Abstract
Background
China has witnessed a greater increase in blood pressure across successive birth cohorts, especially among men, suggesting a widening gender gap in cardiovascular risk profiles. The expanded access to education since the mid-20th century and its differential impact on men’s and women’s cardiovascular health suggest that education may play an important role in this widening gender gap—a topic that remains under-explored.
Objective
This study assessed the mediating effect of education years in the association between birth cohorts and systolic/diastolic blood pressure and examined whether this mediating effect varied by gender.
Methods
Data from the China Health and Nutrition Survey (1991–2015) were analyzed using multilevel moderated mediation analysis among adults born between 1950 and 1975.
Results
Our estimates indicated that (1) education years had a greater increase among women across cohorts. (2) Higher education years were associated with lower systolic/diastolic blood pressure in women, whereas men did not exhibit similar benefits. (3) Consequently, education years mediated the association between cohorts and systolic/diastolic blood pressure differently by gender. More specifically, improving education years curbed the upward cohort trend in systolic/diastolic blood pressure among women but not among men.
Conclusions
Education partially accounts for the widening gender gap in systolic/diastolic blood pressure across cohorts in China. Ensuring educational opportunities could potentially improve cardiovascular health for women. Factors that hinder men from achieving comparable cardiovascular health benefits from increased education years warrant further research.
Supplementary Information
The online version contains supplementary material available at 10.1007/s44197-026-00516-z.
Keywords: Blood pressure, Cardiovascular health, Sex/gender, Education, Birth cohort; china
Introduction
Cardiovascular disease remains the leading cause of death globally [1, 2]. China accounted for over one-fifth of 610 million people living with cardiovascular diseases in the world [3]. Unlike many high-income countries, China’s age-standardized prevalence of cardiovascular diseases has shown limited improvement [3–6]. Monitoring blood pressure, a leading but modifiable risk factor, is essential for cardiovascular disease prevention [5].
Education is often considered a “social vaccine” because it enhances cognitive skills to manage health risks and improves access to health-promoting resources [7–9]. Gender, as a fundamental axis of social stratification, has historically shaped access to education [10]. In China, Confucian and patriarchal traditions historically favored men in education, especially among earlier birth cohorts [11–13].
Paradoxically, despite their educational advantage, men have worse cardiovascular health than women, exhibiting higher blood pressure, greater disease prevalence, and mortality [1]. Moreover, there is evidence suggesting that this male disadvantage in cardiovascular health has widened over time [3, 14, 15]. One explanation lies in gender difference in the health returns to education: while higher education is associated with lower blood pressure among women, the effect is weaker or even reversed for men [16–19].
Additionally, since the mid-twentieth century, global expansions in formal education have substantially narrowed gender disparities in education [11, 12]. China has followed the global trend since the establishment of the People’s Republic of China in 1949. Despite historical setbacks, China’s literacy campaigns, the 1986 Compulsory Education Law, and higher-education expansion in the late 1990 s have progressively narrowed the gender gap in schooling across cohorts [11, 13].
Prior studies have examined how the narrowing female disadvantage in education has contributed to shifting gender disparities in health outcomes such as depression, self-rated health, and life expectancy [20–22]. However, few studies have considered that the health returns to education may differ by gender. Recent evidence shows that education helps reduce women’s disadvantage in cognitive function by accounting for both gender gap in education and in the association between education and cognitive performance [23]. However, it remains unclear whether education shapes gender differences in blood pressure—a cardiovascular health indicator where men are typically disadvantaged.
We hypothesize that education may shape gender differences in blood pressure across successive cohorts through two mechanisms (See Figure S1 in the Supplementary Materials). First, women experience larger improvements in education years across cohorts. Second, they obtain greater cardiovascular health benefits from education. Thereby, women had a less rapid increase in blood pressure throughout cohort succession compared to men, leading to a widening gender gap in blood pressure.
To empirically assess this hypothesis, we used longitudinal data from the China Health and Nutrition Survey (1991–2015) and employed a multilevel moderated mediation approach that jointly evaluated the mediating role of education and its moderation by gender.
Methods
Data Source
We used data from the China Health and Nutrition Survey (CHNS), a longitudinal study that spans from 1989 to 2015 across diverse socioeconomic and geographic regions of China. The original survey in 1989 included eight provinces. The study expanded to include Heilongjiang Province in 1997 and three municipal cities (i.e. Beijing, Shanghai, Chongqing) in 2011. The CHNS utilized a multistage, random cluster sampling method. See Zhang et al. [24] for more details of the CHNS.
