Background
Maternal and infant mortality in the U.S. remain the highest among high-income nations, with disparities worsening after Dobbs v. Jackson Women’s Health Organization. This study examined how abortion bans and partisan political control are associated with maternal and infant mortality, while accounting for sociodemographic, economic, and healthcare factors.
Methods
We conducted a cross-sectional ecological study of all 50 U.S. states, linking maternal mortality ratios (MMR) and infant mortality rates (IMR) with abortion policy classifications (total ban, ban at ≤ 18 weeks, ban after 18 weeks, no gestational ban) and partisan control of governors, legislatures, and senates. Multivariable regressions adjusted for sociodemographic, economic, and healthcare variables.
Results
Unadjusted analyses showed MMR and IMR were higher on average in Republican-led states (27 vs. 20 per 100,000; 6.3 vs. 5.0 per 1,000; all p < 0.01), with political control accounting for up to one-quarter of the variance (R²=0.21–0.25). In adjusted models, total abortion bans were correlated with higher MMR (β = 5.28, p = 0.0315, R²=0.81) and IMR (β = 1.15, p = 0.0014, R²=0.86). Higher mortality correlated with greater fertility and larger Black population shares; protective factors included higher income, state investment, education, life expectancy, Hispanic population, and healthcare access. These patterns clustered by party control.
Conclusions
In this ecological analysis, abortion bans and Republican political control were correlated with higher maternal and infant mortality at the state level. While individual-level causality cannot be inferred, addressing structural disparities may help reduce preventable deaths.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12978-026-02309-w.
Keywords: Maternal mortality, Infant mortality, Abortion policy, Political control, United States, Ecological study
Plain language summary
Maternal and infant deaths remain higher in the United States than in other wealthy countries. These problems have become worse since the Dobbs v. Jackson Women’s Health Organization decision, which allowed many states to ban or restrict abortion. Our study looked at whether abortion laws and political control in each state are related to the number of women who die during or shortly after pregnancy (maternal mortality) and the number of infants who die before their first birthday (infant mortality).
We studied all 50 states uosing publicly available data. We compared maternal and infant death rates with state abortion policies (from total bans to no gestational bans) and with whether Republicans or Democrats controlled the governor’s office and state legislatures. We also considered other important factors such as income, education, access to healthcare, and racial/ethnic makeup of the population.
We found that states with abortion bans and Republican political control had higher maternal and infant death rates than states without bans and with Democratic control. For example, maternal mortality was about one-third higher in Republican-led states. Higher fertility rates and larger Black populations were linked with more deaths, while protective factors included higher income, greater state investment, more education, longer life expectancy, and larger Hispanic populations.
Our findings show that restrictive abortion policies and political control are linked with worse outcomes for mothers and infants. Although this type of study cannot prove direct cause, it highlights the need to address underlying inequalities to save lives.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12978-026-02309-w.
Background
In the wake of the U.S. Supreme Court’s decision in Dobbs v. Jackson Women’s Health Organization (2022), reproductive health policy has undergone rapid and divergent changes across states [1]. While some jurisdictions moved swiftly to ban or severely restrict abortion access, others reaffirmed or expanded legal protections [2]. This shifting legal landscape has placed renewed focus on the public health consequences of reproductive policy, particularly regarding maternal and infant health outcomes [3].
Even before the Dobbs decision, the United States faced a maternal health crisis marked by high mortality rates and significant racial and geographic disparities. The U.S. consistently reports the highest maternal mortality rate (MMR) among high-income countries, and this rate has been worsening over time. Maternal mortality rose from 17.4 deaths per 100,000 live births in 2018 to 32.9 in 2021 an alarming trend that underscores the urgent need for policy and health system interventions [4]. In 2021, non-Hispanic Black women experienced a MMR of 69.9 nearly 2.6 times higher than the rate for non-Hispanic White women (26.6). Mortality also increased significantly with age, from 20.4 among women under 25 to 138.5 among those aged 40 and older [4]. In the same period, the highest MMR were concentrated in Southern states, including Tennessee, Mississippi, Alabama, Arkansas, and Louisiana underscoring the regional nature of this public health crisis [5].
Infant mortality, while improved over past decades, remains a persistent concern in many Southern and Midwestern states. In 2022, the overall U.S. infant mortality rate was 5.6 deaths per 1,000 live births. However, this national average conceals striking disparities. Mississippi reported the highest state-level rate at 9.1 per 1,000, nearly three times higher than Massachusetts, which had the lowest rate at 3.3 per 1,000. Racial disparities are equally concerning: Non-Hispanic Black infants had an infant mortality rate of 10.9 per 1,000 live births, nearly double the national average. These stark geographic and racial differences are not merely biological in origin; they reflect systemic inequities and social determinants of health shaped by longstanding disparities in policy, access to care, and infrastructure [6, 7].
These stark disparities in maternal and infant mortality are not random—they reflect deeper systemic forces shaped by political decision-making. Increasingly, public health researchers have turned their attention to what they term the “political determinants of health.” This concept posits that political ideologies, governance structures, and policy priorities fundamentally influence the organization and outcomes of health systems [8]. In other words, the laws enacted and resources allocated by elected officials help create the structural conditions that drive either health or illness [8]. For instance, one longitudinal study found that state-level infant mortality rates—particularly post-neonatal deaths—were significantly higher under Republican-controlled legislatures than under Democratic or mixed control [9]. Similarly, after the political realignment of the 1960s, maternal mortality rates in the U.S. became consistently higher under Republican presidencies compared to Democratic ones, reversing prior trends [8].
Abortion restrictions rarely standalone, they are embedded within a broader policy landscape defined by disinvestment in public health infrastructure, underinsurance, and inadequate maternal care access. Maternity care deserts are significantly more common in abortion-restrictive states than in abortion-access states, which also have fewer OB/GYNs and certified nurse-midwives per birth [10]. Research indicates that Medicaid expansion is associated with lower maternal mortality, particularly among non-Hispanic Black women [11]. States that expanded Medicaid are also more likely to maintain liberal abortion policies, reflecting broader investments in reproductive and maternal health care [10].
Together, these findings underscore how abortion restrictions frequently coincide with weakened maternal health infrastructure and partisan governance, jointly exacerbating risks to maternal and infant health. While prior studies have explored abortion laws or health disparities independently, few have examined how abortion policy and political control interact to shape health outcomes. Moreover, the alignment between restrictive reproductive policies and broader structural disadvantages, such as higher lack of insurance, poverty, and lower life expectancy remains underexplored in the post-Roe era. This ecological study examines the associations between state-level abortion policy classification, political party control, and maternal and infant mortality, while also accounting for sociodemographic, economic, and healthcare determinants. By analyzing cross-sectional data across all 50 states, we assess how variations in reproductive policy environments and political leadership relate to maternal and infant health outcomes and broader measures of structural disadvantage.
Methods
Study design
We conducted a cross-sectional ecological study using U.S. states as the unit of analysis to examine associations between abortion policy classifications, political party control, and population health outcomes. Primary exposures included state-level reproductive policy categories and partisan control of the executive and legislative branches, reflecting the post-Dobbs v. Jackson Women’s Health Organization landscape. Primary outcomes were maternal mortality rates (MMR) and infant mortality rates (IMR), measured using the most recent publicly available data. Adjusted analyses accounted for socio-demographic, economic, and healthcare access related covariates, including race, poverty rate, median household income, life expectancy, and Children’s Health Insurance Program (CHIP) expenditure. Political control was represented using three variables: governor party, state house majority, and state senate majority. All variables were harmonized to reflect data from 2022 or the nearest available year. Variables obtained from multiple public data sources were harmonized to a common state-level unit of analysis and expressed as rates or per-capita measures where appropriate to ensure comparability across sources. Analyses were conducted entirely at the state level; no individual-level data were used.
Because this study is cross-sectional, exposures and covariates were defined using the most recent data available for each variable. Political control and abortion policy classifications therefore reflect the contemporary state-level policy environment, consistent with other covariates such as insurance coverage, socioeconomic indicators, and public spending. Maternal mortality outcomes were drawn from the most recent finalized data available (2018–2022), as current-year mortality data are not yet released. Political control was conceptualized as a relatively stable policy context rather than a transient exposure, and this approach was chosen to maintain internal consistency across state-level measures.
