Unsafe and Ineffective: Reconsidering the Safety of the Abortion Pill Mifepristone

Since Roe v Wade’s overturn in 2022, the abortion industry has countered by increasing the availability of medical abortion, namely through the drug mifepristone.¹ These efforts build off the Biden-era Food and Drug Administration’s removal of in-person dispensing requirements for mifepristone and the violation of the Comstock Act, a federal statute illegalizing the mailing of “Every article or thing designed, adapted, or intended for producing abortion.” This is noteworthy because medical abortions (like those using mifepristone) account for nearly two-thirds of all abortions nationally.²

Yet the safety of mifepristone is facing scrutiny. The Ethics and Public Policy Center’s (EPPC) post-market study has shown the risk for serious side effects after a woman uses mifepristone is 10.93%, 22 times higher than the FDA’s reported 0.5%.³ This means the abortion industry’s lead effort carries a risk profile far greater than advertised, coming just when the abortion industry is doubling down to increase this method’s availability. This is showing to be deadly not only to the child in the woman’s womb, but the woman herself.

3. U.S. Food and Drug Administration, MIFEPREX (mifepristone) Tablets: Highlights of Prescribing Information, rev. March 2016 (Silver Spring, MD: U.S. Food and Drug Administration), accessed January 24, 2026.

Because death is a tragedy, this article seeks to equip its audience through first detailing how mifepristone works. I will then provide a parsing of the EPPC’s study, where I will respond to critiques against the study by pro-abortion sources. Using the data of the EPPC study, I will then compare mifepristone’s risk with other medications, showing how mifepristone poses a dangerously high threat to public health.

The Action of Mifepristone

To understand how mifepristone induces abortion we must understand normal human development from conception through the stage of embryonic implantation.

Embryonic Implantation

In order to understand how mifepristone disrupts progesterone to end a child’s life, we have to understand some technical anatomy first. When a sperm fertilizes an egg, a zygote is formed. Upon fertilization, the zygote begins to cell-divide as it travels down its mother’s fallopian tube toward the endometrium, where it will implant. At implantation the embryo connects to the mother’s endometrium, a connection lasting until the child’s birth.

This connection is governed by a complex biochemical process, mediated chiefly by the important hormone progesterone. Progesterone suppresses endometrial inflammation (preventing endometrial discharge), and ensures uterine (muscular) relaxation, preventing contractions and expulsion of the child. In other words, if the growing baby is a boat anchored at harbor in a storm, progesterone serves to calm the winds so the ship is safe and anchored. Consistent levels of progesterone for implantation and growth of the baby via the endometrial content is obtained from the mother’s follicle (the ovarian structure which releases the egg) and the embryo’s own secretion of hCG (Human Chorionic Gonadotropin), a hormone which “temporarily rescues ovarian progesterone until the placenta takes over at around 6–8 weeks of pregnancy[.]”⁴ When the placenta takes over it becomes the vital interface between mother and child, producing the progesterone necessary to ensure constant connection to the mother’s uterus, preventing rupture, expulsion, and death. The placenta is a unique and temporary organ, and is part of both the baby and the mother.⁵

Mifepristone: A Pharmaceutical Noose

The mother and child’s joint effort—mediated by the placenta—is maintained by the all-important hormone progesterone. The placenta connects to a progesterone-sensitized endometrium, providing sustained connection between child and mother (whereby nutrients and waste are exchanged). Mifepristone interrupts this progesterone-dependent relationship by invading the same receptors in the progesterone-sensitized endometrium, severing the link between mother and child.⁶

6. Megha Satyanarayana and Mesa Schumacher, “How Medication Abortion with RU-486/Mifepristone Works,” Scientific American, May 17, 2022.

The child’s connection to life—anatomically and biochemically upheld by itself and its mother—is severed. The noose of cellular asphyxiation, inflicted by a chemical separating the child from life, results. Mifepristone, when used to induce an abortion, is nothing more than a pharmaceutical hand placed over the developing child’s mouth, keeping the air out.

