Structural Heteropatriarchy and Maternal Cardiovascular Morbidities

Abstract

The United States has some of the poorest maternal health outcomes of any developed nation. Existing research on maternal cardiovascular morbidities has focused predominantly on individual- and clinic-level drivers, but we know little about community- and structural-level factors that shape these outcomes. We use a composite measure of “structural heteropatriarchy” which includes measures of structural sexism and structural LGB-stigma to examine the relationship between structural heteropatriarchy and three cardiovascular-related maternal morbidities using the National Longitudinal Study of Adolescent to Adult Health (n=3,928). Results using multivariate regressions show that structural heteropatriarchy is associated with increased risk of reporting maternal morbidities. Our findings provide further evidence that sexuality- and gender-based stigma operate together to shape health disparities, including maternal health.

The United States (US) has some of the poorest maternal health outcomes of any developed nation (Hoyert, 2022). While researchers and public health officials have paid increasing attention to the high rates of maternal mortality, they have given less attention to maternal cardiovascular morbidities (such as maternal hypertension and preeclampsia). However, these morbidities are far more common and are associated with the later life cardiovascular health of pregnant people and a variety of adverse fetal outcomes (Kuklina, Ayala, and Callaghan, 2009; Lo, Mission, and Caughey, 2013; Shah, 2020; Zhang, Meikle, and Trumble, 2003). Existing research on maternal morbidities has focused predominantly on individual- and clinic-level drivers, but we know little about community- and structural-level factors that shape these outcomes. Understanding the social conditions associated with maternal cardiovascular morbidities is critical to improving the health of pregnant people both during pregnancy and the postpartum period.

A growing body of work has documented how social conditions can shape a variety of individual health outcomes (Everett et al., 2022; Everett, Hatzenbuehler, and Hughes, 2016; Hatzenbuehler, 2014; Homan, 2019; Philbin et al., 2021). And increasingly, research has documented that maternal and fetal health are sensitive to social contexts; this includes how factors such as structural racism, structural sexism, and LGB-specific policies impact birth outcomes and the health of pregnant people (Everett et al., 2022; Homan, 2017; Wallace et al., 2017).

Research has begun to move beyond exploring the health impacts of a single policy to examining how the policy climate (e.g., multiple LGB-specific policies) shapes health outcomes. However, research on how structural conditions shape maternal and infant health has primarily been considered in isolation from one another (i.e., LGB-specific policies or structural sexism). Recently, Everett et al. (2022) introduced the concept of “structural heteropatriarchy (SHP),” arguing that in line with longstanding feminist theory, structural systems of gender- and sexuality-based exclusion should be considered as co-constitutive and mutually reinforcing systems rather than separately (Butler, 2004; Ingraham, 1994; Rich, 1980; Rubin, 2012). In this study, they created a scale that incorporated indicators of structural sexism, family planning, and LGB-based structural stigma into a single construct. They found that higher levels of structural heteropatriarchy were associated with lower birth weights and increased risk of preterm birth among all women, regardless of sexual orientation identity (Everett et al., 2022).

This study extends this line of research by using the National Longitudinal Study of Adolescent to Adult Health (Add Health) to examine the relationship between SHP and maternal cardiovascular morbidities among women who have had at least one live birth.

Background

Pregnancy places profound stress on an individual’s cardiovascular systems; it triggers physiological changes in blood pressure, glucose and lipid profiles, as well as changes in inflammation and hemostasis, which may extend long past the pregnancy and postpartum period. (Elder et al., 2020; Markovitz et al., 2018; Mosca, Barrett-Connor, and Wenger, 2011; Rich-Edwards et al., 2014). The onset of hypertensive disorders during pregnancy, such as hypertension, preeclampsia, and proteinuria, is associated with maternal and fetal mortality, renal failure, pulmonary edema, ICU admittance, preterm birth, and NICU admissions (Kuklina et al., 2009; Lo et al., 2013; Shah, 2020; Zhang et al., 2003).

