Social Determinants, Blood Pressure Control, and Racial Inequities in Childbearing Age Women With Hypertension, 2001 to 2018

Nonstandard Abbreviations and Acronyms

Hypertension is a major risk factor for eclampsia, preeclampsia, stroke, heart failure, and myocardial infarction, making it one of the most important and modifiable risk factors for pregnancy‐related morbidity and mortality in addition to lifetime cardiovascular disease.^{1, 2} Hypertension is common, affecting 17.6% of childbearing aged women in the United States,³ and has increased in prevalence over the past 10 years.⁴ Black women are more likely than White women to have chronic hypertension and develop hypertensive diseases of pregnancy, including preeclampsia.^{5, 6} Moreover, Black, Hispanic, and Asian women have a higher risk of severe morbidity and stroke during delivery.^{7, 8} Black and Hispanic women who develop pregnancy‐induced hypertension are at least 6 times more likely to die than White women.⁸ Management of hypertension in women of childbearing age is important to mitigate poor outcomes in the perinatal period, inclusive of pregnancy and the year following childbirth. A better understanding of racial and ethnic differences in hypertension control may help advance health equity in the United States.

Risk factors for hypertension and associated adverse maternal outcomes include more than comorbid medical conditions and behavior‐related factors. Multiple studies have shown outcomes of chronic medical conditions, including hypertension, are linked with the conditions of individuals' daily lives and the intersection between multiple identities and life experiences.^{9, 10, 11, 12} Fundamental causes of inequity, such as structural racism and societal bias, lead to disparities in social determinants of health (SDoH).¹³ SDoH, including insurance coverage, experienced racism and discrimination, and health care access, along with race and ethnicity, have been linked to inequities in hypertension control.^{10, 14, 15, 16} Interventions targeting SDoH at the community and individual levels have been shown to improve patient health and decrease risk factors.^{10, 11, 17} SDoH have not been well characterized in women of childbearing age with hypertension, especially in Asian women who are particularly understudied.

We sought to characterize SDoH and blood pressure (BP) control in women of childbearing age with hypertension by race and ethnicity. Our objectives were to (1) examine racial and ethnic differences in SDoH and BP awareness and control, (2) explore the extent to which SDoH, health factors, and modifiable health behaviors contribute to differences in BP control, and (3) examine trends over time in uncontrolled hypertension among women of childbearing age. We focused on the fundamental cause model to explore surface causes that may contribute to racial disparities in BP control to propose fundamental causes of these inequities and potential interventions.¹³ On the basis of prior work studying hypertension in the general population, we hypothesized that rates of hypertension awareness, control, and SDoH would differ between racial and ethnic groups.

METHODS

Study Population

All data and materials are publicly available through the National Health and Nutrition Examination Survey (NHANES) data repository and can be accessed at https://wwwn.cdc.gov/nchs/nhanes/default.aspx.¹⁸ Analytic methods that support the findings of this study are available from the corresponding author on reasonable request. Questionnaire and physical examination data from years 2001 to 2018 were extracted from the NHANES data set, which samples participants making up a representative sample of the US population.¹⁸ NHANES methods were approved by the National Center for Health Statistics Research Ethics Review Board, and participants provided informed consent.¹⁸ This analysis was restricted to nonpregnant women of childbearing age (aged 20–50 years) with hypertension (N=1293). Hypertension was defined as a (1) systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg measured during the physical examination or (2) self‐reported history of antihypertensive medication use. In 2017, BP guidelines expanded to consider hypertension ≥130/80 mm Hg.¹⁹ To conservatively capture hypertension awareness in this sample spanning the guideline change, the prior hypertension definition was used (corresponding to stage 2 hypertension in the 2017 guidelines). We focused on women of childbearing potential to capture the population at risk of maternal morbidity and mortality related to hypertension. Thus, women who were postmenopausal and those who reported a prior hysterectomy, tubal ligation, or removal of both ovaries were excluded. Women were separated into categories (designation in parentheses) based on self‐reported race and ethnicity: non‐Hispanic White race and ethnicity (White), non‐Hispanic Black race and ethnicity (Black), Mexican American (Hispanic), other Hispanic (Hispanic), and non‐Hispanic Asian race and ethnicity (Asian). The non‐Hispanic Asian category was first introduced to NHANES in 2011. Women who selected other race, including multiracial women, were not included because of group heterogeneity (N=64).

