Socioeconomic status, lifestyle, and hypertension prevalence of the older adults in China

Education

4 groups (1=Never attended school; 2=Primary school; 3=Junior middle school; 4=High school or above).

PCHI groups

Three groups divided by the tertile cut-off points, with income adjusted against the 2011 Consumer Price Index.

CUI groups

Three groups by tertile cut-off points of CUI, a composite urbanicity scale of 12 aspects, each scoring between 0 and 10.¹⁴

Smoking status

yes or no, reported by subjects when the answer to the question of “whether there is a history of smoking.”

Drinking status

yes or no, reported by subjects when the answer to the question of “whether you drank last year.”

TEI

Three groups by tertile cut-off points. The average total energy intake per day was calculated against the food ingredient table. The diet data was obtained by 24 hours recall method.

FIT groups

Three groups by tertile cut-off points. The proportion of fat in total energy intake= average calorie from fat per day/average total energy intake per day in the past three days (kcal). One gram of fat is calculated as 9 kcal of calories according to the method proposed by Xue.¹⁵

PA groups

Three groups by tertile cut-off points. We used the “Compendium of Physical Activities 2011”¹⁶ to calculate weekly PA in metabolic equivalent values of the self-reported PAs (sum of “MET/hour of each activity” * “hours lasted for the activities”).

Hypertension diagnosis

Subjects’ right arm blood pressure was measured three times in a sitting position after a 5-minute rest. The mean value of three measurements of the diastolic (and systolic) pressure was used as the diastolic (and systolic) pressure. Systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg and/or taking antihypertensive drugs reported by the subjects is considered as having hypertension, according to the Chinese experts’ consensus for hypertension diagnosis and treatment in the older adults.³

The demographic control variables were age, ethnicity, and family size. For the selected characteristics of the subjects, see Table 1.

The data in this study is nested and the between-group correlation coefficient values (men=0.74, women=0.75) and the design effect values (1108.78, 1250.5) indicated that a multi-level model is fit for use.^17-19

We examined the relationship between SES and hypertension, after adjusting for the demographic and lifestyle factors, in multi-level logistic regression models (generalized linear models). One model for men and one model for women, respectively.

SPSS 23.0 was used for statistical analysis.

In the three rounds of the survey, 4,496 men (on average, 1,499 surveyed per round) and 5,000 women were surveyed (on average, 1,667 per round).

There is little gender difference in PCHI and CUI as subjects in the same households, regardless of gender, have the same PCHI, and subjects who reside in the same community have the same CUI value.

There is a vast difference between genders, and little difference between survey years in terms of education level, age, and family size. More women survived men and lived alone.

Women’s proportion of energy from fat is higher than that of men. More men had a history of smoking and reported drinking last year. Men’s daily cigarette consumption and drinking frequency were higher than women. Men’s smoking rate declined in the three rounds of the survey but was consistently higher than women (Table 1).

From 2006 to 2009, the prevalence of hypertension, the past diagnosis rate, and the rate of treatment in men and women increased (Table 1).

The accurate prediction value of the model increased with the addition of individual, family, community, and lifestyle variables until we obtained a satisfactory model for men and women respectively (Table 2). The model fit was judged based on the -2 log pseudo-fitness data. The smaller the value, the higher the model fit. The same applies for both modeling for men and women.

After adjusting for the demographic and lifestyle factors, men with low income and middle income (compared with high income, OR= 1.90(1.36∼2.67), p < 0.001 and 1.39(1.02∼1.90), p<0.05) and with low and middle CUI (compared with high CUI, OR= 1.70(1.20∼2.42), p < 0.001, and 1.48(1.07∼2.04), p < 0.05) are more likely to have the disease (Table 2).

For women, we found the association between income (PCHI) and hypertension (OR= 2.07***(1.28∼3.11), p < 0.001 for low income, and 1.96***(1.33∼2.90), p < for middle income) and the association between education level. Women with a lower education (primary schools) level are less likely to have hypertension (OR= 0.49(0.26∼0.89), p < 0.05). But it is worth noting that women who never attended school, very probably illiterate, are almost equal probability to have hypertension as those who had high school or above education (OR=0.94(0.50∼1.79, p<0.05) (Table 2).

We found that SES factors—income, urbanicity, and education, are related to hypertension, and that gender differences exist in the relation between SES factors and hypertension in the older adults in China.