National data from 2017 to 2019 show an increase in the prevalence of hypertension during pregnancy from 13.3% to 15.9% (Ford, 2022), and roughly 50% of people with gestational hypertension will develop proteinuria during pregnancy (Magee et al., 2003; Rezk et al., 2016). During a healthy pregnancy, a doubling of protein in urine is normal and expected (Dunlop, 1981). However, levels that exceed this, characterized as proteinuria, are considered a sufficient but not necessary condition for preeclampsia diagnosis (Fishel Bartal, Lindheimer, and Sibai, 2022). In addition, more severe fetal and maternal outcomes have been found among women with proteinuria preeclampsia compared to non-proteinuric preeclampsia (Bramham et al., 2013; Guida et al., 2018; Homer et al., 2008). Even without proteinuria or a preeclampsia diagnosis, hypertension during pregnancy is associated with adverse fetal and maternal outcomes (Magee et al., 2003).

Estimates of the prevalence of preeclampsia in the US range from 5 to 8% of all pregnancies. Preeclampsia during pregnancy is accompanied by additional complications, including renal, hematological, hepatic, or neurologic involvement, fetal growth restriction, and elevated liver enzymes (Hutcheon, Lisonkova, and Joseph, 2011). In sum, hypertension, preeclampsia, and proteinuria represent common and concerning health problems during pregnancy that can have serious and long-term consequences for maternal and infant health. Understanding the factors that increase the risk of these maternal morbidities is therefore essential.

Social Determinants of Maternal Health

A 2021 metareview of social determinants-focused studies from 1990-2018 found that most studies that examine social determinants of maternal morbidities focus on individual-level characteristics (e.g., race and ethnicity, socioeconomic status, insurance coverage), with very few studies looking at broader structural factors (Wang et al., 2020). In particular, data show that Black women in the US are more likely to experience hypertensive disorders in pregnancy than other racial/ethnic groups (Ford, 2022) and have much higher rates of maternal mortality (Hoyert, 2022). In recent years, however, researchers have called for more work on the structural determinants of maternal health and a focus on these upstream processes that center systemic forms of oppression and discrimination and their implications for individual-level health rather than a focus on the individual’s characteristics (e.g., racism vs. an individual’s race and ethnicity) (Clark et al., 2022; Crear-Perry et al., 2021; Ukoha et al., 2022).

To that end, scholars have called for more research that examines how policy, historical forms of discrimination, neighborhood and hospital characteristics shape maternal health. This shift to focusing on upstream factors has the potential to address the root causes of health inequities and move beyond stigmatizing individual persons or subpopulations (Alson et al., 2021; Crear-Perry et al., 2021). For example, studies have found relationships between indicators of structural racism (e.g., redlining, segregation, inequality) and maternal and fetal health outcomes (Grady and Ramírez, 2008; Hailu et al., 2022; Hollenbach et al., 2021; Mayne et al., 2018). However, other dimensions of structural oppression and their relationship to maternal health remain under-studied.

In her seminal 2019 piece, Homan lays out a theory of structural sexism, highlighting that gender inequality functions as a system that operates across institutions (e.g., economic, social, political) and at multiple levels (e.g., state, neighborhood, family) to limit women’s access to power and resources and consequently, can negatively impact their health (Homan, 2019). Subsequent studies have linked indicators of structural sexism to infant and maternal health outcomes, including cesarean section rates and infant mortality (Homan, 2017; Nagle and Samari, 2021). One of the fundamental features of structural sexism is access to abortion, which has received attention in the area of maternal and fetal health research; extant studies show that states with more restrictive abortion policies also have higher rates of preterm and low birthweight births (Wallace, Evans, and Theall, 2017) and maternal mortality (Addante et al., 2021; Vilda et al., 2021).

Everett et al. expanded upon the concept of structural sexism by arguing that researchers should consider the interconnected nature of sex/gender-based discrimination and sexual orientation-based discrimination, a system they call structural heteropatriarchy (Everett et al., 2022). Feminist researchers have long argued that compulsory heterosexuality—the political, cultural, and economic system that reinforces heterosexuality as not only the dominant but the only acceptable, familial, and romantic structure—is integral to the maintenance of the sex/gender system. In fact, some researchers have argued that it is compulsory heterosexuality that undergirds our current patriarchal capitalist society insofar as it institutionalizes traditional gendered arrangements, where women’s primary function is to produce labor through reproduction (Monique, 1981; Rich, 1980). While research has examined the role of gender and sexism in the creation of health disparities, the lack of research explicitly connecting heterosexuality as part of the system is inherently problematic. Ingraham argues that, in fact, the lack of explicit examination of the role of heterosexuality in gender research renders heterosexuality as a feature of patriarchal society that is not just invisible but natural, thereby closing off critical engagement around how these two systems are bound together and mutually constitutive (Ingraham, 1994).

