Area Poverty and Adverse Birth Outcomes: An Opportunity for Quality Improvement

Abstract

Objective:

To evaluate the association between area poverty and adverse birth outcomes in a large health systems’ diverse birthing population.

Methods:

This was a retrospective cohort study utilizing electronic health record and hospital administrative data for pregnant people at nine hospitals within a large health system in the Chicago metropolitan area from 2018 to 2023. Patient addresses were geocoded and categorized by census tract area percent poor households. Unadjusted and adjusted Poisson regression models, controlling for individual-level risk factors, evaluated the independent association between area poverty and birth outcomes to determine the degree to which this association is attenuated by the inclusion of individual-level factors in the model.

Results:

The study included 85,025 pregnant people. Area poverty was associated with sociodemographic factors including young age, non-Hispanic Black race, Hispanic ethnicity, Medicaid insurance coverage, higher parity, and several comorbid conditions. Area poverty was associated with adverse birth outcomes and demonstrated a gradient effect with increasing area poverty on bivariable analyses. On unadjusted regression analyses, residence in areas with ≥5.0% poverty was associated with severe maternal morbidity, preterm birth, and low birth weight, and residence in areas with ≥8.0% poverty was associated with neonatal intensive care unit admission. While these associations persisted on multivariable analysis for severe maternal morbidity and neonatal intensive care unit admission, the associations with preterm birth and low birth weight persisted only for individuals residing in areas of ≥12.0% poverty when controlling for individual-level risk factors.

Conclusion:

Area poverty was associated with adverse birth outcomes in this birthing population even when controlling for individual-level risk factors, highlighting the need for system- and community-level quality improvement interventions.

Précis:

Area poverty was associated with adverse birth outcomes in a diverse birthing population, highlighting the need for system- and community-level quality improvement initiatives.

Introduction

The “weathering hypothesis”, originally described by Arline Geronimus, suggests that socioeconomic conditions influence health outcomes throughout the life course.¹ Recent studies have documented significant associations between community socioeconomic factors and adverse pregnancy outcomes.^2–9 Community-level factors influence maternal and neonatal birth outcomes through complex mechanisms and are thought to contribute to the significant disparities that exist in these health metrics.^7,10 An individual’s community may influence health through several pathways including differences in health care access (prepregnancy and antenatally), access to healthy environmental resources (e.g. grocery stores, parks), and exposure to acute and chronic environmental stressors, including poverty, crime, and discrimination and racism.^11–13

One such community-level exposure that has been linked to adverse pregnancy outcomes is neighborhood deprivation.^3,5 One approach to measuring neighborhood deprivation is to assess neighborhood poverty by calculating the percentage of households living below the federal poverty line within a unit of area such as at the census tract level, which is a smaller, more granular geographic unit than a zip code. When defined in this way, area poverty levels are associated with individual-level socioeconomic and demographic risk factors that influence health, including but not limited to race and ethnicity, age, income, and baseline health status. The relationship between area poverty and adverse birth outcomes may therefore arise from the influence of the accumulation of these individual-level risk factors as well as the independent influence of living in an impoverished neighborhood environment.¹⁴ Understanding this distinction is important for establishing which factors contribute to disparities and for developing mechanisms to potentially mitigate inequities through targeted system- and community-level interventions.

We sought to evaluate this relationship utilizing the diverse birthing population delivering within a large metropolitan health system in Chicago. Our objectives were: 1. To determine the individual-level demographic and socioeconomic factors associated with area poverty in this birthing population and 2. To evaluate the association between area poverty and adverse birth outcomes both without and with adjustment for individual-level risk factors in order to evaluate whether this association is altered by controlling for individual-level factors.

