Introduction

Pregnancy loss, preterm birth (PTB), and low birth weight (LBW) are major public health issues in the United States (US) given their incidence and impact on maternal and infant health. It is estimated that 15–25% of recognized pregnancies end in miscarriage, a life event that has been linked to significant psychosocial consequences and economic burden [72, 73]. Of all infants born in the US, about 10.5% are premature and 8.2% are below normal birthweight [68]. PTB and LBW are leading causes of infant mortality and, for surviving infants, they forecast long-term health and development concerns that pass on substantial economic costs to society [2, 4, 22, 71, 89].

The development of strategies to prevent unwanted pregnancy and birth outcomes hinges partly on understanding their etiologies. The list of risks for pregnancy loss, PTB, and LBW is long and includes genetic and epigenetic influences, maternal and paternal age, placental insufficiency, infection, cervical length, hypertension, diabetes, obesity, smoking, alcohol and substance use, low health care access and quality, chronic stress, violence, discrimination, and racism [30, 35, 45, 74, 77, 81, 91]. Collectively, these and other complex processes contribute to pregnancy loss, PTB, and LBW, though their influence may not manifest equally across outcomes or population subgroups.

To illustrate, there are stark racial and ethnic differences in pregnancy and birth outcomes among American women [2, 65, 75, 76, 82]. For example, Black women are 50–60% more likely than White women to give birth prematurely [8, 68]. Many risks for poor pregnancy and birth outcomes are correlated with race and ethnicity but have not been shown to fully account for racial/ethnic variation. Due partly to their heterogeneous and multifactorial origins, these disparities have proven difficult to explain and redress.

To improve the prediction and prevention of health disparities, scholars have called for approaches that account for the cumulative impact of adverse and inequitable experiences from conception through adulthood [26, 41]. One prominent example is the life course perspective [50, 51], which posits that disparities in birth outcomes partly result from exposure to harmful experiences and conditions during sensitive developmental periods early in life (i.e., early programming) and cumulative stress over time that leads to allostatic load and physiological wear and tear (i.e., cumulative pathways). This model has been widely disseminated, and it serves as a framework that guides federal maternal and child health priorities [3, 26, 28]. Yet, surprisingly few studies have examined whether racial and ethnic disparities in pregnancy and birth outcomes are associated with exposure to stressful life events over the life course [7, 69].

Partial support for the life course perspective can be found in research indicating that poor pregnancy and birth outcomes are associated with increased exposure to adverse childhood experiences (i.e., ACEs), which include harmful events and conditions such as child abuse and neglect, domestic violence, and incarceration of a parent or caregiver [14, 46, 60, 85]. ACEs are more prevalent in the US among non-Hispanic Black than non-Hispanic White individuals [52, 56], and it is plausible that differential exposure to early adversity contributes to disparities in pregnancy and birth outcomes. On the other hand, a series of recent studies have found that ACEs are more prevalent among low-income White adults than low-income Black adults [58, 61, 80]. There also is some evidence to suggest that socioeconomic disparities in birth outcomes vary by race and ethnicity (see [7] for review), though a recent analysis of nearly two million births in California over a 10-year period showed that racial/ethnic disparities in birth outcomes were present across the income distribution [42].

These nuanced findings underscore the point that empirical applications of the life course model should aim to disentangle outcome variance explained by adversity from socioeconomic differences that prevail among racial/ethnic groups [10]. While it is standard practice to account for socioeconomic confounds at the study design phase or through the use of statistical controls, there is growing interest in applying variable- and person-centered techniques to explore between- and within-group heterogeneity [20, 24]. In addition to the use of novel statistical techniques, scholars have attempted to strengthen the ecological and cross-cultural validity of ACE assessments by incorporating indicators of socioeconomic risk [17, 27, 57].

