Abstract
Objective:
To compare rural obstetric patient and neonate characteristics and outcomes by birth location.
Methods:
Retrospective observational cohort study of rural residents’ hospital births from California, Pennsylvania, and South Carolina. Hospitals in rural counties were rural-located, those in metropolitan counties with ≥10% of obstetric patients from rural communities were rural-serving, metropolitan-located, others were non-rural-serving, metropolitan-located. Any adverse obstetric patient or neonatal outcomes were assessed with logistic regression accounting for patient characteristics, state, year, and hospital.
Results:
Of 466,896 rural patient births, 64.3% occurred in rural-located, 22.5% in rural-serving, metropolitan-located, and 13.1% in non-rural-serving, metropolitan-located hospitals. The odds of any adverse outcome increased in rural-serving (aOR 1.27, 95% CI 1.10–1.46) and non-rural-serving (aOR 1.35, 95% CI 1.18–1.55) metropolitan-located hospitals.
Conclusion:
One third of rural obstetric patients received care in metropolitan-located hospitals. These patients have higher comorbidity rates and higher odds of adverse outcomes likely reflecting referral for higher baseline illness severity.
Introduction
One fifth, nearly 60 million, Americans live in rural communities.1 Unfortunately, birthing people and infants living in rural areas suffer higher rates of adverse perinatal outcomes, including severe maternal morbidity (SMM), maternal mortality, neonatal mortality, and infant mortality.2–6 Access to perinatal (obstetric and neonatal) hospital services at the time of childbirth is critical to reducing adverse perinatal outcomes. In the United States (US), in-hospital birth and risk-appropriate perinatal care is associated with better outcomes, particularly for high-risk patients, such as preterm infants.7–10 However, more than half of rural counties in the US do not have local, in-county obstetric services available and even fewer have advanced neonatal services.11,12 With rural hospital closures and the associated loss of perinatal services, obstetric patients living in rural communities may not have local access to routine hospital care for childbirth.11,13 Although geographic access to perinatal care is a determinant of preterm birth, perinatal mortality, and neonatal mortality, little is known about where rural obstetric patients and their neonates receive care at the time of childbirth.14–16
While prior studies have considered hospitals with respect to their minority-serving status and associated outcomes, similar studies of rural populations are limited.17,18 Obstetric patients have some choice regarding where they give birth, though choice is often constrained by how far they are willing or able to travel, whether facilities are covered by their health insurance, and local referral networks. Given that 65% of birth hospitals are in metropolitan areas, these constraints are more pronounced for rural residents who have fewer, if any, local hospitals with perinatal services.12 The goal of regionalized perinatal care is to refer high-risk patients, including those living in rural areas, to hospitals with risk-appropriate services available, yet regionalization policies differ by region and state.19,20 The distribution, patient characteristics and outcomes of rural obstetric patients and neonates across different types of hospitals is unknown. Given that perinatal care encompasses obstetric and infant care, outcomes accounting for the health of both patients is critical when assessing perinatal care delivery. Thus, the objective of this study was to compare rural resident, both obstetric patient and neonate, characteristics and outcomes (any adverse maternal or neonatal outcome) between 1) rural-located, 2) rural-serving metropolitan-located, and 3) non-rural-serving metropolitan-located hospitals.
Methods
Study Population
This was a retrospective observational cohort study of rural residents using obstetric patient hospital discharge data linked to infant birth and death certificates from three states. Births in hospitals with ≥10 births/year in California from 2000–2012, Pennsylvania from 2000–2014, or South Carolina from 2000–2015 were included. These states were chosen because of linked obstetric patient-infant data, diverse populations, and distinct perinatal care systems. The study population was obstetric patients and their neonates who reside in rural communities, which was defined as a rural county of residence and identified using urban-adjacency influence codes (UIC). All non-metropolitan UICs, defined as those containing an urban core of <50,000 residents, were considered rural (UIC 3–12).21 The institutional review boards at the Children’s Hospital of Philadelphia and the departments of health in California, Pennsylvania, and South Carolina approved this study.
