The Association Between Neighborhood Social Vulnerability and Cardiovascular Health Risk among Black/African American Women in the InterGEN Study

Abstract

Background:

Black/African American women in the United States are more likely to live in neighborhoods with higher social vulnerability than other racial/ethnic groups, even when adjusting for personal income. Social vulnerability, defined as the degree to which the social conditions of a community affect its ability to prevent loss and suffering in the event of disaster, has been used in research as an objective measure of neighborhood social vulnerability. Black/African American women also have the highest rates of hypertension and obesity in the U.S.

Objectives:

The purpose of this study was to examine the relationship between neighborhood social vulnerability and cardiovascular risk (hypertension and obesity) among Black/African American women.

Methods:

We conducted a secondary analysis of data from the InterGEN Study that enrolled Black/African American women in the Northeast United States. Participants’ addresses were geocoded to ascertain neighborhood vulnerability using the Centers for Disease Control and Prevention’s Social Vulnerability Index at the Census tract level. We used multivariable regression models to examine associations between objective measures of neighborhood quality and indicators of structural racism and systolic and diastolic blood pressure (BP) and obesity (body mass index > 24.9), and to test psychological stress, coping, and depression as potential moderators of these relationships.

Results:

Seventy-four percent of participating Black/African American women lived in neighborhoods in the top quartile for social vulnerability nationally. Women living in the top 10% of most socially vulnerable neighborhoods in our sample had more than a threefold greater likelihood of hypertension when compared to those living in less vulnerable neighborhoods. Objective neighborhood measures of structural racism (% poverty, % unemployment, % residents > 25 years old without a high school diploma, and % residents without access to a vehicle) were significantly associated with elevated diastolic BP and obesity in adjusted models. Psychological stress had a significant moderating effect on the associations between neighborhood vulnerability and cardiovascular risk.

Discussion:

We identified important associations between structural racism, the neighborhood environment, and cardiovascular health among Black/African American women. These findings add to a critical body of evidence documenting the role of structural racism in perpetuating health inequities and highlight the need for a multifaceted approach to policy, research, and interventions to address racial health inequities.

Hypertension and obesity disproportionately affect Black/African American women in the United States. The prevalence of hypertension among Black/African American women is 56.7% (Ostchega et al., 2020), which is among the highest prevalence rates of hypertension in the world (Beckie, 2017). In addition, Black/African American women develop high blood pressure (BP) earlier in life and are almost twice as likely to die of hypertension-related causes than non-Hispanic White women (Kung & Xu, 2015). Risks of hypertension for Black/African American women include end-stage renal disease, heart disease, stroke, and poor reproductive outcomes (Barcelona de Mendoza et al., 2017; Kung & Xu, 2015). Obesity affects 57% of Black/African American women in the U.S., compared to 40% of White women, and is associated with an increased risk of developing diseases such as hypertension, coronary heart disease, end-stage renal disease, and diabetes (Hales et al., 2020).

A significant root cause of health inequities in the U.S. is structural racism (Bailey et al., 2017). Structural racism describes laws, rules, and practices maintained by policies and institutions, embedded in economic systems and cultural norms, and perpetuates personal biases and broader socioeconomic inequities (Bailey et al., 2021). For example, the discriminatory practice of redlining, which barred mortgage lending in Black/African American neighborhoods for much of the 20th century, served to racially segregate and disenfranchise Black/African American communities with impacts that persist today (Bailey et al., 2021). Structural racism is embedded in policies and inequitable access to resources and opportunities, creating social vulnerability in the form of segregated, socioeconomically disadvantaged, and underresourced neighborhoods. Black/African American women remain more likely to live in more socially vulnerable neighborhoods than White women, even when controlling for individual socioeconomic status (Cozier et al., 2016). Residence in these neighborhoods is associated with increased cardiovascular risk factors, shorter life expectancies, and higher rates of cardiovascular disease morbidity and mortality (Gary-Webb et al., 2020; Xiao & Graham, 2019). Black/African American women who live in socially vulnerable neighborhoods are also at risk for other predictors of poor health outcomes, including higher allostatic load (Wallace et al., 2013), higher levels of the inflammatory markers (Cozier et al., 2016), disruptions in physiologic stress response (Coulon, Wilson, Van Horn et al., 2016), increased hemoglobin A1C (Cozier et al., 2016), lower HDL (Cozier et al., 2016), expedited aging as evidenced by shorter telomere length (Gebreab et al., 2016), and higher prevalence of depressive symptoms (Schulz et al., 2006). Residing in a socially vulnerable neighborhood also has intergenerational effects, as these women experience higher rates of prenatal depression (Giurgescu et al., 2015), small for gestational age birth (Felker-Kantor et al., 2017), and infant mortality (Wallace et al., 2017).