Alongside novel approaches to measuring and modeling ACEs, recent research has drawn attention to adverse adult experiences (AAEs) and their deleterious effects within and across generations [59, 63, 87, 92]. Poor pregnancy and birth outcomes are associated with many adult adversities, including intimate partner violence [35], community violence [29, 54], residential segregation [55], housing insecurity and homelessness [19, 47], and discrimination and racism [1, 45]. Studies have also found that cumulative exposure to AAEs increases the risk of poor physical and mental health above and beyond the effects of ACEs [62, 83]. Yet, no known studies have examined whether the accumulation of adverse experiences throughout childhood and adulthood result in disparate pregnancy and birth outcomes. Research along these lines is needed to further develop life course models of adversity and reproductive health.

Study Aims

To better understand the relationship between life course adversity and disparities in pregnancy and birth outcomes, this study analyzes longitudinal data collected from a sample of 1319 racially and ethnically diverse women from low-income households in Wisconsin who received perinatal home visiting services. The aims of the study are twofold. First, bivariate and multivariate analytic approaches are conducted to test whether ACEs and AAEs are associated with pregnancy loss, PTB, and LBW overall and whether associations differ between self-identified Black, White, and Hispanic (i.e., Latina; Latine; Latinx) participants. Second, latent class analysis is used to identify subgroups that share similar patterns of exposure to life course adversity, and multigroup latent class regressions are performed to explore whether different patterns are linked to pregnancy and birth outcomes within each racial/ethnic group.

Methods

Sample and Data

Data for the current study were collected from participants in the Families and Children Thriving (FACT) Study, an ongoing longitudinal investigation of Wisconsin households that received perinatal home visiting services. All families were served by agencies that are subsidized by the federal Maternal, Infant, and Early Childhood Home Visiting (MIECHV) Program, which allocates funds to states, territories, and tribal communities to support evidence-based programs for priority populations at risk of poor maternal and child health outcomes [25]. In Wisconsin, MIECHV supports four home visiting models: Early Head Start, Healthy Families America, Nurse-Family Partnership, and Parents as Teachers. Illustrating the socioeconomic composition of the sample, 98.3% of participating households were at or below 200% of federal poverty line or eligible for means-tested public benefits.

Launched in August 2015, the FACT Study tracks a cohort of 1967 primary caregivers who completed a baseline survey after the birth of a child associated with a home visiting episode (response rate = 62.8%). Approximately 97.9% of respondents self-identified as women (n = 1923). For the present study, 90 women were removed from the analytic sample because they were less than 19 years old at baseline. Another 234 participants who self-identified as American Indian or Alaska Native, Asian or Pacific Islander, multiracial, or other race/ethnicity were omitted from the sample because their group sample sizes were too small to make valid statistical comparisons.

Another 280 eligible participants were omitted from the analysis via listwise deletion due to missing or incomplete data on adverse experiences or study outcomes. Compared to the 280 participants who were removed from the sample, the analytic sample of 1319 participants was less likely to be Black than White or Hispanic and more likely to live with spouse or partner (p < 0.05); there were no differences in maternal age, number of children, or educational attainment.

Following a data sharing agreement, program records used to create measures of pregnancy loss, PTB, LBW, and ACEs were obtained from an administrative database managed by the Wisconsin Department of Children and Families. Survey data were collected by mail or phone after completing an informed consent procedure. Baseline surveys from the FACT Study supply primary data on participant demographics and AAEs along with supplemental data on pregnancy losses. All study protocols were approved by the instiutional review board at the University of Wisconsin-Milwaukee (FWA00006171).

Measures

Pregnancy and Birth Outcomes

Administrative data were used to create dichotomous measures indicating if participants had ever given birth prematurely (i.e., < prior to 37th week of pregnancy) or to a low birthweight infant (i.e., < 2500 g or 5.5 pounds). Primary data used to create a dichotomous measure of pregnancy loss were drawn from archival program records. A total of 235 participants were missing pregnancy loss data, largely because the information was collected by home visiting staff during a prenatal assessment, meaning that the information was not recorded for most program participants who enrolled postnatally. Missing cases were imputed using data from a baseline survey item from the FACT Study that asked participants to report the number of pregnancy losses since turning age 18.