Hypertension Awareness and Control

Three consecutive BP measurements were taken in the mobile examination center after 5 minutes of seated rest, with a fourth measurement if one was incomplete or interrupted. The average was calculated, excluding the first measurement for women who had >1 BP value recorded. Women with systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg who did not self‐report a hypertension diagnosis were classified as “hypertension unaware.” Uncontrolled BP was defined as systolic BP ≥140 mm Hg or diastolic BP ≥90 mm Hg. Women who self‐reported they were told to take prescribed medicine because of high BP were included in “BP medication prescribed.”

Social Determinants of Health

SDoH were chosen from the American Heart Association's Scientific Statement on Social Determinants of Risk and Outcomes for Cardiovascular Disease, which includes socioeconomic position (education, income, food security, and home ownership), language, and health insurance/access to health care.¹⁰ Education, health insurance, health care access (including having a usual facility to receive health care and health care visits in the past year), and home ownership were self‐reported. For language, women who reported using no amount of English at home were included in the non‐English only group. Poverty/income ratio was calculated by dividing self‐reported family income by the federal poverty guidelines.²⁰ Food security was categorized by NHANES from answers to 10 questions centering around ability to afford food and hunger. Food insecurity questions in 2001 to 2002 and 2003 to 2004 were not asked of high‐income households.¹⁸ Responses for those individuals were marked as “missing,” which NHANES notes should be counted as negative for food insecurity. Individuals who answered affirmatively to ≥3 questions were designated as food insecure. Food insecurity included low (3–5 affirmative responses) and very low food security (6–10 affirmative responses).¹⁸

Other Covariates

Women who reported not smoking at least 100 cigarettes in their lifetime were included as never smokers. Current alcohol use and quantity of daily alcohol consumption were obtained from the dietary questionnaire. Diabetes and hyperlipidemia were defined by a self‐report of diabetes or high cholesterol, respectively. Gravidity was based on number of prior pregnancies. Body mass index was calculated from measured height and weight. Self‐reported health was based on participants' description of their general health as excellent, very good, fair, or poor.

Statistical Analysis

All analyses were performed using the sample weights, primary sampling units, and strata recommended by the National Center for Health Statistics to account for the NHANES complex sampling design. Standard errors were determined for all metrics using the Taylor series (linearization) method. Demographics, hypertension control, and SDoH were compared between White women (reference) and Black, Hispanic, and Asian women. Multivariable logistic regression was used to model the odds of uncontrolled BP by race and ethnicity, in both all women with hypertension and women aware of their hypertension diagnosis. Model 1 was adjusted for age, education, race, poverty/income ratio, insurance, routine place to go for health care, language, home ownership, food security, and non‐US birth. Model 2 was adjusted for model 1 plus health factors and modifiable health behaviors, including diabetes, body mass index, smoking, physical activity, and sodium, fiber, and alcohol intake. Model fitting was checked using c‐statistic, and degrees of freedom were controlled so as not to overfit. Temporal trends in uncontrolled BP were reported over 6‐year periods using disparity ratios. Disparity ratios were created using the method described by Hunt et al.²¹ The percentages of Black, Hispanic, and Asian women with uncontrolled BP were each divided by the percentage of White women with uncontrolled BP (reference). A ratio equal to 1 indicated no disparity. A ratio >1 indicated greater disparity in the non‐White group compared with White women. P‐trend over the pooled years was calculated by χ² test comparing the disparity ratios. P‐trend was determined for Black/White and Hispanic/White disparity ratios, but was not calculated for Asian/White disparity ratios given that there were only 2 time points.

Analyses were performed with Statistical Analysis System software (v.9.4; SAS Institute, Cary, NC). The significance level was set at P<0.05, and all test hypotheses were 2 sided. Several authors had full access to all data in this study and accept responsibility for the data integrity and analysis.

