Geographic access to critical care obstetrics for women of reproductive age by race and ethnicity

Abstract

BACKGROUND:

The goal of risk-appropriate maternal care is for high-risk pregnant women to receive specialized obstetrical services in facilities equipped with capabilities and staffing to provide care or transfer to facilities with resources available to provide care. In the United States, geographic access to critical care obstetrics varies. It is unknown whether this variation in proximity to critical care obstetrics differs by race, ethnicity, and region.

OBJECTIVE:

We examined the geographic access, defined as residence within 50 miles of a facility capable of providing risk-appropriate critical care obstetrics services for women of reproductive age, by distribution of race and ethnicity.

STUDY DESIGN:

Descriptive spatial analysis was used to assess geographic distance to critical care obstetrics for women of reproductive age by race and ethnicity. Data were analyzed geographically: nationally, by the Department of Health and Human Services regions, and by all 50 states and the District of Columbia. Dot density analysis was used to visualize geographic distributions of women by residence and critical care obstetrics facilities across the United States. Proximity analysis defined the proportion of women living within an approximate 50-mile radius of facilities. Source data included the 2015 American Community Survey from the United States Census Bureau and the 2015 American Hospital Association Annual Survey.

RESULTS:

Geographic access to critical care obstetrics was the greatest for Asian and Pacific Islander women of reproductive age (95.8%), followed by black (93.5%), Hispanic (91.4%), and white women of reproductive age (89.1%). American Indian and Alaska Native women had more limited geographic access (66%) in all regions. Visualization of proximity to critical care obstetrics indicated that facilities were predominantly located in urban areas, which may limit access to women in frontier or rural areas of states including nationally recognized reservations where larger proportions of white women and American Indian and Alaska Native women reside, respectively.

CONCLUSION:

Disparities in proximity to critical care obstetrics exist in rural and frontier areas of the United States, which affect white women and American Indian and Alaska Native women, primarily. Examining insurance coverage, interstate hospital referral networks, and transportation barriers may provide further insight into critical care obstetrics accessibility. Further exploring the role of other equity-based measures of access on disparities beyond geography is warranted.

Introduction

Population projections indicate that more than half of all Americans will belong to a minority group by 2044,¹ and with these changing demographics, examining access to care is necessary to ensure equitable distribution of services. A recent review of access to and quality of care in the United States for racial or ethnic minorities, low-income groups, and those residing in rural areas indicates receipt of suboptimal care and offers examples of programs, interventions, and policies to address disparities.² Addressing disparities in care, particularly access to care, is critical for pregnant and postpartum women as nearly a quarter of births in the United States are to women who self-identify as members of racial or ethnic minority groups.³

Maternal morbidity and mortality are influenced by racial and ethnic inequities in access to quality healthcare.^4–11 Pregnancy-related mortality ratios (PRMRs; ie, pregnancy-related deaths per 100,000 live births) from 2007 to 2016 for black women and American Indian/Alaska Native (AI/AN) women were 3.2 and 2.3 times higher, respectively, than the PRMR for white women, and larger disparities existed when examining PRMR by age and race.¹² Higher rates of chronic conditions, infections, and medical complications contribute to disparities by race and ethnicity in severe maternal morbidity and pregnancy-related mortality.^13,14 However, few studies attribute disparities to the differences in the quality of care or to the healthcare delivery systems within which obstetrical care is provided.¹⁵ In 1 study that does examine these differences, Creanga et al¹⁶ (2014), using pooled 2008 to 2011 data from 7 states, found an overall lower performance of black-serving than white- or Hispanic-serving hospitals on delivery-related indicators and higher rates of delivery complications among black than white or Hispanic women in primarily white- and Hispanic-serving hospitals, confirming the critical need for systems to ensure delivery of appropriate obstetrical care.^17,18 To better ensure quality care and decrease such disparities, understanding equity within the care continuum is essential to standardizing perinatal care, including labor and delivery services.^19,20