Methods:

This was a retrospective cohort study using electronic health record (EHR) and administrative hospital data for deliveries from January 2018 to July 2023 from nine hospitals in a Chicago health system. Analysis of the diverse population of reproductive age people within the Chicago metropolitan area affords generalizability to other large, urban populations. Births occurred within a single health system with birthing facilities throughout the Chicago metropolitan region including a large urban academic medical center with >12,000 deliveries per year, three suburban hospitals, and five smaller, exurban community hospitals. We analyzed all unique patient, singleton, live birth deliveries of an infant with gestational age between 23 and 43 weeks. Twenty-three weeks was used as the lower end of the gestational age range as this was generally the earliest gestational age after which a trial of neonatal resuscitation was offered during the study period. If a patient delivered more than once during the study period, only the first birth was included in the analysis. The study was approved by the Northwestern University Institutional Review Board.

Patient addresses were geocoded and matched to Illinois census tracts using 2021 five-year American Community Survey (ACS) census data. The ACS is an annual survey program conducted by the United States (US) Census Bureau which gathers broad demographic and socioeconomic data on a representative sample of the US population.¹⁵ ACS data on the percent of families living at or below the poverty line were used to categorize areas as <1.0%, 1.0%–2.99%, 3.0%–4.99%, 5.0%–7.99%, 8.0%–11.99%, and ≥12.0% poor households. These categories reflect large, comparable proportions of the health system population. Analyzing birth outcomes across this spectrum provides a test of the extent to which the distribution of birth outcomes exhibits a gradient, with different clinical characteristics and birth outcomes seen even in more affluent areas, versus a single high poverty threshold above which an effect is demonstrated. Patients from other states (3.4% of deliveries) were excluded from the analysis.

Several variables were abstracted from electronic medical record data. Maternal age was categorized as ≤19, 20–24, 25–29, 30–34, 35–39, and ≥40 years. Patient race and ethnicity were categorized as Asian, Hispanic, non-Hispanic Black, non-Hispanic White, and none of the above, based on self-reported data from the delivery admissions. Patients’ preferred language was categorized as English versus non-English. Insurance status was categorized as Medicaid versus private or other. Body mass index (BMI) was categorized as normal (BMI ≤23.9 kg/m²), overweight (BMI 25.0–28.9 kg/m²), obese (divided into BMI 30.0–33.9 kg/m², BMI 34.0–39.9 kg/m², and BMI ≥40 kg/m²) or missing. The authors acknowledge that race is a social construct reflective of the experiences and consequences of structural racism. Prior literature has demonstrated that race is an important social determinant of adverse birth outcomes and therefore inclusion of race as a potential confounder was important in this analysis.^4,11,12

Delivery admission International Classification of Diseases (ICD-10) codes were used to characterize the prevalence of any chronic or pregnancy-related condition. Conditions examined individually included pre-existing hypertension, pre-existing diabetes, history of depression or severe mental illness, substance use disorder, anemia, any hypertensive disorder of pregnancy (including gestational hypertension, preeclampsia, and eclampsia), and gestational diabetes. We created a composite variable of “any comorbidity” defined as including one or more of the following: cardiac disease, bleeding disorder, pulmonary hypertension, chronic renal disease, gastrointestinal disease, human immunodeficiency virus (HIV), asthma, autoimmune conditions, neuromuscular disease, thyrotoxicosis, or history of bariatric surgery.

We evaluated seven maternal and neonatal birth outcomes including cesarean birth, vaginal birth complication, cesarean birth complication, severe maternal morbidity (SMM), preterm birth (PTB), low birth weight (LBW), and neonatal intensive care unit (NICU) admission. Route-specific complications were coded using methods adapted from Asch et al.¹⁶ In this paper, the authors outline route-specific complications attainable through ICD-10 coding that can be used to evaluate quality of care of obstetrical residency programs. Vaginal birth complications included hemorrhagic, infectious, or thrombotic complications. Cesarean birth complications included the same as well as operative complications. SMM, as defined by the Center for Disease Control and Prevention (CDC), includes 21 maternal morbidity indicators and corresponding ICD-10 codes.¹⁷ Preterm birth was defined as gestational age at birth before 37 weeks’ and low birth weight was defined as birth weight <2,500 grams.