A number of researchers have identified relationships between objective measures of neighborhood social vulnerability and cardiovascular health. Still, these multiethnic studies have primarily studied men and women older than 50 years, and very few have included Black/African American women of childbearing age (Claudel et al., 2018; Jimenez et al., 2019). Other studies have been limited to older (> 50 years) Black/African American adults (Barber et al., 2016; Coulon, Wilson, Alia et al., 2016). Thus, despite persistent health inequities and elevated risk profiles for hypertension, cardiovascular disease, and maternal mortality, very few studies have explicitly concentrated on a cohort of solely Black/African American women. For instance, in one nationwide study of an all-Black/African American female cohort, Cozier et al. (2007) found an inverse relationship between residential neighborhood median home value and hypertension incidence, even after adjusting for individual risk factors. However, this study included primarily middle-class women and used subjective reports of hypertension diagnosis or antihypertensive medication as the outcome (Cozier et al., 2007).

Black/African American women are at disproportionate risk for poor cardiovascular health outcomes, and these inequities cannot be solely accounted for by individual behaviors or risk factors (Cozier et al., 2016). Given the documented health effects of neighborhood social vulnerability, it is essential to explore associations between the neighborhood environment and cardiovascular health in Black/African American women during the childbearing years. The purpose of this study was to examine whether structural racism in the form of neighborhood social vulnerability is associated with increased cardiovascular health risk (i.e., elevated BP and obesity) among Black/African American women. We hypothesized that living in a neighborhood with higher social vulnerability would be associated with elevated BP and obesity, placing those Black/African American female residents with greater risk factors for cardiovascular disease. We also explored correlations among neighborhood social vulnerability and BP/obesity moderating factors such as stress, depressive symptoms, and coping styles.

Methods

We conducted a secondary data analysis of women enrolled in InterGEN, a longitudinal cohort study of Black/African American women and their children (Crusto et al., 2016; Taylor et al., 2016). The purpose of InterGEN was to examine how environmental (i.e., maternal depression, experiences of racism/discrimination, and parenting stress) and genetic (i.e., DNA methylation) factors influenced BP in mother/child dyads. A community sample of mother–child dyads (N = 250) was recruited in Connecticut from 2014 to 2019. Eligibility criteria included women who: (a) were at least 21 years of age, (b) self-identified as African American or Black, (c) spoke English, (d) did not have a cognitive or psychiatric disorder that could limit accuracy of reporting of study data, and (e) enrolled with a biological child who was 3–5 years old. Per established study protocols, trained research assistants approached women for recruitment in primary care clinics, early childhood centers, and community health fairs, conducted screening to verify eligibility, and obtained written informed consent. The institutional review boards approved the study at participating institutions.

Women and children enrolled in InterGEN completed four study visits over 18–24 months, each approximately 6 months apart. Data used in the present cross-sectional analysis were collected during the baseline (T1) visit. We used audio computer-assisted self-interview (ACASI) software (ACASI LLC, Boston, MA) to collect demographic and psychological data during the visits. Full study methods and procedures have been previously described (Crusto, 2016; Taylor et al., 2016). This secondary analysis of InterGEN data adds objective measures of neighborhood vulnerability and structural racism as environmental variables which may influence BP, which were not available in the original data set.

Variables and Measures

Exposure Variable

Social Vulnerability Index (SVI) 2018:

The Centers for Disease Control and Prevention’s (CDC) Social Vulnerability Index (SVI; Flanagan et al., 2018) is a relative measurement of neighborhood vulnerability, resources, and disadvantage that encompasses four themes:

1. Socioeconomic status;

2. Household composition and disability;

3. Minority status and language; and

4. Housing and transportation (Flanagan et al., 2018).

Scores for each theme are reported individually and as an overall score of neighborhood vulnerability at the county and Census tract levels. The SVI was initially intended to identify the most vulnerable areas during public health and natural disaster emergencies but it has also been used for public health/epidemiology research as an objective measure of neighborhood social vulnerability and disadvantage (An & Xiang, 2015; Dasgupta et al., 2020; Diaz et al., 2021).