In 2015, the American College of Obstetricians and Gynecologists (ACOG) and Society for Maternal-Fetal Medicine (SMFM) published an obstetrical care consensus recognizing the need for levels of maternal care.²¹ The goal of risk-appropriate maternal care is for high-risk pregnant women to receive services in facilities that are equipped to provide the required level of specialized care needed to adequately care for these women.²¹ In 2017, ACOG and SMFM piloted a maternal levels of care verification program involving facility-specific site visits in 3 states. A toolkit to scale up the program is pending.²² Concurrently, the Centers for Disease Control and Prevention developed the Levels of Care Assessment Tool (LOCATe) to provide a standardized assessment of delivery facilities that aligned with the 2015 ACOG and SMFM and 2012 American Academy of Pediatrics guidance on maternal and neonatal risk-appropriate care.²³ To date, 15 states and 1 territory have implemented LOCATe.²⁴

Although previous research has focused on access to care among specific populations, including neonates,²⁵ no study has specifically examined the racial or ethnic distribution of women of reproductive age and distance to facilities that provide the highest level of specialized care for obstetrical patients. Critical care obstetrics (CCO) is care provided during pregnancy or after delivery that may require intensive monitoring by more than what is available in an obstetrical unit, including medical intervention or diagnostic imaging at a facility appropriately equipped for high-risk patients.²⁶ Although no specific guidance exists on transport for CCO, transfer is recommended when a patient is clinically unstable, is deteriorating, or requires specialized services.^21,27 To better understand geographic distance to CCO facilities for women of reproductive age, we conducted a spatial and proximity analysis of CCO to identify where potential gaps in geographic access are occurring in the United States and the impact by race and ethnicity. To do this, we examined population distributions and proximity to CCO by race and ethnicity of women of reproductive age in regions and states in the United States. We then visualized distance to care to determine geographic access.

Materials and Methods

Data sources

We used the publicly available 2015 American Community Survey data from the United States Census Bureau²⁸ to determine the geographic distribution of women of reproductive age, defined as ages from 15 to 44 years. We assumed pregnant women were equally distributed across the population of women of reproductive age. Race and ethnicity were categorized using the United States Census Bureau single-race category definitions²⁸; race was categorized as “white only,” “black only,” “AI/AN” only, and “Asian/Pacific Islander” (A/PI) only. Native Hawaiian was included in the A/PI race category for Hawaii and California. Ethnicity was categorized as “Hispanic” and “non-Hispanic” and was not mutually exclusive because it included women who self-identified race; therefore, analysis of ethnicity was conducted separately. CCO facilities were identified using the 2015 American Hospital Association (AHA) Annual Survey.²⁹ Specifically, the AHA variable OBLev was used because it corresponded to a level III obstetrical unit or CCO defined as “provides services for all serious illnesses and abnormalities and is supervised by a full-time maternal-fetal specialist.”²⁹

Geospatial analysis

A descriptive spatial analysis was used to assess geographic proximity to CCO (ie, distance from residence to facility) for women of reproductive age by race and ethnicity.^30,31 Data were analyzed at 2 geographic levels: national and the 50 states and the District of Columbia. Dot density population distributions were used to visualize the geographic distributions of the subpopulations relative to CCO across the nation, and proximity analysis was used to define the proportion of each subpopulation living within approximately 50 miles of the nearest CCO unit or facility.

Buffer zones of 50 miles in Euclidean distance were created around each CCO facility.^32–34 Euclidean distance buffers use straight-line distance (as the crow flies) and appear on a map as perfect circles around a point. Buffers were translated into census tracts and shaded on maps to indicate access zones. Zones that overlapped were merged to form confluent zones and presented in maps to visualize zones within and across state lines. When the closest facility within 50 miles was across a state line, access zones within the state of interest were included in both analysis and visualization.