Study Definitions, Variables, and Outcomes
Birth hospitals were identified annually (hospital-year) to account for changes in hospital type and closures over time. Birth hospitals were categorized as 1) rural-located, those located in a rural county (UIC 3–12), 2) rural-serving, metropolitan-located, those located in a metropolitan county (UIC 1 or 2) with ≥10% of births from patients living in rural communities, or 3) non-rural-serving, metropolitan-located, those located in a metropolitan county with <10% of births from patients living in rural communities. We examined the distribution of hospitals and associated number of rural residents’ births in the analytic dataset to set the rural-serving threshold and chose the top decile, which corresponded to ≥10% of births from patients living in rural communities. To provide a more detailed description of hospital location, the 12 UIC codes were condensed into 5 categories: 1) large metropolitan area (≥1 million residents; [UIC 1]), 2) small metropolitan area (50,000–1 million residents, [UIC 2]), 3) micropolitan (10,000–49,999 residents; [UIC 3, 5, 8]), 4) non-core and contains a town (2,500–9,999 resident town; [UIC 4, 6, 9, 11]), and 5) non-core and does not contain a town (≤2,500 resident town; [UIC 7, 10, 12]). We examined the availability of obstetric services, indicated by any of the following three variables: provision of any obstetric services, indicating a level of obstetric, and ≥1 obstetric bed, though this approach may miss more than 10% of obstetric unit closures.22 Availability of neonatal services was assessed by the American Hospital Association (AHA) Annual survey variables indicating any intermediate or intensive neonatal care services. Hospital characteristics were obtained from the AHA data and included median number of births per year, hospital ownership, teaching status, community hospital designation, rural referral center designation, critical access hospital designation, core-base statistical area type, and county UIC descriptions.21
Obstetric patient and neonate variables of interest were obtained from the linked data. These obstetric patient demographics and conditions and neonate characteristics included, birth parent age, race/ethnicity, insurance, education, chronic hypertension, pregnancy-induced hypertension, severe pregnancy-induced hypertension/eclampsia, diabetes mellitus, gestational diabetes, placenta previa, multiple gestation, mode of delivery, infant sex, gestational age at birth, and birth weight.
The primary composite outcome was any adverse maternal or neonatal outcome, which encompasses maternal death, SMM, neonatal death, preventable fetal death, unexpected term complications, and preterm complications. Maternal mortality included death within 1 year of delivery. SMM was defined using the 21 indicators described by the Centers for Disease Control and Prevention.23 Neonatal mortality was defined as death before hospital discharge. Preventable fetal death was defined as fetal deaths ≥24 weeks’ gestation or ≥400 grams after hospital admission for management of pregnancy or labor. Unexpected newborn complications, as defined by the National Quality Forum, including both severe and moderate complications.24 Preterm complications included severe (grade III or IV) intraventricular hemorrhage, bronchopulmonary dysplasia, and retinopathy of prematurity. The individual components of the primary composite outcome were examined as secondary outcomes, including SMM with and without blood transfusion.
Statistical analysis
Obstetric patient, neonate, and hospital characteristics were compared by hospital type using Kruskal Wallis tests. The primary composite and secondary outcomes were examined as unadjusted rates and adjusted odds ratios, which were calculated using logistic regression. Models adjusted for obstetric patient and neonate characteristics and conditions (listed above, with the exception of birth weight given collinearity with gestation age at birth), birth year, and state, with hospital as a random effect. Analyses were completed using Stata version 17 (College Station, TX) and associated code is available upon request.
Results
Of 466,896 obstetric patient births, 64.3% occurred in rural-located hospitals, 22.5% in rural-serving, metropolitan-located hospitals, and 13.1% in non-rural-serving, metropolitan located hospitals (Table 1). Compared to non-rural-serving, metropolitan-located hospitals, obstetric patients who gave birth in rural-located hospitals were younger (12.7% vs 8.1% <20 years), more likely Medicaid-insured (54.8% vs 39.7%), and had less education attainment (15.2% vs 9.9% with some high school). The frequency of pre-existing and pregnancy-associated conditions, such as chronic and pregnancy-induced hypertension, was higher in metropolitan-located hospitals, with rates in non-rural-serving, metropolitan-located hospitals twice that of rural-located hospitals. Rates of cesarean birth among rural obstetric patients also varied from 29% in rural-located hospitals to 36% in non-rural-serving, metropolitan-located hospitals.
Table 1.