The total SVI is an overall relative vulnerability score calculated by the CDC for each U.S. Census tract using 15 social factors from U.S. Census tract-level data. Scores range from 0 to 1; scores closer to 1 indicate a Census tract’s greater social vulnerability relative to other Census tracts across the nation. The four themes listed above break down as: (a) socioeconomic status (based on % poverty, employment, income, and educational attainment); (b) household composition and disability (based on % pediatric and advanced-age residents, residents aged > 5 years with a disability, and single-parent households); (c) racial and ethnic minority (based on % residents who do not identify as White, non-Hispanic and % with limited English proficiency); and (d) housing type and transportation (based on % multiunit structures, mobile homes, crowded dwellings, household vehicle access, and institutionalized group quarters; CDC & Agency for Toxic Substances and Disease Registry [ATSDR], 2020). The 2018 SVI uses data from the 2014–2018 American Community Survey (CDC & ATSDR, 2020). The complete list of data used and the CDC’s methods for SVI calculation have been published elsewhere (Flanagan et al., 2018).

We used ArcGIS PRO (ESRI, Redlands, CA) to geocode the T1 home addresses for InterGEN participants (n = 236). We excluded 14 of the study’s 250 participants from analysis due to missing or incomplete addresses. The geocoded points were linked via a spatial join to determine Census tract and obtain objective measures of neighborhood quality, including the CDC Social Vulnerability Index (SVI) overall score and theme scores and the 15 individual variables of tract-level Census data the CDC uses to calculate the SVI.

Structural Racism:

The individual 15 measures of tract-level Census data were included in analyses in addition to the SVI scores because they are indicators of structural racism (Wallace et al., 2017).

Outcome Variables

BP:

According to The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure recommendations (James et al., 2014), BP was measured three times at each study visit by trained research assistants. Mean systolic and diastolic BPs were calculated at each visit for use in regression models. The dichotomous hypertension variable was created based on the 2017 American Heart Association hypertension diagnostic criteria (Whelton et al., 2018). Each participant was categorized as having normal BP (systolic < 120 mmHg and diastolic < 80mmHg) or elevated BP/hypertension (systolic > 120 or diastolic > 80).

Body Mass Index (BMI):

We computed BMI from the participant’s T1 height(m) and weight(kg; kg/m²), which were measured at each study visit by the research assistant. For this analysis, a participant was dichotomously categorized as overweight/obese if their T1 BMI was > 24.9.

Moderating Variables

Stress Overload:

Stress was measured using the Stress Overload Scale (SOS), a 24-item instrument measuring subjective stress overload (Amirkhan, 2012). The SOS, which predicts increased likelihood of becoming physically ill when faced with psychological stress, has high reliability in a diverse sample including Black/African American individuals (α = .95) (Amirkhan, 2012). The SOS includes two subscales: event load and vulnerability. Event load refers to the burden of outside demands, responsibilities, and pressures and the vulnerability subscale captures feelings of powerlessness, inadequacy, debility, and frailty. In addition, participants reported how they felt during the past week using a 5-point Likert scale ranging from “not at all” to “a lot.” The SOS total score, which includes both subscales and had high reliability in our sample (α = 0.95), was used in the current analysis.

Coping Strategies:

Coping was measured using the 33-item Coping Strategy Indicator (CSI; Amirkhan, 1990). For each item on the instrument, respondents indicated the extent to which they used each coping strategy on a Likert scale ranging from “a little” to “a lot” when thinking about a problem they encountered during the last 6 months. Scores were summed and used to create three coping strategy subscales with high reliability: problem-solving (α = .93), seeking social support (α = .91), and avoidance (α = .84). Summed scores for each subscale were used to create categorical variables to indicate the extent to which respondents engage in each coping strategy, ranging from very low to high.

Depressive Symptoms:

Symptoms of depression were measured using the Beck Depression Inventory (BDI; Beck et al., 1988), which has demonstrated validity among low-income African American outpatients (α = .90; Grothe et al., 2005). This 21-item inventory is scored based on the severity of each symptom (0–3). Scores of all items are summed for a total depression score. A score greater than 16 indicates some clinical depression (17–20 = borderline clinical depression, 21–30 = moderate depression, 31–40 = severe depression, score > 40 = extreme depression).