Covariate selection and causal framework
Covariates were selected a priori based on established literature on state-level determinants of maternal and infant mortality and on a prespecified causal framework. We developed a directed acyclic graph (DAG) (Figure S1) to represent hypothesized relationships among political control, abortion policy, structural socioeconomic conditions, healthcare access/investment, and maternal and infant mortality outcomes. The DAG was used to identify a minimally sufficient adjustment set to reduce confounding of the associations of interest while avoiding overadjustment for variables likely to lie on indirect pathways. Accordingly, models adjusted for state-level measures reflecting population composition (race/ethnicity), socioeconomic context (income, poverty, education), health system access and capacity (insurance coverage, primary care availability), and broader population health (life expectancy).
Study measures
Primary exposures
The primary exposures were each state’s legal environment surrounding “abortion access” and its “partisan political control”. Abortion policy was defined using the Guttmacher Institute’s classification of abortion bans in effect as of July 7, 2025 [12]. We utilized the Guttmacher-defined four-tier framework that categorizes states according to the most restrictive gestational policy in place: (1) Total ban: States with a full prohibition on abortion throughout pregnancy, typically allowing only narrow or unworkable exceptions. (2) Ban at or before 18 weeks: States with abortion bans imposed at early gestational points such as 6, 12, or 18 weeks, often before fetal viability and with limited accommodations for rape, incest, or health risks. (3) Ban after 18 weeks: States that permit abortion during early pregnancy but restrict access after 18 weeks gestation, commonly aligning with statutory definitions of fetal viability or late-term thresholds. (4) No gestational ban: States, including the District of Columbia, with no current gestational restrictions on abortion, where access is determined by patient and provider discretion. “Political control” was represented using three variables: governor party, state house majority, and state senate majority. Each was coded as Democratic or Republican based on the party in control as of 2025.
Primary outcomes
The primary outcomes was maternal mortality and infant mortality measures. For maternal health, we examined the maternal mortality ratio (MMR) defined as the number of maternal deaths per 100,000 live births using the standard definition (pregnancy-related death within 42 days of end of pregnancy) from CDC vital statistics. For infant health, the infant mortality rate (IMR), infant deaths before 1 year of age per 1,000 live births were analyzed as the key indicator of infant outcomes [13]. MMR and IMR were analyzed as continuous variables in all descriptive and regression analyses.
Data sources
Maternal Mortality Data and Bayesian Adjustment: Maternal mortality rates (MMRs) were obtained from the National Center for Health Statistics (NCHS) through the National Vital Statistics System, using pooled data from 2018 to 2022 [13]. The CDC intentionally aggregated five years of data to improve statistical reliability, especially for smaller states with fewer maternal deaths. MMRs are reported per 100,000 live births and tabulated by state of residence. However, for 12 states and jurisdictions with fewer than 20 maternal deaths over this period, the CDC suppressed state-level MMRs due to statistical instability and confidentiality concerns. These included: Alaska, Delaware, District of Columbia, Hawaii, Maine, Montana, New Hampshire, North Dakota, Rhode Island, South Dakota, Vermont, and Wyoming. For states with suppressed data, we applied a Bayesian shrinkage method to estimate stabilized MMRs by (i) computing each state’s observed MMR, (ii) shrinking it toward the national pooled mean using a pre-specified prior SD, and (iii) weighting observed and prior means by inverse variances. Observed MMRs were first calculated as:
 |
We then used a Bayesian model that shrinks each state’s observed rate toward the national pooled average (22.3 per 100,000), using a prior standard deviation (SD) of three. Although the observed SD across reportable states was approximately 7, we selected a smaller value to smooth unstable estimates, reduce noise, and avoid over-interpreting extreme values in low-count states while preserving meaningful variation where data allowed. Final adjusted MMRs were computed as weighted averages of observed and prior means, with weights based on inverse variances. In practice, the resulting adjustments for suppressed states were modest (generally within a few deaths per 100,000 live births), and sensitivity analyses using alternative reasonable prior specifications yielded nearly identical estimates. This approach stabilized noisy estimates and enabled full inclusion of all states in regression analyses.
Abortion data and imputation
Abortion rates were obtained primarily from the CDC’s Abortion Surveillance System for 2022 [14]. However, four states—California, New Jersey, New Hampshire, and Maryland—did not report abortion data to the CDC for that year. For these non-reporting states, we imputed abortion rates and out-of-state abortion proportions using modeled estimates from the Guttmacher Institute’s 2017 national study, the most recent and comprehensive source available for these jurisdictions and commonly cited in reproductive health research [15]. The following estimates were applied per 1,000 women aged 15–44: California (16.5), New Jersey (18.7), New Hampshire (9.3), and Maryland (20.4). Corresponding estimated proportions of abortions obtained by out-of-state residents were: California (3%), New Jersey (4%), New Hampshire (16%), and Maryland (7%). These values were selected to approximate 2022 rates and reflect established imputation practices used to address systematic surveillance gaps in non-reporting states. This approach represents deterministic substitution using external surveillance-based estimates.
Political party data source
Political party control data were obtained from the Kaiser Family Foundation (ww.kff.org) State Health Facts platform’s “State Political Parties” indicator, which provides the party affiliation of each state’s governor, state senate majority, and state house majority as of 2025 [16]. To ensure accuracy, we confirmed these classifications using official records from the National Governors Association (NGA) for current gubernatorial affiliations and the National Conference of State Legislatures (NCSL) for legislative partisan composition following the 2024 elections. Each branch was coded as Democratic or Republican; Nebraska’s unicameral, nonpartisan legislature was excluded from party control analyses.
CHIP data and derived variables
State-level Children’s Health Insurance Program (CHIP) “enrollment” data were obtained from the CMS Medicaid & CHIP Performance Indicator dataset for the January 2025 reporting period. This dataset includes monthly CHIP enrollment counts, eligibility determinations, and administrative expenditures. For Rhode Island, which did not report January 2025 data at the time of this study, we used November 2024 data as a proxy [17].
CHIP “expenditure” data was obtained from Kaiser Family Foundation (https://www.kff.org/) State Health Facts data repository [18]. To adjust for population size, we derived two per capita indicators: State CHIP spending per capita defined as 2022 state CHIP expenditures divided by the 2024 state population and Federal CHIP spending per capita defined as 2022 federal CHIP contributions divided by the 2024 state population. Population estimates were obtained from the U.S. Census Bureau’s Vintage 2024 Population Estimates, specifically the Annual Estimates of the Resident Population as of July 1, 2024 (NST-EST2024-POP) [19]. These per capita measures standardized CHIP investments across states and were used in our regression models to explore associations with maternal and infant health outcomes. CHIP expenditures were used as a standardized indicator of state investment in maternal and child health with consistent cross-state availability, rather than as a comprehensive measure of all maternal health–related programs.
Fertility data
State-level fertility rates and total births for 2022 were obtained from the Centers for Disease Control and Prevention (CDC), National Center for Health Statistics (NCHS) [20]. These data reflect the general fertility rate, defined as the number of live births per 1,000 women aged 15–44 years, and are reported by state of residence. Each state’s data were accessed via the NCHS Pressroom’s fertility rate portal, which compiles official birth statistics and demographic indicators used in national surveillance and health planning.
Statistical analysis
All data sources were merged into a single analytic dataset using state as the common identifier. We conducted descriptive and multivariable analyses to examine associations between state-level abortion policy, political party control, and maternal and infant mortality outcomes. The District of Columbia was excluded due to its unique governance structure and non-comparability in legislative partisan control. Minnesota was excluded from state house control analyses because the chamber was evenly divided, and Nebraska was excluded from legislative party control analyses because it has a unicameral, nonpartisan legislature, as detailed in the table footnotes. All remaining states were retained in the analyses, and no states or observations were excluded as outliers based on outcome, exposure, or covariate values. Descriptive statistics summarized exposures, outcomes, and covariates across states. Means, standard deviations, and medians were reported for continuous variables, and proportions for categorical variables. Bivariate comparisons were performed using t-tests to compare group means for normally distributed variables. For non-normal variables, the Kruskal–Wallis test, a nonparametric method that compares ranked group medians, was used to assess unadjusted differences in maternal mortality ratio (MMR) and infant mortality rate (IMR) across abortion policy categories and political party variables. All covariates listed in Table 1—including socio-demographic, economic, and healthcare access related covariates, such as race, poverty rate, median household income, fertility rate, CHIP spending, and life expectancy—were entered into the initial linear regression model. Backward elimination was then applied, sequentially removing the least significant variables based on F statistics until only those meeting the retention criterion of 0.20 remained. A retention threshold of p < 0.20 was used to reduce the risk of residual confounding and overfitting in a small ecological sample, where reliance on conventional p < 0.05 thresholds can lead to unstable estimates and exclusion of conceptually important covariates. Abortion policy and political party control were forced to remain in all models, and covariate inclusion was primarily guided by the DAG-informed causal framework, with statistical criteria used to support model parsimony. Final adjusted models estimated associations between the two primary exposures—abortion policy category and political party control—and the two primary outcomes (MMR and IMR). Abortion policy was modeled as a four-level categorical variable, with “no gestational ban” as the reference group. Political party control was represented using three binary indicators for governor, house, and senate majorities, with Democratic coded as 0 (reference) and Republican coded as 1. Sensitivity analyses compared effect estimates from models including all states (with Bayesian-adjusted maternal mortality estimates for the 11 states with suppressed data). Substantive conclusions and coefficient magnitudes were comparable across specifications, indicating that Bayesian adjustment did not materially influence the primary findings.