A census tract with its population centroid contained in an access zone was defined as a tract with geographic access and appropriately identified on the map. Women of reproductive age living within these tracts were defined as having geographic access to CCO, and proportions of women with and without access were calculated for each racial and ethnic subpopulation. Because distance is considered a barrier to accessing trauma care^35–37 with emergency transport to appropriate facilities based on field triage clinical standards,³⁸ 50 miles was selected based on the concept of the golden hour because no specific guidelines exist for the recommended ground travel to facilities with CCO, although evidence for the golden hour has been debated in recent decades.^39,40

For easier visualization, maps are presented by the US Department of Health and Human Services (HHS) regions. A data table also provides a summary by region and state. For interpretation, we refer to metropolitan areas (ie, areas with population of ≥50,000) as urban areas and micropolitan areas (ie, areas with population of <50,000) as rural areas using the core-based statistical area definitions of the United States Census Bureau.⁴¹ We describe frontier areas as those with <7 persons per square mile.⁴²

Statistical analysis

Numeric counts and proportions are presented for women of reproductive age within and outside of the 50-mile radius concentric circles. Our study is descriptive and uses census data for the total population size, based on location. We did not control for fertility rate by race or ethnicity and for population size. Our study did not require institutional review board approval from the Centers for Disease Control and Prevention because the study did not include human subjects and was considered public health practice. All analysis procedures were conducted using SAS versions 9.3 and 9.4 (SAS Institute, Cary, NC)⁴³ and ArcGIS version 10.5 (Esri, Redlands, CA) software.^33,44

Results

In 2015, >90% of women of reproductive age in the United States had geographic access to CCO, although access varied by race, state, and HHS region (Table). Overall, geographic access, or residence within 50 miles of a CCO facility, was greatest for A/PI women (95.8%), followed by black women (93.5%) and then white women (89.1%). AI/AN women had the least geographic access to CCO at 66%. Most Hispanic women had geographic access to CCO (91.4%). Of note, 3 states based on geographic size and number of CCO facilities, offered 100% geographic access for all racial and ethnic groups; Connecticut, Rhode Island, and the District of Columbia provided geographic access to CCO for all women of reproductive age. By contrast, in the state of Wyoming, no racial or ethnic group had geographic access to CCO because no delivery facilities offered that level of care. Women of reproductive age with high-risk pregnancies residing in Wyoming were required to travel more than 50 miles to a neighboring state for the closest CCO services. When examining states with fewer CCO facilities or facilities in bordering states indicating limited geographic access by race, no state had <50% geographic access to CCO for black women. Arkansas, Montana, and South Dakota had <50% access to CCO for A/PI women, and Montana and South Dakota had <50% geographic access for white women. Alaska, Arkansas, Mississippi, Montana, New Mexico, North Dakota, and South Dakota had <50% geographic access for AI/AN women, and when examining ethnicity, Arkansas, North Dakota, and South Dakota had <50% geographic access for Hispanic women.

Table.

Geographic access to CCO for women of reproductive age residing within a 50-mile radius of a CCO facility by state, race, and ethnicity (2015)^a