Covariates

Covariates were selected via a priori knowledge and included age and smoking status. Additionally, BMI was included as an additional covariate in the models where BP was the outcome.

Statistical Analysis

Statistical analyses were completed using SAS^® software, Version 9.4 (Cary, NC). Descriptive analyses were carried out to assess missing data and normality of variables. Because SVI scores for our participants were skewed toward the most vulnerable neighborhoods in the CDC’s nationwide ranking, we conducted bivariate analyses by dichotomizing the overall SVI score into two groups: the most vulnerable 10% of neighborhoods among our sample (highest 10% of SVI scores = scores > 0.976) and the remaining women who lived in neighborhoods with less vulnerability (remaining 90% of scores). We treated SVI scores as a categorical variable for bivariate analyses and logistic regression models and SVI scores as continuous in linear regression models. Chi-square, t, and analysis of variance tests were used to compare demographic characteristics between the dichotomous groups for overall vulnerability.

We used multivariable linear regression to examine associations between overall SVI score, the four SVI subthemes, and various objective measures of neighborhood vulnerability and mean systolic and diastolic BP and BMI. The SVI scores and objective neighborhood measures were treated as continuous variables for these analyses, and we controlled for age and current smoking status as confounders.

Logistic regression was used to determine the likelihood of hypertension or obesity associated with neighborhood social vulnerability. For the exposure in logistic regression, overall SVI and each of the SVI subthemes were dichotomized into two groups (10% most vulnerable and remaining 90% as a comparator group) as described above.

To test stress, depression, and coping as independent moderators of the relationship between neighborhood vulnerability and cardiovascular health indicators (BP and obesity), we conducted multivariable linear regression as described above. We also added stress (SOS), the three scales for coping styles (CSI), and Beck Depression Inventory (BDI) total score as an interaction term between each moderator and overall SVI and subthemes to each model and included covariates for adjustment.

Results

Study Sample

Women ranged in age from 21–46 years, with a mean age of 31.3 years (Table 1). Most participants were overweight or obese (70.3%) and 21.9 % identified as current smokers. At baseline, 34.5% had elevated BP (> 120/80 mmHg). Mean BMI was 29.8 (SD = 8.3, minimum–maximum 13.7–59.0). More than half of the participants had some college education or more (58.1%), 77.0% received Medicaid or government/Affordable Care Act health insurance, and 46.4% reported an annual household income of less than $15,000. Most women (70.6%) were heads of a single-parent household (i.e., not married or cohabiting).

Table 1.

Characteristics of Women Enrolled in InterGEN Study (N = 250)

		Total Sample	Less Vulnerable Neighborhoods (n=225)	Most Vulnerable Neighborhoods a (n=25)
		N (%)	n (%)	n (%)	p-valueb
Age
	20-29 years	105 (42.0)	92 (40.9)	13 (52.0)	.55
	30-39 years	124 (49.6)	114 (50.7)	10(40.0)
	40-49 years	21 (8.4)	19 (8.4)	2 (8.0)
Education
	Less than high school	13 (5.2)	12 (5.3)	1 (4.0)	.84
	High school graduate	91 (36.7)	80 (35.6)	11 (44.0)
	Some college	82 (33.1)	73 (32.4)	9 (36.0)
	Associate’s degree or higher	62 (25.0)	49 (21.8)	4 (16.0)
Annual household income					.37
	< US $15,000	111 (46.4)	96(45.1)	15 (60.0)
	US $15,000-50,000	102 (42.9)	94(44.1)	8 (32.0)
	> US$50,000	25 (10.5)	23(10.8)	2 (8.0)
Health insurance type
	Private	35 (14.1)	33 (14.7)	2 (8.0)	.07
	Medicaid	154 (62.1)	142 (63.1)	12 (48.0)
	Government provided/ACA	37 (14.9)	28 (12.4)	9 (36.0)
	Other	7 (2.8)	6 (2.7)	1 (4.0)
	None	14 (5.7)	13 (5.8)	1 (4.0)
Current smoker
	Yes	54 (21.9)	48 (21.6)	6(24.0)	.79
	No	193 (78.1)	174 (78.4)	19(76.0)
Marital status
	Single/ Divorced/ Separated	175 (70.6)	161 (71.6)	14 (73.7)	.29
	Married/ cohabiting	73 (29.4)	62 (27.6)	11 (26.3)
Latina ethnicity		22 (8.8)	19 (2.8)	3 (6.0)	.80

Note. SVI = Social Vulnerability Index.