Figure S2 of the Supplementary Materials presents our sample selection process. We restricted our analyses to respondents born between 1950 and 1975, who were not pregnant at the time of the survey. A total of 16,318 respondents, with 70,318 observations, were eligible for our study. We carried out a complete-case analysis, which led us to exclude 1,952 respondents—amounting to 12.0% of 16,318 eligible respondents—due to missing data. Our final sample consisted of 14,029 respondents, with 49,445 observations. These respondents were followed for an average of 9.0 years.
Measures
Dependent Variables
Blood pressure is categorized into two types: systolic and diastolic blood pressure. Systolic and diastolic blood pressure were measured by trained physicians using a mercury sphygmomanometer, after a five-minute rest, and repeated three times at one-minute intervals. We used the mean of three readings as our dependent variables.
Independent Variable
Birth cohorts were measured as a continuous variable by birth year and centered at 1950—the study’s earliest birth year of respondents. We also included the cohort squared term to account for non-linear cohort trends of systolic and diastolic blood pressure, as evidenced by prior research [25].
Mediator
We used education years, the number of years of formal education completed, as a continuous variable.
Modifier
Gender is coded as a dummy variable, with female as the reference category.
Control Variables
We incorporated control variables including age, marital status, hukou, household income per capita, employment, provinces, loss to follow-up, and death (See Table S1 in the Supplementary Materials for measurement details). We included the age squared term to capture the nonlinear age trends of systolic and diastolic blood pressure as documented prior research [16]. Of the 14,029 respondents, 36.0% participated in previous survey waves but were censored in the 2015 wave, and 2.3% had passed away by the 2015 wave. To mitigate potential biases from mortality selection and attrition, we adjusted for dummy variables for the deceased and for respondents lost to follow-up, following the approach used by Chen et al. [26].
We did not adjust for lifestyle factors or health care utilization because these variables are likely on the causal pathway between education and systolic and diastolic blood pressure [7, 9, 27]. Adjusting for them could underestimate the role of education.
Statistical Analyses
Multilevel Moderated Mediation Analysis
To assess the mediating effects of education years in the association between cohorts and systolic and diastolic blood pressure and to determine if the mediating effects were modified by gender, we applied the moderated mediation approach, which combines both mediation and moderation analyses, also known as conditional indirect effects [28].
Given the longitudinal nature of the CHNS data, with observations across waves (Level 1) nested within respondents (Level 2), we applied multilevel moderated mediation analysis, as illustrated in Fig. 1. Time-invariant variables, such as cohorts and gender, are at Level 2. For a time-variant variable like education years, we followed the approach of Zyphur et al. (2019) and Fang et al. (2014) [29, 30]. We differentiated its between-person variance 
from its within-person variance 
. 
is modeled at Level 1, while 
is modeled at Level 2. According to Preacher et al. (2010), if any variable, including predictors, mediators, or outcomes, is at Level 2, then the mediating effect should also be at Level 2. Therefore, we focused on the mediating effect of 
at Level 2, since the independent variable (cohorts) is at Level 2.
Fig. 1.
Conceptual framework of multilevel moderated mediation analysis
The equations for the multilevel moderated mediation analysis are as follows:
Level 1: 
Level 2: 
Level 1: 
Level 2: 
and 
refers to respondent j’s education years and systolic/diastolic blood pressure at age i, respectively. 
and 
are the intercept terms for the education year and systolic/diastolic blood pressure models, respectively. 
and 
denote the error terms for each model. 
and 
indicate the overall average education years and systolic/diastolic blood pressure levels when all predictors are set to their reference levels. 
and 
are the random intercepts for the education year and systolic and diastolic blood pressure models, respectively.
We included the interaction between cohorts and male in the Level 2 education year model to address gender-specific cohort trends of education years, as evidenced by Wu and Zhang [13]. Additionally, we included the interaction between 
and male in the Level 2 systolic/diastolic blood pressure model to account for gender disparities in the association between education years and systolic/diastolic blood pressure, as supported by prior studies [16, 17]. We also included the interaction between cohorts and male in the Level 2 systolic/diastolic blood pressure model to account for the possibility that the direct effect of cohorts on systolic/diastolic blood pressure, after considering the mediating effects of education years, may vary by gender. Details on control variables for each model are introduced in the Supplementary Materials due to word limit constraints.
We conducted the multilevel moderated mediation analysis using the maximum likelihood robust (MLR) estimator in Mplus Version 8.0. Model fit was evaluated using standard fit statistics: Comparative Fit Index (CFI) (> 0.09), Tucker Lewis Index (TLI) (> 0.90), Root mean square error of approximation (RMSEA) (< 0.08), and Standardized Root Mean Square Residual (SRMR) (< 0.08) [31]. For details and example Mplus codes of multilevel moderated mediation analysis, see Zyphur et al. (2019) and Fang (2014) [29, 30].