Table 1.
State-Level Reproductive and Health Indicators by Abortion Policy Category, U.S
| Total Ban (n = 13) |
Ban at 6–18 Weeks (n = 7) |
After 18 Weeks (n = 21) |
No Ban (n = 9) |
||
|---|---|---|---|---|---|
| Indicator | Mean (SD)Median | Mean (SD) Median | Mean (SD) Median | Mean (SD) Median | P-value |
| Infant deaths per 1,000 live births | 6.93 ± 1.01 (6.89) | 6.10 ± 0.81 (5.98) | 5.18 ± 1.17 (5.59) | 5.21 ± 1.05 (4.89) | 0.0008 |
| Maternal deaths per 100,000 live births | 31.49 ± 7.13 (30.9) | 25.04 ± 6.16 (25.1) | 20.9 ± 5.05 (21.9) | 19.82 ± 5.25 (19.10) | 0.0005 |
| Abortions per 1,000 women aged 15–44 | 3.23 ± 1.77 (2.8) | 11.77 ± 6.77 (7.2) | 11.75 ± 5.68 (9.7) | 14.98 ± 6.67 (11.7) | < 0.0001 |
| Percent abortions obtained by out of state residents | 21.78 ± 14.28 (21.0) | 18.29 ± 8.36 (21.0) | 13.86 ± 15.51 (10.0) | 17.78 ± 18.59 (12.0) | 0.1526 |
| Live births per 1,000 women aged 15–44. | 59.95 ± 2.84 (59.7) | 58.71 ± 2.98 (57.6) | 53.90 ± 3.90 (53.3) | 54.32 ± 6.25 (54.0) | 0.0004 |
| Percent of uninsured population | 8.9 ± 2.8 (8.60) | 8.47 ± 2.3 (9.0) | 6.26 ± 2.25 (6.10) | 6.32 ± 2.35 (6.30) | 0.0119 |
| Medicaid eligibility limit for pregnant women (% of FPL) | 206.31 ± 57.59 (210.0) | 221.0 ± 74.27 (201.0) | 212.57 ± 43.60(205.0) | 231.11 ± 35.08 (213.0) | 0.6149 |
| Medicaid eligibility limit (% of FPL) | 189.77 ± 28.67 (201.0) | 216.71 ± 46.86 (213.0) | 227.76 ± 58.32 (215.0) | 243.67 ± 64.66 (217.0) | 0.1571 |
| State CHIP spending per capita (USD) | 11.54 ± 3.18 (11.56) | 10.0 ± 3.01(10.0) | 14.48 ± 10.56 (11.91) | 13.09 ± 13.17 (10.67) | 0.8343 |
| Federal CHIP spending per capita (USD) | 54.38 ± 18.41 (47.09) | 38.68 ± 10.02 (40.91) | 43.36 ± 22.23 (39.31) | 44.40 ± 28.6 (40.29) | 0.1717 |
| Percent of the population below the FPL | 14.71 ± 2.58 (15.2) | 11.84 ± 1.75 (12.3) | 11.1 ± 1.43 (10.9) | 11.29 ± 2.74 (9.9) | 0.0013 |
| Median household income (USD) | 64,671 ± 7157 (62212) | 75,142 ± 8411 (73311) | 82,199 ± 10,136 (80306) | 83,989 ± 12,581 (85086) | < 0.0001 |
| Life expectancy at birth ( years) | 73.42 ± 2.05(72.5) | 76.07 ± 1.88 (76.1) | 77.2 ± 1.63 (77.1) | 76.86 ± 2.05 (77.4) | 0.0003 |
| Percent of population overweight or obese | 69.06 ± 1.61 (69.4) | 63.3 ± 3.26 (66.9) | 64.75 ± 2.81 (64.9) | 64.29 ± 2.65 (64.9) | 0.0001 |
| State spending per capita on services (USD) | 8131.23 ± 2381 (7786.0) | 7095.43 ± 1427 (6812.0) | 10133.48 ± 2409 (9585.0) | 11428.67 ± 4323 (10893.0) | 0.0044 |
| State taxes per capita in revenues (USD) | 3530.39 ± 541.3 (3415.0) | 3427.57 ± 521.46 (3594.0) | 4968.33 ± 1323.0 (4614.0) | 5138.67 ± 1309.0 (4932.0) | 0.0003 |
| Percent of state population White | 68.94 ± 13.77 (71.3) | 64.76 ± 13.02 (62.0) | 64.70 ± 17.85 (66.2) | 62.63 ± 16.55 (64.7) | 0.7399 |
| Percent of state population Black | 13.05 ± 10.85 (10.5) | 13.89 ± 11.37 (14.2) | 7.06 ± 5.77 (5.0) | 7.91 ± 9.01 (3.6) | 0.2479 |
| Percent of state population Hispanic | 9.67 ± 9.73 (7.1) | 13.41 ± 6.95 (11.4) | 14.29 ± 9.95 (11.1) | 15.98 ± 14.23 (12.7) | 0.2308 |
| Primary care practitioner deficit (per 100,000 persons in shortage areas) | 16.85 ± 4.97 (16.44) | 17.26 ± 5.0 (17.77) | 16.92 ± 5.5 (17.1) | 16.07 ± 7.65 (18.16) | 0.9811 |
| Percent with high school or higher | 89.75 ± 1.97 (89.5) | 91.34 ± 1.64 (90.6) | 91.6 ± 2.55 (91.8) | 92.19 ± 2.32 (92.0) | 0.0489 |
| Percent with bachelor degree or higher | 29.45 ± 3.29 (28.9) | 35.03 ± 2.3 (35.3) | 37.04 ± 4.5 (37.0) | 39.09 ± 5.76 (40.0) | < 0.0001 |
P-values are from Kruskal-Wallis non-parametric test comparing ranked medians; statistically significant p-values (p <0.05) are indicated in bold
FPL Federal Poverty Level
Primary care practitioner deficit was calculated as the number of additional primary care practitioners needed to remove Health Professional Shortage Area (HPSA) designations (HRSA), divided by the shortage-area population, per 100,000
To assess model stability in the 50-state ecological sample, we conducted standard regression diagnostics, including evaluation of multicollinearity using variance inflation factors (VIFs) and condition indices. VIFs for primary exposure variables were within conventional ranges across model specifications. Variation in VIFs was primarily attributable to correlated socioeconomic covariates, as expected in state-level analyses. These covariates were retained based on the prespecified directed acyclic graph (DAG) to reduce residual confounding. As a sensitivity analysis, we performed leave-one-out (jackknife) refitting of each fully adjusted model. Primary exposure estimates remained stable across exclusions, with no change in direction or substantive conclusions.
We reported β coefficients (representing the change in outcome per unit change in predictor), standard errors, and p-values. For variables where the β estimate indicated a statistically significant but very small change, coefficients were scaled to reflect a 100 unit increment. Statistical significance was assessed using a two-sided α level of 0.05. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC). This study was deemed exempt from IRB review as it used publicly available, de-identified, aggregate state-level data.
Results
Table 1 shows states with more restrictive abortion policies—particularly those with total bans—consistently demonstrated worse maternal and infant health outcomes, higher fertility, and greater socioeconomic disadvantage compared with states with fewer or no restrictions. The maternal mortality ratio (MMR) was highest in total ban states (31.5 per 100,000 live births) and lowest in no-ban states (19.8; p = 0.0005), while infant mortality showed a similar pattern (6.93 vs. 5.21 deaths per 1,000; p = 0.0008). Abortion rates were markedly lower in total ban states (3.23 per 1,000 women) compared with states with gestational limits (11.7) or no bans (15.0; p < 0.0001), whereas fertility rates were highest in total ban states (59.9 per 1,000; p = 0.0004). Socioeconomic indicators also followed this gradient: total ban states had lower incomes, higher poverty and uninsured rates, shorter life expectancy, and greater obesity prevalence (p < 0.05 for all). No significant differences were observed across policy categories in CHIP spending, Medicaid eligibility thresholds, or the proportion of abortions obtained by out-of-state residents.