Why the Abortion Pill Harms Women, and Why Detractors are Wrong

Further discussion of the EPPC’s study showing the dangers of mifepristone to women is needed to inform those considering using it, those counseling them, and public policy experts. Also, concerns over the study’s methodology have been raised which warrant answers.

Why Mifepristone Harms Women

The EPPC’s 2025 study of mifepristone is the largest-known, analyzing data from an insurance claims database including 865,727 mifepristone abortions from 2017–2023. The data shows 10.93% of women experience sepsis, infection, hemorrhage, or another serious event within forty-five days of receiving mifepristone. This risk is twenty-two times higher than the <0.5% on the drug label. This should prompt public policy experts—notably, the Food and Drug Administration (FDA) and the Health and Human Services Secretary Robert F. Kennedy Jr.—to immediately reinstate earlier patient-safety driven protocols requiring direct physician oversight of women taking mifepristone, and should prompt a renewed investigation of the drug’s actual risk profile.⁷

7. Jamie Bryan Hall and Ryan T. Anderson, The Abortion Pill Harms Women: Insurance Data Reveals One in Ten Patients Experiences a Serious Adverse Event (Washington, DC: Ethics & Public Policy Center, April 28, 2025), 1.

The EPPC study exhibits features that increase the plausibility of its findings and thus strengthen the policy implications it suggests. These include:

1. Its dataset is twenty-eight times larger than all FDA-cited clinical trials cited combined (865,727 women to the FDA’s 30,966 women).

2. Its data is newer than the FDA’s, which relies on data over a decade old.

3. The women its dataset represents are average recipients, not a rigorously screened group. It includes individuals who receive or do not receive pre- and post-abortion care, unlike “the carefully controlled regimen of care that ordinarily prevails in a clinical trial.”⁸

8. Hall and Anderson, The Abortion Pill Harms Women, 2.

4. Its dataset—utilizing the NIH Common Terminology Criteria for Adverse Events (CTCAE version 5)—only counts severe (grade 3) and life-threatening (grade 4) events. It excludes death (grade 5) as it “could not necessarily be observed in the insurance claims data.”⁹ This means the 10.93% risk factor for mifepristone is hard earned, and likely an underestimation.

9. Hall and Anderson, The Abortion Pill Harms Women, 5.

5. The study is methodologically conservative. It includes only “hospitalizations and emergency room visits with abortion-related diagnosis and procedure codes […] in our analysis.” The timeframe for recording a serious side effect associated with mifepristone in the EPPC study was narrower than some timeframes used in FDA studies (forty-five days versus seventy-two days in some studies).¹⁰ Moreover, 72% of emergency room encounters within 45 days of abortion were excluded, and Emergency Room visits for mental health reasons associated with mifepristone only included homicidal and suicidal ideation. This avoids “artificially inflating serious adverse event rates.”¹¹

10. Hall and Anderson, The Abortion Pill Harms Women, 5.

11. Jamie Bryan Hall and Ryan T. Anderson, “Fact Sheet: Excluded Adverse Events in Real-World Study of Mifepristone,” Ethics & Public Policy Center, May 6, 2025.

Why Detractors are Wrong

The EPPC study hasn’t escaped criticism. Instead of attempting to address every critique possible, I will summarize in italics the strongest critiques I’ve seen. Critiques 1–3 come from Agence France-Presse (AFP)—a far-left fact checking organization funded in part by the French government. Critiques 4–9 come from an independent review by pro-abortion U.C. San Francisco epidemiologist Ushma Upadhyay.¹²

12. ANSIRH, “Review of Hall and Anderson: Report on Abortion Safety,” September 2025.

1. Agence France-Presse (AFP): The EPPC paper is neither peer-reviewed nor published in a medical journal.

This critique overlooks the EPPC’s goal, which wasn’t peer-review, but replicability. “We made our study fully replicable for anyone who wants to analyze the insurance claims data.”¹³ Peer-review also doesn’t escape the “replicability crisis” the inability for scientists to obtain the same results as previous investigators. “Peer-review” also assumes an impossibility: unbiased objectivity. Confirming or rejecting a study’s conclusion is subject to the review body’s presuppositions.

