How COVID Skewed Maternal Death Statistics, Fueling False Claims About Abortion

IMAGE CAPTION: Protesters rally in Melbourne, Australia after Roe v. Wade was overturned. Credit: Creative Commons

Federal statistics suggesting that maternal mortality rates plummeted after the overturning of Roe v. Wade are “highly misleading” and being misinterpreted by abortion opponents, according to new research published this week in the journal JAMA Network Open.

“We want to set the record straight,” said co-author Amanda Stevenson, assistant professor of sociology at the University of Colorado Boulder. “Maternal deaths have not, in fact, declined since the Dobbs decision.”

The paper shines a light on how statistical artifacts from the COVID-19 pandemic can skew mortality data and calls into question the way the federal government displays statistics on its website. The authors are now calling on policy makers to practice “extreme caution” when turning to such numbers to inform policy.

“We saw an instance where published data was being used to say something that simply is not true about abortion- and pregnancy-related death,” said co-author Leslie Root, an assistant research professor with the Institute of Behavioral Science. “This paper is about correcting that.”

What the numbers really say

Previous research by Stevenson and others predicted that abortion bans would ultimately increase maternal mortality — which the World Health Organization defines as any death while pregnant or within 42 days of the termination of a pregnancy —because abortion is exponentially safer than remaining pregnant.

However, published data from the National Center for Health Statistics appears to show a precipitous drop in maternal deaths in the months after the June 2022 Dobbs decision, the Supreme Court ruling which ended the constitutional right to an abortion and led to bans in 14 states. 

Abortion opponents have pointed to this data in efforts to expand abortion restrictions, including in a brief submitted to the U.S. Supreme Court in the Emergency Medical Treatment and Active Labor Act (EMTALA) case in June.

As social demographers, Stevenson and Root set out to take a deeper look at the numbers.

Stevenson explains that up-to-date maternal mortality statistics are in high demand, and there are numerous ways to count them: Raw, real-time monthly counts are available upon request, but highly provisional and subject to statistical noise or external variables. As a result, the NCHS displays only “12-month ending sums,” which tally all deaths from the previous 11 months and current month, on its official website.

Root and Stevenson analyzed monthly Centers for Disease Control counts of maternal deaths from 2018 to 2023, as well as counts of deaths involving COVID-19 among women of reproductive age from January 2020 to September 2023.

Then they compared monthly counts to the 12-month ending sums.

They found that, in total, 4,802 pregnant people died from January 2018 to September 2023. From September 2022 to February 2023, shortly after the Dobbs ruling, “12-month-ending sums” did, in fact, show a significant decline. But monthly counts stayed stable.

They also found that exactly one year prior to the seeming plummet in maternal deaths displayed on the NCHS website, the Delta and Omicron waves were surging, with COVID-related deaths on the rise among pregnant women and pandemic-related restrictions making it harder for them to get care.

For instance, based on NCHS 12-month-ending sums, maternal mortality rates appeared to decline by 125 deaths between August 2022 and September 2022 alone (two months after Dobbs). But 93 of the deaths reported in August of 2022 actually occurred 11 months earlier, in September of 2021, at a peak of the pandemic. Similarly high proportions of declines in September 2022 to February 2023 are explained by deaths that occurred 12 months before.

Once those months of COVID surge were no longer counted in the 12-month-ending sums, the maternal death counts dropped.

“We conclude that the rapid decline in the 12-month ending counts after Dobbs is actually driven by events in 2021, specifically the shock to maternal deaths during the Delta and

Omicron waves of the pandemic,” Stevenson said.

In an invited commentary published in the same journal, Paula Lantz, PhD, of University of Michigan Ford School of Public Policy, lauded the authors for filling a “critical need for objective and high-quality abortion policy evaluation research.”

“The findings clearly demonstrate that the observed decline in maternal mortality after the 2022 Dobbs v Jackson Women’s Health ruling is the result of a resolution of the COVID-19 mortality shock, not because of new restrictive abortion laws being passed by state legislatures,” she wrote.

Stevenson and Root caution that similar statistical artifacts from the pandemic may skew other publicly available mortality data and warn policy makers to proceed with “extreme caution” when making decisions based on “seemingly reasonable, but highly misleading” data.