State-Level Prevalence of Maternity Care Deserts: Association With Healthcare Access, Utilization, and Outcomes Among Medicaid Recipients

Abstract

Introduction

More than 2 million women of childbearing age live in U.S. counties without access to maternity care—counties referred to as maternity care deserts. The objective of this study was to identify whether living in a state with a high prevalence of maternity care deserts is associated with maternal and child health access and utilization and contributing to adverse health outcomes among Medicaid recipients.

Methods

This is a descriptive, cross-sectional, secondary data analysis using most recent data available (2021–2023) from multiple existing state-level sources (e.g., Medicaid Health Care Core Data Sets, March of Dimes Peristats, Pregnancy Risk Assessment Monitoring System). State-level analyses included all U.S. states, District of Columbia, and Puerto Rico (N=52). Analyses were stratified by state prevalence of maternity care deserts (high=states where ≥21% of counties have no hospital/birth center providing obstetric services). Multivariate analyses of covariance were used to test associations with 3 domains of dependent variables: healthcare access (i.e., mental health, pediatrics, family medicine), utilization (e.g., timely access to pre and postnatal care, vaccinations), and maternal/child health outcomes (e.g., maternal and infant mortality).

Results

Three multivariate analysis of covariance tests indicated significant main effects for state-level maternity care desert prevalence and all domains (all p<0.009), with large effect sizes (η2≥0.14). Effects were durable, even after controlling for state-level covariates. States/territories with high prevalence of maternity care deserts had more rural counties (t= −4.22, p<0.001), lower levels of educational attainment (t= −2.32, p=0.024), and lower median household income (t=3.09, p=0.004). Those living in states/territories with high prevalence of maternity care deserts had 30% fewer mental healthcare providers (487.50 vs 696.50/100,000 population) and 36% fewer pediatricians (14.25 vs 22.31/100,000 population). They were ∼14% less likely to receive timely prenatal and postpartum care. Infants were 28.62% less likely to receive influenza vaccinations before second birthday. Critically, those in states/territories with high prevalence of maternity care deserts had 34.20% greater risk of maternal mortality, 18.34% greater risk of infant mortality, and 8.92% greater risk of low birthweight.

Conclusions

High state-level prevalence of maternity care deserts is associated with adverse maternal and child outcomes; exacerbating concerns as health during the perinatal period sets the trajectory for health across the lifespan. Policy and practice solutions must address pervasive health inequities associated with maternity care deserts (e.g., expanding access to maternity care through family medicine physicians, doulas, telehealth access, and supporting rural obstetric readiness).

Highlights

• •Restricted access to prenatal, pediatric, and mental healthcare co-occurs.
• •Maternity care deserts are associated with late entry to prenatal and postpartum care.
• •State prevalence of maternity care deserts is related to maternal mortality.
• •Maternity care deserts are related to low birthweight and infant mortality.

Graphical abstract

INTRODUCTION

Despite spending more per capita on health care, maternal and child mortality rates in the U.S. far exceed the rates in other high-income countries—and are rising.¹ An estimated 80% of pregnancy-related deaths in the U.S. are preventable, and infant mortality risks could be reduced with greater accessibility and utilization of perinatal healthcare services.^2,3

Workforce shortages and obstetric unit closures have led to vastly medically underserved geographic areas. This is a growing concern for maternity care access across the U.S., with 52% of rural hospitals and 36% of urban hospitals not offering obstetric care in 2022.⁴ As defined by the nonprofit organization, the March of Dimes, a maternity care desert refers to a county in the U.S. without a hospital or birth center that offers obstetric care nor services from obstetricians, certified nurse midwives, nurse midwives, and family physicians who reported delivering babies.⁵ Nearly one third of counties in the U.S. are classified as maternity care deserts; therefore, 2.2 million women of childbearing age live in counties without access to maternity care. These are predominantly located in Southern and Midwestern states and disproportionately impact rural communities, those underinsured, and people of color.⁵ Systematic factors such as race/ethnicity characterize maternal and child health outcomes within maternity care deserts.⁶ Underlying structural factors such as implicit biases in health care, residential racial segregation, limited insurance, and socioeconomic barriers contribute to health inequities.⁷

Access to and utilization of healthcare services are vital to ensure healthy and safe pregnancies and to potentially address risks for the mother and baby. Delayed care contributes to adverse health outcomes throughout the perinatal period, such as pre-eclampsia, gestational diabetes, and increased risk of admission to neonatal intensive care.^8,9 Women in maternity care deserts in Louisiana had a 91% increased risk of death during pregnancy and up to 1 year after childbirth, regardless of age, race/ethnicity, or SES.¹⁰ Maternal mental health is also a critical issue across the perinatal period, with 1 in 5 patients diagnosed with conditions such as mood or anxiety disorders.¹¹ Despite these concerns, there is limited existing research on the health impacts of living in a maternity care desert.