Sensitivity Analysis
To assess the robustness of our findings against missing data, we re-ran the multilevel moderated mediation analysis using the full information maximum likelihood estimator, known to be less biased and more reliable than complete case analysis [32]. We compared our estimates from the complete case analysis with those obtained using the full information maximum likelihood to verify the robustness of our findings to missing data.
To assess potential residual confounding, we conducted a sensitivity analysis that additionally included antihypertensive medication use as a covariate to examine whether the main findings are robust to adjusting for medication use.
Results
Table S2 in the Supplementary Materials presents the descriptive statistics for variables analyzed for both the entire sample and by gender.
The Widening Gender Gap in systolic/diastolic Blood Pressure across Cohorts
Before assessing the mediating effects of education years, we confirmed the widening gender gap in systolic/diastolic blood pressure across cohorts before adjusting for education years (See Figure S3 in the Supplementary Materials for details). We then investigated the role of education years in this widening gender gap below.
The gender-moderated mediating effects of education years in the association between cohorts and systolic/diastolic blood pressure
Table 1 reports the estimates from the multilevel moderated mediation analysis.
Table 1.
Estimates from the multilevel moderated mediation analysis for systolic/diastolic blood pressure, the China health and nutrition survey (1991–2015)
| Systolic blood pressure | Diastolic blood pressure | |||||
|---|---|---|---|---|---|---|
| Parameter name | Estimate | Standard error | P value | Estimate | Standard error | P value |
| Between-level | ||||||
| Education year model | ||||||
| Education year intercept | −0.26 | 0.02 | < 0.001 | −0.26 | 0.02 | < 0.001 |
| Cohorts→Education year | 0.28 | 0.02 | < 0.001 | 0.28 | 0.02 | < 0.001 |
| Cohort squared→Education year | −0.08 | 0.01 | < 0.001 | −0.08 | 0.01 | < 0.001 |
| Gender→Education year | 0.34 | 0.01 | < 0.001 | 0.34 | 0.01 | < 0.001 |
| Cohort•Gender→Education year | −0.14 | 0.01 | < 0.001 | −0.14 | 0.01 | < 0.001 |
| Systolic/diastolic blood pressure model | ||||||
| Systolic/diastolic blood pressure intercept | 119.05 | 0.60 | < 0.001 | 80.23 | 0.41 | < 0.001 |
| Education year→Systolic/diastolic blood pressure | −0.75 | 0.15 | < 0.001 | −0.47 | 0.10 | < 0.001 |
| Education year•Gender→Systolic/diastolic blood pressure | 0.81 | 0.19 | < 0.001 | 0.69 | 0.13 | < 0.001 |
| Cohorts→Systolic/diastolic blood pressure | 2.48 | 0.22 | < 0.001 | 1.59 | 0.14 | < 0.001 |
| Cohort squared→Systolic/diastolic blood pressure | −0.31 | 0.11 | 0.006 | −0.20 | 0.08 | 0.008 |
| Gender→Systolic/diastolic blood pressure | 4.39 | 0.20 | < 0.001 | 3.40 | 0.13 | < 0.001 |
Cohorts Gender→Systolic/diastolic blood pressure |
1.37 | 0.18 | < 0.001 | 0.41 | 0.12 | 0.001 |
| Within-level | ||||||
| BP model | ||||||
| Education year→Systolic/diastolic blood pressure | −0.42 | 0.19 | 0.02 | −0.25 | 0.13 | 0.04 |
Model fit statistics: Systolic blood pressure model: Comparative Fit Index (CFI) = 0.972; Tucker Lewis Index (TLI) = 0.932; Root mean square error of approximation (RMSEA) = 0.022; Standardized Root Mean Square Residual (SRMR) = 0.004 for Level 1 and 0.024 for Level 2. DBP model: CFI = 0.969; TLI = 0.925; RMSEA = 0.022; SRMR = 0.004 for Level 1 and 0.024 for Level 2
The between-level model for education years’ control variables: the between-person components of age, age squared term, hukou status, household income per capita, and time-invariant controls including cohort squared term, dropout, death, and provincial fixed effects
The between-level model for systolic/diastolic blood pressure’ control variables: the between-person components of age, age squared term, hukou status, marital status, employment status, household income per capita, and time-invariant controls including cohort squared term, dropout, death, and provincial fixed effects
The within-level model for education years’ control variables: the within-person components of age, age squared term, hukou status, and household income per capita
The within-level model for systolic/diastolic blood pressure’s control variables: the within-person components of age, age squared term, hukou status, marital status, employment status, and household income per capita
The estimates from the multilevel moderated mediation including the coefficients of control variables are reported in Table S3 in the Supplementary Materials
The Association between Cohorts and Education Years Varied by Gender
In the Level 2 education year model, a positive coefficient for cohorts and a negative coefficient for cohort squared term indicate that education years increased across cohorts but at a decreasing rate. The positive gender coefficient suggests that men generally had more education years than women. However, the negative coefficient of the interaction between cohorts and gender indicates that women achieved a greater increase in education years across cohorts than men, narrowing the gender gap in education years among younger cohorts (Fig. 2).