Chi square tests were used to evaluate the significance of associations between patient census tract percent poor household categories and sociodemographic characteristics, clinical risk factors, delivery hospital, and maternal and neonatal outcomes. Bivariable analyses estimated the association between census tract poverty and birth outcomes. Poisson regression estimated the association of area poverty with birth outcomes through both unadjusted analyses and adjusted multivariable analyses controlling for individual sociodemographic and clinical characteristics. These analyses evaluate the degree to which this relationship is attenuated by inclusion of individual-level risk factors. Models of NICU admission were performed only on the hospitals with NICUs. All analyses were completed using Stata Version 17 (College Station, TX).

Results

Our study cohort included 85,025 pregnant people with singleton live births between January 2018 and June 2023. Figure 1 maps the multi-county region by census tract poverty levels for pregnant people in the study sample as well as the locations of the nine hospitals within the health system. While there is a concentration of neighborhoods with ≥12.0% poor households on the South Side and West sides of Chicago, there are numerous other areas of poverty in suburban and exurban counties as well.

Fig 1.

Multi-county region by census tract poverty levels for birthing people in the study sample, January 2018-June 2023 (n=85,025 patients). FPL, Federal Poverty Level.

Table 1 shows the sociodemographic characteristics and clinical risk factors of the birthing population by census tract percent poor household categories. There was an age gradient across the areas with the youngest pregnant people living in the higher poverty areas while the most affluent areas had the largest proportion of older pregnant people. Differences were even larger for race and ethnicity with twice the non-Hispanic White population living in more affluent areas and approximately twice the proportion of non-Hispanic Black and Hispanic patients living in the highest poverty areas. In the highest poverty area group, 44.4% of pregnant people had Medicaid insurance delivery coverage as compared to 6.4% in the most affluent areas. Pregnant people from high poverty areas had fewer nulliparous births and there was a sharp increase in the proportion of the highest BMI patients in poorer areas. With exception of previously diagnosed depression and any other comorbidities, there was a consistent gradient by area poverty with higher rates of pre-existing hypertension, pre-existing diabetes, substance use disorder, anemia, and hypertensive disorders of pregnancy in poorer areas.

Table 1.

Patient sociodemographic characteristics and clinical risk factors by census tract percent poor households^*

