The Bad Data Backing Josh Hawley's Attack on Abortion Pills

2 hours ago 2

The abortion pill mifepristone "is not safe and effective," argue the authors of a new study that uses insurance claim data to examine adverse reactions to the pill. They claim to have found a "serious adverse event" rate of 10.93 percent, and they say this finding justifies renewed restrictions on mifepristone.

Sen. Josh Hawley (R–Mo.) seems to agree. The day the Ethics & Public Policy Center (EPPC) released its new mifepristone study, Hawley wrote to Food and Drug Administration (FDA) chief Marty Makary, urging the commissioner to "follow this new data and take all appropriate action to restore critical safeguards on the use of mifepristone."

Makary said in late April that he has "no plans" to restrict access to abortion pills. But he added that "there is an ongoing set of data that is coming into FDA on mifepristone. So if the data suggests something, or tells us that there's a real signal, we can't promise we're not going to act on that data that we have not yet seen."

Taken at face value, that's simply a reasonable statement of scientific uncertainty. But it seems a little odd considering we have decades of data on mifepristone already—this isn't some new or relatively unstudied drug. And things get even dicier when we consider that, a few days after Markary's April 24 comments at the Semafor World Economy Summit, a conservative group released new data that are being used by conservative politicians to suggest it is time for the FDA to act.

The Study

Hall and Anderson base their arguments on their claim that "serious adverse events" are very common with mifepristone—occurring, they say, in slightly more than a tenth of patients.

They based this conclusion on information gleaned from an insurance claim database. It contained "de-identified data for all U.S. patients during the years 2017 to 2023," and they identified 865,727 likely mifepristone-enabled abortions in this database. Of these cases, 10.93 percent ended in a "serious adverse event," they say.

Clinical trial reports of a "less than 0.5 percent" adverse event rate are way off, they conclude—and therefore, they argue, mifepristone is neither safe nor effective.

But the methods they used to reach these conclusions are highly questionable.

First, they identified "mifepristone abortions" by flagging insurance codes indicating the prescription of mifepristone with or without misoprostol, a "medically induced abortion by oral ingestion of medication," or the "elective termination of pregnancy" in conjunction with billing codes that suggest mifepristone was given.

Already there are problems. This study is supposed to concern the safety of people undergoing a drug-induced abortion, right? But people who have miscarried naturally are also prescribed mifepristone, and they seem to be lumped in here, despite the fact that they have different circumstances that could wind up producing more, fewer, or different sorts of adverse events than seen in patients prescribed the drug for elective abortion in the first 10 weeks. Furthermore, the authors do not seem to take into account whether patients followed up the mifepristone with misoprostol, as is the FDA's recommended practice. Taking mifepristone alone might lead to different outcomes than seen with the two-pill protocol.

And then there's a much bigger issue—what the study considers a "serious adverse event."

There's a clear pattern in the study, in which the very specific or clearly dire categories of adverse events occurred in a relatively small number of patients, while the categories that make up the bulk of adverse events flagged are vaguely defined and may represent minor or even benign situations.

Just 0.10 percent of cases resulted in patients getting sepsis, 0.15 ended in patients needing blood transfusions, and 0.22 percent in what the study authors define as other life-threatening events.

The rates of serious and well-defined complications here are all low and well within the parameters advertised by abortion pill manufacturers. The reason this study finds so much higher prevalence of adverse events comes down to the event categories they won't tell us much about.

Emergency Room Visits and 'Other' Complications

Of the 865,727 cases analyzed (some of which involved repeat patients), 94,605 were determined by the authors to have ended in serious adverse events, which they defined as any one of a broad range of outcomes happening within 45 days of taking the pill.

Patients could be coded into multiple categories, so someone admitted to the hospital for an infection might be listed in both the infection and hospitalization categories.

The largest category of adverse events is "other abortion-specific complications," reportedly experienced in 49,169 cases. But the authors give us little idea about what these "other" complications might be. By definition, they do not include sepsis, infection, transfusion, hemorrhage, an ectopic pregnancy, a need for a surgical abortion, or "other life-threatening events." The only explanation the authors give is that "other abortion complications include codes specifically related to an abortion or miscarriage, as well as life-threatening mental health diagnoses, etc."

And this seems all the more questionable when they allude to an undefined category of "mental health" issues that occur within a month and a half after taking the pill.

