For over 50 years, most legal, clinically overseen abortions in the United States were counted.

Thanks to national data collections, we know approximately how many abortions were performed each year by state, by week of pregnancy, by age of the pregnant person, by race and more. The data, like much of our data on public health, has implications for how programs are funded, and how we understand and monitor conditions, according to Gary Young, director of the Northeastern University Center for Health Policy and Healthcare Research.

However, in a post-Dobbs United States, experts say collecting the data is likely to become more complicated, especially in states where abortion is heavily restricted. What’s more, this could, in turn, have implications for public health.

How do we know so much about abortion in the United States? Since The Supreme Court’s 1973 decision on Roe v. Wade, information about abortion has come from two main sources: The Centers for Disease Control and Prevention and the Guttmacher Institute. The CDC publishes data on abortions in the United States each year. They get data from state health agencies, who get it from providers and hand it over voluntarily. In contrast, Guttmacher gets data straight from a census of abortion providers. 

Even before Dobbs, these reports had their limitations. 

“I think everybody has presumed that abortion data has never been fully complete,” says Jean McGuire, Professor of the Practice at Northeastern.

In the case of the CDC report, sometimes states withhold information—in 2019, those states included California, Maryland and New Hampshire. 

Because of this, Guttmacher’s estimates are always higher when compared with the CDC. But Guttmacher’s numbers are still limited by unresponsive facilities. For its September 2019 report, Guttmacher only heard from 59% of known abortion clinics, and filled in the gaps with estimates based on data from previous years. State health department data was also used as a supplement, but the quality of that data is variable. 

Both reports have another crucial design flaw: the data only includes legal abortions overseen by clinicians, not self-managed abortions. And as medication abortions become more popular—in 2022, Guttmacher found that medication abortion accounted for more than half of all abortions overseen by clinicians in the United States—it can be more and more difficult to count them. Pills can be ordered by mail online, and as abortion is restricted in many states following the Dobbs decision, this is becoming a popular—and private—option that can elude researchers.

“Thousands of people used abortion pills on their own before Dobbs, and even more will be doing that now,” says reproductive rights activist and Women Help Women cofounder Susan Yanow. “Of course that will impact our data as these numbers have never been captured and, as the practice is clandestine, can only be estimated.”

Fear of prosecution in states that criminalize abortion is likely to cause underreporting as well. “Providers are going to be more afraid,” says Wendy Parmet, professor at the Northeastern University School of Law. “Whenever you’re dealing with something that is highly stigmatized and the threat of legal enforcement hovers over, it’s very challenging … the greater the perceived threat, and the greater the stigma, the harder it is to get the data.” 

In the digital age, the threat of surveillance exacerbates this problem, says Mia Kim Sullivan, executive director of Collective Power for Reproductive Justice. “The surveillance infrastructure to gather data is already being built by states looking to track and prosecute people seeking abortion care,” says Sullivan. She cites the Every Mother Matters Act, which has already passed in Arkansas and has been introduced in Texas, that requires women seeking an abortion to call a hotline that is run by crisis pregnancy centers. Sullivan says crisis pregnancy centers, which can masquerade as health clinics, are not covered by health information privacy laws. 

Restrictions and stigma have already led to data gaps when it comes to studying abortion worldwide. For a recent Guttmacher study on abortions globally, data came from official sources (most common for high-income countries where abortion is legal), surveys, and studies, but people can be reluctant to report abortions, making the quality of the data variable.

But there is precedent for identifying and filling in these gaps. Public health data in general can be spotty, especially when doctors did not have an incentive to report, Young says. Historically, doctors have underreported more informal parts of the job, like counseling. And today, patients being treated for opioid use disorder (OUD) may be diagnosed with pain due to concerns that employers will discover their diagnosis. 

Normally, gaps like these can be filled by making inferences based on other data, Young says. For abortion, past studies could imply that abortion rates will not be affected by Dobbs laws. One Guttmacher study showed that state policies weren’t the “primary driver” of a decline in abortions in 2017, and that clinic closures don’t impact abortion rates. That’s because “when abortion is illegal, the need for abortion does not disappear,” Sullivan says. Another study looking at the impacts of Texas’ six-week abortion ban shows that people in states that outlaw abortions tend to get them in another state.

However, Sullivan notes, four of the seven states in that study—Oklahoma, Arkansas, Mississippi and Louisiana—have since enacted anti-abortion legislation, and all of their clinics have closed. Alabama is moving to restrict people from leaving the state for abortions. With this “explosion of legislative experimentation,” Sullivan says, getting an abortion is becoming more and more difficult, she says. “There has been a devastating sea change in access to abortion in just the last three months since Dobbs was decided,” Sullivan says. McGuire, for her part, calls the post-Dobbs world “terra incognita.”

Skewed data could impact our understanding of abortion in the United States. It could also be dangerous for public health, as people will be less informed about their healthcare options, Parmet says. “In this country, we put a lot of onus on patients to take care of their own health and to make their own decisions. And in the absence of reliable guidance informed by data, people are actually forced to rely on things they hear from their friends, things they see on TikTok,” Parmet says. “There will be situations where women will be left without the information they need.”

It will also be more difficult to decide what measures need to be taken to help promote public health. “We don’t do things about things we don’t count,” McGuire says.

For media inquiries, please contact media@northeastern.edu.