^aMost vulnerable neighborhoods are those that scored in the top 10% of our sample for overall Social Vulnerability Index (SVI) score (> 0.976). Less vulnerable neighborhoods are the remaining 90% of neighborhoods.

^bChi-square tests used to compare participants living in less vulnerable versus most vulnerable neighborhoods.

Participants lived in 110 Census tracts across Connecticut, most concentrated in 8 urban areas (Figure 1). Overall SVI score (range 0.009–0.999) had a mean of 0.799 (SD = 0.212). Almost three quarters (n = 171, 73.1%) of our sample lived in the most vulnerable quartile of neighborhoods nationwide, indicated by overall SVI scores > 0.750. Women who lived in the most vulnerable 10% of neighborhoods in our sample tended to be younger, less educated, and have a lower income; however, these differences were not statistically significant in bivariate (chi-squared) analyses (Table 1).

Figure 1.

Residential Distribution and Neighborhood Social Vulnerability of InterGEN Participants

Mean Census tract per capita income was $20,781.67 (SD = $8,656.01). The mean poverty rate for our participants’ Census tracts was 25.8% (SD = 12.4), and the unemployment rate was 15.3% (SD = 7.6; Table 2). The mean percentage of residents of minority race/ethnicity (all persons except White, non-Hispanic [CDC, 2020]) was 76.5% (SD = 21.8), 21.2% (SD = 10.1) of residents older than 25 years did not have a high school diploma or equivalent, and 24.9% (SD = 13.1) did not have access to a vehicle. Census tracts had a mean % of single-parent households of 19.3% (SD = 9.8), residents who did not have health insurance were a mean of 12.5% (SD = 6.1), and a mean of 7.8% (SD = 5.8) had limited ability to communicate in English.

Table 2.

Unadjusted and Adjusted Associations between Neighborhood Vulnerability and Blood Pressure and Body Mass Index (N=234)

		Systolic Blood Pressure				Diastolic Blood Pressure				Body Mass Index
		Unadjusted		Adjusted1		Unadjusted		Adjusted1		Unadjusted		Adjusted2
		β (SE)	p- value	β (SE)	p- value	β (SE)	p- value	β (SE)	p- value	β (SE)	p- value	β (SE)	p- value
Social Vulnerability Index (SVI)		0.72 (4.31)	.87	1.34 (3.8)	.73	2.84 (3.40)	.41	4.00 (3.1)	.20	3.19 (2.56)	.21	4.37 (2.6)	.09
SVI Theme 1-Socioeconomic status		4.03 (4.11)	.33	4.50 (3.7)	.22	4.61 (3.24)	.16	5.61 (30)	.06	3.62 (2.44)	.14	4.71 (2.5)	.06
SVI Theme 2- Household composition		−1.26 (3.29)	.70	0.48 (2.9)	.87	−0.27 (2.60)	.92	1.46 (2.4)	.55	1.48 (1.96)	.45	2.52 (2.0)	.21
SVI Theme 3- Minority & language		1.79 (6.76)	.78	0.67 (5.7)	.90	4.54 (5.02)	.37	4.30 (4.6)	.35	4.68 (3.78)	.22	5.35 (3.8)	.16
SVI Theme 4-Housing &transport		−0.91 (3.96)	.82	−1.81 (3.4)	.60	2.38 (3.13)	.45	1.96 (2.8)	.49	1.77 (2.36)	.45	2.01 (2.4)	.39
Selected Neighborhood Characteristics [cohort mean % (SD)]
Poverty	25.8(12.4)	0.13 (0.07)	.08	0.13(0.07)	.06	0.11 (0.06)	.05*	0.13 (0.05)	.02*	0.10 (0.04)	.02*	0.12 (0.04)	.01*
Unemployment	15.3(7.6)	0.15 (0.12)	.24	0.18 (0.11)	.10	0.15 (0.10)	.13	0.18 (0.09)	.04*	0.06 (0.07)	.43	0.07 (0.07)	.31
No high school diploma	21.2 (10.1)	0.16 (0.09)	.07	0.15 (0.08)	.07	0.12 (0.07)	.09	0.12 (0.06)	.07	0.09 (0.05)	.09	0.10 (0.05)	.05*
Single parent household	19.3(9.8)	0.13 (0.09)	.17	0.13 (0.08)	.12	0.06 (0.07)	.41	0.07 (0.07)	.29	0.08 (0.06)	.15	0.10 (0.06)	.08
Minority race residents	76.5(21.8)	0.002 (0.04)	.95	0.01 (0.04)	.77	0.02 (0.03)	.64	0.03 (0.03)	.38	0.02 (0.02)	.39	0.03 (0.03)	.28
Limited English ability	7.8(5.8)	0.12 (0.16)	.27	0.17 (0.14)	.22	0.17 (0.12)	.18	0.16 (0.11)	.15	0.01 (0.09)	.88	0.01 (0.09)	.88
Crowded dwellings	4.1(3.0)	0.25 (0.31)	.42	0.33 (0.27)	.21	0.31 (0.24)	.20	0.38 (0.22)	.09	−0.08 (0.18)	.68	−0.05 (0.18)	.77
No access to vehicle	24.9(13.1)	0.10 (0.07)	.16	0.07 (0.06)	.28	0.08 (0.06)	.15	0.07 (0.05)	.17	0.11 (0.04)	.01*	0.13 (0.04)	.01*
No health insurance	12.5(6.1)	0.13 (0.15)	.38	0.20 (0.13)	.13	0.10 (0.12)	.39	0.15 (0.11)	.16	−0.08 (0.09)	.41	−0.07 (0.09)	.42