Fig. 2.
Predicted education years at the sample’s mean age (41.9 years), based on the China Health and Nutrition Survey (1991–2015). The estimates shown in the figure were derived from a multilevel linear regression analysis on education years. Estimates from the regression are available upon request
The Association between Education Years and systolic/diastolic Blood Pressure Varied by Gender
In the Level 2 systolic/diastolic blood pressure model, the negative coefficient of education years, combined with the positive coefficient of the interaction between education years and gender, suggests that the association between education years and systolic/diastolic blood pressure varied by gender. As illustrated in Fig. 3, in women, higher education years were associated with lower systolic/diastolic blood pressure. Conversely, in men, higher education years tended to be associated with higher systolic/diastolic blood pressure, although the variation across education years was less pronounced.
Fig. 3.
Predicted systolic blood pressure (A) and diastolic blood pressure (B) at the sample’s mean age (41.9 years) by education year and gender, based on the China Health and Nutrition Survey (1991–2015). The estimates shown in the figure were derived from multilevel linear regression analyses on systolic/diastolic blood pressure. Estimates from these regressions are available upon request
The Mediating Effects of Education Year Varied by Gender
Due to the modification effect of gender on both the association between cohorts and education years and the association between education years and systolic/diastolic blood pressure, the mediating effects of education years—also referred to as the indirect effect of cohorts through education years on systolic/diastolic blood pressure—varied by gender. Table 2 reports the total, direct, and indirect effects of cohorts on systolic/diastolic blood pressure.
Table 2.
Total, direct, and indirect effects of birth cohorts on systolic/diastolic blood pressure by gender, estimated from the between-level moderated mediation analysis, the China health and nutrition survey (1991–2015)
| Systolic blood pressure | Diastolic blood pressure | |||||
|---|---|---|---|---|---|---|
| Parameter name | Estimate | Standard error | P value | Estimate | Standard error | P value |
| Female | ||||||
| Direct effect | ||||||
| Cohorts→Systolic/diastolic blood pressure | 2.48 | 0.22 | < 0.001 | 1.59 | 0.14 | < 0.001 |
| Indirect effect | ||||||
| Cohorts→Education year→Systolic/diastolic blood pressure | −0.21 | 0.04 | < 0.001 | −0.13 | 0.03 | < 0.001 |
| Total effect of cohorts on Systolic/diastolic blood pressure (sum of direct and indirect effect) | 2.27 | 0.22 | < 0.001 | 1.46 | 0.14 | < 0.001 |
| Proportion of the total effect of cohorts was mediated by education year | −9.14% | 2.06% | < 0.001 | −8.96% | 2.06% | < 0.001 |
| Male | ||||||
| Direct effect | ||||||
| Cohorts→Systolic/diastolic blood pressure | 3.85 | 0.22 | < 0.001 | 2.01 | 0.15 | < 0.001 |
| Indirect effect | ||||||
| Cohorts→Education year→Systolic/diastolic blood pressure | 0.01 | 0.02 | 0.705 | 0.03 | 0.03 | 0.058 |
| Total effect of cohorts on Systolic/diastolic blood pressure (sum of direct and indirect effect) | 3.86 | 0.22 | < 0.001 | 2.04 | 0.15 | < 0.001 |
| Proportion of the total effect of cohorts was mediated by education year | 0.21% | 0.56% | 0.705 | 1.42% | 0.75% | 0.059 |
| Gender difference in the indirect effect (female’s indirect effect minus male’s indirect effect) | 0.22 | 0.04 | < 0.001 | 0.16 | 0.03 | < 0.001 |
For women, the indirect effects of cohorts through education years on systolic/diastolic blood pressure were negative and statistically significant for both systolic (−0.21, SE = 0.041, P < 0.001) and diastolic blood pressure (−0.13, SE = 0.03, P < 0.001). Direct effects of cohorts on systolic and diastolic blood pressure in women after accounting for education years were 2.48 (SE = 0.22, P < 0.001) and 1.59 (SE = 0.14, P < 0.001), respectively, with total effects being 2.27 (SE = 0.22, P < 0.001) for systolic blood pressure and 1.46 (SE = 0.14, P < 0.001) for diastolic blood pressure. The proportions of the total effects of cohorts mediated by education years were − 9.14% (SE = 2.06%, P < 0.001) for SBP and − 8.96% (SE = 2.06%, P < 0.001) for diastolic blood pressure.