Our study found that higher per capita household income functions as a protective factor of hypertension in the older adults. Similarly, Kaplan et al.²⁰ had reported a significant inverse linear relationship between household income and the hypertension prevalence rate in the United States, but no evidence of such a relationship in Canada, and argued that social disparities and barriers to health care access and primary prevention in the United States may have played a role in it. We agree with Kaplan et al.’s explanation.²⁰ As China had adopted a market-oriented healthcare system since its reform and opening-up at the end of the 1970s, income has become a much more important factor in obtaining living and healthcare resources. Many resources were obtained via a market mechanism, instead of rationing. Market-oriented policies stimulate efficiency in its economy, but at the cost of equity, as demonstrated here in health/disease disparity. To realize health equity, probably the strikingly severe income inequity as demonstrated in this study should be tackled first.

Our study confirms that people living in communities with higher urbanicity tend to have lower hypertension prevalence. Similar findings had been reported by previous studies. Buys et al.²¹ reported that the older adults with poorer community conditions have a higher prevalence of hypertension. Isaac and Laveist²² had reported that older people with convenience stores and recreational facilities have lower rates of hypertension, even if they live in communities of lower and medium SES. One possible mechanism is that with the decline of the mobility of older people, restrictions through neighborhood conditions became more prominent, rendering the older adults less capable of avoiding adverse socioeconomic conditions.²¹ Another possible mechanism might be that better income and living conditions facilitate the older adults to access quality preventive and curative health care and reduce anxiety and stress.²³

Therefore, we argue that geriatric health management, especially the management of chronic diseases, depends on an array of conditions, including transportation, community health care services, built community environment. Future research could test such a hypothesis.

We found that women with a lower education (primary schools) level are less likely to have hypertension; women who never attended school are almost as likely to have hypertension as women who had high school or above education.

The findings of previous research on education level and hypertension in the older adults in China are inconsistent, possibly influenced by the age groups, locality, and ethnic groups studied. Lei et al.²⁴ reported no gradient between the level of education and the prevalence of hypertension with a nationally representative sample data of the middle-aged and older adult population. Zuo et al.²⁵ reported no gradient between education level and hypertension among the older adults of 12 counties in Hubei. However, in Xinjiang, Zhang et al.²⁶ found a reversed correlation between education and hypertension among the older adults. Xinjiang is a northwestern region of China, where the proportion of Han Chinese is lower than that in other regions of China.

As for the possible mechanism underlying our finding, we speculate that older people with lower education have lower health literacy, which is essential in the management of their chronic diseases,²⁷ and those with higher education tend to have lower physical activity level (which we had found in our study, would be published soon in another article), which could be a predictive factor of hypertension in the older adults in China. You et al.²⁸ found that middle-aged and older adults in China who usually participated in the moderate-to-vigorous activity for more than 10 minutes were less likely to have hypertension and Miao et al.²⁹ found similar results for the older adults in Beijing. More research is needed to verify the relationship between education and hypertension in the older adults.

The present study contributes to the literature by providing gender-specific results between SES and hypertension. It has been reported that the older adults with low SES in such countries as China^7,8,30,31, Germany³², UK³³, South Korea³⁴, and the United States^20,35, have worse health. However, previous studies ignore the gender difference in such a connection between SES and health. This study finds that the relationship between SES and hypertension in older men and women differ, which indicates that when considering the relation between social stratification and health of the older adults, gender should not be ignored.

We did not control the genetic variants, trace elements, diet structure, sleep, limitation of daily activities, cognitive abilities (which may affect their medication compliance and lifestyle), healthcare (such as medical security availability and access to medical services), SES of early life, pre-retirement occupation, and comorbidity in the study. The composite PA index that we use may obscure the effect difference of PA of high, medium, and low intensity. Future research could also consider dose, starting age and frequency of smoking and drinking and exposure to secondhand smoke, compare the accuracy of accelerometer-measured and self-recalled PA, and, if possible, use Mendelian randomization to examine the causality between SES and hypertension and the effect of lifestyle in the relationship.

Socioeconomic factors significantly correlate with hypertension in older men and women in China and that there are gender differences.

Declaration

7. Availability of data and materials

The data used by this study is accessible at https://www.cpc.unc.edu/projects/china/data.