Adverse Childhood Experiences

Home visiting providers assessed participant ACEs using the Childhood Experiences Survey, an expanded ACE assessment that has demonstrated good internal consistency, test–retest reliability, and criterion validity [13, 57]. Ten conventional ACEs were coded dichotomously, including five forms of child maltreatment (physical abuse,sexual abuse; emotional abuse; physical neglect; emotional neglect) and five forms of household dysfunction (mental illness; substance use; incarceration/jail; domestic violence; prolonged absence of a parent). The 10 items were summed to create a cumulative score. In addition to the 10 conventional ACEs, five novel adversities were measured dichotomously, including three socioeconomic indicators (frequent financial problems; food insecurity; homelessness) and two extrafamilial indicators of victimization (bullying; crime). These five items were summed with the 10 conventional ACEs to create a 15-item score. For further information about item language and scoring as well as the prevalence of ACEs in the FACT Study, see [57].

Adverse Adult Experiences

At baseline, FACT Study participants completed the Adult Experiences Survey, a measure of AAEs that has demonstrated good internal consistency, test–retest reliability, and criterion validity (for information about item language, scoring, and prevalence, see [62]). Ten adversities that occurred after the respondent’s 18th birthday were assessed, five of which were associated with a spouse, partner, or household member: physical or emotional abuse, sexual abuse, substance use, mental health problem, and incarceration or jail. The remaining five AAEs were sexual assault, crime victimization, frequent financial problems, homelessness, and discrimination. The 10 items were coded dichotomously and summed to create a cumulative score.

Race and Ethnicity

The race and ethnicity of study participants was self-reported and coded into three mutually exclusive categories: non-Hispanic Black, non-Hispanic White, and Hispanic.

Covariates

Survey data from the FACT Study were used to construct four demographic measures that served as model covariates. At baseline, a continuous measure of participant age was calculated along with a categorical measure of educational attainment with the following values: (1) less than high school, (2) some high school, (3) high school diploma or General Educational Development certificate, (4) some college credit, no degree, (5) associate degree, (6) bachelor’s (4-year) degree or higher. Participant responses also yielded a dichotomous variable indicating if they were cohabitating with a partner or spouse as well as a count measure denoting their number of biological children.

Analysis Plan

A descriptive analysis of data was performed, followed by bivariate logistic regressions tests of association between adversity measures and study outcomes for the full sample and for racial/ethnic subgroups. Controlling for covariates, multivariate logistic regressions were performed to examine main effects associated with ACE and AAE scores. Interaction terms were added to a second set of multivariate analyses to test whether race and ethnicity moderated the estimated effects of ACEs and AAEs. The preceding analyses were conducted in IBM SPSS 28.

Mplus 8.8 was used to complete a latent class analysis (LCA) of heterogeneous patterns in adverse childhood and adult experiences while adjusting for covariates. Model fit was judged based on a lower Bayesian information criterion, a significant Lo-Mendell-Rubin adjusted LRT test, and practical significance [67]. Missing data were handled via full information maximum likelihood estimation. Once an optimal number of adversity classes was identified, the LCA was transformed into a multigroup latent class regression with a model constraint command and model test statement [66], which yielded within-group comparisons for Black, White, and Hispanic women in their predicted probabilities of pregnancy and birth outcomes based on their exposure to lifetime adversity.

Results

Table 1 describes the sample demographic characteristics and key study variables. Participants’ mean age was 27.4 years, and their mean level of educational attainment was 3.3; 41.5% had a high school diploma or equivalent, and 40.5% had obtained some postsecondary education (not shown). More than half (57.2%) of participants were living with a partner/spouse, and they averaged 2.0 biological children. There were significant differences among racial/ethnic groups for each demographic indicator.

