Social determinants of health and hypertension in women compared with men in the United States: An analysis of the NHANES study

Abstract Background Social determinants of health (SDH) reflecting social deprivation have been developed for population health management. There is a paucity of data on the prevalence of SDH and its associations with prevalent hypertension in women compared with men. Methods A total of 49 791 participants aged over 20 years from the 1999–2018 National Health and Nutrition Examination Surveys, were included. Information on the SDH, including race/ethnicity, education level, family income, housing, marriage status, employment, were collected. We calculated the prevalence ratio (PR) for each adverse SDH with prevalent hypertension and uncontrolled hypertension by using Cox regression with equal times of follow‐up assigned to all individuals with adjustment for age, diabetes, taking lipid‐lowering medication, and health behaviors. The population‐attributable fractions (PAF) of the SDH were also assessed. Results A lower proportion of low education attainment was observed in women than men (women: 16.8% vs. men: 17.9%, p = .003), but women had higher proportions of low family income (15.3% vs. 12.5%, p < .001), unmarried state (47.3% vs. 40.9%, p < .001), and unemployment (22.7% vs. 10.7%, p < .001). All the SDH was significantly associated with hypertension in women. There were significant dose–response associations between the numbers of adverse SDH with hypertension. The total PAF of SDH for prevalent hypertension was greater in women (22.2%) than in men (13.9%). Conclusions The widely influential SDH is associated with prevalent hypertension and uncontrolled hypertension. To improve hypertension management, health resources should prioritize socioeconomically disadvantaged groups considering gender differences.


| INTRODUCTION
During the recent three decades, the prevalence of elevated systolic blood pressure (SBP ≥ 140 mmHg) substantially increased, and disability-adjusted life-years and deaths associated with elevated blood pressure also increased. 1 In the United States, the prevalence of hypertension (blood pressure ≥130/80 or taking medication to lower blood pressure) decreased from 47.0% in 1999-2000 to 41.7% in 2013-2014 and then increased to 45.4% in 2017-2018. 2 Hypertension is a well-known risk factor for cardiovascular disease (CVD), 3 lowering blood pressure has been shown to decrease the incidences of stroke, heart attack, and heart failure.
Social determinants of health (SDH), including low socioeconomic status, low education, ethnic differences, suboptimal built environment, and social support networks are increasingly being discussed due to their associations with major diseases. 4,5 For example, low socioeconomic status, based on household income, education, and employment status, was associated with hazard ratios of 2.3 for CVD mortality and 1.7 for CVD incidence in the UK Biobank cohort. 6 The individual's social characteristics, including their environment and living conditions, may differ between women and men and performed different effects on developing hypertension. For instance, increased life stressors, work-related anxiety, and depression, typically have a more significant impact on women with hypertension than on men. 7 The other SDH, such as marital status and social support, also have different magnitudes on hypertension among women and men. 7 The PURE study found that there was no gender difference in the association between low education level with incident CVD, the contribution to CVD death by low education was also similar in women and men. 8 However, depression was more strongly associated with the risk of CVD in men than in women. 8 Population-attributable fractions (PAF) represent the percentage of the disease prevalence or incidence in the population that will be prevented by removing a specific risk factor. 9 Given a public health perspective, the PAF helps to prioritize health budgets and the distribution of resources depending on the proportion of outcomes attributed to a particular exposure. A large number of studies consistently found that a substantial proportion of incident CVD was attributable to hypertension. 10 The purpose of the present study was to determine the associations and PAFs of hypertension due to SDH by gender among US adults from 1999 to 2000 through 2017-2018. To accomplish these goals, data from 10 cycles of the US National Health and Nutrition Examination Survey (NHANES) were analyzed.

| Study participants
The NHANES comprises a series of cross-sectional, national, stratified, multistage probability surveys of the civilian, noninstitutionalized US population. NHANES was designed to assess the health and nutritional status of the US general population. Details of the study design, protocols of data collection, and data sets are publicly available (http://www.cdc.gov/nchs/nhanes.htm). Every participant completed a household interview and underwent a physical examination. From 1999 to 2000, the survey had been conducted in 2-year cycles. In the present study, 10 cycles conducted from 1999-2000 through 2017-2018 were used. The study protocols were approved by the institutional review board of the National Center for Health Statistics, and written informed consent was obtained from each participant.

