The Health Divide: The maternity care crisis worsens, plus stubborn gaps in U.S. fetal mortality rates

Author(s)
Published on
September 16, 2024

More than 35% of U.S. counties are ‘maternity care deserts’

More than 2.3 million women of reproductive age live in counties without a single birthing facility or obstetric clinician, according to the March of Dimes’ newly-published 2024 report.

The report offers a useful map highlighting the large swaths of the country without these vital services. States that have the highest percentages of counties without access to maternity care include North Dakota, South Dakota, Alaska, Oklahoma and Nebraska.

"For too many families across the U.S., the ability to have a healthy pregnancy depends on where they live," said Dr. Amanda Williams, March of Dimes’ chief medical officer, in a statement. "Our 2024 report underscores that maternity care is still not prioritized in our country and there is an urgent need for systemic changes to improve outcomes for moms and babies in the U.S. and to ensure that these families have access to the care they need and deserve."

Women living in places without adequate maternity care face a 13% higher risk of preterm birth, among other health complications, the report says. The rates of inadequate prenatal care are higher among women of color and low-income women.  

Recent hospital closures and reductions in obstetric services have accelerated the spread of these deserts. Since the organization last published a report in 2022, more than 100 hospitals have closed their obstetric units. As a result, pregnant people have to travel farther for maternity care, or seek emergency services. Among the solutions the report suggested are mobile health centers, which provide maternal and infant health services to areas otherwise lacking them.

A lack of OB-GYNs is adding to the problem, with physician burnout, lower reimbursement rates, and a desire for better work-life balance among the factors limiting the supply of specialists, according to a 2022-published article in the World Journal of Gynecology & Women’s Health. 

Another more recent factor shaping the OB-GYN workforce  is the shifting abortion landscape, as Alejandra O’Connell-Domenech reported earlier this year for The Hill. The Supreme Court’s decision to overturn federal protections for abortion had led to declines in applications to OB/GYN residency programs “far more significantly in the 13 states that enacted complete abortion bans after the Dobbs ruling," according to American Medical Association research.

CDC: Fetal mortality reaches historic low, but racial inequities persist 

Fetal deaths have reached a historic low, according to the National Center for Health Statistics’ newly released 2022 report.  

With so much public health and media attention focused on infant mortality, the latest measure — which looks at fetal deaths at 20 weeks gestation or more — marks progress on “a major but often overlooked public health issue,” the report notes. 

The fetal mortality rate dropped from 5.73 deaths per 1,000 births in 2021 to 5.48 in 2022, a 4% decline.

But a closer look within the numbers tells a different story of persistent disparities. For example, the most recent fetal mortality rate was highest for Black women and Native Hawaiian and Pacific Islanders remains nearly twice that experienced by white women — about 10 deaths per 1,000 compared to five. The lowest rate was found among Asian females. 

While the report does not explore reasons for the racial disparities, USA Today reporter Eduardo Cuevas wrote that experts have suggested overall “racial health disparities, preexisting conditions, structural discrimination and access to quality care may be factors.”

Usha Ranji, associate director for women’s health policy at the research nonprofit KFF, also pointed to the “weathering effect” of stress and other factors that influence adverse health outcomes for Black mothers and other women of color. 

Fetal mortality rates were highest in Mississippi (6.22), Arkansas (5.38), and Alabama (5.21). New Mexico, Texas, Connecticut and Montana were among the states with the lowest rates.

Among the risk factors described in the report were maternal age, smoking during pregnancy, and pregnancies carrying more than one fetus. The measure does not include abortions. 

For adults on Medicaid, where you live affects your vision coverage 

There are major disparities in vision coverage nationwide for adults with Medicaid health insurance. 

That’s the overarching finding of a new analysis recently published in Health Affairs. The findings offer a state-by-state analysis of adult Medicaid benefits for both fee-for-service and managed care plans.  Among the most striking statistics: About 27% of adult Medicaid enrollees live in states without eyeglass coverage. 

The study, which was based on state Medicaid policies from 2022 to 2023, also noted that 12% of adults with Medicaid, or 6.5 million people, live in states that didn’t cover routine eye exams. Even in the two-thirds of the states that did provide routine eye care, enrollees were required to pay a share of costs in some way. 

While the federal government establishes overarching rules for Medicaid, the national health insurance for low-income adults and children, states run their own programs. That means they can modify their coverage policies, such as whether eye exams and eyeglasses for adults are fully covered. (For children, federal law provides for vision services, regardless of the state.) 

The study highlights a patchwork of coverage nationwide, with seven states offering no coverage for exams or glasses for adults. 

These coverage gaps translate into real-world health impacts. Eye exams are vital in identifying eye diseases early, when treatment can prevent vision loss. Exams are also vital for prescription glasses to correct refractive errors, the leading cause of vision impairment in the country. Paying for your own exam without insurance costs about $485, the study notes. 

“Our study clearly shows that there are opportunities to expand coverage of routine vision services at the state level, and based on previous research, we expect more generous coverage would reduce rates of vision impairment, improve quality of life, and promote health equity,” said Brandy Lipton, the study’s lead author and associate professor of health, society and behavior at the University of California, Irvine in a release on the findings. 

How AI can avoid embedded health bias  

The data being used to build artificial intelligence could quickly compound the racial bias that already plagues the health care system, reports STAT’s Katie Palmer in the latest installment in the outlet’s “Embedded Bias” series.

But, the article asks, could these same new tech tools also deliver a better path forward? 

For example, Palmer describes how Olga Kravchenko, a biomedical informatician at the University of Pittsburgh, used ChatGPT to help identify patterns in ICU care. She found that patients labeled as Native American and Caucasian were more likely to use a ventilator compared to other racial groups. The reasons for the disparity are unknown but experts theorized that Indigenous Americans might arrive to the ICU sicker, while white patients might be given privileged access to ventilators.

“The answers to those questions are rarely clear,” Palmer writes. “But asking why a racial disparity appears in data is a critical step to ensuring that its signal doesn’t get misused in a predictive algorithm, for example, that helps hospitals determine who’s most likely to benefit from a limited supply of ventilators.”

With so much bias already baked into today’s health care data, solutions are complicated. One idea: “a ‘bias glossary’ for every medical dataset, a summary of the data distortions that responsible model developers should be careful to avoid.”

This month, the journal npj Digital Medicine also explored “fundamental shortcomings” in health care AI, and the challenges inherent in new models — such as “bias exhaust,” or the residual biases in AI that emerge from actual systemic discrepancies in care. Even though a new AI tool might appear accurate in a controlled testing environment, that might not translate to patient care at the bedside, the authors explain. They also point to the importance of providing developers “diverse, comprehensive clinical datasets to train models on actual patient populations and scenarios.”

What we’re reading