The role of pregnancy- related comorbidities in maternal health disparities among Asian American/Pacific Islanders

Technical Summary: The Role of Pregnancy-Related Comorbidities in Maternal Health Disparities Among Asian American/Pacific Islanders

Problem Statement
Asian American/Pacific Islander (AAPI) populations are frequently aggregated in health research, a practice that masks significant heterogeneity in social standing, immigration history, and health outcomes. While severe maternal morbidity (SMM) affects AAPI groups, previous estimates suggest substantial variation among subgroups (e.g., Filipina, Indian, Chinese, Pacific Islander) that approaches or exceeds disparities observed in Non-Hispanic Black populations compared to Non-Hispanic Whites. A critical gap exists in understanding the specific drivers of these disparities. Specifically, it is unclear how pregnancy-related comorbidities—namely gestational hypertension, pre-eclampsia, and gestational diabetes—mediate the relationship between race/ethnicity and SMM across different AAPI subgroups. This study seeks to estimate counterfactual disparities by isolating the pathway through these comorbidities to determine if they account for the observed differences in SMM rates.

Methodology
The study utilized a population-based analysis of linked birth certificate and hospital discharge data from California (2011–2020), encompassing 1,849,698 births. The analytic sample was restricted to non-Hispanic individuals identifying as White, Chinese, Indian, Filipina, or Pacific Islander, excluding those with multiple race/ethnicity reports or Hispanic ethnicity.

• Outcome: Severe Maternal Morbidity (SMM) was defined using the CDC's updated SMM Index (21 indicators) captured in delivery and postpartum hospital records, ambulatory files, and emergency room admissions up to 42 days postpartum.
• Exposure: Race/ethnicity, conceptualized as a social construct and marker of exposure rather than a biological trait.
• Mediator: A composite of pregnancy-related comorbidities (gestational hypertension, pre-eclampsia, and gestational diabetes), identified via vital statistics checkboxes or ICD codes.
• Covariates: Maternal age, parity, pre-pregnancy BMI, education level, and insurance type were derived from vital statistics and hospital records.
• Statistical Approach: The authors employed Marginal Structural Models (MSMs) to estimate Controlled Direct Effects (CDEs). This approach allowed for the isolation of the effect of race/ethnicity on SMM that does not operate through the mediator (comorbidities), while accounting for time-varying confounding.
  • Total Effect: Estimated via Poisson regression to determine existing disparities.
  • Counterfactual Scenarios: CDEs were calculated under two hypothetical conditions:
    1. M=0: No individuals in the population have pregnancy-related comorbidities.
    2. M=1: All individuals in the population have pregnancy-related comorbidities.
  • Sensitivity Analyses: Included re-running models with a "non-transfusion SMM" definition, stratifying by nativity (US-born vs. foreign-born) where sample sizes permitted, and calculating mediational E-values to assess robustness against unmeasured confounding.

Key Results

• Prevalence: All AAPI groups exhibited higher prevalences of pregnancy-related comorbidities compared to Non-Hispanic Whites. The burden was highest among Filipinas (28.8% with any comorbidity) and Pacific Islanders (23.9%), followed by Indians (25.2%) and Chinese (17.9%).
• Total Effects: Filipinas and Pacific Islanders experienced the largest disparities in SMM compared to Whites, with Risk Ratios (RR) of 1.64 (95% CI: 1.58, 1.70) and 1.67 (95% CI: 1.55, 1.80), respectively. Chinese and Indian groups showed smaller total effect disparities (RR 1.03 and 1.14, respectively).
• Mediation Analysis (M=1 Condition): When the mediator was set to 1 (everyone has comorbidities), disparities were eliminated for Chinese and Indian groups (CDE RR: 0.75 and <1.0, respectively) and alleviated for Filipinas and Pacific Islanders (CDE RR: 1.21 and 1.21, respectively). This suggests that for Chinese and Indian groups, the disparity is almost entirely driven by the higher prevalence of comorbidities; if comorbidity rates were equalized, these groups would have better or equal SMM outcomes compared to Whites.
• Mediation Analysis (M=0 Condition): When the mediator was set to 0 (no one has comorbidities), disparities for Chinese and Indian groups increased (CDE RR: 1.14 and 1.24), while disparities for Filipinas and Pacific Islanders remained similar to total effects. This indicates that even in the absence of pregnancy-related comorbidities, significant unexplained disparities persist for these groups, and for Chinese/Indian groups, the disparity is actually stronger without the comorbidity pathway.
• Interaction: A negative interaction was observed between race/ethnicity and comorbidities, suggesting that the detrimental effect of comorbidities on SMM is attenuated in these AAPI groups compared to Whites, despite their higher prevalence of these conditions.

Significance and Claims
The paper claims that pregnancy-related comorbidities are a substantial driver of SMM disparities for AAPI groups, but the extent of this contribution varies significantly by subgroup.

• Disaggregation is Critical: The findings reinforce that aggregating AAPI data obscures distinct risk profiles. For Chinese and Indian populations, disparities in SMM are largely explained by the higher prevalence of pregnancy-related comorbidities. Conversely, for Filipinas and Pacific Islanders, comorbidities explain only a portion of the disparity; significant unexplained disparities remain even when comorbidities are controlled for.
• Intervention Implications: The results suggest that while preventing and managing pregnancy-related comorbidities is vital, it is insufficient to eliminate disparities for Filipina and Pacific Islander populations. Tailored interventions addressing other drivers—such as social determinants of health, discrimination, or immigration-related stressors—are necessary for these groups.
• Methodological Contribution: By using counterfactual mediation analysis, the study moves beyond simple association to quantify how much of a disparity is "explained" by a specific biological pathway versus how much remains "unexplained" by other social or structural factors.

The authors conclude that future research must investigate the management of these conditions and other potential drivers, such as discrimination and acculturation, to fully address maternal health inequities within the diverse AAPI population.