J. Wang1, Q. Wei1, X. Wan2
1. School of Management, Wuhan Institute of Technology, Wuhan, China; 2. Institute of Income Distribution and Public Finance, School of Taxation and Public Finance, Zhongnan University of Economics and Law, Wuhan, China
Corresponding Authors: Xin Wan, Associate Professor, Institute of Income Distribution and Public Finance, School of Public Finance and Taxation, Zhongnan University of Economics and Law, Wuhan, China, 430073, Email: firstname.lastname@example.org
J Prev Alz Dis 2020;
Published online November 28, 2020, http://dx.doi.org/10.14283/jpad.2020.67
Objective: This study selects the health indicators of older adults to analyze the impact of tea drinking on health. Design: This is a panel data.
Setting: This study uses data from China Health and Nutrition Survey (CHNS), which covers nine provinces and ten waves, between 1997 and 2015.
Participants: a total of 706 old adults are consistently surveyed in six surveys on issues such as health and nutrition.
Measurements: Health of old adults is assessed by self-reported health (SRH), tea drinking is 0-1 dummy variable, and also analyze with the frequency of tea drinking. This study uses ordered probit model to analyze the influence of tea drinking on SRH.
Results: Findings reveal a significant negative correlation between tea drinking and SRH of older adults. It is shows that the significant positive correlation exists between the tea drinking frequency and SRH, but the quadratic term of tea frequency shows the significant negative correlation. It means drinking tea benefits older adults in terms of improved health, but excessive consumption of tea is not healthy for them. The heterogeneity analyses reveal that there are no significant geographic, tea-drinking pattern or gender differences in the conclusion that tea drinking is good for older adults’ health.
Conclusion: In this study, we find correlation between tea drinking and SRH of older adults, and tea drinking is beneficial toward the improvement of SRH, but drinking tea in excess is not good for older adults’ health.
Key words: Tea drinking, self-reported health, frequency of tea drinking, heterogeneity, older adults.
The proposal of “international tea day” adopted at the 41st session of Food and Agriculture Organization of the United Nations in 2019 has a far-reaching and significant impact on the development of the global tea industry and the promotion of tea culture. In China, with the improvement of residents’ living standards, tea has been widely consumed as a daily social and common beverage. The healthy function of tea has been gradually recognized by the academic circles under the multidisciplinary cross and penetration (1). However, a detailed review of the relevant literature at home and abroad shows no consensus on the health effects of the elders of tea drinking despite the wide range of individual experimental data. The effect of tea drinking on individual health varies when different countries, certain diseases, and other research methods are involved.
Following the year 2006, China became the country with the largest total tea consumption, and the problem of ageing is also becoming more serious and the health of older adults is of great concern. So the effect of tea drinking on health of older adults in China is a meaningful and valuable topic, which complements the current research and analysis on the role of tea drinking. This study selects the health indicators of older adults to analyze the impact of tea drinking on health. The innovations of this study are as follows: First, the research samples used are representative, rich, and effective. The large-scale epidemiological investigation on the influence of tea drinking in China has not been executed, and the conclusions of the existing research on tea drinking in a small range are inconclusive. Data in this study are derived from the CHNS. The data covered cities and towns across all cardinal points of China from 1989 to 2015. CHNS database includes demographic characteristics, economy, public resources, and health indicators, which meet the needs of this research. Second, individual SRH is used as the dependent variable to study the effect of tea drinking. SRH is used as an effective and popular measure of individual health, although other undiscovered or proven effects of tea drinking can also be used. Third, the analytical method is scientific. In this study, the ordered probit model is used for analysis. Considering the possible endogenous problems between tea drinking and SRH, the robustness test is conducted by employing methods, such as lag period. This study also analyzes the heterogeneity of tea drinking effects for different gender groups to enrich the research content.