| Data collection
Participants in the study underwent in-home interviews, as well as visits to a mobile examination center, where they responded to additional questionnaires and underwent physical examinations and blood sample collection. During the in-home interview, personal medical history and medication use for diabetes, hypertension, and other conditions were evaluated. Current smoking was defined as having smoked at least 100 cigarettes in life and smoking at present. Current alcohol drinking was defined as taking at least 12 times drinks of any type of alcoholic beverage in the last 12 months. Physical activity was estimated using the form of the ars.usda.gov/ba/bhnrc/fsrg).
Information on age, gender, race/ethnicity, education level, family income, housing, marriage status, employment, and medical history were gathered using a standard questionnaire. Low education attainment was defined as attaining less than a high school education. The income-to-poverty ratio (annual family income divided by the poverty threshold adjusted for family size and inflation) was used as a measure of family income. 11 Low family income was defined as less than 100% of the income-topoverty ratio. 11 For investigating housing status, the participants were asked "Is this home owned, being bought, rented, or occupied by some other arrangement by you or someone else in your family?" Employed status was dichotomized as unemployed and T A B L E 1 The characteristics of subjects by gender.

| Statistical analysis
The appropriate weights and design factors were invoked in all the analyses to account for the multistage probability sampling design of the survey. Demographic and other characteristics of study partici-  proportional reduction in disease prevalence that would be achieved if the risk factors were theoretically removed from the population.
PAFs were calculated with adjustment for the aforementioned covariates.

| RESULTS
The present study was limited to participants aged 20 years or older (n = 55 081). In addition, those who were pregnant or lactating at the time of examination or with unknown pregnancy status (n = 2639) or did not have hypertension information (n = 2651) were excluded.
After exclusion, a total of 49 791 participants were included in the final analysis sample (  were significantly associated with uncontrolled hypertension both in F I G U R E 2 Associations between social determinants of health with hypertension in women and men. Model 1 was adjusted for age, body mass index, diagnosed diabetes (yes/no), and taking lipid-lowering medications (yes/no); Model 2 was additionally adjusted for current smoking (yes/no), current drinking (yes/no), high level of physical activity (yes/no), sleep duration, and sodium intake. PR, prevalence ratio.

F I G U R E 3
Associations between social determinants of health with uncontrolled hypertension in women and men. Model 1 was adjusted for age, body mass index, diagnosed diabetes (yes/no), and taking lipid-lowering medications (yes/no); Model 2 was additionally adjusted for current smoking (yes/no), current drinking (yes/no), high level of physical activity (yes/no), sleep duration, and sodium intake. PR, prevalence ratio.   (Table 2). Low education attainment, not homeowner, and being unmarried contributed substantially to prevalent hypertension in women and men, the PAFs of these factors were also higher in women than men.
Among the participants with hypertension, the PAFs of SDH for uncontrolled hypertension were similar in women (22.6%) and in men (23.2%) ( Table 2). Not non-Hispanic White contributed substantially to the risk of uncontrolled hypertension in women (8.3%) and men (9.2%). Low educational attainment (5.6%) and low family income