Characteristics	All Patients N=85,025	<1.0% N=16,247	1.0–2.99% N=16,993	3.0–4.99% N=13,090	5.0–7.99% N=13,783	8.0–11.99% N=8,705	≥12.0% N=16,252
Maternal age (years)
≤19	1.3	0.4	0.5	0.9	1.3	2.0	3.1
20–24	8.2	3.2	4.1	6.1	8.0	12.5	17.1
25–29	20.1	15.6	16.1	18.7	21.5	24.4	26.3
30–34	40.5	47.1	45.5	42.4	39.9	34.8	30.8
35–39	24.4	27.9	28.2	25.9	24.0	21.1	18.0
≥40	5.5	5.9	5.6	6.0	5.3	5.2	4.8
Race/ethnicity
Asian	6.4	7.4	6.5	6.9	7.8	5.6	4.2
Hispanic	15.6	7.0	8.6	11.6	16.4	24.6	29.1
Non-Hispanic Black	8.9	3.3	4.0	4.7	6.3	12.8	23.4
Non-Hispanic White	57.8	71.5	70.2	66.2	56.9	45.3	32.3
None of the above	11.3	11.3	10.7	10.6	12.6	11.8	11.0
Parity
Nulliparous	56.1	60.2	57.3	56.8	58.2	53.0	50.1
Parity 1	26.9	26.7	27.6	27.5	25.8	27.2	26.9
Parity ≥2	17.0	13.2	15.1	15.8	16.0	20.0	23.1
Medicaid	20.0	6.4	9.7	13.2	20.0	29.6	44.4
Non-English language	1.7	1.5	1.3	1.7	2.1	2.0	1.8
BMI (kg/m2)
Missing	7.8	8.6	8.9	7.9	6.9	7.1	6.8
≤24	9.2	11.3	10.4	9.0	9.0	7.1	7.3
25–29	34.2	39.1	37.7	35.3	34.0	28.6	27.6
30–33	23.2	21.9	22.3	23.9	24.0	24.1	23.7
34–39	17.2	13.5	14.7	16.4	17.7	21.1	21.6
≥40	8.5	5.5	6.0	7.6	8.4	12.0	13.1
Pre-existing hypertension	7.3	6.6	6.7	7.1	7.1	8.3	8.6
Pre-existing diabetes	3.4	2.7	2.6	3.1	3.5	4.7	4.5
History of depression or severe mental illness	16.0	15.4	16.0	16.7	16.2	17.0	15.1
Substance use disorder	3.9	3.1	3.2	3.8	4.2	4.2	5.2
Anemia	10.7	8.7	8.6	9.8	10.6	12.5	14.6
Any comorbidity†	12.5	13.2	13.4	13.0	12.3	12.3	10.8
Gestational diabetes	6.2	5.2	5.7	6.5	6.5	7.1	6.6
Gestational hypertension	4.9	4.9	4.9	4.6	4.9	5.0	5.2
Hypertensive disorders of pregnancy‡	5.8	5.6	5.8	5.4	5.8	5.9	6.4
Hospital
Academic hospital	60.0	60.2	58.3	53.3	60.3	57.5	67.8
Suburban hospital 1	13.4	15.7	17.3	15.2	12.1	13.7	6.6
Suburban hospital 2	8.7	7.7	7.1	10.6	10.0	11.7	7.4
Suburban hospital 3	6.6	10.1	8.8	7.7	3.6	5.0	3.7
Other five exurban hospitals	11.2	6.3	8.6	13.1	14.0	12.2	14.5

BMI = body mass index

Data are expressed as %

All associations showed a p-value <0.001 with one exception being NICU admission where p = 0.01

^*Census tract percent poor households is defined as percent of families living at or below the poverty line and is based on the 2021 five-year American Community Survey (ACS) census data

^†Any comorbidity includes one or more of the following: cardiac disease, bleeding disorder, pulmonary hypertension, chronic renal disease, gastrointestinal disease, HIV, asthma, autoimmune conditions, neuromuscular disease, thyrotoxicosis, or history of bariatric surgery

^‡Hypertensive disorders of pregnancy includes either gestational hypertension, preeclampsia, or eclampsia

The academic hospital accounted for 60.0% of the total birthing population and the highest proportion (67.8%) of patients from the highest (≥12.0%) poverty area. Suburban hospitals varied in their socioeconomic case mix, generally serving more affluent patients while the smaller exurban community hospitals served disproportionately poorer patient populations.

Figure 2 shows maternal and neonatal birth outcomes by census tract poverty area. All comparisons of outcome differences by census tract area poverty level were p<0.001. Of our total patient population, 28.7% delivered by cesarean with only small variations by area poverty. Complications from vaginal and cesarean birth were higher for pregnant people living in higher (≥8.0%) poverty areas. There was a gradient in rates of SMM, with the highest rates in pregnant people from the poorest areas. PTB rates were also higher in higher poverty areas, reaching 9.4% in areas with ≥12.0% poor households, with the same gradient seen for low birth weight and NICU admission.

Fig 2.