RESULTS

Demographic Characteristics, Comorbidities, and BP

Across women of childbearing age with hypertension (N=1293), the mean (SE) age was 36.4 (0.25) years, 59.2% were White women, 23.4% were Black women, 15.8% were Hispanic women, and 1.7% were Asian women (Table 1). Black and Hispanic women had higher prevalence of obesity than White women (P<0.001 and P<0.05, respectively), and Hispanic women had higher prevalence of diabetes (P<0.05). A total of 1 in 5 women (20%) had never been pregnant; Black and Hispanic women were less likely to have never been pregnant than White women (P<0.001 and P<0.05, respectively). Black, Hispanic, and Asian women had higher mean gravidity than White women (all P<0.05). Of White women, 21% rated their health as fair or poor, which was less than Black (33%), Hispanic (45%), and Asian (31%) women (all P<0.001). Asian women were more likely to be unaware of their hypertension (26% versus 14% of White women; P<0.001). White, Black, and Hispanic women had similar hypertension awareness. More Black women reported being prescribed antihypertensive medication than White women (75% versus 61%; P<0.001). Black and Asian women had higher mean BP and were more likely to have uncontrolled BP than White women (38% and 36% versus 25%; both P<0.001) (Table 1).

Social Determinants of Health

Black and Hispanic women with hypertension reported less education, with 19% and 39% earning less than a high school degree versus 12% of White women (both P<0.001). Asian women were more likely to have higher education, with 77% attaining at least some college compared with 63% of White women (P<0.001). More Hispanic and Asian women spoke no English at home than Black women (34% and 32% versus 2%; both P<0.001). Black and Hispanic women were less likely to have private insurance and more likely to experience poverty, with 59% of Black and 65% of Hispanic women at a poverty/income ratio ≤1.85 versus 34% of White women (all P<0.001). Hispanic and Asian women were more likely to have no place to go for health care (16.2% and 14.6%, respectively) than White women (8.8%; both P<0.05). Hispanic women were also more likely to have seen no health care provider in the past year (18%) compared with White women (7.9%; P<0.001). Food insecurity was more prevalent among Black (25%) and Hispanic (32%) women than White women (13%; both P<0.001) (Table 2).

Trends in Uncontrolled BP Disparity Ratios

The Black/White uncontrolled BP disparity ratio remained >1 throughout all pooled years, indicating higher rates of uncontrolled BP in Black women (Figure 1). Moreover, this increased from years 2007 to 2012 (1.19 [95% CI, 0.75–1.63]) to 2013 to 2018 (2.28 [1.29–3.28]; P‐trend<0.001). The Asian/White disparity ratio was 1.62 (95% CI, 0.74–2.50) in 2011 to 2012 and 1.78 (0.93–2.63) in 2013 to 2018. The Hispanic/White disparity ratio in years 2001 to 2006 and 2007 to 2012 started around 1, then increased in 2013 to 2018 to 1.44 (95% CI, 0.78–2.10; P‐trend<0.001).

• Download figure
• Download PowerPoint

Factors Associated With Uncontrolled BP

Black women had higher odds of uncontrolled BP compared with White women, and this association persisted after adjusting for SDoH and health factors/modifiable health behaviors (odds ratio [OR], 2.31 [95% CI, 1.08–4.92]; Table 3) and when examining only women aware of their hypertension (OR, 3.12 [95% CI, 1.25–7.82]; Table S1). There was no difference in the prevalence of uncontrolled BP between Hispanic and White women. Asian women had higher odds of uncontrolled BP (OR, 2.03 [95% CI, 1.07–3.87]) compared with White women, although this was no longer significant after adjusting for SDoH. After adjusting for SDoH, health factors, and modifiable health behaviors (model 2), poverty/income ratio >1.85 and private insurance were associated with higher odds of uncontrolled BP, whereas lower sodium intake (<2000 mg/d) was associated with lower odds of uncontrolled BP (Figure 2).

• Download figure
• Download PowerPoint

DISCUSSION

In this US population of women of childbearing age with hypertension, we found that compared with White women, Black women had higher odds of uncontrolled BP that persisted after adjusting for SDoH, health factors, and modifiable health behaviors, whereas Asian women had higher odds of uncontrolled BP that was no longer significant after adjusting for SDoH. Racial and ethnic inequities in individual SDoH, including health care access and food security, were also identified. Although SDoH have been shown to impact maternal mortality, minimal research has explored factors beyond socioeconomic status.²² Our findings are informative for strategies aimed at addressing inequities in adverse pregnancy  and cardiovascular disease outcomes throughout the life course.

