Comparing severe maternal morbidity and neonatal outcomes between Somali and non-Somali black birthing cohorts

Abstract

BACKGROUND

Black birthing individuals experience worse perinatal outcomes compared to White counterparts, yet within-group differences among Black populations remain understudied. The U.S. Black population is heterogeneous, encompassing long-established U.S.-born communities and immigrants from across the African diaspora, whose health profiles may differ. Somali immigrants represent the largest African-born population in many U.S. communities and face documented challenges including cultural adaptation and language barriers. However, severe maternal morbidity (SMM) has not been specifically studied in this population.

OBJECTIVE

To compare severe maternal morbidity and neonatal birth outcomes between Somali and non-Somali Black birthing individuals.

STUDY DESIGN

This was a retrospective cohort study from 2016 to 2024, in which Somali and non-Somali Black individuals were identified within an academic-community health system. Somali individuals were identified using self-reported Black/African race combined with documented primary language (Somali) and/or country of origin (Somalia); individuals meeting either language or country criterion with Black/African race were classified as Somali regardless of birthplace. SMM was defined using CDC 21-indicator criteria (excluding blood transfusion) and assessed during birth hospitalization and through 6 weeks postpartum. Neonatal morbidity was defined as a composite of respiratory distress syndrome, transient tachypnea of the newborn, retinopathy of prematurity, bronchopulmonary dysplasia, sepsis, 5-minute Apgar ≤7, hypoxic-ischemic encephalopathy, NICU admission, or neonatal death. Comparisons were performed using chi-square, Fisher’s exact, and t-tests as appropriate. Logistic regression estimated adjusted odds ratios (aOR) controlling for parity, mode of birth, insurance type, maternal age, and hypertensive disorders of pregnancy.

RESULTS

Among 63,435 index births, 7403 (11.7%) were to Black birthing individuals, including 2427 (32.9%) Somali and 4976 (67.1%) non-Somali Black individuals. Somali individuals were older, more frequently multiparous, more likely to have public insurance, and had lower epidural and cesarean birth rates (all P≤.001). SMM during birth hospitalization was similar between groups. Six-week postpartum SMM was lower among Somali births (0.5% vs 1%, P=.021; aOR: 0.53, 95% CI: 0.26–0.99). Neonatal composite morbidity was modestly lower among Somali births compared to non-Somali Black births (15.0% vs 18.7%, P=.001; aOR: 0.85, 95% CI: 0.74–0.98). Subgroup analysis demonstrated no differences in SMM or neonatal composite outcomes between Somali individuals who used interpreter services and those who did not.

CONCLUSION

In this single health system study, SMM during birth hospitalization was comparable between Somali and non-Somali Black birthing individuals, whereas postpartum severe maternal morbidity and neonatal composite morbidity were lower among Somali births. This within-diaspora comparison underscores the importance of examining heterogeneity within Black populations and suggests that maternal and neonatal outcomes may differ across Black subgroups despite shared structural disadvantage.

AJOG Global Reports at a Glance.

Why was the study conducted?

Severe maternal morbidity has not been specifically studied among Somali immigrants, the largest African-born population in many U.S. communities, nor compared to the broader Black population to examine within-group heterogeneity in perinatal outcomes.

Key findings

Severe maternal morbidity during birth hospitalization was comparable between Somali and non-Somali Black birthing individuals, while 6-week postpartum severe maternal morbidity was lower among Somali births after adjustment for confounders (adjusted odds ratio 0.53, 95% confidence interval 0.26–0.99). Neonatal composite morbidity was modestly lower among Somali births (adjusted odds ratio 0.85, 95% confidence interval 0.74–0.98).

What does this study add to what is known?

This study provides the first within-diaspora comparison of severe maternal morbidity between Somali and non-Somali Black populations, demonstrating that the heterogeneous Black population experiences comparable severe maternal morbidity during birth hospitalization, while neonatal outcomes may differ modestly.

