The National Health and Nutrition Examination Survey (NHANES) is a nationally representative program conducted by the National Center for Health Statistics to assess the health and nutritional status of the U.S. population. By employing a complex, multistage probability sampling design, NHANES collects comprehensive data through interviews, physical examinations, and laboratory tests. The dataset provides valuable insights into various health conditions, risk factors, and nutritional trends, making it a widely utilized resource for epidemiological research and public health assessments.

A total of 70,190 participants from the NHANES 2005–2018 dataset were initially considered. After excluding 33,347 individuals due to missing data on social determinants of health, 36,843 participants remained. An additional 7 individuals were excluded due to missing hypertension information, resulting in a final analytical sample of 36,836 participants. Among them, 15,082 were classified as having hypertension, while 21,754 were identified as not having hypertension. Figure 1 shows the flow chart of inclusion and exclusion of participants.

This study utilized the Healthy People 2030 framework to define five core domains of SDoH: economic stability, education access and quality, healthcare access and quality, neighborhood and built environment, and social and community context. Corresponding NHANES survey data were used to assess these domains. We selected eight key indicators representing these domains: (1) employment status, (2) poverty-income ratio (PIR), (3) food security, (4) education level, (5) access to healthcare, (6) type of health insurance, (7) home ownership, and (8) marital status. As shown in Figure 2, Spearman correlation analysis demonstrated mild to moderate associations among these SDoH indicators. In addition, to assess multicollinearity, we calculated the Generalized Variance Inflation Factor (GVIF) and its adjusted form (Adjusted GVIF) for each social determinant of health (SDoH) variable. All variables showed GVIF values close to 1, indicating minimal multicollinearity in our regression models (Supplementary Table S1).

To examine the association between cumulative adverse SDoH and hypertension risk, We constructed a cumulative adverse SDoH score following approaches used in prior research (20, 21), where dichotomized indicators are summed to capture the overall burden of social disadvantage. Among the eight SDoH indicators initially considered, five (unemployment, low PIR, food insecurity, low educational attainment, and non-private health insurance) were retained, as they showed statistically significant associations with hypertension in multivariable logistic regression models. The remaining three indicators (access to healthcare, home ownership, and marital status) were not significantly associated and were therefore excluded from the cumulative score. The final cumulative score ranged from 0 to 5, with higher scores indicating greater social disadvantage.

A standardized blood pressure measurement protocol, consistent with recommendations from the American Heart Association, was employed for data collection between 2007 and 2018. Trained healthcare professionals measured blood pressure using a mercury sphygmomanometer with an appropriately sized cuff. Measurements were taken after the participant remained seated in a resting state for 5 min, with three consecutive readings recorded at 30 s intervals. The average of these three readings was used to determine systolic and diastolic blood pressure levels. Hypertension was defined as meeting any of the following criteria: systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, self-reported history of hypertension, or current use of antihypertensive medication.

This study accounted for several important factors in the analysis. To investigate the link between SDoH and hypertension, age, sex, and race/ethnicity were included as fundamental control variables. Moreover, a range of health-related indicators—such as body mass index (BMI), smoking and drinking habits, participation in physical activities, hyperlipidemia, diabetes, and chronic kidney disease (CKD)—were considered. These indicators were incorporated into sensitivity analyses to assess their potential modifying effect on the relationship between SDoH and hypertension.