In this study, we integrate two dimensions of structural stigma that largely have been considered separate—structural sexism and structural LGB-stigma—and connect them to multiple dimensions of cardiovascular health during pregnancy. This approach extends the research of Everett et al. in two ways. First, we apply the concept and measure of SHP to maternal hypertensive disorders. Given that previous research has linked indicators of structural sexism and structural LGB-stigma to maternal health outcomes, we seek to further test the SHP construct by examining its relationship to additional health-relevant reproductive indicators. No research has applied structural sexism to maternal hypertensive disorders, and only one has examined the link between structural LGB-stigma and maternal hypertension. In addition, preeclampsia and proteinuria have remained unexamined.

We believe that SHP will impact maternal cardiovascular health for several reasons. First, previous studies have shown that exposure to structural forms of stigma increases stress. For marginalized groups, long-term exposure to forms of chronic discrimination, sometimes referred to as “weathering” (Forde et al., 2019; Geronimus, 1992), has been shown to have profound effects on women’s reproductive systems and obstetrical outcomes. This relationship is in part due to behavioral changes in response to stress (e.g., increased risk of smoking, alcohol use, depression) (Brady and Sinha, 2005; Grzywacz and Almeida, 2008; Kassel, Stroud, and Paronis, 2003; Najavits, Weiss, and Shaw, 1997), as well as physiological responses to stress during pregnancy (Boakye et al., 2021; Dunkel Schetter, 2011; Erbetta, Almeida, and Waldman, 2022; Mishra et al. 2020; Zierden et al. 2023). Moreover, the level of SHP in the environment also likely influences access to material resources (e.g., income and education), as well as reproductive health care services, which may also impact maternal cardiovascular health. Understanding the social contexts that increase the risk for maternal cardiovascular disorders is critical, given their implications for both mothers’ and infants’ long- and short-term health.

Data and Methods

Data come from the Add Health (Harris and Urdy, 2012), a large, probability-based study that has tracked the health of a diverse sample of American adolescents as they transition into adulthood. The study was initiated in 1994 and has collected data from respondents for over 25 years. The initial Add Health sample was drawn from 80 high schools and 52 middle schools throughout the US, with unequal selection probabilities in 1994 (Harris, 2011; Harris and Urdy, 2012). A subsample of students (n=20,747) was asked to complete additional in-home interviews and was contacted for follow-up interviews between 2001 and 2002 (Wave III), 2007-2008 (Wave IV), and 2016-2018 (Wave V). Response rates were 77.4% for Wave III and 69.3% for Wave V. Our sample was restricted to women who reported having had at least one live birth by Wave V, had participated in Waves I, III, and V of data collection, and had valid sampling weights (n=3,942). An additional 14 women were excluded from the sample for missing data on variables included in the analysis (0.36%), resulting in a final sample size of 3,928. This study was approved by the Institutional Review Board at the University of Utah (IRB 00114786)

Measures.

Our dependent variables were derived from a survey question at Wave V that asked respondents, “Have you ever had any of these illnesses or complications during a pregnancy? High blood pressure; Protein in your Urine; Preeclampsia?” From these survey items, we created two variables: one that captures any of the three maternal health conditions (1=yes, 0=no) and a count measure that ranges from 0 (no conditions) to 3 (all conditions). We also created a series of dichotomous variables that measure each condition separately.

Independent Variable.