Specific aims are to identify whether living in a state with a higher prevalence of maternity care deserts is associated with (1) access to mental health providers, pediatricians, and family medicine physicians; (2) utilization of postpartum care, mental health services, and pediatric care; and (3) health outcomes among Medicaid beneficiaries. Medicaid coverage is determined at the state level and provides access to care throughout the state; therefore, this geographic unit of analysis is relevant to inform policy development and implementation.¹²

METHODS

Study Sample

This study is a descriptive, cross-sectional, secondary data analysis. Multiple data sources provide comprehensive and multifactorial understanding of associations between state prevalence of maternity care deserts and specified public health outcomes. Data sources and variables are summarized in detail in Table 1,13, 14, 15, 16, 17, 18, 19, 20, 21, 22 with links to the most recent publicly available data (2021–2023). When available, all U.S. states, District of Columbia, and Puerto Rico were included (N=52). Some data sources were unavailable in all states/territories; this are noted in tables and figures. Data were subset to include Medicaid enrollees, representing 42% of U.S. births; however, data for maternal and infant mortality could not be subset to Medicaid enrollees owing to data-use restrictions.

Table 1.

Data Sources and Variables

Data source (years)	Variables	Mean number, percentage, or rate (SD)
Sociodemographic variables
March of Dimes (2023)13	National estimated number of women aged 15–44 years	1,452,421
	State-level population of women aged 15–44 years by race/ethnicity	White: 62.68% (16.87) Black: 12.39% (11.09) American Indian or Alaska Native: 1.73% (3.24) Asian: 5.70% (6.15) Hispanic: 14.43% (11.70)
Centers for Medicare and Medicaid Services (2024)14	National percentage of Medicaid recipients of reproductive age (aged 15–49 years)	26%
Centers for Medicare and Medicaid Services (2024)14	State-level average of Medicaid coverage of births	42%
Data and Analytics Aggregator: PolicyMap, U.S. Census American Community Survey (2018–2022, aggregated)15	State-level educational attainment	High-school diploma or less: 36.97 (5.38) Some college or higher: 63.03 (5.38)
	State-level estimated median household income	$73,770 (14,247)
	State-level percentage of people living in poverty	12.93% (4.89%)
U.S. Department of Agriculture, Rural–Urban Continuum Codes (2023)16	State-level urban–rural classification	Nonmetropolitan (rural) counties: 55.02% (24.30)
Healthcare access
March of Dimes (2023)13	Maternity care access (based on county level)	Maternity care deserts: 1,026 (31.89%) Low or moderate access: 406 (12.62%) Full access: 1,763 (54.80%)
Area Health Resources Files (2017–2022, mental health; 2022–2023 pediatric and family)17,18	State average rate of mental healthcare providers (combined total average: psychiatrists; psychologists; clinical and counseling psychologists; counselors; substance abuse, behavioral disorder, and mental health providers; rehabilitation counselors; social workers; and mental health and substance abuse social workers)	585.85 (per 100,000)
	State average rate of pediatricians	18.72 (per 100,000)
	State average rate of family medicine physicians	32.17 (per 100,000)
Healthcare utilization
Medicaid Child and Adult Health Care Core Sets19 (Federal Fiscal Year 2022) Released by Centers for Medicare and Medicaid Services	Timeliness of prenatal care (percentage of women delivering a live birth with a prenatal care visit in the first trimester, on or before the enrollment start date, or within 42 days of enrollment in Medicaid or CHIP)	76.61% (16.27)
	Timeliness of postpartum care (percentage of Medicaid recipient women delivering a live birth who had a postpartum care visit on or between 7 and 84 days after delivery)	68.73% (14.55)
	Well-child visits in the first 30 months of life	55.72% (10.17)
	Child developmental screening (developmental screening in the first 3 years of life)	39.91% (18.68)
	Percentage infants who had an MMR vaccination by second birthday	80.49% (7.56)
	Percentage infants who had ≥2 influenza vaccinations by second birthday	45.36% (12.39)
PRAMS20a (2020–2021)	Received postpartum mental health screening	18.99% (3.65)
Health outcomes
CDC WONDER21 (2016–2021)	Gestational hypertension	9.49% (1.68)
	Gestational diabetes	8.29% (1.79)
	Infant birthweight, % (<2,500 grams)	10.31% (1.54)
	Infant mortality rate (per 1,000 live births)	5.55 (1.34)
CDC and National Center for Health Statistics22 (2018–2021)	Maternal mortality rate (per 100,000 live births)	25.48 (9.29)
PRAMS20a (2020–2021)	Prenatal depression (diagnosed health problem during pregnancy)	26.20% (7.38)
	Self-reported postpartum depression symptoms: modified PHQ-2 (2020–2021)	18.23% (3.45)