Birth Parent and Infant Characteristics by Hospital-Serving Status
| Rural-located N (%) |
Rural-serving, metropolitan-located N (%) |
Non-rural-serving, metropolitan- located N (%) |
P-value | |
|---|---|---|---|---|
| Births | 300,361 (64.3) | 105,161 (22.5) | 61,374 (13.1) | |
| Birth parent characteristics | ||||
| Age (years) | <0.001 | |||
| <20 | 38,094 (12.7) | 11,208 (10.7) | 4,940 (8.1) | |
| 20–24 | 94,875 (31.6) | 30,333 (28.8) | 14,979 (24.4) | |
| 25–34 | 139,403 (46.4) | 52,444 (49.9) | 32,799 (53.4) | |
| ≥35 | 27,959 (9.3) | 11,171 (10.6) | 8,651 (14.1) | |
| Race/ethnicity | <0.001 | |||
| Asian | 2,395 (0.8) | 528 (0.5) | 446 (0.7) | |
| Black | 42,728 (14.2) | 16,104 (15.3) | 8,448 (13.8) | |
| Hispanic | 48,077 (16.0) | 19,778 (18.8) | 8,439 (13.8) | |
| Other | 8,404 (2.8) | 2,346 (2.2) | 1,205 (2.0) | |
| White | 198,757 (66.2) | 66,405 (63.2) | 42,836 (69.8) | |
| Insurance | <0.001 | |||
| Private | 123,159 (41.0) | 52,453 (49.9) | 34,276 (55.9) | |
| Medicaid | 164,598 (54.8) | 47,586 (45.3) | 24,377 (39.7) | |
| Self Pay | 6,153 (2.1) | 2,943 (2.8) | 958 (1.6) | |
| Other | 6,451 (2.2) | 2,179 (2.1) | 1,763 (2.9) | |
| Education | <0.001 | |||
| No High School | 9,930 (3.3) | 2,481 (2.4) | 1,039 (1.7) | |
| Some High School | 45,661 (15.2) | 11,119 (10.6) | 6,079 (9.9) | |
| High School Diploma/GED | 102,204 (34.0) | 29,208 (27.8) | 16,331 (26.6) | |
| At Least Some College | 75,734 (25.2) | 25,160 (23.9) | 18,874 (30.8) | |
| 4 Year College | 29,545 (9.8) | 11,489 (10.9) | 9,345 (15.2) | |
| >4 Year College | 13,605 (4.5) | 5,457 (5.2) | 4,766 (7.8) | |
| Missing | 23,682 (7.9) | 20,247 (19.3) | 4,940 (8.1) | |
| Birth parent conditions | ||||
| Chronic hypertension | 3,454 (1.2) | 1,789 (1.7) | 1,403 (2.3) | <0.001 |
| Pregnancy-induced hypertension | 10,593 (3.5) | 4,664 (4.4) | 4,530 (7.4) | <0.001 |
| Severe pregnancy-induced hypertension/Eclampsia | 1,588 (0.5) | 902 (0.9) | 1,106 (1.8) | <0.001 |
| Gestational diabetes | 14,826 (4.9) | 5,781 (5.5) | 3,912 (6.4) | <0.001 |
| Diabetes mellitus | 2,113 (0.7) | 1,130 (1.1) | 1,121 (1.8) | <0.001 |
| Placental previa | 1,075 (0.4) | 480 (0.5) | 453 (0.7) | <0.001 |
| Multiple gestation | 4,380 (1.5) | 2,301 (2.2) | 2,798 (4.6) | <0.001 |
| Cesarean birth | 87,217 (29.0) | 32,475 (30.9) | 22,071 (36.0) | <0.001 |
| Infant characteristics | ||||
| Male | 153,608 (51.1) | 53,479 (50.9) | 31,511 (51.3) | 0.12 |
| Gestational age at birth (weeks) | <0.001 | |||
| <28 weeks | 675 (0.2) | 637 (0.6) | 1,076 (1.8) | |
| 28–31 | 1,158 (0.4) | 1,210 (1.2) | 1,950 (3.2) | |
| 32–36 | 20,277 (6.8) | 9,407 (9.0) | 8,667 (14.1) | |
| 37–41 | 268,680 (89.5) | 92,191 (87.7) | 48,382 (78.8) | |
| >41 | 4,868 (1.6) | 996 (1.0) | 687 (1.1) | |
| Missing | 4,703 (1.6) | 720 (0.7) | 612 (1.0) | |
| Birth weight | 3312.95 (533.8) | 3270.66 (605.6) | 3149.63 (750.8) | <0.001 |
Over the study period, there were 412 individual hospitals included in the cohort representing 3,555 hospital-years. The majority (60.0%) of rural-located hospitals were in micropolitan areas with the remaining 40.0% in non-core areas. The majority of rural-serving, metropolitan-located hospital-years (83.9%) were in small metropolitan areas and non-rural-serving, metropolitan-located hospital-years (64.7%) were in large metropolitan areas. The distribution of hospital-years mirrors the distribution of individual hospitals (Table 2). During the study period, more than 80% of hospital-years where rural residents gave birth, regardless of hospital type, were non-profit, non-teaching, community hospitals that were not critical access or rural referral centers (Table 2). Over the study period, availability of intermediate or intensive neonatal care services was limited across all hospital types (7.7% of rural-located, 22.6% of rural-serving, metropolitan-located, and 58% of non-rural-serving, metropolitan-located hospital-years).
Table 2.