Table 2 shows states with Republican-controlled legislatures or governors consistently had worse maternal and infant health outcomes compared with Democratic-controlled states. Across houses, senates, and governors, maternal mortality averaged about 27 vs. 20 deaths per 100,000 and infant mortality about 6.3 vs. 5.0 per 1,000 live births (all p < 0.01). Fertility was higher and abortion rates lower in Republican-led states (both p < 0.0001). Socioeconomic indicators followed the same gradient, with Republican states reporting lower household incomes, shorter life expectancy, and higher poverty, uninsured, and obesity rates, while Democratic states invested more in public services and generated greater tax revenues. Demographic and educational differences were also evident: Republican-led states had a higher share of White residents, whereas Democratic states had more Black and Hispanic residents and higher rates of bachelor’s degree attainment. No significant differences were observed for CHIP spending, Medicaid eligibility thresholds, or the primary care practitioner deficit.
Table 2.
State-Level Reproductive and Health Indicators by Political Control, United States
| State House Control |
State Senate Control |
State Governor Party |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Democrat (n = 19) |
Republican (n = 29) |
Democrat (n = 19) | Republican (n = 29) | Democrat (n = 23) | Republican (n = 27) | ||||
| Indicator | Mean (SD) Median | Mean (SD) Median | P-value | Mean (SD) Median | Mean (SD) Median | P-value | Mean (SD) Median | Mean (SD) Median | P-value |
| Infant deaths per 1,000 live births | 5.01 ± 1.11 (4.54) | 6.31 ± 1.16 (6.61) | 0.0006 | 5.02 ± 1.12 (4.52) | 6.30 ± 1.16 (6.61) | 0.0008 | 5.26 ± 1.12 (5.69) | 6.21 ± 1.27 (6.61) | 0.0079 |
| Maternal deaths per 100,000 live births | 19.98 ± 5.06 (19.6) | 27.07 ± 7.28 (24.20) | 0.0004 | 19.68 ± 5.19 (19.35) | 27.01 ± 7.35 (24.2) | 0.0003 | 20.42 ± 5.83 (19.6) | 27.13 ± 7.23 (24.20) | 0.0008 |
| Abortions per 1,000 women aged 15–44 | 14.62 ± 5.77 (13.0) | 7.26 ± 5.93 (5.0) | < 0.0001 | 14.53 ± 5.68 (12.95) |
7.2 ± 5.85 (5.0) |
< 0.0001 | 13.97 ± 6.43 (13.0) | 6.84 ± 5.1 (5.1) | 0.0001 |
| Percent abortions obtained by out of state residents | 13.63 ± 14.1 (10) | 19.48 ± 15.78 (19.00) | 0.1088 | 13.76 ± 13.76 (9.5) | 19.52 ± 15.76 (19.0) | 0.1034 | 15.39 ± 17.97 (9.0) | 18.81 ± 12.0 (19.0) | 0.0525 |
| Live births per 1,000 women aged 15–44. | 52.56 ± 3.99 (53.1) | 58.31 ± 3.83 (58.7) | < 0.0001 | 52.88 ± 4.09 (53.15) | 58.28 ± 3.86 (58.7) | < 0.0001 | 54.17 ± 3.99 (53.6) | 57.98 ± 4.83 (58.7) | 0.0020 |
| Percent of uninsured population | 5.92 ± 1.93 (6.1) | 8.29 ± 2.68 (8.5) | 0.0034 | 5.78 ± 1.95 (6.1) | 8.32 ± 2.63 (8.5) | 0.0011 | 6.01 ± 1.87 (6.1) | 8.33 ± 2.78 (8.5) | 0.0018 |
| Medicaid eligibility limit for pregnant women (% of FPL) | 227.16 ± 34.64 (214.0) | 205.93 ± 57.6 (201.0) | 0.0297 | 228.95 ± 36.52 (213.5) | 206.62 ± 57.64 (201.0) | 0.0494 | 228.17 ± 41.49 (217) | 204.63 ± 55.36 (202) | 0.0277 |
| Medicaid eligibility limit (% of FPL) | 239.05 ± 60.26 (220.0) | 203.86 ± 45.64 (208.0) | 0.0451 | 241.35 ± 59.72 (220.0) | 203.97 ± 45.68 (208.0) | 0.0351 | 238.26 ± 54.79 (218.0) | 202.96 ± 48.52 (205.0) | 0.0110 |
| State CHIP spending per capita (USD) | 17.23 ± 11.99 (16.73) | 10.67 ± 3.51 (10.57) | 0.0651 | 15.47 ± 13.25 (15.39) | 10.97 ± 3.54 (10.67) | 0.3817 | 15.56 ± 11.95 (14.04) | 10.51 ± 4.26 (10.57) | 0.0749 |
| Federal CHIP spending per capita (USD) | 48.4 ± 25.4 (40.29) | 45.32 ± 18.6 (42.43) | 0.8744 | 46.03 ± 26.31 (39.8) | 45.7 ± 18.3 (42.43) | 0.5552 | 49.26 ± 24.41 (44.71) | 42.77 ± 18.52 (41.12) | 0.4192 |
| Percent of the population below the FPL | 11.22 ± 1.91 (10.5) | 12.98 ± 2.67 (12.6) | 0.0050 | 11.2 ± 1.98 (10.5) | 12.92 ± 2.7 (12.4) | 0.0053 | 11.67 ± 2.15 (10.9) | 12.62 ± 2.79 (12.3) | 0.1888 |
| Median household income (USD) | 86,030 ± 10,466 (84972) | 70,848 ± 9780 (70804) | < 0.0001 | 85,751 ± 10,528 (85029) | 71,008 ± 9789 (70804) | < 0.0001 | 82,414 ± 11,925 (81361) | 72,344 ± 10,705 (71810) | 0.0040 |
| Life expectancy at birth (years) | 77.57 ± 1.68 (77.7) | 74.81 ± 2.18 (74.9) | < 0.0001 | 77.6 ± 1.69 (77.95) | 74.83 ± 2.19 (74.9) | < 0.0001 | 77.13 ± 2.0 (77.4) | 75.03 ± 2.32 (75.0) | 0.0018 |
| Percent of population overweight or obese | 63.74 ± 2.74 (64.5) | 67.42 ± 2.6 (67.8) | 0.0001 | 63.8 ± 2.69 (64.7) | 67.42 ± 2.6 (67.8) | < 0.0001 | 64.46 ± 2.84 (64.9) | 67.32 ± 2.85 (68.0) | 0.0011 |
| State spending per capita on services (USD) | 10544.68 ± 2272 (10893.0) | 8711.03 ± 3382 (7924.0) | 0.0033 | 10455.85 ± 2257 (10724.50) | 8733.172 ± 3379 (7924.000) | 0.0032 | 10488.09 ± 2323 (10556.00) | 8511.444 ± 3311 (7897.000) | 0.0009 |
| State taxes per capita in revenues (USD) | 5334.95 ± 1091.0 (5019.0) | 3772.28 ± 1026 (3594.0) | < 0.0001 | 5341.1 ± 1109.0 (5290.5) | 3791.24 ± 1030.0 (3594.0) | < 0.0001 | 4976.17 ± 1162.0 (4867.0) | 3926.63 ± 1219.0 (3703.0) | 0.0010 |
| Percent of state population White | 58.08 ± 17.6 (58.3) | 69.57 ± 13.09 (73.0) | 0.0193 | 58.94 ± 17.56 (59.85) | 69.59 ± 13.09 (73.0) | 0.0252 | 61.12 ± 16.49 (62.5) | 69.12 ± 14.31 (74.5) | 0.0849 |
| Percent of state population Black | 8.73 ± 7.58 (6.5) | 10.67 ± 10.05 (7.1) | 0.7438 | 8.79 ± 7.45 (6.65) | 10.57 ± 10.04 (7.1) | 0.8388 | 8.67 ± 7.03 (6.5) | 10.62 ± 10.4 (8.7) | 0.9379 |
| Percent of state population Hispanic | 17.98 ± 11.69 (14.9) | 10.43 ± 8.711 (7.4) | 0.0030 | 17.25 ± 11.8 (14.7) | 10.53 ± 8.67 (7.5) | 0.0092 | 16.53 ± 11.23 (13.7) | 10.49 ± 8.9 (7.4) | 0.0072 |
| Primary care practitioner deficit (per 100,000 persons in shortage areas) | 16.54 ± 6.77 (19.02) | 17.12 ± 4.86 (17.77) | 0.8248 | 16.37 ± 6.60 (17.63) | 17.15 ± 4.87 (17.77) | 0.6991 | 17.59 ± 5.135 (18.16) | 16.12 ± 5.87 (16.44) | 0.5270 |
| Percent with high school or higher | 91.04 ± 2.6 (91.4) | 91.12 ± 2.25 (90.6) | 0.9664 | 91.19 ± 2.63 (91.4) | 91.13 ± 2.25 (90.6) | 0.8229 | 91.14 ± 2.28 (91.4) | 91.22 ± 2.49 (91.4) | 0.9922 |