Introduction

Somali immigrants represent one of the largest African-born populations in many U.S. metropolitan areas and one of the fastest-growing immigrant communities in high-income countries, with documented racial disparities for Black birthing populations compared to White counterparts.¹ The U.S. Black population is heterogeneous, encompassing long-established U.S.-born communities and immigrants from across the African diaspora, each with distinct cultural practices, health behaviors, and socioeconomic profiles.2, 3, 4 Among these subgroups, Somali immigrants are of particular interest due to their cultural and linguistic distinctiveness, and concentrated settlement patterns that facilitate community-level study.⁵ Prior work has demonstrated that Somali-born individuals experience higher rates of cesarean births, particularly emergency cesarean births, increased risk of small for gestational age (SGA) infants, and higher rates of stillbirths.6, 7, 8 Additionally, Somali populations face cultural and linguistic barriers in accessing healthcare services, with lower rates of antenatal care utilization.⁹

While existing studies on Black birthing individuals have primarily compared maternal health outcomes to the general population, which consists predominantly of White individuals,6, 7, 8, 9 there is limited literature examining Somalis compared to the broader Black population. Additionally, severe maternal morbidity (SMM), as defined by the Centers for Disease Control and Prevention (CDC) as unexpected outcomes of labor and delivery that can result in short- or long-term health consequences, has not been specifically studied within Somali populations. Furthermore, while the "healthy immigrant effect" suggests that foreign-born individuals may initially experience better birth outcomes that diminish with longer residence,10, 11, 12 whether this phenomenon applies to SMM among Somali immigrants specifically remains unexplored. Collectively, these shortcomings underscore the need for research examining SMM within specific ethnic subgroups of the Black population.

To address this critical knowledge gap, our study aimed to compare SMM and neonatal outcomes between Somali and non-Somali Black birthing individuals. We hypothesized that Somali birthing individuals would demonstrate higher rates of SMM and adverse neonatal outcomes compared to non-Somali Black individuals, given the additional challenges posed by cultural adaptation, language barriers, immigration-related stressors, and potential difficulties in healthcare system navigation.

Methods

We conducted a retrospective cohort study using data from the University of Minnesota Obstetric Measures (UMOMs) database¹³ [IRB approved: STUDY0012822] for births occurring between January 1, 2016, and July 31, 2024 within an academic community health system, which is comprised of nine birthing hospitals across urban, suburban, and rural settings. The study protocol was approved by the University of Minnesota Institutional Review Board [STUDY00023654]. This study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting cohort studies.¹⁴ The IRB waived need for consent.

The study population included pregnant individuals aged 18 to 45 years who had their first singleton pregnancy within our system during the study period, and who identified as Black or African American. To identify Somali birthing individuals, we required self-reported race/ethnicity (Black/African) combined with at least one of the following: primary language preference (Somali) or country of origin (Somalia) as documented in the electronic health record (EHR). Individuals who self-reported Black/African race and met either the language or country criterion were classified as Somali, regardless of whether they were born in the United States or in Somalia; this approach captures both first-generation immigrants and U.S.-born individuals who maintain Somali as their primary language, reflecting continued cultural identity. Similarly, self-reported race and ethnicity were used to identify non-Somali Black birthing individuals. We excluded subsequent births from individuals during the study period, and births occurring before 20 weeks’ gestation. In a preplanned subgroup analysis, we examined SMM and neonatal outcomes between non-English-speaking and English-speaking Somali birthing individuals, using interpreter services as a proxy.

Our primary outcome was SMM, as defined by the CDC as an “unexpected outcome of labor and delivery that results in significant short- or long-term consequences to a woman’s health” and identified using the 21 indicators corresponding to International Classification of Diseases (ICD) codes in hospital discharge data.¹⁵ These indicators encompass potentially life-threatening conditions, complications, and interventions. Blood transfusions were excluded from the composite SMM measure, which aligns with the CDC’s criteria. SMM was assessed separately during the birth hospitalization and 6-week postpartum period to distinguish acute intrapartum complications from delayed postpartum pathology and avoid obscuring group-specific patterns. Secondary outcomes focused on neonatal morbidity, including a composite measure of respiratory distress syndrome (RDS), transient tachypnea of the newborn (TTN), retinopathy of prematurity, bronchopulmonary dysplasia, sepsis of newborn, 5-minute Apgar score ≤7, hypoxic-ischemic encephalopathy (HIE), neonatal intensive care unit (NICU) admission, and neonatal death before hospital discharge. These indicators were selected based on prior literature examining neonatal outcomes in immigrant and Somali populations and represent the spectrum of neonatal morbidity captured in our institutional database.^6,16, 17, 18