For men, the indirect effects of cohorts through education years on systolic/diastolic blood pressure were positive but not statistically significant. The gender disparities in the indirect effects of cohorts through education years were statistically significant, with values of 0.22 (SE = 0.04, P < 0.001) for systolic blood pressure and 0.16 (SE = 0.03, P < 0.001) for diastolic blood pressure. To sum it up, improving education years counteracted the upward cohort trend of systolic/diastolic blood pressure in women but not in men, which can partly explain the widening gender gap in systolic/diastolic blood pressure across cohorts.
Estimates from the Sensitivity Analysis
We re-ran the multilevel moderated mediation analysis using the full information maximum likelihood (See estimates in Table S4 in the Supplementary Materials). The sensitivity analysis produced consistent results with the complete case analysis, confirming the robustness of our findings against missing data. Although the magnitude of the mediating effects of education years in women decreased after using the full information maximum likelihood, the main findings regarding the gender-differentiated mediating effects of education years remain largely unchanged.
We found that estimates from the model, which adjusted for antihypertensive medication use, are consistent with the main results (See estimates in Table S5 in the Supplementary Materials), suggesting that the findings are robust to this adjustment.
Discussion
We found that gender differences in both the increase in education years across successive cohorts and the association between education years and blood pressure together partly explain the divergent cohort trends in blood pressure between men and women, thereby contributing to the widening gender gap. More specifically, among women, rising education years partially offset a cohort-based increase in blood pressure. Without this increase in education years, the upward cohort trend would have been more pronounced. Among men, although education years also rose across cohorts, the increase was smaller and less consistent than that observed among women, consistent with prior evidence [13]. More importantly, the association between education years and blood pressure among men was slightly positive, suggesting that higher education did not confer cardiovascular health benefits for men as it did for women—and may even be associated with higher blood pressure.
The divergent cohort trends in men and women help explain men’s steeper cohort increases in systolic and diastolic blood pressure [15, 25], a more rapid rise in cardiovascular disease prevalence, and a slower decline in cardiovascular mortality over time among men in China [3, 14].
This gender disparity in the association between education years and blood pressure aligns with prior studies in both China or other countries [16–19], which might be attributed to gender norms. Women are often expected to prioritize health goals, potentially enhancing the protective effect of education years on blood pressure [33]. In contrast, men are often expected to be masculine, which may encourage risky behaviors (e.g., smoking, alcohol drinking) and discourage health-promoting behaviors like healthcare utilization and thereby reduce the protective effects of education years [33]. Additionally, beauty standards emphasizing slenderness for women might result in more weight-control efforts among women [34], which can positively impact blood pressure. Further investigation into the factors responsible for this gender disparity is needed but is beyond this study’s scope.
Although our study focus on China, it may apply to other contexts. In Western high-income countries, a more educated population correlated with a more pronounced decline in the prevalence of cardiovascular disease risk factors among women [35], consistent with our results. Furthermore, prior studies indicate that improving gender equality in education across cohorts may reduce the female disadvantage in depression, self-rated health, and cognitive function among younger cohorts [21–23]. Additionally, Pinho-Gomes et al. (2023) found that increased gender equality in education was associated with greater life expectancy gains for women than for men at the country level [20], potentially contributing to the increasing male disadvantage in life expectancy. This finding is particularly relevant to our study because cardiovascular disease is a major cause of death worldwide and an important driver of life expectancy. It is therefore possible that the widening male disadvantage in life expectancy observed in [20] may be partly explained by the growing male disadvantage in cardiovascular health.
We limited our analyses to individuals born in 1975 and earlier to maximize age overlap across cohorts, which may underestimate the mediating effect of education years. This mediating effect may be stronger in more recent cohorts. Although government-led literacy campaigns and the 1986 Nine-year Compulsory Education Law positively improved enrollment and gender equality in primary and lower secondary education, major expansions in higher education did not occur until the late 1990 s [11, 13]. By then, the youngest cohort in our analysis, those born in the early 1970 s, had already surpassed school age and were less likely to benefit from these expansions. Future analyses should include individuals born after 1975 to validate the gender-differentiated mediating effect of education as more data on younger cohorts become available.