13. Jamie Bryan Hall and Ryan T. Anderson, Frequently Asked Questions About the Largest Study on Chemical Abortion (Washington, DC: Ethics & Public Policy Center, May 7, 2025), 3.

2. Agence France-Presse (AFP): The dataset—865,727 mifepristone abortions—counts those who may have taken mifepristone without misoprostol [a drug that expels the child from the uterus after it is aborted], which isn’t evidenced based.

This objection only confirms that women shouldn’t take mifepristone without medical oversight. We can hardly expect the average consumer to know what the “evidenced based” practices are for taking this potent drug!

3. Agence France-Presse (AFP): “The paper also recorded 40,960 visits [nearly half its data] to a hospital emergency department as serious adverse events following medical abortion […but] a visit to the emergency room alone is not evidence of a serious adverse event.”¹⁴

14. Hall and Anderson, “Fact Sheet: Excluded Adverse Events,” 2; ANSIRH, “Review of Hall and Anderson,” 1.

This is a “what if?” objection, not actually engaging with the data or methodology of the EPPC study, instead raising the specter of a problem with the study that isn’t actually proven to exist. It fails to note that the EPPC study took pains to exclude 72% of Emergency Room visits within the first forty-five days to ensure that only those included in the study were counted because of a serious or life-threatening event associated with taking mifepristone. The 72% excluded (98,483 Emergency Room visits) “involved diagnoses not clearly related to the chemical abortion or not medically serious.”¹⁵

15. Jamie Bryan Hall and Ryan T. Anderson, “Fact Sheet: Excluded Adverse Events in Real-World Study of Mifepristone,” Ethics & Public Policy Center, May 6, 2025.

4. Ushma Upadhyay: The authors counted any subsequent procedure to complete the abortion as a serious adverse event. However, these are considered treatment failures and captured in effectiveness rates. It is known and expected that about 3–5% of patients will need additional medications or a procedure to complete the abortion. This is not a serious or urgent complication, but an expected one. This category accounts for 24,563 (one-fourth) of the 94,605 “serious adverse events” in the report.

Upadhyay attempts to dismiss a quarter of the EPPC’s findings through sleight of hand, reclassifying mifepristone’s failure (and need for further medical attention) as an expected feature of “effectiveness rates.” But another EPPC study found that mifepristone fails to cause an abortion 5.26% of the time, and is more properly classified a “failure rate” forcing the woman to seek further medical attention. Combined with the 10.93% risk in their previous study, this means 13.51% of women “experience at least one serious adverse event or repeated abortion attempt within 45 days of first attempting a mifepristone abortion.”

5. Ushma Upadhyay: The EPPC’s largest category, “Other abortion-specific complications,” is unclear, and constituted partially by homicidal and suicidal ideation, which Upadhyay specifically claims “rigorous published research overwhelmingly establishes that serious mental health crises are not a complication of abortion.”¹⁶

16. ANSIRH, “Review of Hall and Anderson: Report on Abortion Safety,” 2.

Upadhyay claims research overwhelmingly proves mental health crises aren’t a complication of abortion, but she doesn’t provide a single study to warrant this. Further, she is dismissive of real mental health ramifications women experience after medication abortion.¹⁷

17. Priscilla K. Coleman found that post-abortive women had an 81% increased risk of developing mental health challenges; see “Abortion and Mental Health: Quantitative Synthesis and Analysis of Research Published 1995–2009,” British Journal of Psychiatry 199, no. 3 (September 2011): 180–86. See also David M. Fergusson, L. John Horwood, and Joseph M. Boden, “Abortion and Mental Health Disorders: Evidence from a 30-Year Longitudinal Study,” British Journal of Psychiatry 193, no. 6 (December 2008): 444–51.