Note: Data were obtained from CDC PRAMS (https://www.cdc.gov/prams/php/methodology/index.html#:∼:text=PRAMS%20Weighting%20Process&text=Thus%2C%20the%20data%20are%20weighted,frame%20of%20eligible%20birth%20certificates).

^aPRAMS data are weighted by state jurisdiction in which each jurisdiction stratifies a random sample of 100–250 new mothers monthly from a group of eligible birth certificates. Therefore, the PRAMS aggregate data are not weighted to provide national estimates.

CDC, Centers for Disease Control and Prevention; CHIP, Children's Health Insurance Program; MMR, measles, mumps, and rubella; PHQ-2, Patient Health Questionnaire 2; PRAMS, Pregnancy Risk Assessment Monitoring System; WONDER, Wide-ranging Online Data for Epidemiologic Research.

Measures

State-level prevalence of maternity care deserts served as the primary exposure variable for all analyses. Analyses were stratified by state prevalence of maternity care access using 2023 data from March of Dimes Peristats, which defines maternity care desert as any county in the U.S. without a hospital or birth center that offers obstetric care nor any services from obstetric providers (e.g., obstetricians, certified nurse midwives, family physicians who deliver obstetric care). The March of Dimes describes 4 categories of access to prenatal care: (1) no access (i.e., maternity care desert), (2) low access, (3) moderate access, and (4) full access.¹³ Taking a conservative analytic approach, the authors evaluated states on the basis of counties with no access versus those with any access (low–medium–full access combined). The authors then dichotomized the states/territories on the basis of the percentage of counties classified as maternity care deserts: low prevalence refers to states/territories where <21% of counties are maternity care deserts (n=25), and high prevalence refers to states/territories where ≥21% of the counties are maternity care deserts (n=27). The cut off of 21% represents the median state-level percentage of counties designated as maternity care deserts. This analytic approach was selected to provide a straightforward method to obtain state-level ORs that are most relevant for clinical guidelines and policy development.

Assuring that the median split provided a valid and reliable designation of maternity care desert prevalence, the authors compared this state-level designation with March of Dimes county-level designation. This validity check affirmed near perfect alignment, with only 4 states moving from high to low prevalence of maternity care deserts (Indiana, Louisiana, North Carolina, and Wyoming).

To operationalize accessibility to other healthcare services—specifically pediatricians, family medicine physicians, and mental health providers—data from the 2022–2023 Area Health Resource Files were used.^17,18 Median split designated high and low state-level prevalence of these care providers: 17 of 100,000 population for pediatricians; 30 of 100,000 for family medicine physicians; and 531 of 100,000 population for mental health providers. Median values of pediatricians and family physicians are in accordance with median rates of pediatricians (16.69 of 100,000 population) and family physicians (31.38 of 100,000 population) on the basis of data from the U.S. Bureau of Labor Statistics.^23,24 Mental health providers included a broad spectrum of providers (psychiatrists; psychologists; clinical and counseling psychologists; counselors; substance abuse, behavioral disorder, and mental health providers; rehabilitation counselors; social workers; and mental health and substance abuse social workers). To assess validity of using the median split for mental health provider access, the authors compared with Health Resources and Services Administration–designated percentage of need met, documenting 89% alignment.²⁵