Asian women were more often unaware of their hypertension and twice as likely to have uncontrolled BP than White women, although the latter finding was no longer significant after adjusting for SDoH. Although higher income and private insurance were independently associated with uncontrolled BP, there was no significant difference in these factors between Asian and White women. Prior research has shown dietary sodium intake, which was associated with uncontrolled BP, is higher among Asian adults than other racial and ethnic groups.²³ The lack of significance after full adjustment for SDoH may be attributable to low power, resulting from the smaller sample size of Asian women (N=44) included in this study. Further research including a larger number of Asian women may help to better illuminate BP control in this population.

On the other hand, the higher prevalence of uncontrolled BP among Black women persisted after adjustment for SDoH and modifiable health behaviors, aligning with prior research.^{14, 24} This uncontrolled BP inequity is likely not explained by underprescription of antihypertensive medications, as more Black women reported antihypertensive prescriptions than White women. Despite being prescribed antihypertensive agents, barriers to medication adherence, including cost, complex regimens, and lack of shared decision‐making, may contribute to inequities in BP control.²⁵ Addressing clinical inertia, or the failure of clinicians to escalate therapy, may improve BP, although studies have shown conflicting results on whether clinical inertia is more, less, or similarly prevalent among Black patients.^{26, 27, 28, 29, 30}

Our findings suggest factors not explored in this analysis, such as experienced racism, social supports, or stress, may drive inequities in BP control. As discussed by the Black Maternal Heart Health Roundtable, SDoH alone do not explain racial inequity in maternal outcomes and addressing structural racism is necessary to achieve health equity.³¹ The consequences of structural racism have been associated with cardiovascular risk; residential segregation is associated with increased risk of cardiovascular disease, whereas perceived neighborhood safety is associated with lower BP, especially among women.^{32, 33} Powell‐Wiley et al suggest a health‐equity driven framework for investigating SDoH, including both structural and daily lived experiences of racism and discrimination.³⁴ In addition, the concepts of weathering and allostatic load suggest cumulative stress and social disadvantage over time lead to decreased health.^{35, 36} Measurements of stress biomarkers and effects, including BP, used to quantify allostatic load have shown Black women are the most affected racial subgroup.^{37, 38} Incorporating measures of racism and discrimination in questionnaire‐based health studies as well as collection of biologic markers of allostatic load may help further our understanding of these factors' impact on disease. Furthermore, inequities in BP control for Black women compared with White women have become more disparate and concerning over this 18‐year period. A similar trend is seen in maternal mortality rates; Black women have 3 to 4 times the maternal mortality rates of White women, and this disparity has persisted over years.^{22, 39, 40} These findings underscore the urgency in identifying and addressing factors contributing to higher BP in Black women.

Aside from Black race, private insurance and income above the poverty line were associated with uncontrolled BP. Private insurance has been associated with higher costs and less health care access than public insurance.⁴¹ This may be attributable to limited availability or discernability of nearby providers who accept individual insurance plans. The increased odds of uncontrolled BP in women above the poverty line may capture women who would benefit from government services, such as Medicaid or the Supplemental Nutrition Assistance Program (SNAP), but do not meet qualifying poverty guidelines. In the United States, 35 million households are unable to afford essentials while remaining above the federal poverty line.⁴² We speculate that these women may be unable to afford health necessities, such as health care visit copays, medications, or healthy and low‐sodium foods, all which may contribute to uncontrolled BP.

Our study found 1 in 4 Black and 1 in 3 Hispanic women with hypertension experience food insecurity, which is 2‐ to 3‐fold the prevalence in the general US population.⁴³ A report by the US Department of Agriculture showed hypertension is more prevalent with greater food insecurity,⁴⁴ which is consistent with other studies' findings associating food insecurity with hypertension, cardiovascular disease, and poorer chronic disease management.^{9, 45, 46} This relationship has also been seen in children, with household and child food insecurity associated with high BP.⁴⁷ Although food security was not independently associated with uncontrolled BP, its prevalence is important when considering the inverse relationship between sodium intake and BP control, as sodium levels are frequently higher in lower‐cost food options, such as canned, ultraprocessed, and fast food. The lower educational attainment reported by Black and Hispanic women compared with White women may contribute to this inequity, as food insecurity is 6 times more likely when the head of a household does not have a high school education versus those with a college degree.⁴⁸ Contributors to racial educational disparities, such as redlining and the high cost of education, may be potential areas of intervention.^{49, 50} Despite the high prevalence of food insecurity, we did not see a difference in BP control between Hispanic and White women.