We employed the structural heteropatriarchy (referred to as SHP in methods and results) scale outlined by Everett et al. in their 2022 paper (Everett et al., 2022). This measure is comprised of 12 measures that captured structural sexism, family planning policy, and LGB policy. This scale was constructed based on longstanding feminist theoretical arguments that gender and sexuality are mutually constitutive systems that reinforce one another. We operationalize this theory by combining three interrelated dimensions—structural sexism, family planning policy, and LGB policy— into a single construct. These three dimensions have previously been linked to women’s and sexual minority individuals’ health and wellbeing (Beccia et al., 2022; Clarke and Mühlrad, 2021; Everett et al., 2022; Hatzenbuehler, 2014; Homan, 2017; McKetta et al., 2022; Nagle and Samari, 2021; Rapp et al., 2022; Stevenson, 2021), are highly correlated, and previously established in the literature as either individual or combined forms of structural stigma.

More specifically, we constructed this measure for Wave III, except for the percent conservative religious, which was limited to Wave I. GPS coordinates at all waves were taken at the respondents’ household and then linked to external contextual-level data sources (e.g., the US Census, the Centers for Disease Control and Prevention, and election results). These contextual indicators came from ancillary contextual files available in the restricted Add Health data set (Fowler, Settle, and Monbureau, 2010; Joyner and Manning, 2019).

Our scale includes five measures of structural sexism: 1) the census tract ratio of men’s to women’s median income; 2) the census tract ratio of men’s to women’s labor force participation; 3) the census tract ratio of men’s to women’s unemployment rate; 4) the county-level proportion of voters for the republican presidential candidate; and 5) the county level proportion of conservative religious adherents (all religions combined) ((Fowler et al., 2010). The income, labor force participation, and unemployment rate measures were standardized so that a value greater than 0 indicates that men earned more than women.

The ratio of men to women’s median income was constructed using two contextual indicators that evaluate the median earnings of men and women, separately, aged 16 or older in Wave III at the census tract level. We divided the median income of men during Wave III by that of women at Wave III. The measure was then standardized such that values greater than 0 indicate a male earning advantage. Similar measurements were also developed to capture the relationship between men’s and women’s participation in the labor force and their unemployment rates.

We also include the standardized proportion of the population that voted for a Republican presidential candidate at Wave III at the county level. Research has shown that the percentage of Republican voters within a community has a negative impact on the mental wellbeing of sexual minority adults (Everett, 2014). Additionally, previous studies have found a relationship between conservative ideology and prejudice against individuals in the LGB community, in part due to the ties between conservative ideology and the endorsement of traditional gender roles (Prusaczyk and Hodson, 2018, 2020). Furthermore, research has shown a connection between sexist attitudes and increased odds of voting for Republican political candidates (Bock, Byrd-Craven, and Burkley, 2017; Schaffner, MacWilliams, and Nteta, 2018).

A standardized measure of conservative religious adherents per capita at Wave I, the only Wave for which this information is available, was included in the scale. Previous research has documented that conservative religious beliefs are associated with bias against the LGB community and endorsement of traditional gender roles, as well as sexist attitudes (Mikołajczak and Pietrzak, 2014; Prusaczyk and Hodson, 2018; Whitehead and Perry, 2019).

We included three state-level measures of family planning policy in Wave III. Abortion access is an essential feature of women’s ability to fully participate in society and to be recognized as full citizens. Previous research has documented the links between abortion policy and maternal and infant health (Clarke and Mühlrad, 2021; Foster, 2020; Krieger et al., 2016; Stevenson, 2021). Our three measures capture: 1) public funding for abortion; 2) mandatory waiting periods and informed consent; and 3) parental consent for abortion. Public funding for abortion is coded as a dichotomous variable that captures whether there is limited or no public funding for abortion (1) or public funding is available (0). Mandatory waiting periods and informed consent for abortion captured whether there was a state-level requirement that pregnant people receive lectures and state-prepared materials on fetal development, prenatal care and adoption, and mandatory waiting period after such information is received. States with no informed consent laws were coded as “0”; states with informed consent laws and an unenforced mandatory waiting period were coded as “1”; and states with informed consent laws and enforced waiting periods were coded as “2”. Parental consent for abortion was measured at the state-level and captures whether states have no parental consent laws (0), unenforced consent laws (1), or enforced consent laws (2).

Finally, we include four indicators of state-level LGB policy derived from Human Rights Council (HRC) data. These data points were released in 2019 as part of an ancillary study of LGB contexts (Joyner and Manning 2019); these data points correspond to the years of the Wave III data collection and the respondent’s location at the time of the Wave III survey. Four dichotomous indicators measured whether a state had: 1) sexual orientation-related policy protections, including employment discrimination; 2) hate crime statute; 3) same-sex marriage; and 4) same-sex/second partner adoption.