| DISCUSSION
Using a nationally representative sample of the adult population, our study obtains three major findings. First, women have a more unfavorable SDH than men. This finding was supported by much higher prevalence of low family income, being unmarried, and unemployed in women than in men. Second, despite gender differences in SDH levels, the magnitudes of the associations with prevalent hypertension and uncontrolled hypertension for most SDH were similar in women and men. Third, the contribution of SDH for prevalent hypertension were higher among women than men, especially for low education attainment, not homeowner, and being unmarried. Our results suggest that one-quarter of hypertensive cases can be substantially avoided with improved SDH for women.
Several studies have reported the impact of SDH on hypertension. Similar to our study, the findings of the Atherosclerosis Risk in Communities Study (ARIC) also found that better individual-level SDH was associated with lower hypertension incidence in later life. 19 The association between education and hypertension prevalence has been relatively consistently reported by previous studies, high educational attainment improves the awareness, treatment, and control of hypertension. 20 Blood pressure in people with high educational attainment might be better controlled than in those with low educational attainment. 21 For example, a study in South Korea reported an association between educational attainment and better awareness of blood pressure among women. 22 Pandit et al. found that those who had higher formal education were more aware of their overall health and more compliant with medical therapy, which ultimately lead to better blood pressure control. 23 Based on a cohort conducted in rural Vietnam, less formal education was associated with a lower likelihood of hypertension in men, but this relationship was completely reversed in women. 24 People with higher educational levels were likely to have higher access to healthcare services and better performance in disease prevention and management. 25 Adults with lower education attainment more generally were less likely to initiate and receive preventive treatments. 26 In our study, we not only confirmed the positive associations between SDH and prevalent hypertension, but also presented the gender differences in the associations and contributions of SDH for hypertension. We found the magnitude of the associations between low educational attainment with prevalent hypertension was slightly higher in women than that in men, the contributions to prevalent hypertension and uncontrolled hypertension of this risk factor were also higher in women. Similar to our findings, the PURE study also found that low education was the largest contributing risk factor for CVD death (PAF, 11.6% in women vs. 10.3% in men). 8 This finding enforced the understanding of the role of gender-related factors in the prevalence and control of hypertension. Another study also provide evidence T A B L E 2 Population-attributable fractions (PAFs) and 95% confidence interval (CI) for six social determinants on prevalent hypertension and hypertension not controlled by gender. from China that lower socioeconomic status was associated with incident hypertension and women were more susceptible. 27 Therefore, programs to reduce hypertension prevalence and improve hypertension control should be given high priority among women with low educational attainment. On the other hand, strengthening education attainment (or addressing health barriers in low-education groups) was important to improve hypertension management.
Marital status has also been shown to play an important role in blood pressure. Unmarried men were reported to be nearly 50% more likely to have hypertension compared to married men, while, in women, being widowed increased the risk of hypertension by 92%. 28 There is little literature that investigates the role of employment as well as other gender roles in association with hypertension. Our study found that unemployment was consistently significantly associated with prevalent hypertension but not hypertension control in both women and men, however, the contribution of unemployment to prevalent hypertension was higher in men than that in women. The cause of these relationships is still unclear, but it may be secondary to lifestyle and cultural differences. For instance, calorie-heavy foods are typically inexpensive and rapid to prepare and consume in high-income countries. Hence, while those with lower income or unemployed in a resource-rich country tend to eat unhealthier fast food and processed sugars, leading to high blood pressure. 7 Several mechanisms may explain our findings of SDH relating to hypertension. There are substantial previous literatures concerning the etiology of hypertension specifically focusing on modifiable risk factors related to diet, inactivity, tobacco and alcohol consumption, and obesity. 29 Social factors, particularly individual socioeconomic status, may affect the prevalence/incidence of hypertension via these behavioral factors. 30 It is well known that education, income, and gender inequality can influence life decisions and resource allocation.
People with low socioeconomic status may have limited access to social and economic resources, recreational facilities, and healthy foods, which may directly or indirectly affect individuals' ability to engage in healthy behaviors. In a multicohort study, Wang et al.
found that low socioeconomic status was significantly associated with an increase of four times in the odds of initiating physical inactivity, an increase of more than two times in the odds of continuing physical inactivity and of continuing smoking. 30 This study had several limitations. First, NHANES comprised a series of cross-sectional surveys, so longitudinal changes in SDH and blood pressure at an individual level could not be evaluated. Second, many important SDH, such as living environment and regional economic level and medical resources, were not measured and could not be included in this analysis. Third, although most risk factors were measured using validated methods, measurement error was possible, especially when data were self-reported.
In conclusion, the widely influential SDH is significantly associated with hypertension and hypertension control. These findings support the need for urgent actions and reinforced efforts to address socioeconomic inequalities in hypertension management.
To improve hypertension management, more health resources should prioritize socioeconomically disadvantaged groups with considering gender differences when designing and implementing secondary prevention programs.

This work was supported by grants from Heilong Jiang Provincial
Health Commission (2017-036).