepristone, is used alongside misoprostol for medication abortion and is FDA-approved for terminating pregnancies through 10 weeks of gestation. As we have written before, serious side effects from medication abortion are rare.
Medication abortions make up most abortions in the U.S. — 63% of them in 2023, according to the Guttmacher Institute, a research organization that supports reproductive rights.
Health and Human Services Secretary Robert F. Kennedy Jr. and FDA Commissioner Dr. Marty Makary have repeatedly promised an FDA review of mifepristone. In a May 14 Senate Health, Education, Labor & Pensions Committee hearing, Kennedy called the April report “alarming” and said that it “indicates that at very least the label should be changed” for the drug. Kennedy and Makary reiterated the promise of a review in a Sept. 19 letter, again citing the April report as reason for concern about the drug’s safety.
“Recent studies—such as the study by the Ethics and Public Policy Center (EPPC), which you highlighted in your letter—indicate potential dangers that may attend offering mifepristone without sufficient medical support or supervision,” Kennedy and Makary wrote, responding to a July letter from 22 Republican state attorneys general. Kennedy shared his response letter in an Oct. 2 post on X that went further to claim that recent studies “already point to serious risks when mifepristone is used without proper medical oversight.” HHS did not reply to a question seeking to clarify what other studies the letter or Kennedy was referencing.
EPPC is a conservative nonprofit that opposes abortion. EPPC was part of the advisory board for Project 2025, the roadmap for the “next conservative President,” which advocated that the FDA take away its approval for medication abortion drugs or otherwise restrict their use.
The EPPC report, based on health insurance claims data from an undisclosed source, claimed that mifepristone came with a serious adverse event rate of 10.93%, “at least 22 times” higher than the rate of less than 0.5% reported on the drug’s FDA label.
But 263 reproductive health researchers wrote in an Aug. 27 response that “the report does not provide reliable evidence to support this claim, nor does any published literature.” The literature, including 22 clinical trials as well as many years of real-world data, indicates a low rate of serious adverse events, in keeping with the FDA label. Mifepristone was first approved for medication abortion in France in 1988, then in other countries, and in the U.S. in 2000.
“Decades of conclusive scientific evidence amassed through more than one hundred rigorous studies based on hundreds of thousands of patient outcomes have overwhelmingly established the safety and effectiveness of mifepristone for medication abortion and management of early pregnancy loss,” the reproductive health researchers wrote in their letter.
The EPPC report authors “clearly misconstrued and used deceptive methods to erroneously inflate the rate of serious adverse events after an abortion,” the letter’s lead signer, Ushma Upadhyay, a public health social scientist at the University of California, San Francisco’s Advancing New Standards in Reproductive Health program, told us.
After the EPPC report was released in April, Dr. Stella M. Dantas, then president of the American College of Obstetricians and Gynecologists, told AFP Fact Check that the report “manipulates data to drive a myth that medication abortion isn’t safe.”
“FDA approval of mifepristone must reflect the rigorous clinical evidence that has proven unequivocally that it is safe and effective for use in medication abortion and miscarriage management,” a coalition of 13 reproductive health organizations, including ACOG, said in a May 22 statement.
Data With Undisclosed Origins
To calculate its serious adverse event rate, the EPPC report counted serious adverse events “within 45 days following a mifepristone abortion” using health insurance claims data. The report states that the team “purchased access to a commercially available all-payer health insurance claims database including de-identified data for all U.S. patients during the years 2017 to 2023.”
However, the report did not specify which claims database EPPC used. Experts said this is unusual and makes it difficult to check the EPPC’s work.
“Data transparency is a hallmark of high quality research,” Alina Salganicoff, a senior vice president and the director of the Women’s Health Policy program at the nonpartisan health policy organization KFF, told us via email. “I have worked with Medicaid claims data and with commercial claims data and the standard practice is to disclose the data source. Without a disclosure, it is impossible to assess the methods and know the reliability and limitations of the data they used, and replicate the findings.”