6. Ushma Upadhyay: “The EPPC study lacks a standardized definition of hemorrhage.” It also overlooks that “A successful medication abortions always induces some bleeding.” Further, “hemorrhage is considered a serious adverse event only when it requires a blood transfusion.”

Despite these claims, the study Upadhyay cites contradicts her, judging “A clinically relevant definition [of hemorrhage] would include both a clinical response to excessive bleeding, such as transfusion or admission, and/or bleeding in excess of 500 mL.” Blood transfusion is only one specified criteria among others (“such as”).

7. Ushma Upadhyay: Other “life-threatening adverse events,” “infections” and chronic issues related to cardiac and pulmonary events aren’t necessarily abortion-related, and likely “over-represented […] because many public and private insurance plans cover abortions only if a pregnancy is causing or exacerbating serious health issues.”

This is another “what if?” critique. Further, Upadhyay’s claims are based upon another study she coauthored seeking to standardize the classification of abortion incidents, which showed adverse events from chronic illness due to pregnancy and abortion should be distinguished. But this study’s population received surgical abortion, which shouldn’t be conflated with medication abortion risk.¹⁸

18. Diana Taylor, Ushma D. Upadhyay, Mary Fjerstad, Molly F. Battistelli, Tracy A. Weitz, and Maureen E. Paul, “Standardizing the Classification of Abortion Incidents: The Procedural Abortion Incident Reporting and Surveillance (PAIRS) Framework,” Contraception 96, no. 1 (July 2017): 1–13.

8. Ushma Upadhyay: Granting the EPPC’s results (the 10.93% risk profile), the policy implications don’t follow. First, the study didn’t examine changes in serious adverse events after telehealth became available. Second, it didn’t prove reverting to in-person dispensing requirements would reduce adverse events. Thus, to cite Upadhyay in her exact words, “The policy implications stated [restoring previous restriction on mifepristone dispensing] do not follow from the authors’ own analyses, even if one were to take them at face value.”

Upadhyay overstates her case here. A need for future study doesn’t nullify the present implications of a past study. In fact, the whole point of the EPPC study was to promote further investigation of the safety of mifepristone. Further, the EPPC’s results undermine the FDA’s original data, the foundation upon which dispensing liberties were taken. Upadhyay’s suggested avenues of study are secondary concerns, unnecessary for restoring historical dispensing practices. But they are needed to confirm the actual safety of mifepristone.

9. Ushma Upadhyay: EPPC’s analysis violates “Standard practices in epidemiology emphasize the importance of transparency in methodological approaches to ensure the reliability, reproducibility, and credibility of research findings.” Their analysis cannot be verified or reproduced. The EPPC didn’t disclose their all-payer insurance claims database, preventing independent verification. Further, the study doesn’t disclose the diagnosis and treatment codes used, which is standard practice when analyzing insurance claims data.

Upadhyay’s critiques here are unfounded. The EPPC’s analysis can be verified through acquiring the same or a similar dataset and performing the same study. As cited by the EPPC, “The dataset is available for purchase and our methodology is public. This study is fully replicable, and we encourage others to replicate it.”¹⁹

19. Jamie Bryan Hall and Ryan T. Anderson, “Frequently Asked Questions About the Largest Study on Chemical Abortion,” Ethics & Public Policy Center, May 7, 2025, question 7.

Upadhyay’s review exhibits a striking—regrettably common—one-sidedness. A persuasive argument steelmans opposing views, then deposes them. Upadhyay’s review—though rhetorically charged—fails here. Her pro-abortion bias has outrun her ability for dispassionate assessment. When considering the AFP critiques, one finds a similarly jaded one-sidedness and failure to reckon with the EPPC study, instead relying on expert opinion and recycled data from decades ago (namely, the FDA studies which are “foundational” for mifepristone being considered “safe and effective”).

We must ask next, given mifepristone has a nearly 11% severe/life-threatening risk to the women, what other medications are prescribed which have so high a risk factor?