States	All women at the age of 15–44 y by race and ethnicity			White		Black		AI/AN		Asian/PI		Hispanic
	Total countb	Access countc (N)	%	N	%	N	%	N	%	N	%	N	%
United States	63,127,294	56,942,218	90.2	39,178,629	89.1	8,401,869	93.5	369,556	66.0	3,921,068	95.8	11,389,910	91.4
Region I	2,861,031	2,813,528	98.3	2,186,898	98.0	222,720	99.6	8534	92.6	179,058	99.6	354,783	99.7
Connecticut	686,245	686,245	100	493,084	100	84,710	100	2356	100	40,111	100	126,725	100
Maine	236,084	218,603	92.6	204,842	92.7	2868	94.7	1256	67.5	4184	97.3	4457	92.5
Massachusetts	1,366,223	1,353,048	99.0	1,009,329	98.8	113,563	99.5	3106	100	112,626	99.6	176,104	99.7
New Hampshire	244,823	238,265	97.3	219,256	97.2	3680	98.8	338	90.1	8743	99.5	9717	98.9
Rhode Island	210,986	210,986	100	161,617	100	16,270	100	1113	100	10,584	100	35,288	100
Vermont	116,670	106,381	91.2	98,770	91.0	1629	93.9	365	89.9	2810	95.9	2492	93.2
Region II	5,774,663	5,670,915	98.2	3,399,224	90.3	956,919	99.3	18,134	93.2	584,526	99.5	1,220,807	99.4
New Jersey	1,726,126	1,699,577	98.5	1,061,906	98.3	264,300	98.3	3524	98.7	189,930	98.9	382,108	98.8
New York	4,048,537	3,971,338	98.1	2,337,318	97.1	692,619	99.8	14,610	92.0	394,596	99.7	838,699	99.7
Region III	5,989,170	5,773,813	96.4	3,929,492	95.4	1,162,409	98.5	13,841	96.4	338,091	99.2	521,784	98.7
Delaware	179,986	174,637	97.0	111,578	96.1	44,493	99.1	501	78.9	8501	98.3	18,773	97.7
District of Columbia	174,954	174,954	100	81,512	100	70,655	100	755	100	10,072	100	17,482	100
Maryland	1,200,200	1,188,493	99.0	620,301	98.5	392,193	99.6	3418	99.5	86,503	99.8	124,305	99.6
Pennsylvania	2,418,792	2,291,658	94.7	1,766,911	93.9	303,053	97.9	4273	93.9	103,088	98.3	186,533	97.9
Virginia	1,677,523	1,641,117	97.8	1,068,019	97.5	341,265	97.7	4523	98.3	126,326	99.7	169,653	99.1
West Virginia	337,715	302,954	89.7	281,171	89.4	10,750	94.2	371	95.4	3601	93.5	5038	92.0
Region IV	12,379,983	11,125,169	89.9	7,433,573	89.9	2,760,968	89.1	45,960	78.0	349,643	94.3	1,560,925	94.7
Alabama	961,054	806,699	83.9	517,670	83.8	245,665	83.8	4007	80.4	14,735	89.6	35,962	85.2
Florida	3,661,231	3,366,389	91.9	2,375,051	91.9	662,693	91.2	9410	86.3	119,180	94.8	986,789	96.7
Georgia	2,099,365	1,884,577	89.8	1,032,684	89.1	658,484	89.9	4972	90.5	90,021	96.3	194,303	92.9
Kentucky	854,349	708,319	82.9	606,385	82.3	64,333	86.8	1363	77.1	13,474	85.5	25,980	84.7
Mississippi	601,391	418,660	69.6	237,866	73.3	167,500	65.5	970	32.8	4862	63.9	13,343	75.5
North Carolina	1,974,618	1,834,425	92.9	1,213,758	94.0	433,461	91.2	18,674	72.2	63,496	95.4	183,486	91.7
South Carolina	937,777	891,265	95.0	560,995	94.4	279,374	96.5	2987	96.5	16,602	95.9	52,209	94.7
Tennessee	1,290,197	1,214,835	94.2	889,164	92.6	249,458	99.1	3577	93.4	27,273	97.1	68,853	96.2
Region V	10,140,197	9,392,249	92.6	7,110,672	91.3	1,296,722	98.5	35,287	75.5	413,565	96.0	902,770	97.2
Illinois	2,604,061	2,423,761	93.1	1,623,816	91.0	401,922	97.3	5664	94.8	158,504	98.3	488,410	98.3