Table 1 Description of sample characteristics, adverse experiences, and birth outcomes
Full size table

On average, participants reported 3.5 adversities on the 10-item ACE score and 4.7 adversities on the 15-item ACE score. The sample mean on the 10-item AAE score was 3.4. Significant racial/ethnic differences in ACEs and AAEs were observed, with White women reporting the highest scores and Hispanic women reporting the lowest scores.

Approximately 31.1% of participants experienced a pregnancy loss, 19.5% experienced a PTB, and 14.4% gave birth to a LBW infant. The prevalence of pregnancy loss did not differ significantly among racial/ethnic groups (Black = 32.2%; Hispanic = 30.2%; White = 31.2%). The prevalence of PTB did vary significantly by race/ethnicity (p < 0.05), with higher rates reported by Black women (26.7%) than White women (18.3%) and Hispanic women (17.0%). LBW also varied by race/ethnicity, with higher rates for Black women (20.1%) than Hispanic women (13.7%) and White women (12.7%).

Study Aim 1

Table 2 presents unadjusted bivariate associations between ACEs and study outcomes in the full sample and by racial/ethnic subgroup. Pregnancy loss was significantly correlated with all ACEs except for indicators of socioeconomic concern (financial problems; food insecurity; homelessness). Associations with PTB and LBW were less robust; only sexual abuse was significantly correlated with all three outcomes. The 10-item and 15-item ACE scores were associated with an increased odds of pregnancy loss, but not PTB and LBW. Descriptively, patterns of association between ACEs and study outcomes varied by race and ethnicity. For instance, pregnancy loss was correlated with 11 out of 15 ACEs for White women and 8 out of 15 ACEs for Hispanic women, but for Black women no ACEs correlated with pregnancy loss in the expected direction. The 10-item and 15-item ACE scores were associated with a greater likelihood of pregnancy loss in White and Hispanic women but not Black women.

Table 2 Univariate odds of association between adverse childhood experiences and pregnancy and birth outcomes
Full size table

As shown in Table 3, there were significant unadjusted associations between pregnancy loss and all AAEs except for partner/spouse mental health problems. Associations between AAEs and PTB and LBW were less consistent. Only incarceration/jail of a partner or spouse was significantly associated with all three outcomes. The 10-item AAE score was associated with all three outcomes in the full sample.

Table 3 Univariate odds of association between adverse adult experiences and pregnancy and birth outcomes
Full size table

Model 1 coefficients in Table 4 are from multivariate logistic regressions that estimate associated effects of ACE scores and AAE scores on pregnancy and birth outcomes. Controlling for demographic characteristics, participant ACE scores were significantly associated with an increased risk of having had a pregnancy loss (b = 0.13, SE = 0.02; p < 0.001) and a LBW infant (b = 0.06, SE = 0.03; p = 0.048), but not PTB (b = 0.05, SE = 0.03; p = 0.104). The 10-item AAE score also was significantly associated with pregnancy loss (b = 0.13, SE = 0.03; p < 0.001) and LBW (b = 0.08, SE = 0.03; p = 0.020), but not PTB (b = 0.05, SE = 0.03; p = 0.117).

Table 4 Multivariate analysis of pregnancy and birth outcomes
Full size table

Table 4 also presents coefficients from tests of interaction effects between race/ethnicity and adversity scores (model 2). Results indicated that the strength of association between ACEs and pregnancy loss was significantly greater for White and Hispanic women than for Black women. Moderation test results trended in the same direction for other outcomes, but no other statistically significant differences were present.

Study Aim 2

Figure 1 displays results from a latent class analysis of 15 ACEs and 10 AAEs. Four classes were identified (sample percentages in parentheses): (1) low life course adversity (32.8%), (2) moderately high childhood adversity and moderately low adult adversity (22.1%), (3) moderately low childhood adversity and moderately high adult adversity (27.5%); (4) high life course adversity (17.6%). Results (not shown) from an analysis of the full sample indicated that, when compared to women who experienced low life course adversity (class 1), women in classes 2, 3, and 4 were more like to have had a pregnancy loss, and women in classes 3 and 4 also were more likely to have had a PTB and LBW infant.