Maternal (A) and neonatal (B) birth outcomes by census tract poverty levels. Vaginal birth complications include hemorrhagic, infectious, or thrombotic complications, for vaginal births only. Cesarean birth complications include infectious, hemorrhagic, thrombotic, or operative complications, for cesarean births only. Severe maternal morbidity defined per Center for Disease Control and Prevention criteria including hysterectomy, blood transfusion, sepsis, shock, eclampsia, disseminated intravascular coagulation, severe anesthesia complication, acute renal failure, amniotic fluid embolism, air or thrombotic embolism, acute myocardial infarction, cardiac arrest, conversion of cardiac rhythm, pulmonary edema/heart failure, adult respiratory distress syndrome, aneurysm, sickle cell disease with crisis, cerebrovascular accident, temporary tracheostomy, or ventilation. Preterm birth was defined as gestational age at birth before 37 weeks of gestation. NICU, neonatal intensive care unit; FPL, Federal Poverty Level.

Table 2 shows the regression analyses for the association of census tract percent poor household with birth outcomes. Living in areas of ≥5.0% poverty was significantly associated with SMM on both unadjusted and adjusted analyses. Adjusted incidence rate ratios were lower when including sociodemographic variables, however, even when controlling for these factors, there was a higher risk of SMM in areas of ≥5.0% poverty. Neonatal outcomes were also associated with area poverty with the greatest effects seen in areas with the highest poverty rates. Living in areas of 5.0–7.99%, 8.0–11.99%, and ≥12.0% poverty was significantly associated with PTB on bivariable analyses (IRR 1.11, 95% CI 1.02–1.21; IRR 1.18, 95% CI 1.08–1.30; IRR 1.38, 95% CI 1.28–1.50, respectively) but only living in areas of ≥12.0% poverty remained significantly associated on multivariable analysis (aIRR 1.20, 95% CI 1.10–1.31). A similar pattern was seen with LBW with only areas of ≥12.0% poverty having higher risk on both unadjusted (IRR 1.53, 95% CI 1.40–1.67) and adjusted analyses (aIRR 1.23, 95% CI 1.12–1.36), and again the effect was smaller on adjusted analyses. Living in areas of ≥8.0% poverty was significantly associated with NICU admission on both unadjusted and adjusted analyses with only small attenuation on multivariable analysis.

Table 2.

Bivariable and multivariable regression results for association of birth outcomes with census tract percent poor households^*

Outcome	<1.0% N=16,247	1.0–2.99% N=16,993	3.0–4.99% N=13,090	5.0–7.99% N=13,783	8.0–11.99% N=8,705	≥12.0% N=16,252
Maternal Outcomes
Cesarean Birth
Unadjusted model	Reference	1.02 (0.98–1.06)	1.04 (0.99–1.08)	1.00 (0.96–1.04)	1.02 (0.97–1.07)	0.97 (0.93–1.01)
Adjusted model	Reference	1.02 (0.98–1.06)	1.02 (0.98–1.07)	1.02 (0.98–1.07)	1.03 (0.98–1.08)	1.03 (0.99–1.08)
Vaginal Birth Complication†
Unadjusted model	Reference	1.07 (1.01–1.15)	1.02 (0.95–1.09)	1.06 (0.99–1.13)	1.11 (1.03–1.21)	1.15 (1.08–1.23)
Adjusted model	Reference	1.09 (1.01–1.16)	1.04 (0.97–1.11)	1.02 (0.95–1.10)	1.07 (0.99–1.14)	1.06 (0.98–1.16)
Cesarean Birth Complication‡
Unadjusted model	Reference	1.01 (0.92–1.11)	0.96 (0.87–1.07)	0.98 (0.89–1.08)	1.01 (0.90–1.13)	1.13 (1.03–1.24)
Adjusted model	Reference	1.02 (0.93–1.12)	0.97 (0.88–1.07)	0.94 (0.85–1.04)	0.97 (0.87–1.09)	0.99 (0.89–1.09)
Severe Maternal Morbidity§
Unadjusted model	Reference	1.11 (0.92–1.33)	1.04 (0.85–1.27)	1.38 (1.15–1.66)	1.50 (1.22–1.84)	1.51 (1.27–1.80)
Adjusted model	Reference	1.09 (0.91–1.32)	1.00 (0.82–1.22)	1.29 (1.07–1.56)	1.30 (1.05–1.60)	1.22 (1.01–1.48)
Neonatal Outcomes
Preterm Birth||,¶
Unadjusted model	Reference	1.01 (0.93–1.09)	1.06 (0.97–1.16)	1.11 (1.02–1.21)	1.18 (1.08–1.30)	1.38 (1.28–1.50)
Adjusted model	Reference	0.99 (0.91–1.07)	1.05 (0.96–1.15)	1.09 (1.00–1.19)	1.08 (0.98–1.19)	1.20 (1.10–1.31)
Low Birth Weight (<2500 grams)
Unadjusted model	Reference	1.01 (0.91–1.11)	1.04 (0.94–1.15)	1.11 (1.01–1.23)	1.21 (1.08–1.35)	1.53 (1.40–1.67)
Adjusted model	Reference	1.00 (0.91–1.10)	1.04 (0.94–1.15)	1.06 (0.96–1.18)	1.08 (0.97–1.21)	1.23 (1.12–1.36)
NICU admissionf
Unadjusted model	Reference	1.03 (0.96–1.11)	0.98 (0.09–1.07)	1.00 (0.92–1.09)	1.12 (1.03–1.23)	1.23 (1.15–1.33)
Adjusted model	Reference	1.02 (0.95–1.11)	0.99 (0.91–1.08)	1.02 (0.93–1.10)	1.10 (1.00–1.20)	1.22 (1.13–1.33)