Demographic and clinical characteristics were collected from the EHR using standardized criteria. Race, ethnicity, and country of origin were self-reported by individuals. Parity was determined from the obstetric history documented at the first prenatal visit, and gestational age at birth was calculated from the best obstetric estimate using last menstrual period or earliest ultrasound. Insurance status was categorized as public (including Medicaid and other state/federal programs), private (employer-sponsored or self-purchased insurance), or uninsured. Birth hospital location was classified as urban, suburban, or rural based on U.S. Census classifications. Mode of birth was categorized as spontaneous vaginal birth, operative vaginal birth (assisted), or cesarean birth based on birth documentation. Epidural use was determined through review of anesthesia documentation. Hypertensive disorders of pregnancy (HDP) were defined according to ACOG criteria and included chronic hypertension, gestational hypertension, preeclampsia without severe features, preeclampsia with severe features, and superimposed preeclampsia.¹⁹

Categorical variables were compared between groups using Fisher's exact tests. Continuous variables were analyzed using Student's t-tests. Logistic regression models were used to estimate unadjusted and adjusted odds ratios and 95% confidence intervals (CI) for the Somali cohort for SMM during the birth hospitalization or within 2 days and at 6 weeks postpartum. Adjusted model for neonatal composite outcome controlled for parity, birth type, and insurance type. In addition to these variables, the adjusted model for SMM also controlled for maternal age and HDP. Gestational age was not included as a covariate in the adjusted model for neonatal composite outcome as it was considered a potential mediator on the causal pathway between Somali ethnicity and this outcome, and adjustment would risk introducing mediator bias. Subgroup analyses examined differences between English-speaking and non–English-speaking Somali birthing individuals, as well as between Somali and African American individuals. P values <.05 were considered statistically significant. All statistical analyses were performed using R (version 4.2.2, “Innocent and Trusting”)

Results

During the study period from January 2016 to July 2024, we identified 65,368 index births in our system. After excluding non-Black/African American/African birthing individuals (n=56,032), multiple gestation (n=1,096), and births at less than 20 weeks’ gestation (n=837), our final cohort included 7403 Black birthing individuals: 4976 (67.1%) non-Somali Black and 2427 (32.9%) Somali individuals (Figure 1). The Black birthing control cohort was a heterogeneous population, though the majority reported US birth and English as the primary language, it also included individuals born in Ethiopia, Kenya, and other African and Caribbean nations (Supplemental Table 1)

Figure 1.

Study population selection flowchart

Abdullahi et al. Comparing severe maternal morbidity and neonatal outcomes between Somali and non-Somali black birthing cohorts. Am J Obstet Gynecol 2026.

Significant demographic differences were observed between groups (Table 1). Somali individuals were older at birth (mean age 31.5 vs 29.0 years, P<.001), more likely to be multiparous (73.6% vs 57.2%, P<.001), and more likely to have public insurance (79.4% vs 66.7%, P<.001). Somali individuals had lower rates of epidural use (38.7% vs 49.2%, P<.001), rural hospital birth (0.1% vs 2.2%, P<.001), and cesarean birth (30.2% vs 39.4%, P<.001). Hypertensive disorders of pregnancy (HDP) were less frequent among Somali individuals (12.1% vs 19.7%, P<.001), including lower rates of chronic hypertension (0.8% vs 3.7%, P<.001).

Table 1.