Hospital Characteristics by Hospital-Serving Status
| Rural-located N (%) |
Rural-serving, metropolitan-located N (%) |
Non-rural-serving, metropolitan- located N (%) |
P-value | |
|---|---|---|---|---|
| Hospital-years | 896 (25.2) | 499 (14.0) | 2,160 (60.8) | |
| Hospital birth volume, median (IQR) | 295 (179–497) | 585 (306–1158) | 1509 (793–2612) | <0.001 |
| Ownership | <0.001 | |||
| Government | 149 (16.6) | 54 (10.8) | 225 (10.4) | |
| Non-Profit | 650 (72.5) | 352 (70.5) | 1,707 (79.0) | |
| Profit | 82 (9.2) | 93 (18.6) | 215 (10.0) | |
| Missing | 15 (1.7) | 0 (0.0) | 13 (0.6) | |
| Teaching Status | <0.001 | |||
| Non-teaching | 844 (94.2) | 426 (85.4) | 1,748 (80.9) | |
| Minor teaching | 37 (4.1) | 66 (13.2) | 239 (11.1) | |
| Major teaching | 0 (0.0) | 7 (1.4) | 160 (7.4) | |
| Missing | 15 (1.7) | 0 (0.0) | 13 (0.6) | |
| Community Hospital | 881 (98.3) | 498 (99.8) | 2,133 (98.8) | <0.001 |
| Critical Access Hospital | 164 (18.3) | 20 (4.0) | 1 (0.1) | <0.001 |
| Rural Referral Center | 168 (18.8) | 54 (10.8) | 90 (4.2) | <0.001 |
| Core Base Statistical Area Type | <0.001 | |||
| Metropolitan | NA | 407 (81.6) | 2,054 (95.1) | |
| Micropolitan | 586 (65.4) | 85 (17.0) | 66 (3.1) | |
| Rural | 277 (30.9) | NA | NA | |
| Missing | 33 (3.7) | 7 (1.4) | 40 (1.9) | |
| Urban Adjacency Influence Description | <0.001 | |||
| Large Metropolitan | NA | 54 (10.8) | 1,043 (48.3) | |
| Small Metropolitan | NA | 445 (89.2) | 1,117 (51.7) | |
| Micropolitan | 570 (63.6) | NA | NA | |
| Noncore with town | 299 (33.4) | NA | NA | |
| Noncore no town | 27 (3.0) | NA | NA | |
| Available perinatal services | ||||
| Obstetric services | <0.001 | |||
| Yes | 831 (92.8) | 449 (90.0) | 2,043 (94.6) | |
| No | 26 (2.9) | 0 (0.0) | 10 (0.5) | |
| Missing | 39 (4.4) | 50 (10.0) | 107 (5.0) | |
| Number of bassinets, median (IQR) | 10 (6–12) | 13(8–20) | 20 (12–30) | <0.001 |
| Neonatal intermediate or intensive care | 75 (8.4) | 103 (20.6) | 1,123 (52.0) | <0.001 |
Abbreviations: NA – not applicable.
In the study population overall, 42,905 (9.2%) rural maternal-neonatal dyads experienced an adverse outcome. The unadjusted rates of the primary composite outcome (any adverse outcome) differed by hospital type, increasing between rural-located (7.5%), rural-serving, metropolitan-located (9.1%), and non-rural-serving, metropolitan-located (17.8%) hospitals (Table 3). The most frequent secondary outcome and driver of the primary outcome was unexpected newborn complications, which occurred among 38,731 (8.3%) of rural residents overall and was 6.6% in rural-located hospitals, 8.2% in rural-serving, metropolitan-located hospitals, and 16.8% in non-rural-serving, metropolitan-located hospitals. Rates of maternal mortality and preventable fetal deaths to were too few to report by hospital type (<10 observations) and were not further examined as individual secondary outcomes.
Table 3.
Adverse Perinatal Outcomes by Hospital-Serving Status
| All rural residents N (%) |
Rural-located N (%) |
Rural-serving, metropolitan-located N (%) |
Non-rural-serving, metropolitan located N (%) |
P-value | |
|---|---|---|---|---|---|
| Any adverse birth parent or neonatal outcome | 42,905 (9.2) | 22,425 (7.5) | 9,528 (9.1) | 10,952 (17.8) | <0.001 |
| Birth parent components of the composite outcome | |||||
| Birth parent mortality | 27 (0.01) | Too few to report | 0.20 | ||
| SMM | 2,833 (0.6) | 1,418 (0.5) | 720 (0.7) | 695 (1.1) | <0.001 |
| SMM without blood transfusions | 5,854 (1.3) | 3,372 (1.1) | 1,246 (1.2) | 1,236 (2.0) | <0.001 |
| Neonatal components of the composite outcome | |||||
| Neonatal mortality | 1,027 (0.2) | 387 (0.1) | 237 (0.2) | 403 (0.7) | <0.001 |
| Preventable fetal deaths | 37 (0.01) | Too few to report | <0.001 | ||
| Unexpected newborn complications | 38,731 (8.3) | 19,792 (6.6) | 8,630 (8.2) | 10,309 (16.8) | <0.001 |
| Preterm complications | 2,186 (0.5) | 496 (0.2) | 543 (0.5) | 1,147 (1.9) | <0.001 |
Abbreviations: SMM-severe maternal morbidity
Too few to report indicates cells have <10 observations.