Hunter Estes, communications director for EPPC, told us in an email that the organization made legal agreements not to disclose the name of the data vendor. However, “over a dozen brokers sell this data, and using it for studies of this kind is standard practice in the medical literature,” he said, referring us to an FAQ posted on the EPPC website. “There is nothing unique about our dataset, which contains nearly all U.S. insurance claims from 2017 to 2023, both public and private.”
Experts said they struggled to identify what data EPPC had used.
“Signatories to this letter, which includes numerous researchers who regularly analyze insurance and medical claims data in their work, do not know of a single ‘commercially available all-payer health insurance claims database’ that includes both private and public health insurance ‘data for all U.S. patients’ as described by EPPC,” the letter from the reproductive health researchers said. It matters what data were used, the researchers said, as there are “known methodological limitations of some databases for analyzing abortion claims data.”
“There are many data aggregators that license access to their products, but I am not familiar with any that claim comprehensive all-payer coverage across the US,” Shirley V. Wang, a pharmacoepidemiologist at Harvard Medical School and Brigham and Women’s Hospital, told us via email. Wang studies drug safety and effectiveness and has worked on improving transparency standards in her field.
“Most of the data aggregator vendors that we work with have in their data use agreement/contract that their company name must be listed in publications (it is part of their advertising for the data product), so it is surprising that such an incredible data asset would go unnamed,” Wang added.
The EPPC report also lacked other details, the letter from reproductive health researchers said. For example, it claimed to have been “conducted and validated by a team of data scientists, analysts, and engineers, with assistance from our clinical team of board-certified obstetricians and gynecologists.” However, the only listed authors were two EPPC leaders, one with a master’s degree in statistics and the other with a doctorate in political philosophy.
EPPC also did not fully report which diagnosis and procedure codes it used to determine whether people who received mifepristone went on to suffer serious adverse events, the reproductive health researchers said.
The EPPC report was published three days after the April 25 publication of a similar report from the Restoration of America Foundation, which is part of a family of conservative organizations and PACs. The reports appear to share underlying data and methods — and even include identical numbers in multiple places, despite some differences — but Estes, the EPPC spokesperson, told us the groups have no affiliation.
We reached out to the Restoration of America Foundation for more details on its report, which has no listed authors, and whether there is any connection to the EPPC report, but did not receive answers to our questions.
Report Appears to Inflate Mifepristone’s Risks
The methods the EPPC report did disclose raised further questions for reproductive health researchers, who said the team was likely counting situations as serious adverse events that shouldn’t have qualified.
The “largest issue,” Upadhyay told us, was that the study appeared to overcount emergency department visits as serious adverse events, even though it is known that people receiving medication abortions often visit the emergency department simply due to having questions about symptoms that turn out to be normal or wanting follow-up to see if they remain pregnant. “Oftentimes, people will go to an ER when they don’t have another source of immediate care,” she said.
The EPPC report stated that 4.73% of women receiving a medication abortion had related emergency room visits that qualified as serious adverse events. A fact sheet from EPPC said that the report “excluded a majority of emergency room visits to avoid overstating risks” (the emphasis is EPPC’s). But the letter from reproductive health researchers said that “it is impossible to verify how this was done” based on the information disclosed in the report.
Upadhyay co-authored a 2018 study showing that more than half of emergency department visits for abortion-related issues resulted just in observation of the patient, without a need for treatment. And not all cases where some treatment was needed would qualify as serious adverse events.
Adverse events are negative experiences that occur after using a medical product. While they can be caused by the product — and therefore be a true side effect — they may also be coincidental. According to the FDA, such events are generally considered serious if they are life-threatening or result in hospitalization, disability, permanent damage or death. In the context of medication abortion, Upadhyay said, adverse events are serious when someone has such significant bleeding that they need a transfusion, for example, or when major surgery such as a hysterectomy is required.
This does not include treatments that are sometimes provided in cases in which the medication abortion was not fully effective, the letter from reproductive health researchers said. The authors wrote that “characterizing follow-up treatment for an incomplete or failed medication abortion as a serious adverse event is inconsistent with published literature and FDA guidance.” About 3% to 5% of people getting a medication abortion need further treatment to complete the abortion, such as getting more misoprostol or a procedure to evacuate the uterus, which is not a major surgery, Upadhyay noted.