In our analysis, we accounted for the complex multistage sampling design of NHANES and applied the corresponding survey weights to ensure nationally representative estimates. Following the NHANES Analytic and Reporting Guidelines, we used combined sampling weights (1/7  ×  WTMEC2YR) appropriate for the 14-year survey cycle, which allowed us to adjust for unequal probabilities of selection, nonresponse, and oversampling. Descriptive statistics were calculated to characterize the study population, where continuous variables were expressed as weighted means ± standard errors (SEs), while categorical variables were presented as weighted frequencies and percentages. Group differences in baseline characteristics were assessed using t-tests for continuous variables and chi-square tests for categorical variables. To examine the association between individual and cumulative adverse SDoH factors and hypertension risk, survey-weighted logistic regression models were employed. Three models were constructed: an unadjusted model, Model 1 (adjusted for age, sex, and race), and Model 2 (which further adjusted for seven additional SDoH variables). A cumulative SDoH score was created by summing five key adverse SDoH factors, and its impact on stroke risk was evaluated using these regression models. Additionally, a trend analysis was performed to determine whether there was a linear relationship between the cumulative number of adverse SDoH factors and hypertension risk. To further explore potential sex-specific effects of SDoH, stratified analyses by sex were conducted. Sensitivity analyses were also performed by adjusting for potential confounders, including BMI, smoking and alcohol consumption, physical activity levels, hyperlipidemia, diabetes, and CKD, to assess the robustness of the results. Furthermore, mediation analysis was performed to evaluate whether depression partially mediated the relationship between SDoHs and hypertension risk, considering their potential dual roles as outcomes of adverse SDoHs and risk factors for hypertension. All statistical tests were two-sided, with a significance threshold set at P < 0.05, and survey weights were applied to ensure representativeness of the findings.

The study included 36,836 participants, with 40.94% diagnosed with hypertension (Table 1). Significant differences in demographic and socioeconomic characteristics were observed between participants with and without hypertension (P < 0.001). Hypertensive participants were more likely to be older, with 37.62% aged 60–79 years and 8.81% aged ≥80 years, compared to younger age groups. Males had a slightly lower prevalence/proportion (49.60%) than females (50.40%). Racial disparities were also evident: among hypertensive participants, 69.96% were White and 13.71% were Black, whereas 5.67% were Mexican American.

Among hypertensive participants, 20.72% were unemployed, 52.64% had a lower PIR, and 18.08% had less than a high school education, these proportions were higher than in the non-hypertensive group (p < 0.001). Additionally, a higher proportion of hypertensive participants had government insurance or were uninsured (39.27%) and were more likely to own their homes (73.26%). Most demographic and socioeconomic characteristics differed between participants with and without hypertension (p < 0.001), except marital status and food security. These findings emphasize notable disparities in SDoH factors that may contribute to hypertension risk.

The logistic regression analysis revealed significant associations between multiple SDoH and hypertension risk (Table 2). In the crude model, unemployment, lower PIR, lower education level, government or no insurance, and renting a home were all significantly associated with higher odds of hypertension. However, after adjusting for key covariates in Model 1 (age, sex, and race) and further controlling for other SDoH variables in Model 2, some associations weakened while others remained robust.

Unemployment was consistently associated with a higher risk of hypertension, with the adjusted odds ratio (AOR) in Model 2 at 1.27 (95% CI: 1.17–1.37, P < 0.0001). Similarly, individuals with a PIR below 300% had an increased hypertension risk (AOR: 1.20, 95% CI: 1.10–1.31, P < 0.001). Food insecurity showed no significant association in the crude model, but after adjustment, those experiencing marginal to very low food security had higher odds of hypertension (AOR: 1.25, 95% CI: 1.14–1.36, P < 0.0001).

Education level also played a role, as participants with less than a high school education had a higher likelihood of hypertension (AOR: 1.09, 95% CI: 1.03–1.17, P = 0.02). Notably, individuals without healthcare access exhibited a lower hypertension risk in all models (AOR: 0.60, 95% CI: 0.55–0.66, P < 0.0001). Furthermore, reliance on government or no insurance was associated with an increased risk (AOR: 1.08, 95% CI: 1.01–1.15, P = 0.01).

In contrast, homeownership and marital status showed weaker or non-significant associations. While renting was initially linked to lower hypertension risk, the association became non-significant in Model 2 (AOR: 0.99, 95% CI: 0.91–1.08, P = 0.83). Similarly, not being married was only marginally associated with hypertension in Model 1 (AOR: 1.12, P = 0.002) but lost significance in Model 2 (AOR: 1.06, P = 0.08).