Birth parent mortality: Birth parent death during the birth hospitalization, or on readmission with 6 weeks of delivery, or with 1 year of delivery
SMM: 21 Centers for Disease Control SMM indicators (calculated with and without blood transfusion)
Neonatal mortality: neonatal death before hospital discharge
Preventable fetal death: fetal deaths ≥24 weeks’ gestation or ≥400g after hospital admission for management
Unexpected newborn complications: Unexpected newborn complications, both severe and moderate complications (National Quality Forum definition)
Preterm complications: Severe (grade III or IV) intraventricular hemorrhage, bronchopulmonary dysplasia, retinopathy of prematurity
Compared to rural-located hospitals, the adjusted analysis revealed a significant increase in the odds of any adverse maternal or neonatal outcome in rural-serving (adjusted odds ratio [aOR] 1.27, 95% confidence interval [CI] 1.10, 1.46) and non-rural-serving (aOR 1.35, 95% CI 1.18, 1.55) metropolitan-located hospitals (Figure 1). The odds of an unexpected newborn complication were higher in both rural-serving (aOR 1.33, 95% CI 1.15, 1.55) and non-rural-serving (aOR 1.42, 95% CI 1.23, 1.65) metropolitan-located hospitals. Additionally, the odds of preterm complications were significantly higher in non-rural-serving, metropolitan-located hospitals (aOR 2.09, 95% CI 1.61, 2.70).
Figure 1. Unadjusted and Adjusted Odds of Adverse Perinatal Outcomes by Hospital-Serving Status.
Abbreviations: SMM-severe maternal morbidity
Birth parent mortality and preventable fetal death are included in the primary composite outcome, but are not individually reported given the small numbers.
Variables in the adjusted model include: birth parent age, race/ethnicity, insurance, education, chronic hypertension, pregnancy-induced hypertension, severe pregnancy-induced hypertension/eclampsia, gestational diabetes, diabetes mellitus, placenta previa, multiple gestation, cesarean birth, infant sex, infant gestational age at birth, birth year, state, with hospital as a random effect.
Discussion
With less than two thirds of rural obstetric patients giving birth in rural-located hospitals, this study highlights the frequency with which rural patients receive perinatal care at urban hospitals, located outside of their local community. Rural obstetric patients who gave birth in a rural-located hospital had fewer complicating conditions and a lower odds of adverse outcomes. The higher rates of complicating conditions and higher odds of any adverse outcome among rural patients giving birth at non-rural-serving, metropolitan-located hospitals, likely reflects the referral of higher risk patients to metropolitan-located hospitals where higher-acuity services are available, reflective of the higher rates of preterm birth in these patients. Rates of any adverse outcome were largely driven by unexpected newborn complications, which illustrates the need to provide and coordinate obstetric and neonatal care within and across geographic areas. This analysis highlights the variation in the clinical needs of rural obstetric patients and neonates; notably, those giving birth in metropolitan areas have 2.4-fold higher rates of complicating conditions and adverse outcomes.
The proportion of rural obstetric patients, 35.6% giving birth in a metropolitan area is higher than previously reported. A study published on this topic using data from 2010 and 2012, which defined local births as those within 30 miles by road from the patient’s residence, found that 25.4% of rural residents gave birth in non-local hospitals.18 While conceptually similarly, these studies examined obstetric patients in different states (California, Pennsylvania, and South Carolina in the current study vs Colorado, Iowa, Kentucky, New York, North Carolina, Oregon, Vermont, Washington, and Wisconsin in the prior study). Each state has a different geography and rurality level, which influences the physical proximity of obstetric patients to hospitals with perinatal services.25 Differences in rurality (e.g., very remote regions vs micropolitan areas) likely inform the local distribution of rural- vs. metropolitan-located rural resident births, for which the expected or optimal distribution is unknown.
Many rural hospitals provide basic obstetric and neonatal care, whereas hospitals with subspecialty services are often located in metropolitan areas.12,27 Thus, the higher rate of rural residents with pre-existing or pregnancy-associated conditions giving birth in metropolitan-hospitals, is an expected finding. This pattern supports the provision of regionalized, risk-appropriate care, in which the needs of the obstetric patient and neonate match hospital capabilities. Risk-appropriate care has been shown to improve neonatal outcomes, though similar obstetric data are less clear.7–10,28,29 These differences in care capabilities are reflected in differences in patient case mix, with subspecialty hospitals caring for more complex patients who have an inherently higher risk of complications. These differences are reflected in the unadjusted and adjusted rates of any adverse outcome. We adjusted for conditions present before or at the time of birth, yet there may be other unmeasured variables reflective of the obstetric patient or neonate’s status prior to birth or factors during the birth hospitalization,10 such as hospital processes, which may be contributing to differences in outcomes between rural- and metropolitan-located hospitals. Future work to examine potential drivers of these differences in outcomes will be important in optimizing care for rural residents regardless of birth location.
Geographic access to perinatal care includes both the presence of a local hospital with perinatal services and physical distance (mileage to a hospital). Given trends in closures of rural hospitals and associated loss of obstetric services, geographic access may dictate if rural residents even have the option to give birth in the community where they live.11,13 The location of a birth is the result of several distinct and potentially competing factors. Beyond geography, obstetric patients often consider insurance coverage and in-network hospitals, suggestions received from a family members or friends, and referrals made by their obstetric clinician (e.g., to seek subspecialty services at a metropolitan-located hospital).26 While this study cannot differentiate between these factors, it is important to consider the many drivers that influence birth location and culminate in a rural-located vs metropolitan-located birth as some of these factors may impact perinatal outcomes. Future studies designed to understand how obstetric patients move through the healthcare system may help to elucidate the influence of these and other factors on birth location.