We then summed these indicators (i.e., structural sexism, family planning, LGB factors) to create a continuous SHP measure ranging from 0 to 11 with a Cronbach’s alpha of 0.75; Table 1 also provides the individual alpha scores for each measure included in the scale. All contextual measures were assessed at the wave indicated in Table 1 at the time of in-home interviews using GPS coordinates taken at the respondents’ households. These were then linked to external contextual-level data sources (e.g., the U.S. Census, the Centers for Disease Control and Prevention, and election results).

Table 1.

Indicators used to construct the heteropatriarchy scale including overall Cronbach’s alpha for the scale and individual alpha values for each indicator

	Level	Wave	Alpha
Structural Sexism Measures
Ratio of men’s / women’s median income	Tract	3	0.74
Ratio of men’s / women’s labor force participation	Tract	3	0.75
Ratio of women’s / men’s unemployment rate	Tract	3	0.75
% of votes cast for Republican president	County	3	0.71
Conservative denomination adherents per capita	County	1	0.75
Abortion Policy Measures
Public funding for abortion	State	3	0.7
Mandatory waiting periods and informed consent	State	3	0.75
Parental consent for abortion	State	3	0.69
LGB Policy Measures
Employment discrimination for SO	State	3	0.72
Hate crime statute for SO	State	3	0.74
Same-sex marriage / union /etc.	State	3	0.73
Same-sex adoption	State	3	0.74
Test Scale Alpha			0.751

Covariates

Our analyses included a series of control measures that may be associated with increased risk for adverse maternal health conditions. Sexual orientation at Wave III was measured as a dichotomous variable that captured whether respondents identified as “exclusively heterosexual” (referent) or with a sexual minority identity (e.g., mostly heterosexual, bisexual, gay/lesbian, other). Race/ethnicity was a categorical variable measured at Wave III that captured whether respondents identified as non-Hispanic white, non-Hispanic Black, Hispanic, or “other” race/ethnicity. We include a measure of the number of years of education at Wave III that ranges from six to twenty-two. Finally, respondent relationship status at Wave V was coded as a categorical variable that captured whether respondents reported being married (referent), divorced, widowed or separated, or never married.

Because the maternal health condition survey item captured whether the respondent has “ever” had these conditions, we included two measures of age to assess the relationship between age and these conditions. The first measures how old the respondent is at Wave V and ranges from 33 to 43 years of age. We also included a measure that captures whether the respondent was over the age of 35 at the time of their first birth, the clinical cutoff for “advanced maternal age.” An additional variable was generated that captures whether a respondent was missing this information. We also include a categorical variable that measures whether the respondent had had one live birth (referent), two live births, or three or more by Wave V.

We also include a series of contextual county-level variables measured at Wave III that research suggests correlate with maternal morbidities. The first is the number of low birthweight infants per 1,000 births at the county level. This variable was recoded into deciles and ranges from 1 to 10. The second is a continuous measure of the proportion of respondents living under the federal poverty line, which ranges from 0 to .81. Third, because Black women are more likely to face maternal morbidities, we also include a measurement of the proportion of the population that is Black in the county [range: 0 to 1]. Finally, we included a measure that captured whether respondents lived in an urban area, defined as residing in a county in a metro area of more than 250,000 people.

Analyses

We first present descriptive statistics for the total sample. Next, we present the results from a series of multivariate generalized linear models for any condition and the number of conditions. We calculate the risk ratios for any maternal condition and the incidence rate ratio for our count variable. The first model examines the relationship between SHP at Wave III and maternal health, and the second added all our control variables. Next, using the same modeling strategy, we examine the links between SHP and each maternal health condition individually. Finally, we conducted supplemental analyses examining whether there were interactions between SHP and sexual minority status and between SHP and race and ethnicity. All models adjust for Add Health’s population weight and cluster on the primary sampling unit identified in the data set.