NICU = Neonatal intensive care unit

Data are expressed as adjusted incidence rate ratios and 95% confidence intervals

Bolded text indicates significant associations

^*Adjusted Poisson regression models included age, race and ethnicity (as a social construct), parity, prior cesarean birth, pre-existing hypertension, pre-existing diabetes, other medical comorbidity, anemia, history of depression, severe mental illness or substance use, body mass index, and census tract level of percent of poor households

^†Vaginal birth complications based on ICD10 codes of hemorrhagic, infectious, or thrombotic complications; for vaginal deliveries only

^‡Cesarean birth complications based on ICD10 codes of infectious, hemorrhagic, thrombotic, or operative complications; for cesarean deliveries only

^§Severe maternal morbidity defined per Center for Disease Control and Prevention criteria including hysterectomy, blood transfusion, sepsis, shock, eclampsia, disseminated intravascular coagulation, severe anesthesia complication, acute renal failure, amniotic fluid embolism, air or thrombotic embolism, acute myocardial infarction, cardiac arrest, conversion of cardiac rhythm, pulmonary edema/heart failure, adult respiratory distress syndrome, aneurysm, sickle cell disease with crisis, cerebrovascular accident, temporary tracheostomy, or ventilation

^||Preterm birth was defined as gestational age at birth before 37 weeks’ gestation

^¶Note that suburban hospital 1, suburban hospital 3, and the 5 exurban hospitals do not have NICU services and therefore transfer anticipated preterm births to the academic hospital and suburban hospital 2

Discussion

In this large, metropolitan birthing population from a single health system in Chicago, area poverty was associated with individual sociodemographic factors including young maternal age, non-Hispanic Black race, Hispanic ethnicity, Medicaid insurance coverage, higher parity, obesity and chronic medical comorbities. We did not demonstrate an association between area poverty and diagnosed mental illness, though given that our study relies on administrative coding of an established diagnosis by a clinician this may be due to underdiagnosis. On bivariable analysis, our results demonstrated a continuous gradient effect across all levels of area poverty for some maternal birth outcomes and an even more pronounced poverty level gradient for neonatal outcomes, rather than simply a ‘high’ poverty threshold. This association between area poverty and several adverse birth outcomes was attenuated but remained significant on multivariable analyses controlling for individual-level risk factors.