"Let's say you go see your doctor for anxiety or insomnia a month after having an abortion," writes Jessica Valenti in the Abortion, Every Day newsletter. "Because EPPC is counting any event that happens within 45 days of taking mifepristone—and they're not defining 'life-threatening'—your sleep problem or anxiety could be counted as a 'serious adverse event' of abortion pills."

It also seems possible that some situations—sexual assault, relationship troubles, a wanted pregnancy with fetal anomalies, health problems, money problems—could trigger both a desire for an abortion and some mental health problems. In those cases, it would hardly be the pill itself that is to blame. And even if the mental health issues flagged were directly related to feelings about the abortion, it still hardly seems fair to describe them as serious complications stemming from the pill itself.

The study's second biggest category of adverse events is listed as emergency room visits, experienced in 40,960 cases. This category also suffers from some lack of information, as an emergency room visit alone does not actually tell us whether a serious complication took place.

Women who take mifepristone often start experiencing the pill's effects outside of the normal business hours of an doctor's office or health clinic. Some might go to the emergency room to check on levels of pain or bleeding that turn out to be normal.

Without actually knowing the nature and outcome of the emergency room visits in this study, we can't conclude much of anything. The fact that there were more than 40,000 trips to the emergency room but far fewer hospitalizations—5,699—suggests that many of these visits did not result in serious risks or harms.

For what it's worth, the FDA has said that emergency room visits in themselves are not enough to count as serious adverse events. The agency defines serious adverse event as death, hospitalization, anything life-threatening, anything resulting in disability or permanent damage, anything requiring medical or surgical "intervention to prevent permanent impairment or damage," and anything that "may require medical or surgical intervention…to prevent one of the other outcomes."

Drilling Down Further

The third largest category of adverse events in the EPPC study is "hemorrhage," which occurred in 28,658 cases. Hemorrhages can be quite serious, of course. But "even that [hemorrhage] data is suspect because the EPPC doesn't define what it means!" writes Valenti. "Doctors I spoke with say bleeding is expected during medication abortion, and it's only considered a serious complication if a blood transfusion is needed."

Transfusions were needed in far fewer cases (1,257) than there were cases of hemorrhage. The discrepancy in numbers here draws into question whether all of the events counted in the hemorrhages category should really be deemed serious adverse events. (Again, we could be looking at women who sought medical attention for what turned out to be normal amounts of bleeding.)

The fourth largest category of adverse events was needing a surgical abortion after taking the pill, happening in 24,593 of the cases. But it's a well-advertised fact that the two-pill regimen is successful in only about 87 to 98 percent of cases, depending on when in pregnancy the mifepristone and misoprostol are taken.

If fewer than three percent of people in the study needed surgical abortions after taking mifepristone, that is well within the expected range and not necessarily a sign of anything unusual. It simply means these drugs, like many drugs, are not 100 perfect effective.

The smallest category of adverse events in the EPPC analysis was sepsis, a life-threatening infection that occurred in 824 cases.

Far more people suffered from some sort of infection—11,707 cases, or 1.34 percent.

Again, with infections, we have a category where it's hard to know what exactly is meant. Infections here may include yeast infections, urinary tract infections, or minor vaginal bacterial infections—things that may be unpleasant but aren't generally serious unless left untreated and aren't necessarily the fault of taking the pill per se. Having sex, using tampons, or engaging in other not-recommended behaviors too soon after an abortion or miscarriage can result in infections.

The remaining categories were ectopic pregnancy (something not caused by abortion pills but a situation in which abortion pills will not work and would not be recommended), in 3,062 cases; and "other life-threatening adverse events," in 1,956 cases.

Bad Data Makes Bad Policy

Perhaps aware that total revocation of FDA approval isn't in the cards (yet), Hawley told Makary it "is time to revisit and restore the FDA's longstanding safety measures governing mifepristone."

Over the past decade, the FDA has relaxed its rules around mifepristone prescribing, now allowing it to be prescribed up to 10 weeks into a pregnancy (instead of the former seven weeks), to be prescribed remotely, and to be shipped by mail. Those last two changes seem especially to anger anti-abortion advocates, since they help women in states where abortion pills are illegal in obtain them from doctors and pharmacies in states where they are not.

Advocates previously tried to get the courts to change those requirements. The matter went all the way to the U.S. Supreme Court, but it rejected the challenge, holding that those who brought the case didn't have standing to do so. And while the same issues could be raised by a different group of plaintiffs, the standing problem could prove difficult to overcome.