Baseline demographic and clinical characteristics of Somali and non-Somali black birthing individuals

Characteristic	Somali (n=2427)	Non-Somali Black (n=4976)	P value
Maternal age at birth (y)	31.5±5.6	29.0±6.2	<.001a
Multiparous	1785 (73.6)	2827 (57.2)	<.001b
Insurance type—public	1926 (79.4)	3318 (66.7)	<.001b
Epidural use	940 (38.7)	2446 (49.2)	<.001b
Birth location—rural hospital	2 (0.1)	111 (2.2)	<.001b
Cesarean birth	733 (30.2)	1959 (39.4)	<.001b
Chronic hypertension	20 (0.8)	182 (3.7)	<.001b
Hypertensive disorder of pregnancy	293 (12.1)	982 (19.7)	<.001b

Data are mean ± SD or n (%) unless otherwise specified.

y, years.

^aStudent’s t-test

^bFisher’s Exact Test.

Abdullahi et al. Comparing severe maternal morbidity and neonatal outcomes between Somali and non-Somali black birthing cohorts. Am J Obstet Gynecol 2026.

SMM during birth hospitalization was similar between groups (0.7% Somali vs 0.8% non-Somali Black; P=.482). Unadjusted 6-week postpartum SMM rates were lower among Somali individuals (0.5% vs 1.0%; P=.021), and this difference was seen after adjustment for parity, mode of birth, insurance type, maternal age, and HDP (aOR: 0.53, 95% CI: 0.26–0.99) (Table 2). The low frequency of individual SMM indicators limited meaningful comparisons of specific components during both the birth hospitalization and postpartum period (Figure 2; Supplemental Figure 1; Supplemental Tables 2–3).

Table 2.

Severe maternal morbidity among Somali and non-Somali individuals during birthing hospitalization and 6-week postpartum period

SMM indicator	Birthing hospitalization					6-week postpartum period
	Somali (n=2427)	Non-Somali Black (n=4976)	P valuea	Unadjusted OR (95% CI)	Adjusted ORb (95% CI)	Somali (n=2427)	Non-Somali Black (n=4976)	p-valuea	Unadjusted OR (95% CI)	Adjusted ORb(95% CI)
SMM	16 (0.7)	41 (0.8)	.482	0.80 (0.43–1.40)	1.04 (0.55–1.89)	12 (0.5)	51 (1.0)	0.021	0.48 (0.24–0.87)	0.53 (0.26–0.99)

Data are n (%).

SMM, severe maternal morbidity; OR, odds ratio; CI, confidence interval.

^aFisher’s Exact Test

^bLogistic regression models adjusted for parity, mode of birth, insurance type, maternal age and hypertensive disorder of pregnancy.

Abdullahi et al. Comparing severe maternal morbidity and neonatal outcomes between Somali and non-Somali black birthing cohorts. Am J Obstet Gynecol 2026.

Figure 2.

Severe maternal morbidity indicators between Somali and non-Somali black individuals during birthing hospitalization

Abdullahi et al. Comparing severe maternal morbidity and neonatal outcomes between Somali and non-Somali black birthing cohorts. Am J Obstet Gynecol 2026.

In terms of neonatal outcomes, neonates of Somali individuals had higher birth weights (mean 3313 g vs 3195 g, P<.001) and higher gestational age at birth (39.5 vs 38.7 weeks, P<.001) (Supplemental Table 4). However, the distribution of infant size for gestational age did not differ significantly between groups (P=.849), with comparable rates of SGA (12.1% Somali vs 12.3% non-Somali), appropriate-for-gestational-age (80.7% vs 80.3%), and LGA infants (7.1% vs 7.5%).

The composite neonatal morbidity (Supplemental Table 5) was lower among infants born to Somali individuals (15.0% vs 18.7%, P=.001). This included lower rates of retinopathy of prematurity (0.5% vs 1.0%, P=.010), 5-minute Apgar scores ≤7 (6.3% vs 7.7%, P=.036), and NICU admission (11.9% vs 15.3%, P<.001). This difference (Table 3) remained statistically significant after adjusting for parity, mode of birth, and insurance type (adjusted OR: 0.85, 95% CI: 0.74–0.98).