Results

Descriptive Statistics

Table 2 presents the descriptive statistics for our sample. Twenty-six percent of the sample reported having had at least one form of a maternal hypertensive disorder during pregnancy. The mean number of hypertensive disorders per pregnancy was .42. Maternal hypertension was the most commonly reported condition (20.4%), while approximately 11% of the sample reported having protein in their urine, and 12% reported preeclampsia.

Table 2.

Descriptive Statistics for Sample (n=3,928)

	%/M	SE
Any Hypertensive Disorder (%)	25.72
No. of Hypertensive Disorders (M)	0.42	0.02
Maternal Hypertension (%)	20.04
Protein in Urine (%)	10.83
Preeclampsia (%)	11.93
SHP Scale (Wave III)	7.32	0.2
Sexual Minority Identity (Wave III)	12.79
Race/Ethnicity, Wave III
Non-Hispanic white	70.28
Non-Hispanic Black	14.14
Hispanic	11.43
Other race/ethnicity	4.16
Age (Wave V)	37.72	0.11
Maternal Age at first birth (Wave V)
<35	89.12
≥35	6.52
Missing	4.37
Number of live births (Wave V)
One	24.84
Two	43.26
Three or more	31.91
Education (Wave III)	13.23	0.09
Relationship Status (Wave V)
Married	65.9
Divorced/Widow/Separated	17.68
Never Married	16.42
Contextual Variables (Wave III)
Census tract poverty	0.15	0.01
Proportion Black	0.14	0.01
Low birth weight (deciles)	5.70	0.17
Urban Metro	67.63

Source: National Longitudinal Study of Adolescent to Adult Health

Notes: SHP= Structural Heteropatriarchy; M= mean; SE=Standard Error

The mean score of the SHP scale at Wave III was 7.32 (out of 11). Almost 13% of the sample reported a sexual minority identity, 70% were non-Hispanic white, and the average number of years of education at Wave III was 13.2. Most women had their first birth before age 35 (89%), and the mean age at Wave V was 37.7 years.

The mean proportion of individuals living in poverty at the census tract level was .15, and the mean proportion of Black residents in the census tract was .14. Sixty-seven percent of respondents live in an urban metro area.

Multivariate Results

Table 3 presents the results from our multivariate models. Panel A, Model 1 shows that a one-unit increase in the SHP scale was associated with a 4% increase in the risk of reporting a maternal cardiovascular condition (RR= 1.04, p=.00). These results were robust to the inclusion of all our control measures in Model 2. In addition, exposure to SHP in Wave III was significantly associated with any condition (RR= 1.03, p=.04).

Table 3.

Results from multivariate logistic and ordered logistic models assessing the relationship between SHP and any/total maternal cardiovascular conditions

	Panel A: Any Adverse Maternal Health Outcome (n=3,928)						Panel B: Total # Conditions (0 to 3) (n=3,919)
	Model 1			Model 3			Model 1			Model 2
	RR	95% CI	P	RR	95% CI	P	IRR	95% CI	P	IRR	95% CI	P
SHP Scale (Wave III)	1.04	(1.02, 1.07)	0.00	1.03	(1.00, 1.05)	0.04	1.05	(1.02, 1.07)	0.00	1.03	(1.00, 1.06)	0.04
Sexual Minority Identity (Wave III)				1.00	(0.81, 1.23)	0.98				1.11	(0.87, 1.41)	0.40
Race/Ethnicity (Wave III)
Non-Hispanic Black				1.19	(0.92, 1.53)	0.18				1.11	(0.85, 1.46)	0.71
Hispanic				0.98	(0.75, 1.28)	0.88				0.91	(0.68, 1.21)	0.51
Other race/ethnicity				0.70	(0.43, 1.14)	0.15				0.58	(0.36, 0.94)	0.03
Maternal Age at First Birth (Wave V)
Over 35 for first birth				1.25	(0.95, 1.65)	0.11				1.31	(0.94, 1.82)	0.12
Missing				0.63	(0.38, 1.06)	0.10				0.70	(0.39, 1.25)	0.43
Number of live births (Wave V)
Two births				1.15	(0.92, 1.43)	0.23				1.29	(0.99, 1.69)	0.05
Three or more				1.52	(1.20, 1.91)	0.00				1.73	(1.33, 2.24)	0.00
Education (Wave III)				0.99	(0.94, 1.05)	0.82				0.98	(0.92, 1.04)	0.43
Relationship Status (Wave V)
Divorce/Separated/Widow				1.04	(0.85, 1.65)	0.69				0.97	(0.78, 1.20)	0.78
Never Married				0.92	(0.71, 1.18)	0.49				0.78	(0.58, 1.05)	0.10
Contextual Variables (Wave III)
Census tract poverty				0.92	(0.45, 1.86)	0.81				1.04	(0.40, 2.72)	0.93
Proportion Black				0.89	(0.61, 1.32)	0.58				0.90	(0.53, 1.49)	0.66
Low birth weight (deciles)				1.01	(0.97, 1.04)	0.73				1.01	(0.97, 1.04)	0.77
Urban				0.89	(0.74, 1.08)	0.23				0.85	(0.68, 1.06)	0.14