Note: SVI, Social Vulnerability Index; BMI, Body Mass Index; SE, Standard Error; CI, Confidence Interval; SD, Standard Deviation

¹Models adjusted for age, BMI, and smoking status.

²Model adjusted for age and smoking status.

^*p ≤ .05

Mean total stress (SOS) score was 60.9 (SD = 24.04, minimum–maximum 24–119). Mean depression score (BDI) was 6.4 (SD = 7.3, minimum–maximum 0–44). Depression and coping strategies employed by the InterGEN cohort have previously been described in-depth (Brown et al., 2019; Millender et al., 2021; Wright et al., 2020).

In unadjusted and adjusted models (Table 2), we found significant associations between selected neighborhood characteristics and cardiovascular risk factors. The percent of residents living in poverty (β = 0.13, p = .02) and percent of unemployed residents in a Census tract (β = 0.18, p = .04) were significantly associated with higher diastolic BP. BMI significantly increased as percent poverty (β = 0.12, p = .01), percent of residents > 25 years without a high school diploma (β = 0.10, p = .05), and percent of residents without access to a vehicle (β = 0.13, p = .01) increased in adjusted models.

In adjusted models, women living in the most vulnerable neighborhoods were 3.29 times more likely to have elevated BP than women living in less vulnerable neighborhoods (aOR 3.29; 95% CI [1.30, 8.32]; Table 3). Thus, we did not find a significant association between neighborhood vulnerability and BMI in our cohort.

Table 3.

Associations between Neighborhood Vulnerability and Blood Pressure and Obesity (N=234)

	Hypertension1				Overweight/ Obese2
	OR	95% CI	aOR a	95% CI	OR	95% CI	aOR b	95% CI
Residence in 10% Most Vulnerable Neighborhoods
Social Vulnerability Index (SVI)	2.58*	1.10, 6.05	3.29*	1.30, 8.32	1.67	.64, 4.32	2.14	.76, 5.99
SVI Theme 1-Socioeconomic status	1.96	.85, 4.52	2.16	.87, 5.36	1.41	.57, 3.48	1.63	.62, 4.27
SVI Theme 2- Household composition	0.87	.36, 2.08	0.97	.38, 2.49	1.15	.48, 2.75	1.35	.53, 3.43
SVI Theme 3- Minority & language	1.24	.56, 2.76	1.33	.56, 3.16	0.89	.40, 2.02	0.83	.55, 2.05
SVI Theme 4-Housing & transportation	0.63	.24, 1.66	0.53	.19, 1.51	1.57	.60, 4.10	1.51	.57, 3.96

Note: CI = Confidence Interval; OR = Odds Ratio; BP = Blood Pressure; BMI = Body Mass Index

¹Hypertension (HTN): based on 2017 AHA HTN categories, (systolic BP > 120 or diastolic BP >80) .