| Percent with bachelor degree or higher | 39.49 ± 4.83 (39.0) | 32.14 ± 3.81 (32.7) | < 0.0001 | 39.39 ± 4.86 (39.5) | 32.23 ± 3.86 (32.9) | < 0.0001 | 38.09 ± 4.79 (37.5) | 32.66 ± 4.81 (32.2) | 0.0003 |
P-values are from Kruskal-Wallis non-parametric test comparing ranked medians; statistically significant p-values (p <0.05) are indicated in bold
FPL Federal Poverty Level
Primary care practitioner deficit was calculated as the number of additional primary care practitioners needed to remove Health Professional Shortage Area (HPSA) designations (HRSA), divided by the shortage-area population, per 100,000
Minnesota was excluded from house control because it was divided. Nebraska was excluded from both house and senate control analyses because it has a unicameral, nonpartisan legislature
Table 3 shows that abortion restrictions were far more prevalent in Republican-controlled states across houses, senates, and governors. Total abortion bans and bans at 6–18 weeks were observed exclusively in Republican-led states, while Democratic-led states were more likely to allow abortions after 18 weeks or have no bans (all p < 0.001). Republican states were significantly more likely to mandate ultrasound requirements, restrict medication abortion to physicians, and prohibit advanced practice clinicians from providing medication abortion (all p < 0.001). Exceptions were also more limited in Republican-led states: they were less likely to allow abortion for threats to the pregnant person’s life or health, general health, pregnancies resulting from rape or incest, or diagnoses of lethal fetal anomalies (p < 0.05 for all).
Table 3.
State-Level Abortion Restrictions by Political Control, United States
| Variable | State House Control |
State Senate Control |
State Governor Party |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Democrat (n = 19) |
Republican (n = 29) |
Democrat (n = 19) | Republican (n = 29) | Democrat (n = 23) | Republican (n = 27) | ||||
| Abortion ban | n (%) | n (%) | P-value | n % | n (%) | p-value | n % | n (%) | p-value |
| Total ban | 0 (0.0) | 13 (100.0) | 0.0001 | 0 (0.0) | 13 (100.0) | < 0.0001 | 1 (7.69) | 12 (92.31) | 0.0005 |
| Ban at 6–18 weeks | 0 (0.0) | 6 (100.0) | 0 (0.0) | 6 (100.0) | 1 (14.29) | 6 (85.71) | |||
| After 18 weeks | 13 (61.9) | 8 (38.1) | 12 (57.14) | 9 (42.86) | 14 (66.67) | 7 (33.33) | |||
| No ban | 6 (75.0) | 2 (25.0) | 8 (88.89) | 1 (11.11) | 7 (77.78) | 2 (22.22) | |||
| Ultrasound Requirements | |||||||||
| Abortion Banned | 0 (0.0) | 12 (100.0) | < 0.0001 | 0 (0.0) | 12 (100.0) | < 0.0001 | 1 (8.33) | 11 (91.67) | 0.0006 |
| No | 19 (79.17) | 5 (20.83) | 20 (80.0) | 5 (20.0) | 18 (72.0) | 7 (28.0) | |||
| Yes | 0 (0.0) | 12 (10.0) | 0 (0.0) | 12 (100.0) | 4 (30.77) | 9 (69.23) | |||
| Medication Abortion Restricted to Physicians Only | |||||||||
| Abortion Banned | 0 (0.0) | 13 (100.0) | < 0.0001 | 0 (0.0) | 13 (100.0) | < 0.0001 | 1 (7.69) | 12 (92.31) | 0.0004 |
| No | 17 (77.27) | 5 (22.73) | 19 (82.61) | 4 (17.39) | 17 (73.91) | 6 (26.09) | |||
| Yes | 2 (15.38) | 11 (84.62) | 1 (7.69) | 12 (92.31) | 5 (35.71) | 9 (64.29) | |||
| Advanced Practice Clinicians Allowed to Provide Medication Abortion | |||||||||
| Abortion Banned | 0 (0.0) | 13 (100.0) | < 0.0001 | 0 (0.0) | 13 (100.0) | < 0.0001 | 1 (7.69) | 12 (92.31) | 0.0004 |
| No | 2 (15.38) | 11 (84.62) | 1 (7.69) | 12 (92.31) | 5 (35.71) | 9 (64.29) | |||
| Yes | 17 (77.27) | 5 (22.73) | 19 (82.61) | 4 (17.39) | 17 (73.91) | 6 (26.09) | |||
| Exception: Threat to the life of the pregnant person | |||||||||
| No | 6 (75.0) | 2 (25.0) | 0.0248 | 8 (88.89) | 1 (11.11) | 0.0012 | 7 (77.78) | 2 (22.22) | 0.0347 |
| Yes | 13 (32.5) | 27 (67.5) | 12 (30.0) | 28 (70.0) | 16 (39.02) | 25 (60.98) | |||
| Exception: Threat to the physical health of the pregnant person | |||||||||
| No | 18 (66.67) | 9 (33.33) | < 0.0001 | 20 (71.43) | 8 (28.57) | < 0.0001 | 18 (64.29) | 10 (35.71) | 0.0034 |
| Yes | 1 (4.76) | 20 (95.24) | 0 (0.0) | 21 (100.0) | 5 (22.73) | 17 (77.27) | |||
| Exception: Threat to the general health of the pregnant person | |||||||||
| No | 8 (22.86) | 27 (77.14) | 0.0001 | 9 (25.0) | 27 (75.0) | 0.0002 | 13 (35.14) | 24 (64.86) | 0.0093 |
| Yes | 11 (84.62) | 2 (15.38) | 11 (84.62) | 2 (15.38) | 10 (76.92) | 3 (23.08) | |||
| Exception: Pregnancy resulting from rape | |||||||||
| No | 19 (48.72) | 20 (51.28) | 0.0071 | 20 (50.0) | 20 (50.0) | 0.0058 | 22 (55.0) | 18 (45.0) | 0.0107 |
| Yes | 0 (0.0) | 9 (100.0) | 0 (0.0) | 9 (100.0) | 1 (10.0) | 9 (90.0) | |||
| Exception: Pregnancy resulting from incest | |||||||||
| No | 19 (47.5) | 21 (52.5) | 0.0121 | 20 (48.78) | 21 (51.22) | 0.0102 | 22 (53.66) | 19 (46.34) | 0.0204 |
| Yes | 0 (0.0) | 8 (100.0) | 0 (0.0) | 8 (100.0) | 1 (11.11) | 8 (88.89) | |||
| Exception: Diagnosis of a lethal fetal anomaly | |||||||||
| No | 17 (48.57) | 18 (51.43) | 0.0367 | 18 (50.0) | 18 (50.0) | 0.0295 | 20 (54.05) | 17 (45.95) | 0.0539 |
| Yes | 2 (15.38) | 11 (84.62) | 2 (15.38) | 11 (84.62) | 3 (23.08) | 10 (76.92) | |||
P-value is from chi-square test; statistically significant p-values (p <0.05) are indicated in bold
Table 4 shows that states with total abortion bans had higher maternal mortality (β = 5.28, p = 0.0315) and higher infant mortality (β = 1.15, p = 0.0014) compared with no-ban states. Fertility rates increased both maternal and infant mortality, while longer life expectancy decreased maternal mortality (p < 0.001). Economic factors were protective: higher median household income, greater state spending, and higher tax revenues all reduced infant mortality. Population composition also played a role—a larger White population decreased both maternal and infant mortality, a larger Black population increased infant mortality, and a larger Hispanic population decreased infant mortality. Higher educational attainment reduced maternal mortality, whereas primary care shortages were associated with increased maternal mortality.