Existing literature on the association between area deprivation and adverse pregnancy outcomes is highly heterogenous and has shown mixed results. A systematic review and meta-analysis demonstrated an increased risk of PTB, small for gestational age (SGA), and stillbirth in areas with greater deprivation, however the majority of these studies were from European populations limiting generalizability to the US.³ In contrast, a recent study of births in the Portland, OR metropolitan area failed to find an association between living in a deprived area and PTB, LBW, SGA, or NICU admission.^9,18 The association between neighborhood deprivation and maternal outcomes has been less studied though some studies suggest an increased risk of hypertensive disorders of pregnancy, with mixed results for risk of gestational diabetes.^18–20 We similarly showed an increased risk for adverse perinatal outcomes in areas with ≥12.0% poverty including PTB, LBW, and NICU admission. These associations persisted, though were attenuated, with inclusion of individual-level risk factors in the multivariable models. Additionally, we demonstrated increased risk for SMM in areas with ≥5.0% poverty which persisted in multivariable models, an association that has not been evaluated previously in the literature. These findings demonstrate that although individual demographic and socioeconomic factors are known to be aggregated in impoverished neighborhoods and therefore contribute to these relationships, there remain additional unexplained community-level effects on birth outcomes that likely reflect both unmeasured individual characteristics and the influence of ecological area characteristics common to all residents.¹⁰

These findings support the importance of understanding the location, characteristics, and birth outcomes of the most impoverished pregnant people in order to best design quality improvement interventions targeting both the individual and the community through health systems interventions and broader community interventions. Such system-wide interventions might include screening for critical individual social needs during pregnancy (such as food insecurity, housing insecurity, medication affordability, and transportation and childcare needs), access to dedicated social workers to facilitate linkage to services in the community and to screen for eligibility for state-based assistance programs such as WIC (Women, Infants, and Children), and patient navigators to ease the burden of frequent and complex care needs during and after pregnancy.^11,21–24

Health systems should also design and implement formal mechanisms for monitoring disparities in a standard list of important birth outcomes, such as those evaluated in this study. Such mechanisms may include the establishment of maternal morbidity reports that utilize EHR data to allow individual clinicians and hospitals within a health system to monitor and more easily compare birth outcomes and the development of maternal morbidity review boards.²⁵ Additionally, hospitals and health systems can use these monitoring tools to target interventions to improve specific birth metrics.

Broader community outreach interventions should also be undertaken with community members to promote longevity of implemented programs through community ownership. Community programs might include group prenatal exercise and nutrition classes, group pregnancy and birthing classes, and community donation drives for newborn and postpartum supplies.^26–28 Larger health system community goals might include the establishment of clinic sites integrated into neighborhoods of the highest poverty to reduce transportation costs and travel times and promote community engagement. Another intervention that has been shown to improve pregnancy-related and long-term maternal and neonatal health outcomes is visiting nurse programs geared towards the most socially vulnerable populations.²² Additionally, health systems can advocate for evidence-based social and economic change interventions, such as paid family and medical leave, child tax credit and other guaranteed income programs, affordable housing, and comprehensive health insurance, all known to have major implications for population health.^29–31

Our study has several strengths including the large sample size of diverse pregnant people delivering within a US health system. We also report on easily replicable outcome variables, which can serve as both future research variables and quality improvement targets, including several maternal variables that have not yet been studied. Important limitations of our study include the reliance on EHR and administrative data, which is subject to misclassification bias, though we suspect the effect of this to be small given the large sample size. Additionally, our birthing population is largely affluent, limiting our ability to further stratify outcomes by higher poverty areas and the generalizability of our results.

In summary, these data contribute to the literature highlighting the association between an individual’s community socioeconomic context and their pregnancy health outcomes. Given this well-established association, health systems have a responsibility to critically evaluate the social context in which they deliver health services to improve the provision of equitable care to their patients.