Fig. 1
figure 1

Predicted probability of latent class assignment for adverse childhood and adult experiences

Full size image

Results shown in Table 5 are from multigroup latent class regressions that generated within-group comparisons for Black, Hispanic, and White women. The coefficients represent the predicted probability that an adversity class was associated with a poor outcome, with low life course adversity (class 1) serving as the reference group. One pairwise difference was observed among Black women: Compared to participants assigned to class 1, participants in class 3 had a significantly higher probability of having had a pregnancy loss. To illustrate for ease of interpretation, Black women in class 1 had a 27% chance of having had a pregnancy loss whereas Black women in class 3 had a 51% chance of having had a pregnancy loss. For further context, adjusted probabilities shown in Table 5 can be compared to unadjusted outcome means in Table 1. For instance, the unadjusted prevalence of pregnancy loss among all Black women was 32% (Table 1), which was comparable to the adjusted probability of pregnancy loss among Black women who were exposed to high life course adversity (Table 5).

Table 5 Predicted probability of pregnancy and birth outcomes associated with life course adversity classes, stratified by race/ethnicity
Full size table

For Hispanic women, high life course adversity (class 4) was significantly associated with a greater risk of ever having a pregnancy loss, while adversity class 2 was marginally associated with an increased risk of prior pregnancy loss. For White women, assignment to classes 2, 3, and 4 was significantly associated with a higher risk of pregnancy loss; class 3 assignment was marginally associated with a higher risk of PTB; class 4 assignment was significantly associated with a higher risk of PTB; class 3 and 4 assignments were marginally associated with a higher risk of LBW.

Discussion

Informed by the life course perspective [50, 51], the primary aim of this study was to assess whether exposure to childhood and adult adversity contributed to pregnancy and birth outcome disparities among Black, White, and Hispanic women. Basic and advanced statistical methods were used to analyze data collected from a racially and ethnically diverse sample of low-income women who received perinatal home visiting services. What follows is a brief summary of key findings, a discussion of study implications and limitations, and a set of future directions for research on life course adversity and reproductive disparities.

Summary of Key Findings

Most ACEs and AAEs were correlated with at least one pregnancy or birth outcome in the full sample, but there was marked variation by race/ethnicity. For example, risk of pregnancy loss was correlated with 11 ACEs for White women, eight ACEs for Hispanic women, and no ACEs for Black women in the expected direction. Higher ACE scores were linked to some outcomes in White and Hispanic women but not Black women. Higher AAE scores were associated with all three outcomes in White women, pregnancy loss and LBW in Hispanic women, and only pregnancy loss in Black women. Interaction tests revealed that the association between ACE scores and pregnancy loss was greater for White and Hispanic women than for Black women; moderation effects trended in the same direction for PTB and LBW but were not statistically significant. Four latent classes of life course adversity were identified, and within-group contrasts showed that, compared to White women who experienced low lifetime adversity, White women in higher-adversity classes were often at risk of poorer pregnancy and birth outcomes. Comparatively, the estimated effects of assignment to higher-adversity classes were not as robust for Hispanic women, and they were even less robust for Black women.

Implications

Corroborating disparities that have been documented in the US for over a century [79], we found that Black women were at an elevated risk of PTB and LBW. By a mean age of 27, 1 in 5 Black women had given birth prematurely and more than 1 in 4 had given birth to a LBW infant. Comporting with population estimates [8, 68], the unadjusted prevalence of PTB and LBW in this sample was anywhere from 46 to 58% higher for Black women than White and Hispanic women. Yet, pregnancy loss rates did not differ significantly by race/ethnicity, contradicting the findings of Mukherjee et al. [65] that showed Black women were 1.57 times more likely than White women to have had a miscarriage. Our findings more closely align with results from the National Survey of Family Growth [75], which indicated that the percentage of pregnancies ending in loss were similar among Black and White women (20.7% vs. 21.3%) but lower among Hispanic women (15.0%). Unlike PTB and LBW, racial/ethnic differences in the epidemiology of pregnancy loss have received limited attention and deserve further investigation.