Table 4.
Adjusted Linear Regression Models of Maternal and Infant Mortality by Abortion Policy and Political Control, United States
| Variable | MMR (R2 = 0.81) |
IMR (R2 = 0.86) |
MMR (R2 = 0.76) |
IMR (R2 = 0.84) |
MMR (R2 = 0.78) |
IMR (R2 = 0.84) |
MMR (R2 = 0.78) |
IMR (R2 = 0.84) |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Beta (SE) | P-value | Beta (SE) |
P-value | Beta (SE) | P-value | Beta (SE) | P-value | Beta (SE) | P-value | Beta (SE) | P-value | Beta (SE) | P-value | Beta (SE) | P-value | |
|
Total ban (Ref. no ban) |
5.28 (2.36) |
0.0315 |
1.15 (0.34) |
0.0014 | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
|
Ban at 6–18 weeks (Ref. no ban) |
3.90 (2.15) |
0.0787 |
0.30 (0.33) |
0.3698 | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
|
Ban after 18 weeks (Ref. no ban) |
3.52 (1.64) |
0.0386 |
0.12 (0.23) |
0.5936 | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
|
State house control (Ref. Democrats) |
-- | -- | -- | -- |
0.01 (1.87) |
0.9937 | -0.59 (0.33) | 0.0773 | -- | -- | -- | -- | -- | -- | -- | -- |
|
State senate control (Ref. Democrats) |
-- | -- | -- | -- | -- | -- | -- | -- | 1.65 (1.86) | 0.3800 |
-0.33 (0.31) |
0.3069 | -- | -- | -- | -- |
|
Governor party (Ref. Democrats) |
-- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
1.97 (1.45) |
0.1822 |
-0.08 (0.23) |
0.7386 |
| Abortions per 1,000 women aged 15–44 | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
| Percent abortions obtained by out of state residents | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
| Live births per 1,000 women aged 15–44. |
0.47 (0.17) |
0.0111 | -- | -- |
0.47 (0.17) |
0.0095 |
0.07 (0.03) |
0.0159 | 0.29 (0.18) | 0.1123 |
0.06 (0.03) |
0.0420 |
0.38 (0.15) |
0.0173 |
0.04 (0.02) |
0.0659 |
| Percent of uninsured population |
-0.49 (0.33) |
0.1518 | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
| Medicaid eligibility limit for pregnant women (% of FPL) |
-0.02 (0.01) |
0.1007 | -- | -- |
-0.02 (0.01) |
0.0895 | -- | -- |
-0.02 (0.01) |
0.0705 | -- | -- |
-0.02 (0.01) |
0.1677 | -- | -- |
| Medicaid eligibility limit (%FPL) | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
| State CHIP spending per capita (USD) | -- | -- |
0.04 (0.02) |
0.0292 | -- | -- | -- | -- | -- | -- |
0.03 (0.02) |
0.1465 | -- | -- |
0.03 (0.02) |
0.1446 |
| Federal CHIP spending per capita (USD) |
0.05 (0.03) |
0.1508 |
-0.03 (0.01) |
0.0006 | -- | -- |
-0.02 (0.00) |
0.0003 | -- | -- |
-0.03 (0.01) |
0.0033 |
0.05 (0.03) |
0.0571 |
-0.03 (0.01) |
0.0031 |
| Percent of the population below the FPL |
-1.02 (0.68) |
0.1453 | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
| Median household income (USD) |
-0.02 (0.01) |
0.1677 |
-0.01 (0.00) |
< 0.0001 | -- | -- |
-0.01 (0.00) |
< 0.0001 | -- | -- |
-0.01 (0.00) |
0.0001 | -- | -- |
0.01 (0.00) |
0.0002 |
| Life expectancy at birth (years) |
-3.36 (0.60) |
< 0.0001 | -- | -- |
-2.45 (0.48) |
< 0.0001 | -- | -- |
-2.47 (0.48) |
< 0.0001 |
-0.15 (0.08) |
0.0903 |
-2.60 (0.42) |
< 0.0001 |
-0.16 (0.08) |
0.0651 |
| Percent of population overweight or obese | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
| State spending per capita on services (USD) | -- | -- |
0.01 (0.00) |
< 0.0001 | -- | -- |
0.01 (0.00) |
0.0036 | -- | -- |
0.01 (0.00) |
0.0176 | -- | -- |
0.01 (0.00) |
0.0291 |
| State taxes per capita in revenues (USD) | -- | -- |
-0.03 (0.01) |
0.0029 | -- | -- |
-0.03 (0.01) |
0.0012 | -- | -- |
-0.03 (0.01) |
0.0038 | -- | -- |
-0.03 (0.01) |
0.0055 |
| Percent of State Population White |
-0.15 (0.06) |
0.0254 |
-0.03 (0.01) |
0.0038 | -- | -- | -0.02 (0.01) | 0.0856 |
-0.09 (0.06) |
0.1515 |
-0.04 (0.01) |
0.0005 |
-0.11 (0.05) |
0.0379 |
-0.04 (0.01) |
< 0.0001 |
| Percent of State Population Black | -- | -- |
0.06 (0.01) |
< 0.0001 | -- | -- | 0.03 (0.01) | 0.0382 | -- | -- | -- | -- | -- | -- | -- | -- |
| Percent of State Population Hispanic | -- | -- | -- | -- | -0.15 (0.08) | 0.0755 | -0.03 (0.01) | 0.0188 |
-0.19 (0.09) |
0.0312 |
-0.05 (0.01) |
0.0001 |
-0.11 (0.08) |
0.1440 |
-0.05 (0.01) |
< 0.0001 |
| Primary Care Practitioner deficit (per 100,000 persons in shortage areas) |
-0.25 (0.11) |
0.0275 | -- | -- | -0.22 (0.11) | 0.0660 | -- | -- |
-0.21 (0.11) |
0.0735 | -- | -- |
-0.20 (0.11) |
0.0816 | -- | -- |
| Percent with high school or higher | -- | -- |
0.21 (0.06) |
0.0019 | -0.93 (0.42) | 0.0347 | -- | -- |
-0.72 (0.47) |
0.1330 | -- | -- | -- | -- | -- | -- |
| Percent with bachelor degree or higher |
0.72 (0.25) |
0.0064 | -- | -- | 0.51 (0.20) | 0.0162 | -- | -- |
0.47 (0.20) |
0.0251 |
0.07 (0.04) |
0.0658 |
0.42 (0.20) |
0.0370 |
0.08 (0.04) |
0.0529 |
| R-square value for the model | 81% | 86% | 76% | 84% | 78% | 84% | 78% | 84% | ||||||||
• P-values are from adjusted linear regression models; statistically significant p-values (p <0.05) are indicated in bold
• FPL Federal Poverty Level, MMR Maternal Mortality Ratio, IMR Infant Mortality Rate
• Primary care practitioner deficit was calculated as the number of additional primary care practitioners needed to remove Health Professional Shortage Area (HPSA) designations (HRSA), divided by the shortage-area population, per 100,000
• All variables listed in column 1 were included in the initial selection process. “--” indicates the variable was not retained in the final model after backward elimination. β coefficients are shown with standard errors (SE) and p-values. For median household income, CHIP spending, and tax revenues, β values are scaled per $100
Table 5 shows results from both unadjusted and adjusted models. In unadjusted analyses, total abortion bans were associated with higher maternal mortality (β = 11.67, p < 0.0001, R²=0.41) and infant mortality (β = 1.72, p = 0.0005, R²=0.35). Political control of the governor, house, and senate also showed significant associations, each explaining about one-fifth to one-quarter of the variance (R²=0.21–0.25). After adjustment for sociodemographic, economic, and healthcare covariates, model fit improved substantially (R²=0.76–0.86). Associations remained significant for total bans and bans after 18 weeks, whereas political control variables attenuated and lost significance, with their explanatory power absorbed by broader structural factors.
Table 5.