That means that changing the national rules regarding the prescription of mifepristone and misoprostol may have to rely on persuading the FDA to change its policy. So that's what the EPPC is trying to do.

"The FDA should reinstate the original patient safety protocols that were required when mifepristone was first approved," write the study's authors, Jamie Bryan Hall and Ryan T. Anderson. "Doing so will likely reduce the harms to women and permit better monitoring to determine whether this drug should remain on the market."

They argue that getting mifepristone "should require at least three in-person office visits by the patient," that it should be prescribed only by physicians (some states allow prescription by nurse practitioners, physician's assistants, or certified midwives), that it should have to be taken in a medical facility, that it should be allowed only through the first seven weeks of pregnancy, and that prescribing physicians "must be able to provide surgical intervention in cases of incomplete abortion or severe bleeding, or have made plans to provide such care through others, and be able to assure patient access to medical facilities equipped to provide blood transfusions and resuscitation, if necessary." Rule changes such as these would make it highly difficult for women in states where abortion is illegal to access abortion-inducing drugs outside of the black market.

The fact that so much of this study was worded so vaguely should give everyone pause. If all of the serious adverse events cited here were undeniably serious, the researchers should have no problem defining them concretely or making their datasets public. That's the only way to reassure readers that they aren't inflating their count with less serious health problems.

It would be silly to argue that medication abortion is 100 percent safe. But to put the serious complications in perspective, we should also look to complication rates in pregnancy and in miscarriages. No matter how a pregnancy ends, a pregnant woman incurs some risk. (For the record, "maternal sepsis occurred in 0.04% of deliveries in the United States," according to research published in the Journal of the American Medical Association in 2019.)

It's also silly to contrast mifepristone's current complication rates with some mythical future in which far fewer of these abortions would take place. The alternative to legal and accessible mifepristone isn't necessarily "no abortion" and "no complications." It's also abortion with black market mifepristone or misoprostol pills, abortion attempts through more dangerous means, and abortions in which women feel afraid to seek care if they experience complications.

More Sex & Tech News

Jury selection starts today in the OneTaste trial. Nicole Daedone, founder of the orgasmic meditation company OneTaste, and Rachel Cherwitz, the company's former head of sales, are charged with one count of conspiracy to commit forced labor and no underlying forced labor charge. Read more about the weak, weird case against them here.

More nuanced findings about teens and tech. Politicians and anti-tech activists like to pretend that research overwhelmingly shows negative findings when it comes to young people and social media. But more and more we've been seeing studies that challenge this conventional wisdom, aloing with mixed messages and nuanced feelings from young people themselves.

So it goes in a recent survey from the Pew Research Center. "Our latest survey of U.S. teens ages 13 to 17 and their parents finds that parents are generally more worried than their children about the mental health of teenagers today," Pew reports.

Moreover, kids are more worried about the way social media might effect other kids' health than their own—a sign while they might buy into the dire warnings about teens and tech that are all around them, they don't necessarily find this concern validated by their own experiences:

Roughly half of teens (48%) say these sites have a mostly negative effect on people their age, up from 32% in 2022. But fewer (14%) think they negatively affect them personally.

Meanwhile, 28 percent of the teens surveyed said social media has a mostly positive effect on their own lives, and the majority—58 percent—said it's neither negative nor positive. And while teens see some ways social media affects them negatively, most notably in terms of sleep and productivity, they also cite it as a source for socialization, creativity, and learning about mental health.

The return of the App Store Accountability Act. This newsletter covered a previous version of this bill last year. Like last year's iteration, this one would make digital marketplaces, such as Google Play and Apple's App Store, verify user agers and require parental consent before minors can download apps. Several social media companies, including Meta, Snap, and X, "have endorsed the idea of making app stores a sort of central clearinghouse for age verification," reports CNN. Apple and Google are opposed.

NetChoice sues over Georgia's social media age-check mandate. Georgia's Senate Bill 351 would require social media platforms to "make commercially reasonable efforts to verify the age of account holders," require parental consent before anyone under age 16 can have an account, and ban targeted advertising to minors. In a new lawsuit, the tech trade group NetChoice argues that the law is unconstitutional. "NetChoice has already permanently blocked similar laws in Arkansas and Ohio because these heavy-handed government mandates violate the First Amendment," the group explains. "Georgia's law is no different."