It is well documented that obstetric patients and infants living in rural communities experience higher burdens of death and disease in the US.2–6,30–32 Rural communities face a number of unique changes as they manage rising mortality risk, depopulation, and diversity, which have implications for rural resident health and healthcare delivery.33 These changes are exacerbated by structural urbanism,34 in which the allocation of investments and resources within healthcare systems disadvantages rural populations. This unequal allocation or maldistribution of perinatal resources is likely in rural birth hospitals, which may be a potential driver of rural-urban disparities.35,36 A study of birth hospitals in New York City found that a large portion of racial and ethnic disparities in adverse outcomes was explained by the birth hospital.37 One study examining Black-White disparities and Black- and non-Black-serving birth hospitals, found that in non-teaching hospitals (like the majority of those where rural residents give birth) a large portion of the disparity in in-hospital maternal morality was attributable to the birth hospital.17 Our definition and examination of rural-serving hospitals is a starting point for similar studies of how birth hospitals may or may not be contributing to rural-urban disparities in perinatal outcomes and future opportunities to consider the intersectionality of structural urbanism and structural racism.
Ideally, birth results in a good outcome for both birthing obstetric patients and neonates, which is reflected in an increasing appreciation for dyadic outcomes in the literature.38 Perinatal care delivery can be fragmented and professional organizations often separate the organization and provision of obstetric and neonatal care.39 This study in which any adverse obstetric patient or neonatal outcome was driven by the rate of unexpected newborn complications highlights this point. It is important for hospitals to be equip to manage both the obstetric patient and their neonate. Geospatial studies have found that intensive care services for obstetric patients and neonates are not always co-located, with as many as 20% of hospitals with high-risk services for one half of the birth parent-infant dyad not within 10 miles of high-risk services for the other half of the dyad.40 A recent study reports that between 2014 and 2019 the number of states in the US with policies for maternal or neonatal transport or reimbursement for transport increased, suggesting that efforts to provide risk-appropriate care to the dyad are improving.41 Ensuring the availability of services needed by both patients requires coordination of clinically appropriate, rural-relevant care and associated policies to facilitate local care or when appropriate.
Limitations and strengths
This study has limitations. First, patients captured in this study are not a nationally representative sample, though the geography and healthcare systems in California, Pennsylvania, and South Carolina are distinct and capture diverse patient experiences. Second, although this is a large cohort, maternal death and preventable fetal death frequency was low and could not be reliably examined as secondary outcomes. Third, given these data reflect the birth location, not the admitting hospital, medical transfers for advanced obstetric or neonatal care may explain results. Fourth, the variables accounted for in the models may be incomplete due to unmeasured clinical confounders, such as those that capture relevant hospital resources which are not available in these data. However, the differences in the point estimates between the unadjusted and adjusted results show that many relevant factors are indeed accounted for and relevant to the outcomes of interest.
This study also has strengths. First, this study highlights the distribution of rural resident births across rural- and metropolitan-located hospitals, emphasizing the need for studies of rural perinatal health to look at births both inside and outside of the local community and across birth locations. Second, this study considers the concept of rural-serving and introduces a potential, data-informed definition (hospitals in the top decile). This concept offers new opportunities and perspectives for future studies to consider how rural- and non-rural-serving hospitals may contribute to disparities in outcomes between obstetric patients and infants who reside in rural and urban communities. Additionally, our primary outcome captures the dyadic nature of perinatal care, emphasizing the importance of the interrelated needs and health of obstetric patients and their infants.
Conclusion
While 64.3% of rural residents give birth in rural-located hospitals, the remainder travel to metropolitan-located areas to give birth. Rural residents with more complicated clinical conditions are more likely to give birth outside of the rural community and are at increased risk of adverse outcomes. In order to optimize outcomes for all rural residents and their infants, future studies must look at the location of all rural resident births, seek to identify factors that influence if a birth occurs locally or non-locally, and consider how birth hospitals may be contributing to disparities in rural-urban perinatal health and outcomes.
Footnotes
Conflict of Interest: The authors have no conflicts of interest to disclose.
Data availability
Data generated and analyzed during the current study are not publicly available due to data use agreements, but may be available from the senior author on reasonable request.
References
- 1.Ratcliffe M, Burd C, Holder K, Fields A. Defining Rural at the U.S. Censusu Bureau; 2016. http://www2.census.gov. Accessed April 5, 2021.
- 2.National Advisory Committee on Rural Health and Human Services. Maternal and Obstetric Care Challenges in Rural America: Policy Brief and Recommendations to the Secretary. 2020. https://www.hrsa.gov/sites/default/files/hrsa/advisory-committees/rural/publications/2020-maternal-obstetric-care-challenges.pdf. Accessed July 14, 2021.