Indiana	1,291,074	1,191,802	92.3	965,483	91.6	133,344	97.7	2628	96.7	29,514	86.7	91,310	96.7
Michigan	1,892,788	1,788,066	94.5	1,329,936	93.3	304,608	99.4	8155	73.2	68,960	97.9	102,562	96.8
Minnesota	1,049,651	914,760	87.1	731,622	85.9	68,099	96.8	6725	56.5	64,566	96.0	55,643	89.5
Ohio	2,214,451	2,137,832	96.5	1,693,613	95.9	306,352	99.6	4075	95.1	57,017	97.9	86,166	98.8
Wisconsin	1,088,172	936,028	86.0	766,202	84.5	82,397	98.9	8040	75.0	35,004	88.8	78,679	95.2
Region VI	8,247,445	6,891,241	83.6	4,799,060	82.3	1,080,720	87.8	88,940	62.8	323,118	92.9	2,325,177	86.0
Arkansas	575,660	301,976	52.5	218,977	51.1	67,299	64.9	1432	35.9	5039	39.0	14,910	32.6
Louisiana	939,821	850,836	90.5	498,715	89.8	305,725	91.9	4103	69.8	19,105	94.0	42,403	92.4
New Mexico	398,836	237,979	59.7	171,368	61.8	5218	66.8	13,907	32.1	5857	75.4	135,462	64.8
Oklahoma	755,547	609,010	80.6	426,506	80.7	53,191	86.3	44,759	74.2	19,058	89.2	64,908	79.3
Texas	5,577,581	4,891,440	87.7	3,483,494	86.2	649,287	89.5	24,739	87.9	274,059	96.1	2,067,494	89.1
Region VII	2,681,337	2,187,595	81.6	1,789,233	80.1	216,945	92.5	13,440	81.0	73,579	89.9	159,418	78.4
Iowa	584,642	518,534	88.7	462,285	88.6	19,119	86.4	2110	91.9	16,781	91.0	34,399	90.8
Kansas	558,931	428,660	76.7	350,984	75.6	31,503	89.6	4896	82.7	18,113	85.3	49,209	68.7
Missouri	1,175,395	953,975	81.2	735,446	78.5	147,688	93.6	3630	74.6	28,814	91.9	43,515	83.3
Nebraska	362,369	286,426	79.0	240,518	77.1	18,635	96.2	2804	79.7	9871	91.4	32,295	77.8
Region VIII	2,292,969	1,742,672	76.0	1,460,894	75.5	53,480	92.1	23,411	38.8	67,410	89.2	319,843	85.8
Colorado	1,072,351	952,364	88.8	773,940	87.8	43,302	97.8	9013	77.9	40,219	95.9	227,467	89.9
Montana	183,607	82,487	44.9	72,812	45.5	400	60.2	4712	33.9	1266	49.7	3853	53.2
North Dakota	139,429	71,170	51.0	63,051	51.9	1610	60.4	3232	38.2	1511	58.9	1645	36.5
South Dakota	156,853	59,462	37.9	52,653	40.8	1700	62.0	1578	9.9	1044	39.6	2412	40.0
Utah	631,168	577,189	91.4	498,438	91.3	6468	95.7	4876	62.8	23,370	96.1	84,466	93.5
Wyomingd	109,561	0	0	0	0	0	0	0	0	0	0	0	0
Region IX	10,141,212	9,230,141	91.0	5,456,083	90.3	579,304	95.9	90,411	66.1	1,395,798	95.1	3,731,361	90.2
Arizona	1,296,676	1,136,974	87.7	865,797	89.5	58,952	96.3	34,185	50.5	50,963	95.4	406,624	87.0
California	8,017,349	7,371,604	91.9	4,216,079	90.6	462,448	95.5	50,767	82.6	1,197,028	96.8	3,129,920	90.6
Hawaii	266,402	191,711	72.0	40,545	65.8	5583	93.1	383	57.7	94,386	76.7	20,750	67.6
Nevada	560,785	529,852	94.5	333,662	92.9	52,321	99.3	5076	73.4	53,421	99.0	174,067	96.0
Region X	2,619,288	2,114,895	80.7	1,613,500	79.3	71,682	94.9	31,598	56.7	196,280	93.3	293,042	81.2
Alaska	146,929	87,276	59.4	58,992	63.2	3541	78.8	6145	29.3	8819	70.6	7322	66.2
Idaho	311,315	184,698	59.3	166,784	59.1	1573	78.1	2389	53.5	4656	74.7	23,558	55.3
Oregon	770,666	526,802	68.4	419,275	66.3	13,129	86.7	5841	60.0	37,373	83.3	85,937	75.8
Washington	1,390,378	1,316,119	94.7	968,449	94.3	53,439	99.2	17,223	84.0	145,432	99.1	176,225	91.0

