US pregnancy and postpartum deaths receive substantial news coverage, and reporting is frequently alarmist. This summer, a LiveScience article claimed that “pregnancy is deadlier in the US than in other wealthy countries.” The article stated that there were 19 maternal deaths for every 100,000 live births in the US, or more than twice as many maternal deaths in the US as in Canada, and nearly four times as many as in the UK.
A JAMA study prompted similar news earlier this year, with headlines reporting a “Massive Spike in Pregnancy-Related Deaths in U.S.” and that “Pregnancy-Related U.S. Death Rates Have Jumped in Recent Years.” The JAMA study found that the age-standardized rate of pregnancy-related mortality increased by 27.7% from 2018 to 2022 and concluded that “the concerning rates in the US should be an urgent public health priority.”
Public focus on the US’s comparatively high—and increasing—maternal mortality is not new. Previous reporting has emphasized that US maternal mortality is higher than that of other nations and has grown since the early 2000s. A brief review of OECD publications and Our World In Data visualizations paints a similar picture (Figures 1 and 2): US maternal mortality is substantially higher than that of comparable countries and has risen over time.
Figure 1. Maternal Mortality Ratio Estimates, 2020
Note: Estimates are three-year averages.
Source: Maternal and infant mortality: Health at a Glance 2023 | OECD
However, the figures and reporting on pregnancy and postpartum death often obscure more than they illuminate. This is due to two separate challenges.
First, because pre-2003 data missed some maternal deaths, Center for Disease Control’s National Center for Health Statistics (NCHS) urged states to add a pregnancy checkbox to death certificates in an effort to identify missed deaths. States began adding a pregnancy checkbox in 2003, and implementation of the pregnancy check box on death certificates continued over the next fourteen years.
All states adopted a version of the checkbox by 2018, which constituted a major change to the collection of pregnancy and postpartum death data. However, this gradual adoption of the pregnancy checkbox led to the impression that US maternal mortality increased steadily over the same period (Figure 2).
Figure 2. Maternal Mortality Ratio, 1990 to 2023
Note: Our World in Data defines the maternal mortality ratio as the estimated number of women who die from maternal conditions per 100,000 live births. Recent editions of this chart and data include a note that U.S. “values from 2003-2017 [are] affected by measurement change.”
Source: Maternal Mortality Ratio | Our World in Data
Although the pregnancy checkbox was intended to improve maternal death identification, the addition of the pregnancy checkbox led to a large number of misclassified deaths and an overestimation of maternal deaths. For example, subsequent NCHS studies showed data errors as a consequence of the new checkbox, including hundreds of women 60 years or older classified as “pregnant.”
If the death certificate’s pregnancy checkbox is checked, which means that the person was pregnant at the time of death, then it is considered a maternal death under the NVSS/NCHS maternal death definition. However, using a precise definition of maternal deaths, which requires pregnancy to be mentioned by the medical certifier as a cause of death on the death certificate, a major 2024 study found that maternal mortality was stable between the periods 1999-2002 and 2018-2021, with a little over 10 deaths per 100,000 live births in each period. Using this estimate, US maternal mortality is comparable to reported UK mortality, and lower than reported Canadian mortality (compare to Figure 1, above). It’s also comparable to the OECD average.[1]
The study further found a “substantial reduction in direct obstetric causes of death,” which suggests that maternity care in the US is improving, an encouraging finding. The authors conclude that the high maternal mortality in the US and the apparent increase in maternal mortality are likely due to the misclassification of incidental and non-maternal deaths as maternal deaths under current checkbox use.[2]
Previous studies come to similar conclusions, including an NCHS report which found that previous maternal mortality increases up until 2017 “were largely due to the staggered implementation of the checkbox,” and an earlier study which found the checkbox produced large false-positives among women 40 or older and drove most of the observed increase in maternal mortality between the years 1998-2002 and 2008–2012.
In addition to changes to data collection, another factor complicating interpretation is that the US collects additional pregnancy and postpartum measures, which differ from the standard international maternal mortality measure.[3]
While maternal deaths include deaths up to 42 days postpartum from causes related to or aggravated by pregnancy, pregnancy-related mortality captures deaths from causes related to or aggravated by pregnancy up to 1 year postpartum, and pregnancy-associated mortality captures deaths from any cause up to 1 year postpartum.[4] The latter two measures are far broader than maternal deaths, either in terms of the period over which they track deaths or the deaths they count as “maternal.”
This difference can lead to confusion. For example, a recent The Week article states there were “32.6 deaths per 100,000 live births in 2022” in the US, which is the same pregnancy-related mortality figure that appears in the JAMA study. The Week article apparently compares this figure to the UK, Canada, and Norway’s maternal mortality rate, which results in an erroneous and overstated cross-country disparity.[5]
Another headline from earlier this year describes homicide/suicide as the #1 cause of “maternal death.” However, homicide and suicide do not qualify as maternal deaths under the NCHS or World Health Organization definitions, and instead are pregnancy-associated deaths.
The figures reported in both the LiveScience article and the JAMA study rely on death certificate data, so they are subject to the relevant data challenges.[6] Additionally, the JAMA study reports pregnancy-related deaths, which—even setting false positives related to the pregnancy checkbox aside—are higher than maternal mortality rates for the same years.
Pregnancy and postpartum death measures have policy implications. If pregnancy and postpartum deaths far outpace those in other nations, or have grown quickly, then policymakers and healthcare providers will be eager to identify reforms and prioritize resources accordingly.
Alternatively, if pregnancy and postpartum deaths are in line with rates in other nations and have been stable over time, then policymakers’ approach will be more measured. Fortunately, compelling evidence suggests that maternal deaths have been largely stable long-term. For various reasons, policymakers and the public should remain cautious when interpreting figures related to pregnancy and postpartum death.
[1] International comparisons should be treated cautiously, as data-collection methods vary across countries.
[2] Where non-maternal deaths are defined as “deaths of non-pregnant women” and incidental deaths are deaths “by a cause unrelated to and unaffected by the pregnant state and its management.”
[3] Figure 1 is not affected by this issue.
[4] Consistent with the Pregnancy Mortality Surveillance System; Centers for Disease Control and Prevention definition of pregnancy-associated mortality.
[5] The article cites 2022 international numbers, but the figures line up best with international maternal mortality numbers in a Commonwealth chart where data is pulled from a variety of dates.
[6] The JAMA study notes these limitations and advises that coding issues may still exist despite the authors limiting the study to 2018 and after (by 2018, all states finished pregnancy checkbox implementation and NCHS changed their coding rules in an effort to mitigate errors and overestimation).