Source: National Longitudinal Study of Adolescent to Adult Health, Waves III & V

Notes: SHP= Structural Heteropatriarchy; RR= Risk Ratio; IRR= Incidence Rate Ratio; CI=Confidence Interval

Panel B examines the total number of maternal conditions using ordered Poisson regression. Like Panel A, the results show that SHP in Wave III is associated with the total number of conditions (IRR=1.05, p=0.000). After the inclusion of our control variables, a one-unit increase in SHP at Wave III remained associated with an increased risk of reporting more maternal health conditions (IRR=1.03, p=.04).

Table 4 examines the relationship between SHP and each maternal health condition individually. In model one, a one-unit increase in the SHP scale was associated with a 4% increase in the risk of reporting maternal hypertension (RR=1.04, p=.01), a 5% increase in the risk of reporting preeclampsia, and 7% increase in the risk of reporting proteinuria. After the controls in model 2, only the association between SHP and proteinuria remained statistically significant (RR=1.05, p=.03).

Table 4.

Results from multivariate logistic regressions examining the relationship between SHP and individuals’ maternal cardiovascular conditions

	Maternal Hypertension (n=3,938)
	Model 1			Model 2
	RR	95% CI	P	RR	95% CI	P
SHP Scale (Wave III)	1.04	(1.01, 1.06)	0.01	1.02	(0.99, 1.05)	0.21
	Preeclampsia (n=3,922)
	Model 1			Model 2
	RR	95% CI	P	RR	95% CI	P
SHP Scale (Wave III)	1.05	(1.01, 1.09)	0.01	1.02	(0.99, 1.07)	0.25
	Proteinuria (n=3,912)
	Model 1			Model 2
	RR	95% CI	P	RR	95% CI	P
SHP Scale (Wave III)	1.07	(1.03, 1.11)	0.00	1.05	(1.00, 1.09)	0.03

Source: National Longitudinal Study of Adolescent to Adult Health, Waves I-V

Notes: SHP= Structural Heteropatriarchy; M= mean; RR= Risk Ratio; CI=Confidence Interval

Model 1 adjusts for SHP at Wave III; Model 2 additionally adjusts for age, maternal age, race/ethnicity, sexual identity, parity, education, relationship status, and all other contextual level control variables.

We conducted supplemental analyses (not shown) examining potential interactions between SHP and sexual minority status to determine if the relationship between SHP and maternal cardiovascular health was moderated by sexual identity. These interactions were also not significant. Additionally, we conducted a series of interactions between the SHP scale and race/ethnicity for all our outcomes. These interactions were also not statistically significant. Taken together, these results suggest a relationship between SHP and maternal cardiovascular health, regardless of sexual orientation or race/ethnicity.

Discussion

Hypertensive conditions during pregnancy have increased dramatically in recent years (Kuehn, 2021). Given both the acute risks to pregnant people and infants associated with hypertensive disorders in pregnancy (Bellamy et al., 2007; Ferreira, Peeters, and Stehouwer, 2009), as well as the long-term health implications for women’s cardiovascular health (Elder et al., 2020; Quist-Nelson et al., 2023; Rich-Edwards et al., 2014; Stuart et al., 2022) improving maternal health is one of the most pressing health issues in the US. The results from this study add to the growing literature on the relationship between structural forms of stigma and individual health (Everett et al., 2022; Everett et al., 2016; Hatzenbuehler, 2014; Homan, 2019; Philbin et al., 2021). Specifically, our results extend recent research on the relationship between SHP and birth outcomes, including cardiovascular disorders during pregnancy.