Higher Risk: Mifepristone, or Cancer Meds?

How does mifepristone’s risk compare to other medications? Recall that the EPPC study discussed above found 10.93% of mifepristone users experienced at least one severe or life-threatening event within 45 days of taking it. These “severe” or “life-threatening” events were graded using a standard, the NIH Common Terminology Criteria for Adverse Events (CTCAE version 5). “Severe” events were graded a 3, and “life-threatening” a 4. So, 10.93% of women experienced a grade 3 or 4 event (such as hemorrhage, infection, suicidal or homicidal thoughts, blood clots, anaphylaxis, or surgery).

The question is, what drugs exist today that have a percentage of its users adversely affected near mifepristone’s rate? If 10.93% of women experience an adverse event that is “severe” or “life-threatening,” are there other drugs which harm their population at or near the same percentage, and to the same severity (grade 3–4 CTCAE)?

Table 1 compares mifepristone with medications affecting 9% or higher of its recipients with a CTCAE of 3 or higher (severe or life-threatening event). This shows mifepristone standing shoulder-to-shoulder beside medications with a well-documented high risk for severe or life-threatening events.²⁰

20. In the following tables, mifepristone is placed in parentheses to note that it was not part of the studies, but stands in the tables as comparison.

What do these comparisons mean?

First, the baseline morbidity (health) of mifepristone recipients vs. immune modulator recipients vary remarkably. Mifepristone recipients are generally younger, healthy women. Immune therapy recipients have cancer, are generally older, and likely have a higher morbidity burden (disease/dysfunction of their bodies). Second, immune therapy recipients are safeguarded by receiving the medication only under direct supervision of medical professionals, often in a hospital. The patient has been pre-screened and their physiology reasonably optimized before treatment. Mifepristone is taken privately with no medical oversight, with minimal pre-screening. Patients get mifepristone in the mail.

The actual risk of mifepristone is reasonably higher “pound-for-pound” compared to immune modulators. Though mifepristone recipients are generally healthier, severe or life-threatening adverse events are equivalent to drugs used to treat diseases among society’s sickest. This suggests mifepristone’s risk score is deceptively potent, given the recipients’ health and administration location.

Finally, we must consider what is an acceptable “baseline” risk for a drug relative to its perceived benefit (“perceived” because abortion is falsely considered beneficial, rather than a tragedy). While no studies could be found detailing what an acceptable “serious” or “life-threatening” incident rate is in the population of its users, the FDA does give guidance. The FDA suggests weighing the risk/benefit of a drug and whether this is acceptable based upon “therapeutic context.” As the FDA notes, “FDA is likely to have a lower tolerance for potential serious risks or toxicities when a drug is intended to treat conditions for which many treatment options with lesser risks are available, or when it evaluates preventative medicines, where the intended population may be healthy people.”²⁴ Giving birth to the child or the “preventative medicine” of lifestyle choices which keep oneself from becoming pregnant are markedly preferable to the dangers of mifepristone, especially considering the FDA’s lower tolerance for medications with a serious risk where the “intended population may be healthy people.”

24. U.S. Food and Drug Administration, Benefit-Risk Assessment for New Drug and Biological Products: Guidance for Industry (Silver Spring, MD: U.S. Department of Health and Human Services, Center for Drug Evaluation and Research and Center for Biologics Evaluation and Research, October 2023).

Conclusion

Mifepristone is unsafe and ineffective. The availability of mifepristone without in-person medical oversight poses a public health crisis to women. Access to mifepristone through telehealth, and the ability to receive it by mail and take it privately, amplifies this. The EPPC’s findings are a welcomed, divergent view to old data propping up the “safety and efficacy” of mifepristone. Its results should be taken seriously by women considering using mifepristone, catalyze new investigation into this drug’s actual risk profile, and raise public policy experts’ concern over the hidden risks this prevalent drug poses to one of society’s healthiest populations: women in the physiological prime of their lives.