AI/AN, American Indian/Alaskan Native; CCO, critical care obstetrics; PI, Pacific Islander.

^aGeographic access is defined as residential access within 50 miles of a CCO facility;

^bTotal counts represent all women of reproductive age in that state;

^cAccess counts represent all women of reproductive age in that state who have geographic access to or live within a 50-mile radius of a CCO facility;

^dThere are no women with access to a CCO facility within the state of Wyoming.

By region, >90% of black women of reproductive age had geographic access to CCO in 8 of 10 regions. Similarly, >90% of A/PI women in 8 regions had geographic access to CCO. Half of all regions (5/10) had >90% geographic access for white women, and few regions (3/10) had >90% access for AI/AN women. For Hispanic women, 6 of 10 regions had >90% geographic access.

Geospatial visualization of access by race and ethnicity

Overall, when visualizing geographic access by tract, a larger proportion of white women had no geographic access to CCO than other races, represented by density dots (Figure 1). However, all racial and ethnic groups had some areas of no access (Figure 2, A and B; Figure 3, A and B). More specifically, groups of white women had no access to CCO in rural areas of all states except Connecticut and Rhode Island (Figure 2, A and B). Groups of black women in region IV (Alabama, Georgia, Florida, Kentucky, Mississippi, and North Carolina) and region VI (Arkansas, Louisiana, Oklahoma, and Texas) had no geographic access to CCO. Among AI/AN women, groups had no geographic access in regions IV (Mississippi and North Carolina), V (Michigan, Minnesota, and Wisconsin), VI (New Mexico and Oklahoma), VIII (Colorado, Montana, North Dakota, South Dakota, Utah, and Wyoming), IX (Arizona and California), and X (Alaska, Idaho, and Washington). In all regions, women with no geographic access to CCO primarily resided in rural or frontier areas of states or on nationally recognized reservations.

FIGURE 1. Proximity of women of reproductive age to hospitals with CCO services by race and ethnicity, United States (2015).

The top map depicts distribution of race, and the bottom map depicts distribution of Hispanic ethnicity. Top map: Black women are represented by dark blue dots; white women are represented by light blue dots; 1 dot=500 women; except for AI/AN women, represented by pink dots, where 1 dot=250 women. A/PI women are represented by green dots. CCO hospitals, that is, hospitals with critical care obstetrics services, are represented by a red cross. Tracts refer to United States Census Tracts with population centroid within 50 miles of CCO hospital, represented in tan. Bottom map: Hispanic women are represented by dark blue dots; non-Hispanic women are represented by light blue dots; 1 dot=500 women. Note: Each racial and ethnic population layer is drawn separately in a descending order based on the proportion that the population represents in the total population of women of reproductive age; therefore, in urban areas where the populations are heavily clustered, the lower layers (layers with the higher total number of women) may be partially obscured.

A/PI, Asian and Native Hawaiian/Pacific Islander; AI/AN, American Indian/Alaska Native; CCO, critical care obstetrics.

FIGURE 2. Proximity of women of reproductive age to hospitals with CCO services by race, HHS regions (2015).

A, The top map depicts distribution of race in HHS regions I and II, and the bottom map depicts distribution of race in HHS regions III and V. B, The top map depicts distribution of race in HHS region IV, and the bottom map depicts distribution of race in HHS regions VI, VII, VIII, IX, and X. Black women are represented by dark blue dots. White women are represented by light blue dots; 1 dot=500 women; except for AI/AN women represented by pink dots where 1 dot=250 women. A/PI women are represented by green dots. CCO hospital refers to hospital with critical care obstetrics services, represented by a red cross. Tracts refer to United States Census Tracts with population centroid within 50 miles of CCO hospital, represented in tan. Each racial and ethnic population layer is drawn separately in a descending order based on the proportion that the population represents in the total population of women of reproductive age; therefore, in urban areas where the populations are heavily clustered, the lower layers (layers with the higher total number of women) may be partially obscured.

A/PI, Asian and Native Hawaiian/Pacific Islander; AI/AN, American Indian/Alaska Native; HHS, Department of Health and Human Services; CCO, critical care obstetrics.

FIGURE 3. Proximity of women of reproductive age to hospitals with CCO services by Hispanic ethnicity, HHS regions (2015).