The direct effects of cohorts on blood pressure, after adjusting for the mediating effects of education years, remained significant and differed by gender. These findings highlight the need for further research into other factors that can explain the widening gender gap in blood pressure across cohorts.
Our findings have both theoretical and practical implications. Theoretically, this study identifies two mechanisms through which how education shapes gender differences in blood pressure: the narrowing gender gap in educational attainment across cohorts, and the stronger negative association between education and blood pressure among women than among men. By formally testing these two pathways, this study helps clarify why men’s educational advantage has not translated into better cardiovascular outcomes. These findings suggest that reducing gender disparities in health requires both narrowing gender gaps in education and addressing gender differences in the health returns to education.
Practically, these results underscore the need for gender-sensitive interventions to curb the unfavorable trend in blood pressure. Enhancing educational opportunities for women can help reduce their cardiovascular disease risk by lowering blood pressure. Special attention should be given to ensuring that less educated women are not left behind in the gains of cardiovascular health improvement. For men, interventions should consider addressing social norms and behavioral factors that may constrain the cardiovascular benefits of education. However, the underlying reasons why men appear to benefit less from education warrant further investigation.
Our study has several limitations. First, the CHNS are not nationally representative, so our findings cannot be generalized to the entire country. However, the CHNS covers a large sample size and includes geographically and economically diverse regions, ensuring the validity of our study. Second, although blood pressure is a key indicator of cardiovascular health, our focus on this single outcome limits the ability to generalize findings across other health domains. Third, men are more likely to experience premature death than women, and those with better cardiovascular health are more likely to survive, which may overestimate the widening gender gap in blood pressure due to the mortality selection effect. However, since our analyses focused on individuals under 65 years old, any bias from mortality selection is likely minimal. Fourth, we did not include the period variable due to perfect collinearity with age and cohort variables in longitudinal data.
Despite these limitations, this study provides important insights into how education shapes gender disparities in blood pressure across cohorts. Our findings suggest that education mitigates increases in blood pressure among women not only because the gender gap in education has narrowed across cohorts, but also—more importantly—because the health returns to education are not uniform across genders. These results caution against explaining gender disparities in health solely in terms of gendered exposure and highlight the importance of considering gender differences in health responses to social exposures. The social and behavioral mechanisms underlying the limited protective association between education and blood pressure among men warrant further investigation.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
This research used data from China Health and Nutrition Survey (CHNS). We are grateful to research grant funding from the National Institute for Health (NIH), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) for R01 HD30880, National Institute on Aging (NIA) for R01 AG065357, National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) for R01DK104371 and R01HL108427, the NIH Fogarty grant D43 TW009077 since 1989, and the China-Japan Friendship Hospital, Ministry of Health for support for CHNS 2009, Chinese National Human Genome Center at Shanghai since 2009, and Beijing Municipal Center for Disease Prevention and Control since 2011. We thank the National Institute for Nutrition and Health, China Center for Disease Control and Prevention, Beijing Municipal Center for Disease Control and Prevention, and the Chinese National Human Genome Center at Shanghai.
Abbreviations
- CHNS
China Health and Nutrition Survey
- CFI
Comparative Fit Index
- TLI
Tucker Lewis Index
- RMSEA
Root mean square error of approximation
- SRMR
Standardized Root Mean Square Residual
Author Contributions
JW, JZ, and RL contributed to the conception and design of the study. JW, BJ, and JM conducted the data cleaning and analysis. JZ and RL reviewed the statistical methods and contributed to the interpretation of results. JW drafted the initial manuscript, and all authors provided feedback on earlier drafts. All authors have read and approved the final manuscript.
Funding
JW received funding from the Shanghai Social Science Foundation (Grant No. 2020ESH003). RL received support from the Start Fund for Specially Appointed Professors of Jiangsu Province and the Seeds Funding of Nanjing Medical University (NMUR20220001). The funding agency had no involvement in the design of the study, data collection, analysis, interpretation, manuscript writing, or the decision to submit the paper for publication.
Data Availability
The data from the China Health and Nutrition Survey is publicly available and can be accessed at: https://www.cpc.unc.edu/projects/china.
Declarations
Ethics Approval and Consent to Participate
This study used publicly available, anonymized secondary data, and as such, ethics approval and consent were not required.
Consent for Publication
Not applicable.
Competing Interests
The authors declare no competing interests.