Although racial/ethnic differences in reproductive outcomes are well known, the causes remain incompletely understood. The life course perspective posits that these disparities result from differential exposure to stressors during developmentally sensitive periods and ongoing stressors that increase allostatic load and cumulative wear and tear. We found little support for this conceptual model in this analysis of 25 major life stressors. Paradoxically, both early adversity and cumulative pathways of adversity better explained poor pregnancy and birth outcomes for non-Hispanic White and Hispanic women than Black women. These findings were robust across basic and advanced statistical analyses of between- and within-group differences.

Specific explanations are not readily apparent, but generally the findings imply that the observed disparities in pregnancy and birth outcomes are due to influences that went unmeasured in this study. Pregnancy loss, PTB, and LBW are correlated with many environmental risks that are experienced unequally by racial/ethnic groups in the US, including inadequate prenatal and preconception care, low social support, and stressful neighborhood conditions to name but a few (for review, see [8]. Moreover, a rapidly growing body of research suggests that health disparities are rooted in exposure to racism, a chronic stressor facing Black, Indigenous, and other People of Color [32, 70, 86]. Racism, manifested interpersonally through racial discrimination and structurally through intersecting and mutually reinforcing systems and policies (e.g., housing,employment; criminal justice), has been shown in most studies to be strongly associated with adverse pregnancy and birth outcomes [1, 12, 39, 45, 84]. There is a need to further specify the causal pathways that link racism to disparities while moving toward consensus on definitions and measures of racism along with methods for differentiating the effects of racism from the effects of other adverse conditions and social determinants [31, 33, 40, 53]. Toward that end, there is a commensurate need to explore alternative measures of ACEs and AAEs that incorporate experiences of discrimination and racism [6].

Unmeasured biological and genetic processes that contribute to health disparities also may have attenuated the effects associated with adversity on pregnancy and birth outcomes in Black women. Downstream correlates of discrimination and racism such as hypertension, obesity, and infection and associated biological processes such as immune response and inflammation are plausible sources of reproductive disparity [8, 15, 18, 88]. In addition, while the amount of variance explained by genetics is thought to be small, theory and evidence suggest that epigenetic changes affecting fetal programming are likely to contribute more substantially to differential outcomes [8, 44, 53, 93]. Research on transgenerational epigenetic inheritance indicates that the effects of historical traumatic events, including ACEs, can be passed across generations [48]. Thus, inquiries informed by the life course perspective may be complemented by epigenetic studies that identify modifiable mechanisms of intergenerational transmission.

Taken together, all signs point to the conclusion that reproductive inequities stem from highly complex, syndemic origins. Although the adversities we measured did not help to explain why unwanted pregnancy and birth outcomes disproportionately affect Black women, it must be emphasized that exposure to early adversity and cumulative life course adversity increased the risk of poor outcomes in the full sample and in each racial/ethnic group. Thus, the fact that ACEs and AAEs were not linked to racial/ethnic disparities should not be misconstrued as implying that these events and conditions lacked predictive validity. Rather, we verified that adverse experiences are associated with poorer pregnancy and birth outcomes (e.g., [1, 15, 35, 46, 55, 60, 85]). Pending replication, our results suggest that widespread prevention of ACEs and AAEs may reduce pregnancy loss, preterm birth, and low birth weight overall but may not mitigate the persistent disparities that continue to plague the population and puzzle the field.