²Overweight /Obese = BMI > 24.9.

^aAdjusted for age, body mass index, smoking status.

^bAdjusted for age and smoking status.

^*p ≤ .05

Moderating Effect of Stress on the Association Between Neighborhood Social Vulnerability and Cardiovascular Risk

In adjusted models, stress significantly moderated the association between overall neighborhood social vulnerability and systolic BP (Table 4). Stress also moderated the relationships between the SVI socioeconomic theme and systolic BP, the SVI household composition and disability theme and diastolic BP, and overall SVI score and BMI (Table 4). Neither coping nor depression moderated the association between neighborhood vulnerability and cardiovascular risk (results not shown).

Table 4.

Moderating Effect of Stress Interaction on the Association between Residential Neighborhood Vulnerability on Blood Pressure and Body Mass Index (N=222)

	Systolic Blood Pressure1		Diastolic Blood Pressure1		Body Mass index2
	β (SE)	p-value	β (SE)	p-value	β (SE)	p-value
Social Vulnerability Index (SVI)
SVI	25.12 (12.9)	.05*	17.05(10.8)	.12	17.38 (8.8)	.04*
Stress Overload Score	0.30 (0.2)	.07	0.17 (0.1)	.22	0.17 (0.1)	.12
SVI*Stress Overload	−0.39 (0.2)	.04*	−0.22 (0.2)	.19	−0.21(0.1)	.13
SVI Theme 1- Socioeconomic Status
SVI Theme 1	30.78 (11.3)	.01*	17.78 (9.4)	.06	11.98 (7.8)	.12
Stress Overload Score	0.33 (0.1)	.02*	0.16 (0.1)	.19	0.10 (0.1)	.31
SVI Theme 1*Stress Overload	−0.44(0.2)	.01*	−0.21 (0.2)	.15	−0.12 (0.1)	.32
SVI Theme 2- Household Composition
SVI Theme 2	12.37(8.6)	.15	14.14 (7.1)	.04*	8.94(5.8)	.13
Stress Overload Score	0.10 (0.1)	.27	0.12 (0.1)	.31	0.07 (0.1)	.24
SVI Theme 2*Stress Overload	−0.19 (0.1)	.15	−0.21(0.1)	.05*	−0.11 (0.1)	.24
SVI Theme 3- Minority Status and Language
SVI Theme 3	30.26 (20.1)	.13	19.63 (16.7)	.24	21.68 (13.7)	.11
Stress Overload Score	0.40 (0.3)	.13	0.22 (0.2)	.32	0.23 (0.2)	.21
SVI Theme 3*Stress Overload	−0.50 (0.3)	.10	−0.27(0.3)	.29	−0.26 (0.2)	.21
SVI Theme 4- Housing and Transportation
SVI Theme 4	−2.89 (10.9)	.79	−0.52 (9.0)	.95	12.12 (7.4)	.10
Stress Overload Score	−00.03 (0.1)	.83	−0.04 (0.1)	.70	0.11 (0.1)	.16
SVI Theme 4*Stress Overload	0.01 (0.2)	.98	0.04(0.1)	.78	−0.15 (0.1)	.16

Note. BP= Blood Pressure; BMI = Body Mass Index; SVI = Social Vulnerability Index

¹Adjusted for age, BMI, smoking status.

²Adjusted for age and smoking status.

^*p ≤ .05

Discussion

This study adds to a significant body of evidence regarding the harmful effects of structural racism on cardiovascular disease among Black/African American women. We explored associations between objective indicators of structural racism, including social vulnerability and neighborhood quality, and hypertension and obesity in Black/African American women. We found that living in a more socially vulnerable neighborhood is associated with a greater than threefold increase in the likelihood of hypertension, and various individual measures of neighborhood quality are also associated with higher cardiovascular risk. Specifically, living in a neighborhood with higher rates of poverty and unemployment was associated with higher diastolic BP, and living in a neighborhood with increased poverty and lower educational attainment was associated with higher rates of obesity.

To our knowledge, our study is the first to report associations between objective measures of neighborhood quality and objective measures of cardiovascular health in a cohort of Black/African American women of childbearing age. This study is also one of the first to explore these relationships among women living in the Northeast, rather than in the Southern United States. It adds to an emerging body of literature documenting a significant relationship between the quality of the residential neighborhood environment and BP for Black/African American women (Barber et al., 2016; Claudel et al., 2018; Sprung et al., 2019).

