Have U.S. deaths from pregnancy complications tripled?

CDC pushes back on study claiming overestimates

A new study questioned skyrocketing rates of maternal mortality in the United States. The Centers for Disease Control and Prevention is pushing back.

Experts agree that the U.S. maternal mortality rate is unacceptably high. And year after year of data show that disadvantaged groups, particularly Black and Native American women, die at even higher rates than women in the United States overall during pregnancy and childbirth.

But controversy broke out last week over just how bad the situation is, when a paper by academic epidemiologists published in the American Journal of Obstetrics and Gynecology (AJOG) provoked unusual pushback from the U.S. Centers for Disease Control and Prevention (CDC). The paper suggested a widely reported tripling in the U.S. maternal mortality rate (MMR) over the past 2 decades was in fact largely due to a CDC-led recording change on death certificates, the addition of a “pregnancy checkbox.”

That change was introduced to correct for previous underreporting of maternal deaths by as much as 50%. But the authors concluded this has led to the misclassification of many deaths as due to pregnancy when they were in fact incidental to it. These researchers instead ignored the pregnancy checkbox, counting as deaths only those with an explicit, pregnancy-related cause listed on the death certificate. Their analysis found that the maternal mortality rate hasn’t changed in 2 decades.

It was an explosive finding that generated lots of headlines and within 2 days, the agency shot back with a statement contesting the findings. The authors’ methods, CDC wrote, “are known to produce a substantial undercount of maternal mortality. There are maternal deaths occurring that would not otherwise be identified if the death certificate didn’t include a pregnancy checkbox.”

The American College of Obstetricians and Gynecologists (ACOG) jumped into the fray, too, calling the paper’s conclusions “irresponsible.”

K. S. Joseph, first author of the AJOG paper and an epidemiologist at the University of British Columbia (UBC), shot back in comments to Science, noting that even if up to half of maternal deaths were previously not captured, the maternal mortality rate in 2021 was threefold higher than 20 years earlier. “These increases far outweigh any [previous] undercounting,” he said.

The conflict illuminates the thorny challenges involved in defining and reporting maternal mortality, which is classically understood as deaths caused or hastened by pregnancy, delivery, or events in the 6 weeks after a pregnancy ends. For instance, no one disputes that a death from eclampsia, a dangerous condition provoked by pregnancy, qualifies as a maternal death and is properly included when calculating the mortality rate. But numerous causes of death—think of a preexisting, aggressive cancer in a person who happens to be pregnant—are harder to parse, and the problem is compounded by imperfect reporting on death certificates that CDC relies on.

“Capturing an accurate maternal mortality ratio is complex,” says Marie Thoma, a perinatal epidemiologist at the University of Maryland. “The checkbox actually did improve reporting. … But it led to this other issue of overreporting.”

CDC has reported an alarming increase in the U.S. maternal mortality rate—defined as maternal deaths per 100,000 live births—from 9.65 at the turn of the century to 32.9 in 2021. The rate has nearly doubled since just 2018, when it was 17.4. The stark numbers have generated widespread media coverage, including in Science, about the growing U.S. MMR and how poorly it compares with rates in European and other peer countries, typically in the single digits. The White House has previously led calls to attack the problem.

The analysis published last week challenged the CDC numbers showing that U.S. maternal deaths have spiraled out of control. Epidemiologists at UBC, Rutgers University, and elsewhere found the MMR remained essentially unchanged between 1999 and 2021, hovering at just over 10 deaths for every 100,000 live births. Their method of requiring explicit pregnancy-related causes on death certificates led them to conclude that the CDC estimates represent a substantial overcount.

The “pregnancy checkbox” was inserted on death certificates starting in 2003 to address what was at the time a widely acknowledged, substantial underreporting of maternal mortality: At the time, as many as 50% of physicians completing death certificates failed to report that a woman was, or was recently, pregnant. On death certificates, physicians now are asked to check a box indicating a person was pregnant when they died, or within 42 days of the end of the pregnancy. Doctors are not to check the box if a person died of accidental or incidental causes unrelated to pregnancy, for instance, in a car crash or from a gunshot wound. Although the agency rolled out the feature in 2003, it took 14 years before all 50 states adopted the surveillance tool. After that happened in 2017, the agency began to compute the nationwide rate using the checkbox.

“The high and rising rates of maternal mortality in the United States are a consequence of … reliance on the pregnancy checkbox leading to an increase in misclassified maternal deaths,” the AJOG authors wrote.