Although maternal mortality over the past 25 years has declined globally, it has increased in the United States.⁵¹ As hypertension is a leading cause of maternal mortality that disproportionately affects Black women, SDoH contributing to uncontrolled hypertension represent potential targets for interventions aimed at preventing severe maternal morbidity and mortality. There are actions that can be taken throughout health systems to address SDoH and that may contribute to improved hypertension outcomes. On the provider level, practices should employ communication strategies that maximize patient understanding, encourage questions, and use shared decision‐making. Visual aids, the “teach‐back method,” and interactive lessons have been shown to be helpful communication tools.^{52, 53} Routine screening for food insecurity in practices should be prioritized, with direction toward resources, such as local food banks or federal assistance for those in need. At the local level, interventions specific to culture and community, such as through church and barber shop partnerships, has been shown to improve cardiovascular risk factors and BP control.^{54, 55} This approach has garnered Centers for Disease Control and Prevention support through the Racial and Ethnic Approaches to Community Health program, which has successfully decreased cardiovascular risk factors in many communities.⁵⁶ At the public health level, SNAP and the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) should be easily accessible and allow healthy, low‐sodium options appropriate for women with hypertension. Expansion of public health insurance eligibility may contribute to improved BP control by making health care more accessible.

Our study has several limitations. The results of this study are limited in generalizability to the United States only, as the complex interplay of race, ethnicity, social inequity, structural racism, and health care systems is unique to the country of the population studied. Because our analysis was limited to the NHANES questionnaire, we were unable to assess other important SDoH, such as food deserts, neighborhood safety, availability of nearby health care facilities, rural versus urban living, and experienced racism, that may mediate outcomes. The broad categorizations of race and ethnicity as White, Black, Hispanic, and Asian do not capture the heterogeneity of the individuals within these groups and are thus limited, and certain subgroups may be masked by inclusion in the larger ethnic group. Questionnaire data were self‐reported and, thus, subject to recall bias. Because no question on medication adherence was included, we were unable to assess if antihypertensive medications were taken as prescribed. Participants had their BP measured at 1 visit only, meaning participants who had higher or lower BPs than their baseline recorded that day may have been inaccurately captured or omitted.¹⁹ Finally, body mass index is not always reflective of obesity.

Our study also has many strengths. The large sample size of childbearing age women with hypertension allows us to extrapolate to the general US population of childbearing age women. The NHANES sampling procedures are specifically designed to gather data from a nationally representative sample of participants, which also contributes to the generalizability of this study. Moreover, the BP evaluation done as part of NHANES allows us an objective measure of hypertension control.

Despite effective hypertension diagnostics and therapy, hypertension awareness and control differs among women of childbearing age by race. Asian women are more likely to be unaware of their hypertension. Black women have higher odds of uncontrolled hypertension, even when SDoH are controlled for. This potentially indicates a greater role for alternative factors, including racism and discrimination, which have affected daily life and opportunities for generations. These differences in BP control may contribute to the known racial disparity in maternal mortality. Although not independently associated with BP control, racial and ethnic inequities in food security are important to address, especially given the association of lower sodium with controlled BP. Individual and population‐wide strategies to address these inequities should be instituted to improve hypertension control, reduce maternal morbidity and mortality, and improve lifelong cardiovascular health.

Sources of Funding

Dr Kovell is supported by the National Center for Advancing Translational Sciences, National Institutes of Health, through grant KL2TR001454.

Disclosures

Dr McManus reports grants and personal fees from Bristol Myers Squibb and Pfizer, grants from Boehringer Ingelheim, nonfinancial support from Apple, personal fees and nonfinancial support from Samsung and Flexcon, personal fees from Avania, grants and personal fees from Heart Rhythm Society, and personal fees and nonfinancial support from Fitbit, outside the submitted work. The remaining authors have no disclosures to report.

Footnotes