Adult and Child Health Care Quality Core Set Measurements from the Centers for Medicare and Medicaid Services provided data on healthcare utilization for Medicaid recipients. This includes timeliness of prenatal care, timeliness of postpartum care, well-child visit within the first 30 months of life, child developmental screening in the first 3 years of life, and child vaccinations (i.e., MMR [measles, mumps, rubella], influenza).¹⁹ Postpartum depression screening was obtained from the Pregnancy Risk Assessment Monitoring System (PRAMS), and data were stratified to include only Medicaid recipients.²⁰ Further details are provided in Table 1.13, 14, 15, 16, 17, 18, 19, 20, 21, 22

Gestational hypertension, diabetes, and maternal mortality (rate per 100,000 live births) were collected through the Centers for Disease Control and Prevention (CDC) National Vital Statistics System.²² Patient Health Questionnaire 2 from PRAMS was used to assess postpartum depressive symptoms.²⁰ Natality data included live births, stratified by state and Medicaid as the payment source; infant mortality (rate per 1,000 live births); and low birthweight from CDC WONDER (Wide-ranging Online Data for Epidemiologic Research).²¹ Variables were subset to only include Medicaid enrollees, except for the maternal mortality rate and infant mortality rate, owing to data-use restrictions.

State-level social and demographic variables included median household income; percentage of American Indian/Alaska Native and Asian females aged 15–44 years; high-school diploma or less/some college education or more; and nonmetropolitan counties. These data were obtained from March of Dimes Peristats,¹³ American Community Survey (aggregated 2018–2022),¹⁵ and the U.S. Department of Agriculture (urban/rural designation).¹⁶

Statistical Analysis

Bivariate associations were conducted to evaluate the associations between state-level maternity care desert prevalence designation and each sociodemographic, access, utilization, and health outcome variable. Multivariate analysis of covariance (MANCOVA) was performed, with state prevalence of maternity care deserts (high versus low prevalence) as the independent variable and access, utilization, and health outcome variables as dependent variables within each corresponding model. MANCOVA is used with a dichotomous independent variable and multiple dependent variables, providing good estimation of associations within the most parsimonious models.²⁶ Covariates for the multivariate analyses were selected on the basis of bivariate state-level differences between states with high and low prevalence of maternity care deserts. p<0.05 indicated statistically significant main effects for MANCOVA, and eta-squared threshold values >0.14 represented large effects.²⁷ This analytic approach has been used in prior maternal and child health studies.²⁸ SAS, Version 9.4 (SAS Institute, Cary, NC), was used with ArcGIS (Esri, Redlands, CA) to develop maps illustrating access.

RESULTS

States/territories with high prevalence of maternity care deserts have more rural counties (t= −4.22, p<0.001), lower levels of educational attainment (t= −2.32, p=0.024), and lower median household income (t=3.09, p=0.004) than states/territories with low prevalence. American Indian/Alaskan Native women were 4 times more likely to be living in states with high prevalence of counties as designated maternity care deserts (t= −2.45, p=0.020); Asian women were 50% less likely (t=2.36, p=0.026). On the basis of bivariate state-level differences by high versus low prevalence of maternity care deserts, variables that were statistically significant were included as covariates in each of the 3 MANCOVA models. The control variables included median household income, percentage of American Indian/Alaskan Native and Asian females aged 15–44 years, education, and nonmetropolitan (i.e., rural) counties.

Low maternity care access refers to states/territories where ≥21% of the counties are maternity care deserts; 27 states/territories were designated low access, and 25 were designated high access (IQR=7.08–44.43). Maps illustrate that patterns of maternity care, mental health, and pediatrician shortages co-occur and are primarily located in Southern and Midwestern states (Figure 1). The family medicine shortage pattern differs, with the greatest gaps in the Northeast, West, and South. Figure 2 provides a visual summary of healthcare access across the 52 geographic areas.

Figure 1.

Access to healthcare services in the U.S., based on the median value of the rate of providers for each of the designated healthcare specialties.

Note: Light blue=low access, dark blue=high access, and gray=unavailable data.

Figure 2.

State variability of access to healthcare services in the U.S.

Note: Light blue=low access, dark blue=high access, and gray=unavailable data. Four states have high access to all types of healthcare services assessed: New Hampshire, Oregon, Vermont, and Washington. Seven states have low access to all types of healthcare services assessed: Alabama, Kentucky, Mississippi, Nevada, Oklahoma, Tennessee, and Texas.