Footnotes
Publisher’S Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Ruiyun Li, Email: ruiyun.li@njmu.edu.cn.
Jiaying Zhao, Email: zjy789@hotmail.com.
References
- 1.Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammirati E, Baddour LM, et al. Global burden of cardiovascular diseases and risk Factors, 1990–2019: update from the GBD 2019 study. J Am Coll Cardiol. 2020;76:2982–3021. 10.1016/j.jacc.2020.11.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Walli-Attaei M, Joseph P, Rosengren A, Chow CK, Rangarajan S, Lear SA, et al. Variations between women and men in risk factors, treatments, cardiovascular disease incidence, and death in 27 high-income, middle-income, and low-income countries (PURE): a prospective cohort study. Lancet. 2020;396:97–109. 10.1016/S0140-6736(20)30543-2. [DOI] [PubMed] [Google Scholar]
- 3.Global Burden of Disease Collaborative Network. Global Burden of Disease Study 2021 (GBD 2021) Results. Inst Health Metr Eval IHME 2021. https://ghdx.healthdata.org/organizations/global-burden-disease-collaborative-network Accessed 8 Aug 2024.
- 4.Wang F, Yu Y, Mubarik S, Zhang Y, Liu X, Cheng Y, et al. Global burden of ischemic heart disease and attributable risk factors, 1990–2017: a secondary analysis based on the Global Burden of Disease Study 2017. Clin Epidemiol. 2021;13:859–70. 10.2147/CLEP.S317787. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Zhang M, Shi Y, Zhou B, Huang Z, Zhao Z, Li C, et al. Prevalence, awareness, treatment, and control of hypertension in China, 2004-18: findings from six rounds of a national survey. BMJ. 2023;380:e071952. 10.1136/bmj-2022-071952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zhang J, Tong H, Jiang L, Zhang Y, Hu J. Trends and disparities in China’s cardiovascular disease burden from 1990 to 2019. Nutr Metab Cardiovasc Dis. 2023;33:2344–54. [DOI] [PubMed] [Google Scholar]
- 7.Baker DP, Leon J, Smith Greenaway EG, Collins J, Movit M. The education effect on population health: a reassessment. Popul Dev Rev. 2011;37:307–32. 10.1111/j.1728-4457.2011.00412.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jones PR, Bhatta L, Howe LJ, Vinueza-Veloz MF, Davies NM, Davey Smith G, et al. Education and cardiovascular disease: a within-family mendelian randomization analysis. Int J Epidemiol. 2025;54:dyaf144. 10.1093/ije/dyaf144. [DOI] [PubMed] [Google Scholar]
- 9.Lutz W, Kebede E. Education and health: redrawing the Preston curve. Popul Dev Rev. 2018;44:343–61. 10.1111/padr.12141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Read JG, Gorman BK. Gender and health inequality. Annu Rev Sociol. 2010;36:371–86. 10.1146/annurev.soc.012809.102535. [Google Scholar]
- 11.Treiman DJ. Trends in educational attainment in China. Chin Sociol Rev. 2013;45:3–25. 10.2753/CSA2162-0555450301. [Google Scholar]
- 12.UNESCO Institute for Statistics. Global Education Digest 2010: Comparing Education Statistics Across the World. Montreal: UNESCO Institute for Statistics; 2010. [Google Scholar]
- 13.Wu X, Zhang Z. Changes in educational inequality in China, 1990–2005: Evidence from the population census data. In: Hannum E, Park H, Goto Butler Y, editors. Res. Sociol. Educ. Emerald Group Publishing Limited; 2010;17:23–52. 10.1108/S1479-3539(2010)0000017007
- 14.Liu S, Li Y, Zeng X, Wang H, Yin P, Wang L, et al. Burden of cardiovascular diseases in China, 1990–2016: findings from the 2016 global burden of disease study. JAMA Cardiol. 2019;4:342–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wu J, Jiao B, Zhao J. Gender disparities in blood pressure and the role of body mass index: a birth cohort analysis in China. J Epidemiol Glob Health. 2023;13:485–94. 10.1007/s44197-023-00127-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Loucks EB, Abrahamowicz M, Xiao Y, Lynch JW. Associations of education with 30 year life course blood pressure trajectories: Framingham offspring study. BMC Public Health. 2011;11:139. 10.1186/1471-2458-11-139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ma H, Liu F, Li J, Chen J, Cao J, Chen S, et al. Sex differences in associations between socioeconomic status and incident hypertension among Chinese adults. Hypertension. 2023;80:783–91. 10.1161/HYPERTENSIONAHA.122.20061. [DOI] [PubMed] [Google Scholar]