The EPPC report, however, counted instances in which medication abortion had not fully worked as serious adverse events, listing “Repeated (surgical) abortion” as a serious adverse event affecting 2.84% of women. Despite the name, surgical abortions are procedures that are typically performed in clinics and do not involve incisions.
The reproductive health researchers also said that the report did not “sufficiently define hemorrhage.” The purpose of medication abortion pills “is to cause vaginal bleeding,” Upadhyay said, and a common issue with studies is counting vaginal bleeding as hemorrhage when it should not qualify. The EPPC report said that 3.31% of women experienced hemorrhage qualifying as a serious adverse event. “Without a standardized definition, EPPC is likely misclassifying many cases of normal bleeding that occur with a medication abortion,” the letter from the reproductive health researchers said.
As we’ve said, other research on serious adverse events following medication abortion, including research looking at health insurance claims data, shows a far lower rate of major complications from medication abortions. For example, a 2015 study that looked at Medicaid claims data found a serious adverse event rate of 0.23%, in keeping with the FDA-reported rate of less than 0.5%.
Data Support Safety of Telehealth Abortions
The letter from Kennedy and Makary promised that the FDA would review mifepristone’s Risk Evaluation and Mitigation Strategy, a set of restrictions placed on the drug when it was originally approved that have been modified over the years. During the COVID-19 pandemic, for example, the FDA stopped enforcing previous requirements to dispense the drug in-person and in 2023 formally allowed the drug to be sent via mail, based on data showing this had not led to increased risks for patients.
On X, Kennedy specifically referenced the decision to begin allowing telehealth abortions. “The Biden administration removed mifepristone’s in-person dispensing rule without studying the safety risks,” he said. “We are filling that gap.”
A Q&A on the FDA’s website, however, says that the agency “conducted a comprehensive review” of the available evidence before modifying mifepristone’s REMS in 2023, finding that “there did not appear to be a difference in adverse events between periods when in-person dispensing was and was not enforced.” Studies since then have continued to show that telehealth abortion is safe.
The FDA letter cited the EPPC report as indicating “potential dangers that may attend offering mifepristone without sufficient medical support or supervision.”
To be clear, the “FDA never recommended that the mifepristone be prescribed or dispensed without medical support or supervision,” KFF’s Salganicoff said. Those prescribing the drug must be certified, and it must be dispensed by a certified prescriber or pharmacy.
Nor did the EPPC report present evidence indicating whether the reduced restrictions on mifepristone affected its safety, the letter from the reproductive health researchers said. There was no breakdown of the data by location of mifepristone dispensing, they wrote, “so there is no evidence that adverse events would be different based on in-person versus remote dispensing of mifepristone.”
Upadhyay also pointed out that the data EPPC used was limited to insurance claims, and therefore didn’t include people who paid for abortion pills in cash or otherwise did not use insurance coverage. The data only go through 2023, she said, and in the first years after in-person dispensing requirements were relaxed, telehealth abortions were provided by virtual clinics that didn’t accept insurance. “There’s very, very few if any telehealth abortions represented in this dataset,” Upadhyay said.
Other studies indicate that telehealth abortions are not associated with elevated risks compared with those disclosed on the FDA label for mifepristone. Upadhyay co-authored one such study, published in 2024, on more than 6,000 telehealth abortions, finding a rate of serious adverse events in keeping with the FDA label.
“It makes sense because the way a medication abortion works is you go to a clinic, you get the medications and then you go home, and the process occurs at home anyway,” Upadhyay said. “So theoretically, there should be no difference” between getting the pills from a clinic or through the mail. Regardless, she said, the patient is evaluated for medical eligibility and informed of the risks and the possibility that the pill will not be effective.
“All of the largest studies out there have found that telehealth is just as safe and effective as in-person care,” she said.
Editor’s note: FactCheck.org does not accept advertising. We rely on grants and individual donations from people like you. Please consider a donation. Credit card donations may be made through our “Donate” page. If you prefer to give by check, send to: FactCheck.org, Annenberg Public Policy Center, P.O. Box 58100, Philadelphia, PA 19102.