Binary associations and mean difference between state-level prevalence of maternity care deserts and healthcare access variables reaffirm that states/territories with high levels of maternity care deserts also had significantly lower access to mental health providers and pediatricians (both p≤0.001) (Table 2). Specifically, those living in states/territories with high prevalence of counties designated as maternity care deserts had 30% fewer mental healthcare providers (487.50 vs 696.50 per 100,000 population) and 36% fewer pediatricians (14.25 vs 22.31 per 100,000 population). These results were stable even after controlling for covariates in the MANCOVA (Wilks’ Lambda: F[8,42]=0.47, F=5.95, p<0.001). There was a large effect size (η2≥0.14).²⁷ There was no association for family medicine access.

Table 2.

Bivariate Analyses and MANCOVA Estimating Impact of Maternity Care Desert State Prevalence^a

	Bivariate analyses				MANCOVAb
	Low prevalence of counties designated as maternity care deserts, mean (SD), n	High prevalence of counties designated as maternity care deserts, mean (SD), n	t-value	p-value	F-value	p-value	η2
Accessc
Mental health provider rate (per 100,000)	696.50 (209.90) n=24	487.50 (116.30) n=27	4.32	<0.001	19.93	<0.001	0.29
Pediatric healthcare access rate (per 100,000)	22.41 (10.66) n=25	14.25 (3.40) n=27	3.66	0.001	13.28	0.001	0.21
Family healthcare access rate (per 100,000)	29.76 (10.17) n=25	34.40 (9.28) n=27	−1.72	0.092	—	—	—
Utilizationd
Timeliness of prenatal care (n=44)	81.70 (8.93) n=25	70.56 (20.71) n=21	2.29	0.030	5.22	0.028	0.13
Timeliness of postpartum care (n=48)	73.53 (11.30) n=24	62.97 (16.14) n=20	2.55	0.015	13.11	0.001	0.27
Postpartum mental health screening (n=43)	7.99 (2.84) n=21	19.94 (4.12) n=22	−1.80	0.079	—	—	—
Well-child visits first 30 months of life (n=48)	57.79 (10.92) n=24	53.65 (9.11) n=24	1.42	0.161	—	—	—
Child developmental screening (n=37)	44.16 (19.64) n=18	35.89 (17.29) n=19	1.36	0.182	—	—	—
MMR vaccination by their second birthday (%) (n=45)	80.63 (8.13) n=22	80.37 (7.15) n=23	0.11	0.909	—	—	—
≥2 influenza vaccinations by their second birthday (n=45)	49.62 (13.93) n=22	41.29 (9.30) n=23	2.37	0.022	8.74	0.006	0.20
Health outcomese
Maternal mortality rate (per 100,000 live births) (n=37)	19.89 (6.51) n=17	30.23 (8.72) n=20	−4.02	<0.001	16.20	<0.001	0.32
Prenatal depression (n=37)	25.82 (8.83) n=20	26.55 (5.97) n=22	−0.31	0.755	—	—	—
Self-reported postpartum depression symptoms (n=37)	17.29 (2.80) n=21	19.12 (3.82) n=22	−1.78	0.083	—	—	—
Gestational hypertension (rate) (n=51)	9.14 (1.79) n=24	9.79 (1.55) n=27	−1.40	0.168	—	—	—
Gestational diabetes (rate) (n=51)	8.67 (1.47) n=24	7.95 (1.99) n=27	1.45	0.155	—	—	—
Low birthweight (n=52)	9.80 (1.44) n=24	10.76 (1.51) n=27	−2.32	0.024	8.63	0.006	0.20
Infant mortality rate (per 1,000 live births) (n=49)	4.94 (1.06) n=22	6.05 (1.36) n=27	−3.16	0.003	7.66	0.009	0.18

^aData represent percentage values for study variables subset to only Medicaid enrollees (except for maternal mortality rate and infant mortality rate where subsetting was restricted). Degrees of freedom for each model=1.

^bOn the basis of state-level differences by maternity care desert, control variables include median household income and percentage of AIAN and Asian females aged 15–44 years, high-school diploma or less/some college education or more, nonmetropolitan counties.

^cAnalyses restricted on the basis of data availability: N=52, except for mental health provider rate (n=51, Puerto Rico unavailable).