Table 3.

Neonatal composite outcome odds ratios among Somali and non-Somali black individuals

	Unadjusted OR (95% CI)	Adjusted OR (95% CI)a
Neonatal composite outcomeb	0.77 (0.67–0.88)	0.85 (0.74–0.98)

OR, odds ratio; CI, confidence interval.

^aLogistic regression models adjusted for parity, mode of birth, and insurance type

^bRDS, Transient Tachypnea of Newborn, Retinopathy of Prematurity, Bronchopulmonary Dysplasia, Sepsis of Newborn, Hypoxic-Ischemic Encephalopathy, 5-minute Apgar ≤7, NICU admission, Neonatal death prior to hospital discharge.

Abdullahi et al. Comparing severe maternal morbidity and neonatal outcomes between Somali and non-Somali black birthing cohorts. Am J Obstet Gynecol 2026.

Since language barriers may increase SMM, we performed a subgroup analysis among Somali individuals requiring interpreter services (983 [40.5%]) compared to those not requiring interpreter services (1444 [59.5%]) (Supplemental Table 6). There were no significant differences in SMM during birthing hospitalization (0.3% vs 0.9%, P=.122) and through the 6-week postpartum period (0.5% vs 0.5%, P=1.00). Moreover, neonatal composite morbidity rates did not differ significantly between interpreter-requiring and non–interpreter-requiring groups (14.5% vs 15.4%, P=.562).

Comment

Principal findings

This study examines SMM among Somali birthing individuals compared to non-Somali Black populations in the United States. We found that both groups experienced comparable SMM rates during birthing hospitalization, with this similarity persisting after adjustment for demographic and clinical confounders. Somali individuals also demonstrated lower 6-week postpartum SMM rates, a difference that persisted after adjustment for potential confounding variables. Infants born to Somali individuals experienced modestly better neonatal outcomes, with lower rates of composite morbidity including respiratory complications, Apgar scores, and NICU admissions.

Results in the context of what is known

SMM has emerged as an essential complement to maternal mortality surveillance, which is hindered by the rarity of events and the insufficient number of cases for robust analysis.²⁰ The CDC's standardized approach captures life-threatening conditions requiring significant intervention and provides approximately 100 times more cases annually than maternal mortality, enabling comprehensive assessment of healthcare quality.^20,21 By employing the 21-indicator composite criteria, our study captures the spectrum of severe complications systematically rather than examining isolated outcomes. Acknowledging that the majority of maternal mortality occurs in the postpartum period, we also extended our evaluation of SMM to encompass this high-risk time.

Our combined SMM rates (1.2% for Somali birthing individuals and 1.8% for non-Somali black birth individuals) are consistent with, though not directly comparable to, national estimates for the Black birthing diaspora that use a combined delivery-through-postpartum timeline (2.1%–2.2%).^22,23 Both groups in our study likely experienced elevated SMM rates compared to White populations, as national data demonstrates that Black individuals have 1.7 to 2.2 times higher SMM rates than White individuals across insurance types.^22,24 These findings underscore that systemic inequities affect Black individuals regardless of nativity or subpopulation.

The modestly lower neonatal composite morbidity among Somali births warrants cautious interpretation. Although gestational age and birth weight were statistically higher among Somali infants, the clinical significance of these differences is uncertain, given a gestational age difference of less than one week at term and comparable size-for-gestational-age distributions between groups. Furthermore, the adjusted confidence interval approaching 1.0 suggests a modest effect size susceptible to residual confounding, and the use of multiple comparisons increases the risk of Type I error. These findings should therefore be considered hypothesis-generating.