Figure 3 shows a clear dose-response association between cumulative unfavorable SDoH and hypertension risk. The odds ratios increased progressively with each additional unfavorable SDoH: 1.19 (1.10–1.29) for 1 factor, 1.13 (1.03–1.24) for 2, 1.17 (1.06–1.28) for 3, 1.31 (1.20–1.44) for 4, and 1.46 (1.28–1.67) for 5 factors. After adjusting for age, sex, and race, the association between accumulation of adverse SDoH and the risk of hypertension remained consistent. Notably, in both models, a gradual increase in the risk of hypertension was observed as the number of adverse SDoH increased (P for trend < 0.0001).

Our findings revealed significant sex differences in the association between SDoH and hypertension (Table 3). While unemployment increased hypertension risk in both sexes, the effect was more pronounced in males (AOR = 1.39, 95%CI: 1.22–1.58) than females (AOR = 1.23, 95%CI: 1.11–1.37). Similarly, low PIR (<300%) elevated hypertension risk across sexes, though the association was stronger in females (AOR = 1.40, 95%CI: 1.23–1.59) compared to males (AOR = 1.03, 95%CI: 0.92–1.16).

Distinct patterns emerged for other SDoH factors. Lower education significantly predicted hypertension only in females (AOR = 1.24, 95%CI: 1.09–1.41), while showing no effect in males (AOR = 0.95, 95%CI: 0.85–1.07). Conversely, government or no insurance demonstrated a stronger association in females (AOR = 1.12, 95%CI: 1.00–1.25) than males (AOR = 1.06, 95%CI: 0.95–1.19). Marital status showed particularly divergent effects, with unmarried women exhibiting higher hypertension risk (AOR = 1.04, 95%CI: 0.94–1.14) while unmarried men showed a potential protective trend (AOR = 0.97, 95%CI: 0.88–1.08).

Notably, as shown in Figure 4, the cumulative burden of unfavorable SDoH exhibited a dose-response relationship with hypertension risk in both sexes (P for trend < 0.001). However, the risk gradient appeared steeper in females than males at higher SDoH counts, suggesting women may be particularly vulnerable to the compounding effects of multiple social disadvantages.

To validate the robustness of our findings, we performed comprehensive sensitivity analyses by additionally adjusting for potential confounding factors including BMI, smoking status, alcohol consumption, recreational activity, hyperlipidemia, diabetes, and CKD (Table 4). In the fully adjusted model, all key unfavorable SDoH maintained significant associations with hypertension risk: unemployment (AOR = 1.20, 95% CI: 1.10–1.30), low PIR (AOR = 1.06, 95% CI:1.01–1.17), food insecurity (AOR = 1.11, 95% CI: 1.00–1.23), low educational attainment (AOR = 1.08, 95% CI: 1.03–1.19), and government or no insurance (AOR = 1.02, 95% CI: 1.00–1.07).

Furthermore, our analysis revealed a significant dose-response association between cumulative unfavorable SDoH and hypertension risk (Figure 5). Compared to individuals with no unfavorable SDoH, those with increasing numbers of unfavorable SDoH showed progressively higher hypertension risk: 1 SDoH (AOR = 1.04, 95% CI: 0.94–1.15), 2 SDoH (AOR = 1.09, 95% CI: 1.02–1.20), 3 SDoH (AOR = 1.19, 95% CI: 1.05–1.35), 4 SDoH (AOR = 1.28, 95% CI: 1.13–1.45), and 5 SDoH (AOR = 1.56, 95% CI:1.30–1.88). The trend test for this graded association was highly significant (P for trend < 0.001), demonstrating that the accumulation of multiple unfavorable SDoH substantially elevates hypertension risk in a dose-dependent manner.

Given that adverse SDoH are associated with an increased risk of depression, we conducted a mediation analysis to examine whether depression partially mediated the relationship between adverse SDoH and hypertension. As shown in Figure 6, depression significantly mediated the associations between several SDoH and hypertension risk. Specifically, depression accounted for 12.26% of the effect of unemployment, 9.39% of the effect of low PIR, 11.28% of the effect of food insecurity, and 12.61% of the effect of government or no insurance on hypertension risk. Notably, the mediating role of depression in the association between low educational attainment and hypertension risk was not significant (P > 0.05).