A, The top map depicts distribution of ethnicity in HHS regions I and II, and the bottom map depicts distribution of ethnicity in HHS regions III and V. B, The top map depicts distribution of ethnicity in HHS region IV, and the bottom map depicts distribution of ethnicity in HHS regions VI, VII, VIII, IX, and X. Hispanic women are represented by dark blue dots. Non-Hispanic women are represented by light blue dots; 1 dot=500 women. CCO hospital refers to hospital with critical care obstetrical services, represented by a red cross. Tracts refer to United States Census Tracts with population centroid within 50 miles of CCO hospital, represented in tan. Each racial and ethnic population layer is drawn separately in a descending order based on the proportion that the population represents in the total population of women of reproductive age; therefore, in urban areas where the populations are heavily clustered, the lower layers (layers with the higher total number of women) may be partially obscured.

HHS, Department of Health and Human Services; CCO, critical care obstetrics.

A larger proportion of Hispanic women had geographic access by tract than non-Hispanic women (Figure 3, A and B). Among Hispanic women, some had no geographic access in regions VI (Arkansas, New Mexico, Oklahoma, and Texas), VII (Iowa, Kansas, Missouri, and Nebraska), VIII (Colorado, Utah, and Wyoming), IX (Arizona, California, and Nevada), and X (Alaska, Idaho, Oregon, and Washington).

Comment

Principal findings

Although most women of reproductive age, regardless of race or ethnicity, have geographic access to CCO, our findings indicate that a lower proportion of white and AI/AN women have geographic access than A/PI and black women. These findings may be attributed to a larger proportion of white and AI/AN women residing in rural and frontier areas with smaller, widely distributed populations who are often lacking access to a wide range of services readily accessible to those in urban settings.^45,46 Geographic access may be especially vital for AI/AN women living in isolated, disenfranchised areas, similar to other rural populations with high rates of poverty and overall poorer health outcomes,^45,47,48 who forego medical care owing to limited service availability.⁴⁹ Understanding differences in geographic access for AI/AN women and the impact on delivery of care is critical given the higher risks for severe maternal morbidities than other racial or ethnic groups.⁵⁰

Clinical implications

Research on access to healthcare in rural areas indicates that disparities in care for those residing in these areas are associated with poverty, age, overall health status, or lack of insurance coverage, rather than race or ethnicity.^47,48 However, racial and ethnic disparities exist among adults who identify a usual source of routine healthcare in rural areas,⁴⁹ and rural isolation may contribute to further inequities.⁵¹ This urban and rural disparity is similar to our previous study²⁵ and other studies concluding that rurality is associated with limited access to a range of healthcare services and resources^45,52 resulting in potential care receipt in low-volume hospitals and increased risks of maternal morbidities, such as postpartum hemorrhage, severe perineal lacerations, and wound infections,⁵³ and out-of-hospital preterm births (ie, preterm births that occur in birth centers or home births).⁵⁴ Moreover, infants born to women residing in rural areas have a higher risk of perinatal mortality and neonatal intensive care unit admissions.⁵⁵

In addition, provision of maternity care in rural areas is affected by hospital closures, shortages of specialty physicians, and lack of transportation options.^56,57 Several efforts proposed to enhance access to care for rural residing women include increasing the number of programs that provide incentives for physicians to practice in rural areas,⁵⁶ expanding the role of nurse-midwives and other midlevel provider types,⁵² utilizing telemedicine or telehealth programs,^52,58–60 developing state policies to provide reimbursement for maternal transport,⁶¹ and implementing strategies to support risk-appropriate care.²¹ The passage of federal efforts, such as the Improving Access to Maternity Care Act (H.R. 315)⁶² enacted in 2018, further supports the maternity care needs of women residing in geographically isolated areas to improve outcomes by identifying maternity care provider shortage areas and working to add care providers to those areas. Although adding maternity care providers to rural areas may not address the lack of CCO facilities, it may offer a network of care provision for identifying and monitoring obstetrical conditions of high-risk women in need of critical services before childbirth. Moreover, the reach of this network may provide regionalized clinical care linkages for women at the highest risk of severe maternal morbidities with obstetrical specialists who can monitor health conditions and arrange transfer to CCO facilities for labor and delivery.