- 3.Maron D Maternal Health Care is Disappearing in Rural America. Sci Am. 2016. https://www.scientificamerican.com/article/maternal-health-care-is-disappearing-in-rural-america/#. Accessed March 29, 2022
- 4.Ely DM, Hoyert DL. Differences Between Rural and Urban Areas in Mortality Rates for the Leading Causes of Infant Death: United States, 2013–2015. NCHS Data Brief. 2018;(300):1–8. [PubMed] [Google Scholar]
- 5.Womack LS, Rossen LM, Hirai AH. Urban-Rural Infant Mortality Disparities by Race and Ethnicity and Cause of Death. Am J Prev Med. 2020;58(2):254–260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ehrenthal DB, Daphne Kuo HH, Kirby RS. Infant Mortality in Rural and Nonrural Counties in the United States. Pediatrics. 2020;146(5). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Grünebaum A, McCullough LB, Orosz B, Chervenak FA. Neonatal mortality in the United States is related to location of birth (hospital versus home) rather than the type of birth attendant. Am J Obstet Gynecol. 2020;223(2):254.e1–254.e8. [DOI] [PubMed] [Google Scholar]
- 8.Phibbs CS, Baker LC, Caughey AB, Danielsen B, Schmitt SK, Phibbs RH. Level and volume of neonatal intensive care and mortality in very-low-birth-weight infants. N Engl J Med. 2007;356(21):2165–2175. [DOI] [PubMed] [Google Scholar]
- 9.Lasswell SM, Barfield WD, Rochat RW, Blackmon L. Perinatal regionalization for very low-birth-weight and very preterm infants: a meta-analysis. JAMA. 2010;304(9):992–1000. [DOI] [PubMed] [Google Scholar]
- 10.Lorch SA, Baiocchi M, Ahlberg CE, Small DS. The differential impact of delivery hospital on the outcomes of premature infants. Pediatrics. 2012;130(2):270–278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kozhimannil KB, Interrante JD, Tuttle MKS, Henning-Smith C. Changes in Hospital-Based Obstetric Services in Rural US Counties, 2014–2018. JAMA. 2020;324(2):197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Handley SC, Passarella M, Herrick HM, et al. Birth Volume and Geographic Distribution of US Hospitals With Obstetric Services From 2010 to 2018. JAMA Netw Open. 2021;4(10):1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kaufman BG, Thomas SR, Randolph RK, Perry JR, Thompson KW, Holmes GM, et al. The Rising Rate of Rural Hospital Closures. J Rural Healthl. 2016;(32):35–43. [DOI] [PubMed] [Google Scholar]
- 14.Lorch SA, Srinivas SK, Ahlberg C, Small DS. The impact of obstetric unit closures on maternal and infant pregnancy outcomes. Health Serv Res. 2013;48(2 PART1):455–475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kozhimannil KB, Hung P, Henning-Smith C, Casey MM, Prasad S. Association Between Loss of Hospital-Based Obstetric Services and Birth Outcomes in Rural Counties in the United States. JAMA. 2018;319(12):1239–1247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Aubrey-Bassler F, Cullen R, Simms A, Asghari S, Crane J, Wang PP, et al. Population-based cohort study of hospital delivery volume, geogrpahic accessibility, and obstetric outcomes. Int J Gynecol Obstet. 2019;146:95–102. [DOI] [PubMed] [Google Scholar]
- 17.Burris HH, Passarella M, Handley SC, Srinivas SK, Lorch SA. Black-White disparities in maternal in-hospital mortality according to teaching and Black-serving hospital status. Am J Obstet Gynecol. 2021;225(83):e1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kozhimannil KB, Casey MM, Hung P, Prasad S, Moscovice IS. Location of childbirth for rural women: implications for maternal levels of care. Am J Obstet Gynecol. 2016;214(5):661.e1–661.e10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ryan GM. Toward improving the outcome of pregnancy Recommendations for the regional development of perinatal health services. Obstet Gynecol. 1975;46(4):375–384. [PubMed] [Google Scholar]
- 20.Okoroh EM, Kroelinger CD, Lasswell SM, Goodman DA, Williams AM, Barfield WD. United States and territory policies supporting maternal and neonatal transfer: review of transport and reimbursement. J Perinatol. 2016;36(1):30–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.USDA ERS - Urban Influence Codes. https://www.ers.usda.gov/data-products/urban-influence-codes/. Accessed February 15, 2021.