- 18.Moreno Velásquez I, Peters SAE, Dragano N, Greiser KH, Dörr M, Fischer B, et al. Sex differences in the relationship of socioeconomic position with cardiovascular disease, cardiovascular risk factors, and estimated cardiovascular disease risk: results of the german national cohort. J Am Heart Assoc. 2025;14:e038708. 10.1161/JAHA.124.038708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pacca L, Irish AM, dP Duarte C, Riley AR, Pletcher MJ, Bailey ZD, et al. Differential associations between education and blood pressure by gender and race. BMC Public Health. 2025;25:2278. 10.1186/s12889-025-23409-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pinho-Gomes A-C, Peters SA, Woodward M. Gender equality related to gender differences in life expectancy across the globe gender equality and life expectancy. PLoS Glob Public Health. 2023;3:e0001214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Platt JM, Bates LM, Jager J, McLaughlin KA, Keyes KM. Changes in the depression gender gap from 1992 to 2014: cohort effects and mediation by gendered social position. Soc Sci Med. 2020;258:113088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Volken T, Wieber F, Rüesch P, Huber M, Crawford RJ. Temporal change to self-rated health in the Swiss population from 1997 to 2012: the roles of age, gender, and education. Public Health. 2017;150:152–65. 10.1016/j.puhe.2017.07.007. [DOI] [PubMed] [Google Scholar]
- 23.Wu J, Yang M, Ma J, Yu S, Li R. Education narrows gender disparities in cognitive function in china: evidence from a cross-cohort comparison study. J Aging Health 2025. 10.1177/08982643251371012 [DOI] [PubMed]
- 24.Zhang B, Zhai FY, Du SF, Popkin BM. The China health and nutrition survey, 1989–2011. Obes Rev Off J Int Assoc Study Obes. 2014;15(Suppl 1):2–7. 10.1111/obr.12119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wu J, Muennig PA, Keyes K, Wu J. Generational differences in longitudinal blood pressure trajectories by geographic region during socioeconomic transitions in China. Int J Public Health. 2019;64:1375–87. 10.1007/s00038-019-01276-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Chen F, Yang Y, Liu G. Social change and socioeconomic disparities in health over the life course in China: a cohort analysis. Am Sociol Rev. 2010;75:126–50. 10.1177/0003122409359165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Baker DP, Smith WC, Muñoz IG, Jeon H, Fu T, Leon J, et al. The population education transition curve: education gradients across population exposure to new health risks. Demography. 2017;54:1873–95. 10.1007/s13524-017-0608-9. [DOI] [PubMed] [Google Scholar]
- 28.Preacher KJ, Rucker DD, Hayes AF. Addressing moderated mediation hypotheses: theory, methods, and prescriptions. Multivar Behav Res. 2007;42:185–227. 10.1080/00273170701341316. [DOI] [PubMed] [Google Scholar]
- 29.Zyphur MJ, Zhang Z, Preacher KJ, Bird LJ. Moderated mediation in multilevel structural equation models: Decomposing effects of race on math achievement within versus between high schools in the United States. The Handbook of Multilevel Theory, Measurement, Analysis, American Psychological Association. 2018:473–94. https://quantpsy.org/pubs/zyphur_zhang_preacher_bird_2019.pdf
- 30.Fang J, Wen Z. The analyses of multilevel moderated mediation model. J Psychol Sci. 2023;46:221–9. [Google Scholar]
- 31.Byrne BM. Structural equation modeling with Mplus: Basic concepts, applications, and programming. routledge; 2013. [Google Scholar]
- 32.Enders C, Bandalos D. The relative performance of full information maximum likelihood estimation for missing data in structural equation models. Struct Equ Model Multidiscip J. 2001;8:430–57. 10.1207/S15328007SEM0803_5. [Google Scholar]
- 33.Gupta GR, Oomman N, Grown C, Conn K, Hawkes S, Shawar YR, et al. Gender equality and gender norms: framing the opportunities for health. Lancet. 2019;393:2550–62. [DOI] [PubMed] [Google Scholar]
- 34.Sato K. Unhappy and happy obesity: a comparative study on the United States and China. J Happiness Stud. 2021;22:1259–85. 10.1007/s10902-020-00272-2. [Google Scholar]
- 35.Harper S, Lynch J, Smith GD. Social determinants and the decline of cardiovascular diseases: understanding the links. Annu Rev Public Health. 2011;32:39–69. 10.1146/annurev-publhealth-031210-101234. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data from the China Health and Nutrition Survey is publicly available and can be accessed at: https://www.cpc.unc.edu/projects/china.