Study Limitations

Results should be interpreted while considering the study’s selective sample of women from low-income households receiving perinatal home visiting services in Wisconsin. Although there is little evidence to suggest that home visiting programs enhance pregnancy and birth outcomes [37, 64], the data are from a selective sample and should be compared with caution against populations with dissimilar characteristics. Along with questions of generalizability, home visiting enrollment may have corresponded with unmeasured differences between racial/ethnic groups (i.e., selection effects). Another limitation is the study’s modest sample size, which restricted statistical power and likely increased the Type-II error rate. Future studies with larger and more diverse samples are needed to assess reproductive disparities in other racial/ethnic groups, especially American Indian and Alaska Native populations given their elevated rates of PTB and LBW [9]. In addition, data on adverse experiences were drawn from retrospective assessments that may have introduced recall biases. Finally, while racial/ethnic disparities are of great importance, it is well understood that race and ethnicity are socially constructed phenomena that may be misclassified or oversimplified when they are operationally defined as mutually exclusive, binary categories.

Conclusion and Future Directions

Compared to other high-income countries, the US has higher rates of pregnancy loss, PTB, and LBW [21, 23], and it is also home to unrelenting disparities in reproductive health. In the interest of developing prevention and intervention strategies to remediate these disparities, there is a need to better understand the root causes of these complex outcomes. Two decades ago, Lu and Halfon [50] put forward a life course perspective that links health disparities to early and ongoing environmental stressors. This model has helped to shape maternal and child health policy in the US, though empirical studies of its assumptions are in short supply. The current study aimed to help fill this gap.

In this study of women from low-income households, we confirmed that greater exposure to early adversity and cumulative life course adversity increased the risk of poor pregnancy and birth outcomes. However, instead of contributing to disparities that disproportionately affect Black women and children, early and ongoing exposure to adverse experiences better explained pregnancy loss, PTB, and LBW among non-Hispanic White and Hispanic participants. We look forward to replication studies with larger and more representative samples in the US and worldwide given that similar racial/ethnic disparities have been observed in many high- and middle-income countries [78]. Bearing in mind that socioeconomic inequalities often correlate with race and ethnicity, researchers who interrogate nationally representative datasets are encouraged to unpack racial/ethnic differences across socioeconomic strata.

Further progress toward understanding how adverse environments affect reproductive health may be possible through advances in measurement and modeling. Future work should explore additional perinatal outcomes such as small for gestational age, recurrent pregnancy loss, timing and type of pregnancy loss, and various birth weight thresholds. The study of life course adversity can be enriched by extending recent efforts to develop ACE measures that are chronologically specific and more ecologically and cross-culturally valid [6, 34, 57] while also employing complementary, broad-based assessments of adverse adult experiences. We encourage scholars to push beyond the limits of the current study by measuring other factors that may increase allostatic load and contribute to physiological wear and tear, including conditions that affect the developing fetus. Along with everyday discrimination, racial discrimination, institutional racism, and structural racism warrant special attention, as do processes that may contribute to intergenerational transmission [31, 33, 88].

Finally, while there are many known correlates of reproductive health disparities, longitudinal studies into causal mechanisms have great potential to inform prevention efforts. Considerable funding and industry has been put toward improving pregnancy and birth outcomes via medical and clinical intervention, producing many practice recommendations [16, 38, 90] but no appreciable impact on population-level disparities. Less attention has been paid to psychosocial interventions that may improve pregnancy and birth outcomes overall and reduce disparities. Research has linked continuous support from health providers (e.g., nurses; midwives; doulas) to many benefits for birthing people [36], and one observational study found that doula care was associated with a lower rate of PTB [43]. Group-based models of care show some promise, though the evidence to date has not been uniform [11, 49]. Further trials are needed to examine the effects of these alternative and complementary health practices on pregnancy and birth outcomes, as are policy solutions that remove cost barriers to bringing these models to scale [5, 43]. Yet, if reproductive disparities are borne out of stress-related processes from conception to birth, as the life course perspective suggests, preconception and prenatal interventions may be necessary but insufficient remedies. Thus, we call for research on large-scale programs, policies, and systems change initiatives that may reduce inequities wholesale among disadvantaged and marginalized groups that bear the brunt of this wicked problem.