Unadjusted and Adjusted Linear Regression Equations of Maternal and Infant Mortality by Abortion Policy and Political Control, United States
| Un-adjusted models | Adjusted models | ||||||
|---|---|---|---|---|---|---|---|
| Exposure | Outcome | β | p-value | R-squared | β | p-value | R-squared |
| Total ban vs. no ban | MMR | 11.67 (2.53) | < 0.0001 | 0.41 | 5.28 (2.36) | 0.0315 | 0.81 |
| Ban at 6–18 weeks vs. no ban | MMR | 5.22 (2.94) | 0.0827 | 0.41 | 3.90 (2.15) | 0.0787 | 0.81 |
| Ban after 18 weeks vs. no ban | MMR | 1.08 (2.33) | 0.6454 | 0.41 | 3.52 (1.64) | 0.0386 | 0.81 |
| Total ban vs. no ban | IMR | 1.72 (0.46) | 0.0005 | 0.35 | 1.15 (0.34) | 0.0014 | 0.86 |
| Ban at 6–18 weeks vs. no ban | IMR | 0.88 (0.54) | 0.1073 | 0.35 | 0.30 (0.33) | 0.3698 | 0.86 |
| Ban after 18 weeks vs. no ban | IMR | -0.03 (0.42) | 0.9410 | 0.35 | 0.12 (0.23) | 0.5936 | 0.86 |
| State house control (Republican vs. Democrat) | MMR | 7.09 (1.92) | 0.0006 | 0.21 | 0.015 (1.87) | 0.9937 | 0.76 |
| State house control (Republican vs. Democrat) | IMR | 1.30 (0.34) | 0.0003 | 0.25 | -0.59 (0.33) | 0.0773 | 0.84 |
| State senate control (Republican vs. Democrat) | MMR | 7.34 (1.91) | 0.0004 | 0.24 | 1.65 (1.86) | 0.3800 | 0.78 |
| State senate control (Republican vs. Democrat) | IMR | 1.26 (0.33) | 0.0004 | 0.24 | -0.33 (0.31) | 0.3069 | 0.84 |
| Governor party (Republican vs. Democrat) | MMR | 6.71 (1.88) | 0.0008 | 0.21 | 1.97 (1.45) | 0.1822 | 0.78 |
| Governor party (Republican vs. Democrat) | IMR | 0.96 (0.34) | 0.0072 | 0.14 | -0.08 (0.23) | 0.7386 | 0.84 |
MMR maternal mortality ratio
IMR infant mortality rate
P-values are from linear regression models; statistically significant p-values (p <0.05) are indicated in bold
Adjusted models were, adjusted for covariate listed in Table 4 for each model
Discussion
This study examined how abortion restrictions, political control, and sociodemographic characteristics influence maternal and infant mortality across U.S. states. States with total abortion bans had significantly higher maternal mortality (β = 5.28, p = 0.0315) and infant mortality (β = 1.15, p = 0.0014) compared with states without bans. These findings align with prior research showing that restrictive reproductive policies are linked to poorer maternal and child health outcomes, often through reduced access to timely reproductive healthcare and delays in obtaining services [10, 21–23].
In the unadjusted models, political control of the governor’s office, state house, and senate each explained a meaningful portion of variation in maternal and infant mortality, with R² values ranging from 0.14 to 0.25—suggesting that single political variables alone accounted for up to one-quarter of the variance. However, once broader sociodemographic, economic, and healthcare access variables were included, model fit improved substantially (R² ≈ 0.80–0.86). This shift suggest that much of the apparent association between political control and mortality is explained by structural determinants such as income, insurance coverage, healthcare infrastructure, and educational attainment rather than reflecting an independent effect of political control itself. The directed acyclic graph (Supplement Fig. 1) supports this interpretation by positioning legislative control as an upstream influence on abortion policy, economic conditions, and healthcare access, which subsequently shape maternal and infant mortality outcomes. Political control variables are interpreted as contextual indicators of state governance rather than independent causal exposures. Because abortion policy is downstream of political control and governance shapes health through multiple structural pathways, mediation analysis was not undertaken in this cross-sectional ecological study. Coefficients for political control therefore reflect shared structural conditions rather than discrete policy effects. The attenuation of political control associations after adjustment indicates that its relationship with maternal and infant mortality is largely attributable to downstream economic, educational, and healthcare conditions, consistent with political control operating through these structural pathways rather than exerting a direct effect. Consistent with this framework, no significant differences were observed in CHIP spending across abortion policy or political control categories, suggesting that child insurance investment alone does not explain the observed mortality gradients and reinforcing the central role of reproductive policy and broader structural conditions.
Economic indicators exerted strong protective effects. States with higher median household income, greater per-capita tax revenue, and larger health expenditures consistently exhibited lower infant mortality. These results echo ecological studies demonstrating that fiscal investment in public health infrastructure contributes to improved outcomes [24–26]. Population composition also shaped risks: a higher proportion of White residents was associated with lower mortality, whereas a larger Black population was linked to increased infant mortality. By contrast, a larger Hispanic population share was protective for infant outcomes, consistent with the “Hispanic paradox” previously documented in maternal-child health research [27, 28]. These associations reflect structural and historical inequities, including differential access to healthcare, socioeconomic disadvantage, and the effects of structural racism, rather than biological or intrinsic differences between racial groups. Educational attainment also emerged as an important determinant, as states with higher proportions of college graduates had lower maternal mortality—consistent with literature linking education to improved health literacy, preventive care utilization, and healthier behaviors [29]. Additionally, states with greater primary care shortages experienced elevated maternal mortality, corroborating evidence that healthcare workforce availability is critical for maternal health [30].Finally, longer average life expectancy at the state level was strongly protective against maternal mortality (p < 0.001), underscoring the influence of broader structural determinants of population health. Fertility rate likely captures multiple, overlapping mechanisms relevant to maternal and infant mortality, including population composition, demand on obstetric and neonatal services, and broader reproductive norms and policy environments. Higher fertility may increase strain on healthcare systems in resource-limited settings or reflect contexts with reduced access to reproductive and preventive care.
A key strength of this analysis is the integration of up-to-date, state-level data across political, economic, demographic, and healthcare domains, providing a comprehensive picture of structural determinants of maternal and infant health. The model selection approach further ensured that retained covariates reflected the strongest predictors of outcomes. Several limitations should be noted. This analysis was conducted at the state level, so findings reflect population-level patterns rather than individual risk. As a state-level ecological analysis, this study is subject to cross-level bias and cannot account for within-state heterogeneity; observed associations may therefore reflect compositional rather than contextual effects. Because the study is cross-sectional, temporality cannot be established, and associations may reflect bidirectional or feedback relationships between political environments, policy contexts, and population health outcomes. States with restrictive abortion policies may have also enacted other contemporaneous health and social policies during the study period, which could not be fully disentangled in this analysis and may contribute to residual confounding. Interpretation of the findings should also consider the temporal alignment of exposures and outcomes, as political control and abortion policy classifications reflect the contemporary policy environment, whereas maternal mortality outcomes necessarily lag due to delays in national mortality surveillance and data finalization. Despite this temporal mismatch, the use of contemporaneous political control and abortion policy classifications is unlikely to materially alter the substantive interpretation of the findings. This study was designed as a cross-sectional analysis using the most recent data available across state-level covariates, including insurance coverage, socioeconomic indicators, and public spending, for which historical alignment to earlier political periods is not feasible. Moreover, state political control and reproductive policy environments tend to be durable over time rather than rapidly fluctuating, and therefore likely capture longer-standing governance contexts relevant to maternal and infant health. Accordingly, political control is best interpreted here as a contextual indicator of the broader policy environment rather than a precise temporal exposure. Despite these limitations, our findings underscore that abortion bans and restrictive reproductive health policies are associated with higher maternal and infant mortality. Importantly, socioeconomic investments and higher educational attainment mitigated risks, suggesting that strengthening social safety nets, improving healthcare access, and protecting reproductive autonomy are critical for improving maternal and child health. Policymakers should recognize that political structures influence health indirectly through these mediating pathways, a conclusion consistent with the DAG framework.
Conclusion
Maternal and infant mortality in the United States are shaped not only by clinical and demographic factors but also by political decisions that determine access to care and investment in public health. A novel contribution of this study is the integration of political control, abortion policy, socioeconomic conditions, and healthcare access within a single analytic framework, allowing reproductive policy effects to be evaluated in their broader structural context. Our analysis shows that abortion bans directly increase mortality risks, while partisan control influences outcomes indirectly through structural conditions such as income, education, and healthcare access. Efforts to safeguard reproductive autonomy, expand health coverage, and strengthen social and economic supports are critical to reversing preventable maternal and infant deaths.