^dAnalyses restricted on the basis of data availability; the following states/territories were excluded from utilization MANCOVA: Alabama, Alaska, Arkansas, District of Columbia, Florida, Idaho, Illinois, Indiana, Iowa, Maine, Maryland, Montana, and Puerto Rico.

^eAnalyses restricted on the basis of data availability; the following states/territories were excluded from outcomes MANCOVA: Alaska, Delaware, District of Columbia, Hawaii, Idaho, Maine, Montana, New Hampshire, Puerto Rico, Rhode Island, South Dakota, Vermont, West Virginia, and Wyoming.

AIAN, American Indian and Alaska Native; MANCOVA, multivariate analysis of covariance; MMR, measles, mumps, and rubella.

Bivariate analyses demonstrated statistically significant and clinically meaningful association between maternity care desert prevalence and percentage of Medicaid enrollees receiving timely prenatal (70.56% vs 81.70%; t=2.29, p=0.030) and postpartum (62.97% vs 73.53%; t=2.55, p=0.015) care (Table 2). Children enrolled in Medicaid in states/territories with more maternity care deserts were 28.62% less likely to receive timely influenza vaccinations (41.29 vs 29.62; t=2.37, p= 0.022). There were no differences in bivariate analyses regarding associations of state prevalence of maternity care deserts with Medicaid enrollees receiving postpartum mental health screenings, well-child visits in first 30 months of life, child developmental screenings, and MMR vaccinations. MANCOVA for utilization variables demonstrated significant main effects for state prevalence of maternity care deserts on delayed prenatal and postpartum care (both ∼14% greater likelihood) and childhood influenza vaccination, controlling for sociodemographic covariates (Wilks’ Lambda: F[9,28]=0.50, F=3.17, p=0.009), with a large effect size (η2≥0.14).²⁷

States/territories with high prevalence of maternity care deserts had 34.20% greater risk of maternal mortality (30.23 vs 19.89 per 100,000 population; t= −4.02, p<0.001), 18.34% greater risk of infant mortality (6.05 vs 4.94 per 1,000 population; t= −3.16, p=0.003), and 8.92% greater risk of low birthweight (<2,500 grams; 10.76% vs 9.80%; t= −2.32, p=0.024) (Table 2 and Figure 3). On the basis of data-use restrictions, rates of maternal and infant mortality include the entire state population. Medicaid (versus private insurance) was associated with increased risk of maternal and infant morbidity and mortality; therefore, this is a more conservative estimate.²³ There were no differences in bivariate analyses regarding association of state prevalence of maternity care deserts with prenatal depressive symptoms, gestational hypertension, or gestational diabetes. There were significant main effects for state prevalence of maternity care deserts on maternal and child health outcomes, controlling for sociodemographic covariates in the MANCOVA (Wilks’ Lambda: F[9,27]=0.29, F=7.34, p<0.001) with a large effect size (η2≥0.14).²⁷

Figure 3.

Maternal and infant mortality rates and low infant birthweight based on state prevalence of maternity care deserts.

^aSubsetting to only include Medicaid enrollees was restricted owing to data availability for the maternal mortality rate and infant mortality rate.

^bOn the basis of state-level differences by maternity care desert, control variables include median household income and percentage of AIAN and Asian females aged 15–44 years, high-school diploma or less/some college education or more, and nonmetropolitan counties.

^cAnalyses restricted on the basis of data availability; the following states/territories were excluded from outcomes MANCOVA: Alaska, Delaware, District of Columbia, Hawaii, Idaho, Maine, Montana, New Hampshire, Puerto Rico, Rhode Island, South Dakota, Vermont, West Virginia, and Wyoming.

AIAN, American Indian and Alaska Native; MANCOVA, multivariate analysis of covariance.

DISCUSSION

Median proportion of counties designated as maternity care deserts states was 21%, representing >1 million women in the U.S. States/territories with a higher prevalence of maternity care deserts have disproportionately lower rates of mental health providers and pediatricians, indicating a pattern of restricted healthcare access and likely contributing to observable differences in healthcare utilization—such as timely access to pre and postnatal care—and health outcomes. Of greatest concern are associations between state prevalence of maternity care deserts and maternal and infant mortality rates: an increase of 34% and 18%, respectively. In nationwide analyses, the authors document for the first time that women and infants living in states/territories with higher prevalence of maternity care deserts are associated with a greater risk of dying than those who live in states/territories with a lower prevalence. These results have direct implications for critical practices, programs, and policies to improve maternal and child health.