Some in the field bristled at the findings. ACOG’s interim CEO, Christopher Zahn, said in a 13 March statement: “To reduce the U.S. maternal mortality crisis to an ‘overestimation’ is irresponsible and minimizes the many lives lost and the families that have been deeply affected.”

Experts in science communication worry the study could minimize what remains a substantial problem, no matter which numbers are correct. “What concerns me the most is that this will take away from concern that people rightly have about how we care for pregnant people in this country,” says Robin Jensen, a reproductive communications scholar at the University of Utah.

CDC acknowledged from the outset that its new checkbox system had growing pains. In one agency analysis that linked death certificates with records of hospitalized patients, it found false-positive checkbox rates of 54% and 56% in 2014 and 2016, respectively, where there was no evidence of a pregnancy-related hospital encounter in the deceased patient’s hospital records for the previous year. In another analysis, CDC scientists found positive checkboxes on the 2013 death certificates of more than 300 women aged 65 and older, nearly half of whom were older than 84. CDC changed its method in 2018 to capture only people who died at ages 10 to 44.

But since 2018, the agency has continued to label as maternal deaths any nonaccidental death in this younger age group with the pregnancy box checked. If the death certificate lists a pregnancy-related cause such as eclampsia or bleeding, it is counted as a maternal death. But CDC also counts many generally unrelated causes of death such as brain cancer as maternal mortality.

The AJOG authors found that deaths listing cancer on the death certificate and counted as maternal deaths by CDC grew 46-fold between the turn of the century and the years 2018–21. That’s evidence errors are occurring, Joseph says.

Experts say the disagreement comes down to whether, and to what extent, CDC, in correcting one problem, created another. In its statement last week, CDC acknowledged that “the AJOG report confirms a prior CDC analysis, which found that the pregnancy box is sometimes mistakenly checked on death certificates,” contributing to “some overcounting.” In a paper CDC published in 2020, its scientists reported that adding the pregnancy checkbox doubled the maternal mortality rate between 1999 and 2017. Announcing the finding, the agency wrote that most of the rise was “not likely due to a true increase” in maternal mortality, but “to changes in data collection methods,” meaning the gradual adoption of the checkbox.

However, after the recent study came out, the agency defended its method in a statement to Science. “Capturing … these otherwise unrecorded maternal deaths is critical to understanding the scope of maternal mortality in the United States and taking effective public health action to prevent these deaths.”

For instance, a death of a pregnant person whose preexisting hypertension was aggravated by pregnancy would be caught with CDC’s pregnancy checkbox method. But it might be missed with the AJOG authors’ method, if the physician filling out the death certificate wrote simply “hypertension” and not “hypertension in pregnancy.”

Joseph concedes that possibility. He also agrees it’s likely parts of the increase since 2018 are real. For example, pregnant women who got COVID-19 were at higher risk of death than uninfected pregnant women.

But his team’s finding remains solid, Joseph said in response to CDC’s statement last week. He added: “We are absolutely not saying this problem has been solved. Every maternal death is a serious and extreme tragedy,” and many are preventable, Joseph says.

On one point both sides agree: Too many people of color die during pregnancy, delivery, and its aftermath in the United States. The AJOG authors found that, between 2018–21, 23.8 Black women died per 100,000 live births as opposed to 10.4 women overall, and 8.22 white women. For the year 2021, CDC reported 69.9 deaths per 100,000 live births for Black women, 32.9 for women of all races, and 26.6 for white women.

“There’s a very, very large gap, no matter how you cut it,” says Lyman Stone, a demographer at the Institute for Family Studies and the population consulting company Demographic Intelligence. “Maternal mortality rates among non-Hispanic Black women are egregiously high,” says Stone, who was not involved with the AJOG study and who has independently argued for a more conservative calculation of the maternal mortality rate.

CDC says it’s continually striving to improve the quality of maternal mortality reporting and statistics, including requiring states to verify pregnancy checkbox information on some death certificates beginning last year—a requirement it’s expanding to all pregnancy-related death certificates.

At the same time, some communications experts called on CDC to do better in communicating the nuances of a very complex issue. Kaiping Chen, a scholar of science communication at the University of Wisconsin-Madison, notes that CDC has been transparent on its website about the big bump in maternal mortality rates that was caused by the pregnancy checkbox. “Making the information transparent on the agency’s website is a first step,” she says. “But that doesn’t mean the public, the journalists, will get that. Better communication needs to [go] a step further.”

Sourced from Science