Like findings of past studies, we found that stress had an exacerbating effect on the associations between neighborhood vulnerability and cardiovascular disease risk. For example, in a recent study of Black adults with hypertension in Philadelphia, participants identified the stress of living in an unsafe neighborhood as an explanatory factor for their hypertension (Koehler et al., 2018).

Neither coping nor depression moderated the associations between social vulnerability and cardiovascular risk. Though, in past studies, high-effort, active coping in the form of John Henryism has been associated with increased cardiovascular risk for those living in highly disadvantaged neighborhoods (Booth & Jonassaint, 2016). While we did not find evidence of John Henryism in our sample, we also did not find evidence to suggest that coping protects against the harmful effects of structural racism and a vulnerable residential environment. Therefore, our findings suggest that rather than focusing on individual coping strategies, intervention efforts targeted at dismantling structural racism and its harmful effects—including racial segregation and neighborhood disadvantage—will be a more effective strategy for improving cardiovascular health among marginalized women.

While 80% of the Connecticut population is White, our participants lived in neighborhoods composed of a high percentage (77%) of racial/ethnic minority residents, indicating significant racial segregation (U.S. Census Bureau, n.d.). Efforts to quantify structural racism are still early in development (Groos et al., 2018) but the legacy of historically discriminatory policies like redlining suggests that the socioeconomic neighborhood factors we examined are indeed indicators of structural racism (Bailey et al., 2017; Wallace et al., 2017). The overwhelmingly minority neighborhoods where our participants live also have significantly higher rates of poverty (25.8% vs. 7.86%) and unemployment (15.3% vs. 3.8% in November 2019), lower per capita income ($20,781.67 vs. $44,496), and lower high school completion rates (21.2% vs. 0.8%) than the state of Connecticut (U.S. Bureau of Labor Statistics, 2021; U.S. Census Bureau, n.d.). This is further evidence that the systems and policies intended to serve the populace are failing its most vulnerable and marginalized residents and preventing them from improving their lives and the lives of their families.

Strengths and Limitations

This study is strengthened by examination of an exclusive sample of Black/African American women within one U.S. state. It provides a unique opportunity to analyze the influences of environment and structural racism at the intersection of race and gender while holding certain state-level, health-influencing factors such as Medicaid policies, unemployment benefits, and high-school graduation criteria constant.

We conducted this study as a secondary analysis of previously collected data, and thus we were limited to subjective measures of stress and the residential addresses at the T1 visit. We lacked information about how long the participants had lived at that address and were, therefore, influenced by that specific environment. We also did not measure women’s perception of neighborhood quality. Past research with Black/African American women has shown that perceived lower neighborhood quality and safety was associated with higher rates of depressive symptoms in pregnancy (Giurgescu et al., 2015; Sealy-Jefferson et al., 2016) and in general (Schulz et al., 2006). Future research could explore whether perceptions of neighborhood vulnerability align with objective measures to better tailor interventions to improve both objective and subjective neighborhood vulnerability and associated health outcomes.

While our sample had a higher rate of mother-only households than the national rate of 55% for Black/African American children (National Center for Education Statistics [NCES], 2020), the household income of our sample mirrored the fact that nationally, almost half of Black/African American children who live in mother-only households live in poverty (NCES, 2020). A similar analysis with a sample with greater diversity of family structure and socioeconomic status would remove some of the confounding influences of poverty and single-parent stress that may have been overrepresented in our sample.

Because the effects of structural racism and neighborhood environment may be cumulative across the lifespan and likely across generations, other directions for future research may include a longitudinal study of environmental influences of neighborhood vulnerability and structural racism on cardiovascular health. Further examination of sociobiological outcomes of structural racism, such as epigenetic changes associated with neighborhood vulnerability or indicators of structural racism, is also an important direction for future research. Notwithstanding these limitations, our results provide further evidence that the influence on health risk and outcomes of neighborhood environment and one measure of structural racism is significant.

Conclusion

Our results are an essential addition to the literature on cardiovascular health in Black/African American women and structural racism. Understanding and documenting the impact of structural racism on the residential environment and associated health inequities can help guide a multifaceted approach to policy, research, and interventions to address health disparities and improve health equity for Black/African American women.