The association between state prevalence of maternity care deserts and timeliness of postpartum care is one important contributor. Fifty-three percent of pregnancy-related deaths occur between 7 days and 1 year after childbirth. To address the underlying health conditions that contribute to pregnancy-related deaths, the American College of Obstetricians and Gynecologists recommends that postpartum women have an initial assessment with their obstetrician-gynecologist within the first 3 weeks after childbirth, followed by a comprehensive postpartum visit within 12 weeks after birth. Many perinatal complications, such as hypertension or diabetes, worsen over time, with short- and long-term impacts. Patients with cesarean or perineal wounds are at higher risk for infection and other complications.²⁹ A large proportion of U.S. women do not attend at least 1 postpartum visit, with Medicaid patients less likely to receive postpartum care.³⁰ Given that Medicaid-insured women living in maternity care deserts are less likely to receive timely postpartum care, they are at greater risk for potential complications that may arise during the postpartum period.

Limited access to maternity care increases stress for pregnant patients traveling long distances to receive care.³¹ Untreated perinatal mental health conditions contribute to pre-eclampsia, challenges in adhering to perinatal guidelines, tobacco and alcohol exposure, poor nutrition, and reduced quality of family relationships.³² The American College of Obstetricians and Gynecologists recommends depression and anxiety screening during pregnancy and after birth: at first obstetric visit, during 24–28 weeks gestation, and at comprehensive postpartum visits. It is crucial to detect and treat mental health conditions early after childbirth, when depression onset most often occurs.³³

High state prevalence of maternity care deserts was significantly associated with low birthweight, a leading cause of infant mortality and recognized as a reference point to understand current and future health. Therefore, findings may portend poorer child health trends within these states/territories over time, including respiratory, cardiovascular, developmental, and cognitive problems.³⁴ Patterns of diminished access to maternal, mental health and pediatric care pose challenges to address overlapping health concerns for mothers and children.

Associations of state prevalence of maternity care deserts with infant mortality and low birthweight were durable, even after controlling for state-level covariates such as percentage of nonmetropolitan counties and median household income. Concerns are growing that the healthcare workforce will be unable to meet the growing demands of the pediatric population. Pediatric inpatient units decreased by 19% between 2008 and 2018, impacting nearly one quarter of children in the U.S., with children in rural areas facing the greatest declines.³⁵

Maternity care deserts are most often located within low-income communities, rural communities, and those with greater proportion of American Indian/Alaskan Native heritage.⁵ These social, economic, and racial structural inequities impact quality of care across the perinatal period and are associated with adverse maternal and child health outcomes.³⁶ When the state-level sociodemographic covariates were included as covariates in the multivariate models, findings remained statistically significant for many of the access, utilization, and outcomes variables. This suggests that factors related to race/ethnicity, rurality, and SES contribute to the prevalence of maternity care deserts, although they are not sufficient to account for all of the variance in adverse outcomes.

The prevalence of maternity care deserts continues to increase, according to the 2024 March of Dimes report “Nowhere to Go: Maternity Care Deserts Across the United States.”³⁹ There are several ways to consider service expansion in states/territories with high prevalence of maternity care deserts. Among those with limited access to maternity, mental health, and pediatric healthcare, the authors found no difference for family physicians. Family physicians are trained and credentialed to pair maternal and infant healthcare services, yet fewer than 7% of family physicians deliver babies.^40,41 Likewise, not all nurse practitioners work in obstetrics; however, they are an important part of the prenatal care workforce who could be mobilized to provide maternity care services, particularly in states that have limited access to prenatal care. Expanding access to doula services could be another strategy, particularly among socioeconomically vulnerable and minoritized communities. Currently, 15 state/territory Medicaid programs cover doula costs⁴²; yet, these services are lacking in states with high prevalence of maternity care deserts.