- 22.Interrante J, Carroll C, Handley S, Kozhimannil K. An Enhanced Method for Identifying HospitalBased Obstetric Unit Status. December 2021. https://3pea7g1qp8f3t9ooe3z3npx1-wpengine.netdna-ssl.com/wp-content/uploads/2022/01/UMN-OB-Unit-Identification-Methods_7.pdf. [Google Scholar]
- 23.How Does CDC Identify Severe Maternal Morbidity? | CDC. https://www.cdc.gov/reproductivehealth/maternalinfanthealth/smm/severe-morbidity-ICD.htm. Accessed September 7, 2021. [Google Scholar]
- 24.California Maternal Quality Care Collaborative. NQF # 716: Unexpected Newborn Complications Update (aka Healthy Term Newborn) Summary of Measure Updates. March 2013. [Google Scholar]
- 25.Interrante JD, Tuttle MS, Basile Ibrahim B, Admon LK, Kozhimannil KB. Policy Brief: State and Regional Differences in Access to Hospital-Based Obstetric Services for Rural Residents, 2018. August 2021. https://3pea7g1qp8f3t9ooe3z3npx1-wpengine.netdna-ssl.com/wp-content/uploads/2021/09/UMN-State-Regional-Differences-in-OB_Policy-Brief_8.16.21_508.pdf [Google Scholar]
- 26.Declercq ER, Sakala C, Corry MP, Applebaum S, Herrlich A. Major Survey Findings of Listening to Mothers III: Pregnancy and Birth. J Perinat Educ. 2014;23(1):9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Handley SC, Passarella M, Srinivas SK, Lorch SA. Identifying individual hospital levels of maternal care using administrative data. BMC Health Serv Res. 2021;21(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Srinivas SK, Kozhimannil K, Hung P, Attanasio L, Jou J, Grobman WA. Do Hospitals with a Higher Level of Maternal Care Designation Have Better Maternal Outcomes? Am J Perinatol. 2019;36(6):653–658. [DOI] [PubMed] [Google Scholar]
- 29.Vanderlaan J, Rochat R, Williams B, Dunlop A, Shapiro SE. Maternal Health Associations Between Hospital Maternal Service Level and Delivery Outcomes. Women’s Heal Issues. 2019;29(3):252–258. [DOI] [PubMed] [Google Scholar]
- 30.Hansen AC, Slavova S, O’Brien JM. Rural residency as a risk factor for severe maternal morbidity. J Rural Health. 2022;38(1):161–170. [DOI] [PubMed] [Google Scholar]
- 31.Cameron NA, Molsberry R, Pierce JB, Perak AM, Grobman WA, Allen NB, et al. Pre-Pregnancy Hypertension Among Women in Rural and Urban Areas of the United States. J Am Coll Cardiol. 2020;76(22):2611–2619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kozhimannil KB, Interrante JD, Henning-Smith C, Admon LK. Rural-Urban Differences In Severe Maternal Morbidity And Mortality In The US, 2007–15. Health Aff. 2019;38(12):2077–2085. [DOI] [PubMed] [Google Scholar]
- 33.Saint Onge JM, Smith S. Demographics in Rural Populations. Surg Clin N Am. 2020;(100):823–833. [DOI] [PubMed] [Google Scholar]
- 34.Probst J, Eberth JM, Crouch E. Structural Urbanism Contributes To Poorer Health Outcomes For Rural America. Health Aff. 2019;38(12):1976–1984. [DOI] [PubMed] [Google Scholar]
- 35.Clapp MA, Srinivas SK, Kozhimannil KB, Grobman WA, Kaimal AJ. Comparison of Maternal and Neonatal Subspecialty Care Provision by Hospital. Am J Perinatol. 2021. doi: 10.1055/s-0041-1727223 [DOI] [PubMed] [Google Scholar]
- 36.Kozhimannil KB, Henning-Smith C. Improving Health Among Rural Residents in the US. JAMA. 2021;325(11):1033–1034. [DOI] [PubMed] [Google Scholar]
- 37.Howell EA, Janevic T, Hebert PL, Egorova NN, Balbierz A, Zeitlin J. Differences in morbidity and mortality rates in black, white, and hispanic very preterm infants among New York City Hospitals. JAMA Pediatr. 2018;172(3):269–277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Villar J, Ariff S, Gunier RB, Thiruvengadam R, Rauch S, Kholin A, et al. Maternal and Neonatal Morbidity and Mortality Among Pregnant Women With and Without COVID-19 Infection: The INTERCOVID Multinational Cohort Study. JAMA Pediatr. 2021;175(8):817–826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Handley SC, Srinivas SK, Lorch SA. Regionalization of Care and the Maternal-Infant Dyad Disconnect. JAMA. 2019;322(6):503–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Brantley MD, Davis NL, Goodman DA, Callaghan WM, Barfield WD. Perinatal regionalization: a geospatial view of perinatal critical care, United States, 2010–2013. Am J Obstet Gynecol. 2017;216(2):185.e1–185.e10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.DeSisto CL, Okoroh EM, Kroelinger CD, Barfield WD. Summary of neonatal and maternal transport and reimbursement policies -- a 5-year update. J Perinatol. 2022. doi: 10.1038/s41372-022-01389-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data generated and analyzed during the current study are not publicly available due to data use agreements, but may be available from the senior author on reasonable request.