Our analysis also revealed no significant differences in SMM or neonatal outcomes between Somali individuals who required versus did not require interpreter services, challenging the assumption that language barriers are the primary drivers of adverse outcomes. This contrasts with literature demonstrating that inadequate communication increases risk of adverse outcomes, including inadequate prenatal care and severe maternal morbidity among migrants.^25,26 However, when professional interpreter services are utilized effectively, they improve outcomes. A large cohort study found interpreter use was associated with decreased odds of small for gestational age status (adjusted OR: 0.67, 95% CI: 0.53–0.84),²⁷ while comprehensive reviews demonstrate that professional interpretation improves communication quality, reduces errors, and enhances care outcomes including patient comprehension and satisfaction.²⁸ An alternative analytical approach would examine language barriers as the primary exposure across the entire study population; however, only 7.0% of non-Somali Black individuals in our cohort required interpreter services, limiting statistical power for such an analysis.

Clinical implications

These findings carry several implications for clinical practice and health system policy. First, clinicians should avoid the assumption that immigration status, language barriers, or cultural differences inherently confer higher perinatal risk. Our data demonstrate that within the Black population, Somali immigrants experienced comparable SMM and modestly better neonatal outcomes despite facing additional structural challenges, which should prompt reflection on unconscious biases in clinical risk assessment. Second, these results underscore the importance of disaggregating data within the Black population. Treating “Black” as a monolithic category obscures meaningful within-group variation in demographics, clinical characteristics, and outcomes; health systems and researchers should invest in granular ethnicity data collection to identify subgroup-specific patterns and needs.

Research implications

Future research should incorporate prospective designs with acculturation measures, duration of U.S. residence, qualitative exploration of culturally protective factors, and multisite comparisons across healthcare systems with varying levels of culturally responsive infrastructure to better understand the mechanisms underlying within-group variation in perinatal outcomes among Black populations. Future studies with larger populations of non-English-speaking Black birthing individuals from diverse language backgrounds could better address the role of language barriers as a primary exposure.

Strengths and limitations

Study strengths include our multisite investigation spanning multiple hospital settings, use of standardized CDC criteria ensuring national comparability, and unique within-diaspora comparison addressing a critical literature gap. Conducting this study in Minnesota, home to the largest U.S. Somali population,²⁹ maximizes relevance and representation.

Important limitations warrant consideration. The retrospective design limits causal inference between observed factors and outcomes. Findings may have limited generalizability to other immigrant populations or regions with different healthcare systems. Our health system has established Somali interpreter services and culturally informed care practices, and Minnesota's large Somali community provides substantial social support infrastructure. Results may differ in settings with smaller Somali populations, limited interpreter services, or less developed cultural competency. Reliance on health system data introduces potential misclassification by ethnicity, and we lacked data on the duration of U.S. residence, acculturation measures, specific immigration-related stressors, and generational status. These are factors that literature on the "healthy immigrant effect" and "Hispanic paradox" suggest may significantly modify health outcomes over time and could introduce heterogeneity within our Somali cohort.³⁰ The non-Somali Black comparator was likewise heterogeneous: although the majority reported English as their primary language and U.S. birth, it also included individuals born in Ethiopia, Kenya, and other African and Caribbean nations (Supplemental Table 1). Given our primary interest in comparing Somali to non-Somali Black individuals, who are a heterogeneous population, we did not disaggregate this control population for this study, which may attenuate or obscure between-group differences. Finally, we may have incompletely captured SMM events if individuals received care outside our network, potentially underestimating true maternal morbidity burden.

Conclusions

This study demonstrates comparable severe maternal morbidity rates between Somali and non-Somali Black individuals during birth hospitalization, with lower postpartum severe maternal morbidity and neonatal composite morbidity among Somali births, in an academic-community health system. These findings suggest that the relationship between immigration, cultural factors, and perinatal outcomes within Black populations is complex and warrants specific investigation

CRediT authorship contribution statement

Isahaq Abdullahi: Writing – original draft, Investigation, Data curation, Conceptualization. Lensa Toka: Writing – review & editing, Investigation, Data curation. Katelyn M. Tessier: Writing – review & editing, Formal analysis. Cresta W. Jones: Writing – review & editing, Conceptualization. Bethany A. Sabol: Writing – review & editing, Conceptualization. Sarah A. Wernimont: Writing – review & editing, Supervision, Resources, Project administration, Methodology, Investigation, Conceptualization.