Community health centers⁴³ and the Rural Maternity and Obstetrics Management Strategies Program⁴⁴ should be expanded to improve access to maternity care in states/territories with high prevalence of maternity care deserts. Telemedicine can be effectively integrated into obstetric practices to facilitate prenatal and postpartum care.⁴⁵ Policies and programs enacted on national and state levels support the infrastructure for expanded access to maternity telehealth services. One such initiative is the Data Mapping to Save Moms’ Lives Act, which allows researchers to identify and map areas where there is a need for both increased maternal care and technology services across the U.S. Supporting investments in telehealth services could expand maternity care access.⁴⁶

Transdisciplinary training in mental health should be strengthened to improve care across the perinatal period. In February 2024, the Health Resources and Services Administration released goals intended to support the White House Blueprint for Addressing the Maternal Health Crisis, including increased access to comprehensive high-quality maternal health services that incorporate behavioral health to treat perinatal mental health conditions.⁴⁷ Expanding training and educational opportunities to maternity care providers is vital to improve maternal health, particularly in states/territories that lack access to maternity and mental health care.

Other policy solutions include implementation and sustainability of maternal mortality review committees that review underlying causes of pregnancy-related deaths, racial/ethnic or socioeconomic inequities, and a more in-depth understanding of perinatal complications.⁴⁸ State Perinatal Quality Collaboratives are evidence-based state or multistate networks of teams of perinatal health professionals who work toward improving the quality of care for mothers and babies.⁴⁹ The CDC currently supports 36 state-based collaboratives. The White House Blueprint suggests that these collaboratives should operate at full capacity within all states.⁴⁷ Finally, hospitals without obstetric units must ensure that guidelines are enforced to be obstetric ready, especially within rural communities.

Limitations

This paper focused on state-level prevalence of maternity care deserts rather than county level, which diminishes the ability to identify specific factors that influence maternal and child healthcare access, utilization, and outcomes at a more local level. Although this is a coarser measurement, the authors identified important associations that are statistically significant and clinically meaningful. Moreover, state-level analysis is advantageous when considering broad policy implications (e.g., Medicaid eligibility based on state residence). State-level social and economic policies have the potential to reduce adverse perinatal and infant health outcomes and health disparities in the U.S.³⁷ Nonetheless, smaller units of analysis (e.g., county, town) would be useful to identify structural inequities, target healthcare shortages, and allocate essential resources.

Geospatial analysis was not conducted; therefore, the authors did not evaluate how distance to hospitals providing obstetric services affects access, utilization, and outcomes. A recent study found that longer driving distances to delivery hospitals is associated with a greater risk of adverse maternal health outcomes and admission to the neonatal intensive care unit.³⁸ Unfortunately, the authors did not have access to neonatal intensive care unit data, so this was not assessed in this study. Future research should evaluate the association between distance to the nearest care facility and perinatal outcomes.

CONCLUSIONS

Results provide substantial insight toward understanding the impact of living in a state with a high prevalence of maternity care deserts and important implications to improve maternal and child health within the U.S. Future research should examine the underlying structural factors that contribute to limited access to care and corresponding health outcomes within states/territories with high prevalence of maternity care deserts. Community-level characteristics reflect social and environmental contexts where families live and how these factors influence maternal and child healthcare access, utilization, and outcomes. Future research could explore how health-promoting factors such as walkability, green space, and other neighborhood-level factors are associated with maternity care deserts and, consequently, with healthcare access, utilization, and outcomes.

Following the Dobbs v. Jackson Women’s Health Organization court decision to overturn Roe v. Wade in 2022, experts recognize that abortion restrictions impact many other reproductive health outcomes.⁵⁰ Abortion restrictions are associated with adverse maternal health outcomes, including serious complications such as eclampsia, anxiety, remaining with abusive partners, and maternal death.⁵¹ States with the most restrictive abortion access also have high prevalence of maternity care deserts (Idaho, Texas, Oklahoma, Tennessee). Furthermore, obstetricians are leaving states with restrictive abortion laws that will result in expanded or new maternity care deserts, further restricting care within states that already have limited access.⁵²

This study documents lack of vital healthcare resources to support maternal and child health. Evaluating gaps in access and utilization of healthcare services and differences in health outcomes based on state prevalence of maternity care deserts provides insight into pervasive maternal and child health inequities in the U.S. and disproportionately higher maternal and infant morbidity and mortality rates than in all other high-income nations. This is the first study to examine the nationwide state-level associations between maternity care deserts and maternal and infant mortality. Therefore, the authors identify policy strategies and solutions to address these urgent concerns across the U.S. Healthcare during the perinatal period is essential to ensure safe and healthy perinatal experiences and to set the trajectory for health across the lifespan.