Unfortunately, these results come at a critical period where there have been sharp increases in restrictions to abortion access in the US, most dramatically the repeal of Roe v Wade in 2022, the fifty-year precedent that secured the constitutional right to an abortion at the federal level. At the same time, we have also observed an unprecedented number of legislative attacks on the civil and human rights of LGBTQ populations, using both targeted and universal policies (Philbin et al., 2023). In line with longstanding feminist theory, we argue that compulsory heterosexuality, structural sexism, and structural LGTBQ stigma are interrelated and reinforcing systems of oppression designed to maintain a capitalist patriarchal society (Ingraham, 1994; Monique, 1981; Rich, 1980). Indeed, heterosexuality may be considered the bedrock of the current political system in so far as it reinforces that the primary role of women in society is to reproduce within the confines of heterosexual marriage (Rich, 1980). We argue that restricting individuals’ ability to control their reproduction and their ability to live safely as sexual minority individuals are related forms of enforcing compulsory heterosexuality and motherhood.

It is likely that the larger SHP system trickles down to impact maternal health in at least three ways. The first is the impact SHP has on women’s access to material resources, such as income and education, that are associated with better maternal health outcomes. Second, a fundamental piece of the SHP scale is related to family planning; making abortion more difficult to access means that many pregnant people, for a variety of reasons, including health-related, are forced to carry pregnancies to term that may put them more at risk for adverse outcomes. Third, structural forms of stigma send symbolic messages to individuals about their value and worth in society, and multiple studies have linked structural stigma to population disparities in health and wellbeing (Gee and Ford, 2011; Hatzenbuehler, 2014; Homan, 2019).

This paper has several limitations, including the fact that our indicators of maternal cardiovascular morbidities measure whether an individual “ever” had a cardiovascular disorder during pregnancy. Thus, we cannot adjust for pregnancy-specific health behaviors and preconception health status for women with more than one pregnancy. Moreover, due to the confidentiality processes in place for Add Health data, we were unable to include measures of gender identity-based discriminatory policies, which are a critical piece of sex and gender-based discrimination. Third, Add Health included measures of gender identity in Wave V, however, only six respondents who had had a pregnancy reported an “other” gender identity, and none identified as male. We therefore did not include these participants in the sample. Finally, while we did not find significant interactions between our scale and race/ethnicity or sexual identity in this sample, it is possible that different dimensions of our scale (e.g., abortion, LGB policies) may differentially impact individuals who occupy multiple marginalized statuses differently (Everett and Agénor, 2022). Future work should continue to expand and refine the measurement of structural heteropatriarchy in other data sets where there is more flexibility with the types of policies and other indicators available at the state and local levels.

Despite these limitations, this work adds to research demonstrating the role of upstream factors in (re)producing individual health disparities. Moreover, this work has important theoretical implications for medical sociology as the findings provide more evidence for the consideration of sex and gender-based stigma as a single system of oppression. Living and reproducing in a society that systematically devalues women and sexual and gender minority populations has direct health consequences, and those consequences are levied upon those with the least access to resources to change the power structures that impact them. These findings on structural oppression and maternal health outcomes add to the growing body of literature demonstrating the need for upstream change. For example, state-level ballot initiatives to protect access to abortion have been successful in protecting reproductive rights, even in more conservative states (e.g., Kansas). Moreover, academics should continue to use their scientific expertise to work with and inform policymakers to advocate for inclusive policies that protect the rights of the LGBTQ community. As maternal morbidity and mortality continue to rise in the US (Hoyert, 2022), systematic, collective efforts at the local, state, and federal levels are necessary to change policies and environments to become more inclusive and affirming in order to improve population health.

Highlights.

• Structural sexism and LGB-stigma are related systems that should be considered together.

• We call these combined systems “structural heteropatriarchy”

• Structural heteropatriarchy is associated with maternal cardiovascular morbidities.

• Social environments need to be more inclusive to improve reproductive health.