Research implications

Although most black women have proximal geographic access to CCO, which should result in improved care and birth outcomes,¹⁷ they nonetheless have the highest rates of maternal morbidity, mortality, and adverse birth outcomes.^63,64 These higher rates occur even among college-educated black women than high school–educated white women and persist by socioeconomic status, with black women delivering more preterm and low birthweight infants than white women.^65,66 To better understand the disparities in health outcomes for black women and infants, future research may examine other equity-based measures beyond geographic proximity to CCO, such as social accessibility, focusing on the effects of institutional racism or distribution of social service programs. Finally, although most women of reproductive age have geographic access, the quality and resultant outcomes may be influenced by a range of interrelated and complex social, economic, and environmental factors (ie, education, community resilience).^67,68 Health systems play a role in the quality of care and potential outcomes; hospitals primarily serving black women performed worse on most delivery-related indicators (ie, complicated vaginal delivery, complicated cesarean delivery, obstetrical trauma) than hospitals primarily serving white women.¹⁶ Hospitals serving a disproportionate number of black women were also found to have worse patient safety records for both black and white patients.⁶⁹ Other studies suggest that provider racial bias, poor patient-provider communication, and perceived discrimination are associated with poorer quality maternity care and outcomes.^70,71 A number of strategies proposed to reduce provider bias, including cultural competence training,⁷² implicit bias training for medical students and physicians,⁷³ and increasing efforts to diversify the healthcare workforce,^74,75 may potentially increase quality care for all women.

Strengths and limitations

Although our analysis represents a novel approach to visualizing geographic access to CCO, some limitations exist in our study. First, we defined physical, ready access to CCO as a travel distance within 50 miles of a facility. We did not account for traffic flow, facility location, or other geographic barriers such as rivers or mountains; with minimal research on distance, the 50-mile boundary may or may not be the largest distance a patient⁷⁶ or emergency medical services would travel in an emergency.³⁹ Second, we did not consider the full spectrum of interrelated social, environmental, economic, and structural factors or referral networks that may influence access to services. We determined that most women are not inhibited by a lack of physical access to services unless residing in a rural or frontier area. Third, we did not examine the impact of insurance eligibility when crossing state lines for CCO. Depending on public or private insurance policies, the coverage of care may be affected by receipt of care in a different jurisdiction, affecting our categorization of access to CCO facilities when within 50 miles of residence, but across a state border. Fourth, we used women of reproductive age as a proxy for any women who may need geographic access to CCO; we do not have data on all high-risk pregnant and postpartum women who require direct and immediate access to CCO facilities to evaluate gaps in access. Finally, misclassification of critical care designations may have occurred because levels of maternal care designation policies vary across states, and women may live closer to a neighboring state CCO facility, so they may travel across state lines to receive care based on proximity. Geocoding of facility location is also estimated based on mailing address of facility in the AHA data and may not reflect the exact location of the facility. However, our results provide an initial comprehensive review of access for all women of reproductive age residing in the United States.

Conclusion

Our work indicates that most women of reproductive age have geographic access to CCO, although differences in physical access exist by race in rural and frontier areas of the United States. Given the variation in geographic access to health services, including CCO, providers of services in rural and frontier areas must account for distance in developing treatment plans for high-risk women, whereas providers of services in urban areas must consider factors other than physical access, particularly for black women. Further examination of the social determinants of health and other measures of equity or accessibility among this population may identify additional contributors to disparities. CCO is an essential service affecting maternal and infant outcomes and should be readily accessible to all women regardless of race, ethnicity, or residence.

AJOG at a Glance.

Why was this study conducted?

This study aimed to assess geographic access to critical care obstetrics for all women of reproductive age in the United States by race and ethnicity.

Key findings

Almost all women in the United States have geographic access to critical care obstetrics available within 50 miles of residence; however, some racial, ethnic, and regional differences exist.

What does this add to what is known?

We document racial, ethnic, and geographic disparities in proximity to critical care obstetrics and identify opportunities for improvement.