Supplementary Information
Acknowledgements
The author thanks the CDC’s National Center for Health Statistics, the Guttmacher Institute, the Kaiser Family Foundation, the Centers for Medicare & Medicaid Services, and the U.S. Census Bureau for maintaining the public data used in this study, and acknowledges the mothers, infants, and families whose lives are represented in these data.
Authors’ information (optional)
SR is an Assistant Professor of Population Health at the University of Health Sciences and Pharmacy in St. Louis. His research focuses on the social, political, and policy determinants of health outcomes.
Authors’ contributions
SR conceived the study, collected and analyzed the data, interpreted the findings, and drafted the manuscript.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Data availability
The datasets analyzed during the current study are publicly available from the CDC National Center for Health Statistics, the Guttmacher Institute, the Kaiser Family Foundation, the Centers for Medicare & Medicaid Services, and the U.S. Census Bureau. All sources are cited in the manuscript.
Declarations
Ethics approval and consent to participate
This study was deemed exempt by the University of Health Sciences and Pharmacy in St. Louis Institutional Review Board. The analysis used publicly available, de-identified state-level data.
Consent for publication
Not applicable. This manuscript does not contain any individual person’s data.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Desai A, Holliday R, Wallis M, Thornewill A, Bahraini NH, Borges LM. Policy Changes as a Context for Moral Injury Development in the Wake of Dobbs v Jackson Women’s Health Organization. Obstet Gynecol. 2023;141:15–21. [DOI] [PubMed] [Google Scholar]
- 2.Gostin LO, Wetter S, Reingold RB. One Year After Dobbs—Vast Changes to the Abortion Legal Landscape. JAMA Health Forum. 2023;4:e233091–233091. [DOI] [PubMed] [Google Scholar]
- 3.Singh P, Gallo MF. National Trends in Infant Mortality in the US After Dobbs. JAMA Pediatr. 2024;178:1364–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hoyert DL. Maternal mortality rates in the United States, 2021. NCHS Health E-Stats. 2023. 10.15620/cdc:124678
- 5.Centers for Disease Control and Prevention, National Center for Health Statistics. (n.d.). Maternal deaths and mortality rates: Each state, the District of Columbia, United States, 2018–2022. Retrieved July 10. 2025, from https://www.cdc.gov/nchs/maternal-mortality/MMR-2018-2022-State-Data.pdf
- 6.Centers for Disease Control and Prevention. (2024, September 16). Infant mortality. U.S. Department of Health & Human Services. https://www.cdc.gov/maternal-infant-health/infant-mortality/index.html
- 7.Werner EP, Haselow DT. Racial Disparities in Infant Mortality Remain but are Decreasing in Arkansas and the US, 1995–2022. J Neonatol Clin Pediatr. 2025;12:136. 10.24966/NCP-878X/100136. [Google Scholar]
- 8.Rodriguez JM, Bae B. Political Ideology Direction of Policy Agendas and Maternal Mortality Outcomes in the U.S., 1915–2007. Matern Child Health J. 2024;28:865–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Rodriguez JM, Geronimus AT, Bound J, Wen R, Kinane CM. Partisan Control of U.S. State Governments: Politics as a Social Determinant of Infant Health. Am J Prev Med. 2022;62:1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eugene D, The U.S. Maternal Health Divide, et al. The Limited Maternal Health Services and Worse Outcomes of States Proposing New Abortion Restrictions. Commonwealth Fund; Dec. 2022. 10.26099/z7dz-8211.
- 11.Eliason EL. Adoption of Medicaid Expansion Is Associated with Lower Maternal Mortality. Womens Health Issues. 2020;30:147–52. [DOI] [PubMed] [Google Scholar]
- 12.Guttmacher Institute. (2025, July 7). State Laws and Policies. State bans on abortion throughout pregnancy. Accessed 7.11.2025. Available at: https://www.guttmacher.org/state-policy/explore/state-policies-abortion-bans
- 13.National Center for Health Statistics. (2024). Maternal deaths and mortality rates: Each state, the District of Columbia, United States, 2018–2022 [Data file]. National Vital Statistics System. Accessed 7.11.2025. Available at: https://www.cdc.gov/nchs/maternal-mortality/MMR-2018-2022-State-Data.pdf
- 14.Ramer S, Nguyen AT, Hollier LM, Rodenhizer J, Warner L, Whiteman MK. Abortion Surveillance - United States, 2022. MMWR Surveill Summ. 2024;73:1–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Jones RK, Witwer E, Jerman J. (2019). Abortion incidence and service availability in the United States, 2017. Guttmacher Institute. Accessed 7.11.2025. Available at: https://www.guttmacher.org/sites/default/files/report_pdf/abortion-incidence-service-availability-us-2017.pdf
- 16.Kaiser Family Foundation (KFF). (2025). State political parties. KFF State Health Facts. Accessed 7.11.2025. Available at: https://www.kff.org/other/state-indicator/state-political-parties/
- 17.Centers for Medicare & Medicaid Services. (2025, June 27). State Medicaid and CHIP applications, eligibility determinations, and enrollment data: January 2025 reporting period. Accessed 7.11.2025. Available at: https://data.medicaid.gov/dataset/6165f45b-ca93-5bb5-9d06-db29c692a360
- 18.Kaiser Family Foundation. (2024). Total CHIP spending by state, FY 2023. KFF State Health Facts. Accessed 7.11.2025. Available at: https://www.kff.org/medicaid/state-indicator/total-chip-spending/
- 19.Bureau USC. Population Division. (2024, December). Annual Estimates of the Resident Population for the United States, Regions, States, District of Columbia, and Puerto Rico: April 1, 2020 to July 1, 2024 (NST-EST2024-POP). Vintage 2024 data release. Accessed 7.11.2025. Available at: https://www.census.gov/data/tables/timeseries/demo/popest/2020s-state.html
- 20.National Center for Health Statistics. (2023). Fertility rates and birth data by state: 2022. Centers for Disease Control and Prevention. Accessed 7.11.2025. Available at: https://www.cdc.gov/nchs/pressroom/sosmap/fertility_rate/fertility_rates.htm
- 21.Gemmill A, Franks AM, Anjur-Dietrich S, Ozinsky A, Arbour D, Stuart EA, Ben-Michael E, Feller A, Bell SO. US Abortion Bans and Infant Mortality. JAMA. 2025;333:1315–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Muliyil S. Potential impact of abortion bans on infant health. Nat Med. [DOI] [PubMed]
- 23.Vilda D, Wallace ME, Daniel C, Evans MG, Stoecker C, Theall KP. State Abortion Policies and Maternal Death in the United States, 2015–2018. Am J Public Health. 2021;111:1696–704. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Sowmyan SJ, Hirai AH, Kaufman JS. State and local government expenditures and infant mortality. Pediatrics. 2024;154(6):e2023063571. 10.1542/peds.2023-063571. [DOI] [PubMed]
- 25.Goldstein ND, Palumbo AJ, Bellamy SL, Purtle J, Locke R. State and local government expenditures and infant mortality in the united states. Pediatrics 2020;146(5):e20201134. 10.1542/peds.2020-1134. [DOI] [PubMed]
- 26.Junior JA, Lee LK, Fleegler EW, Monuteaux MC, Niescierenko ML, Stewart AM. Association of State-Level Tax Policy and Infant Mortality in the United States, 1996–2019. JAMA Netw Open. 2023;6:e239646. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Mathews TJ, Driscoll AK. Trends in Infant Mortality in the United States, 2005–2014. NCHS Data Brief 2017:1–8. [PubMed]
- 28.Xenos MA, Becker AB, Anderson AA, Brossard D, Scheufele DA. Stimulating Upstream Engagement: An Experimental Study of Nanotechnology Information Seeking. Soc Sci Q. 2011;92:1191–214. [Google Scholar]
- 29.Mensch BS, Chuang EK, Melnikas AJ, Psaki SR. Evidence for causal links between education and maternal and child health: systematic review. Trop Med Int Health. 2019;24:504–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wallace M, Dyer L, Felker-Kantor E, Benno J, Vilda D, Harville E, Theall K. Maternity Care Deserts and Pregnancy-Associated Mortality in Louisiana. Womens Health Issues. 2021;31:122–9. [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.
Supplementary Materials
Data Availability Statement
The datasets analyzed during the current study are publicly available from the CDC National Center for Health Statistics, the Guttmacher Institute, the Kaiser Family Foundation, the Centers for Medicare & Medicaid Services, and the U.S. Census Bureau. All sources are cited in the manuscript.