Data and model
Because the dependent variable involves four kinds of discrete values, the ordered probit model should be adopted, which is widely used to deal with multiple kinds of discrete data. The empirical model is set as follows:
healthi,t=α+βteai,t +θXi,t +γϕt +ηδi+μi,t, (1)
wherehealthi,t denotes the SRH for individual i surveyed in wave t, which is measured by ordered qualitative labels, such as “very good” (the value is 1), “good” (the value is 2), “fair” (the value is 3), and “poor” (the value is 4). SRH is a commonly used health measurement index, which reflects the comprehensive condition of personal subjective and objective health and can provide supportive information for the decision makers of disease prevention to an extent (2). SRH measure is inexpensive, can easily collect data, and is synthetic by construction. teai,t is the key variable, which is a dummy indicating the tea drinking of individual i in wave t. This study also examines the effects of the frequency of tea drinking on health. tea frequencyi,t denotes the tea drinking frequency of individual i in wave t, which is measured by ordered qualitative labels, such as “almost every day” and “four to five times a week.” The values are 1−7 in order.
Xi,t represents a series of control variables, including personal basic information, chronic disease, individual habits, dietary and activity preferences, healthy exercise and nutritional diet cognition. Personal basic information is the control variable for health economics (2, 3), including age, gender, height, weight, marital status, education, medical insurance, household registration type, income and household sanitation. Medical insurance and household registration are the dummy indicating. Income is measured by personal annual income deflator. The variable of household sanitation is measured by two dummy variables of having running water and flush toilet. Chronic diseases are important factors affecting the elders’ health and harm the brain, heart, kidney, and other important organs. Having hypertension, diabetes, myocardial infarction, stroke, and fracture are five dummy variables indicating chronic diseases. Individual habits are measured by two dummy variables of smoking and drinking. Diet and activity preferences also have indirect effects on health, which are measured by six dummy variables of preferences. The cognitions of healthy exercise and diet indirectly affect health, which are dummy variables. φt is wave fixed effects, δi is individual fixed effects, and μi,t is the error term.
The data used in this study were obtained from the CHNS, which is an ongoing international collaborative project among the Carolina Population Center at the University of North Carolina at Chapel Hill, the National Institute of Nutrition and Food Safety, and the Chinese Center for Disease Control and Prevention. The survey covered nine provinces, including Liaoning, Heilongjiang, Shandong, Jiangsu, Henan, Hubei, Hunan, Guangxi and Guizhou, including the waves 1989, 1991, 1993, 1995, 1997, 2000, 2004, 2006, 2009, and 2015. When the survey was conducted, a multistage, random cluster process was used to draw the sample surveyed in each province. Approximately 4,400 households were listed in the overall tracking survey, covering 19,000 individuals. Rich personal and family microdata of older adults are 65 years of age and older provided rich and detailed data for the research. The study constructed panel data and omitted data prior to years 1997 and the year of 2015 because of data of tea drinking missing. Tea drinking was derived from the question “Do you normally drink tea?” SRH was obtained from the question “Right now, how would you describe your health compared to that of other people your age?” This study excluded individuals who reported “unknown” from the sample. The results of variable statistical characteristics are presented in Table 1.
Results and related tests
Basic regression results
Column (1) and (2) in Table 2 shows the regression results of the ordered probit model with wave fixed effect, individual fixed effect, and different control variables that are successfully added. No matter how the control variables change, the significance and sign of the estimated coefficients of dummy variables of tea drinking remain unchanged. The results indicate that the estimated coefficient of dummy variable of tea drinking is significantly negative, which means drinking tea can significantly improve SRH of older adults compared with non-tea drinkers.
Considering the possible influence of chronic diseases, individual habits, dietary and activity preferences, cognitions of healthy exercise and diet on individual health status, this study further extends the basic regression and adds the above variables into analysis respectively. The results are shown in Columns (3), (4), (5) and (6) of Table 2. All regressions in Table 2 show that regardless of how the variables are added, the estimated coefficients of dummy variables of tea drinking are significant and negative, which are correlated with the dependent variables in all regressions. Therefore, tea drinking is beneficial to SRH of older adults. The reliability of the conclusion is further verified by various robustness tests.
Note: 1. ***, **, and * indicate 1%, 5%, and 10% significance levels. 2. p-value in parentheses.
To verify the robustness of the basic conclusion, the following analysis is considered. First, change the income variable. In certain families old adults may not earn income, but they consume total family income. In many rural families, personal income cannot be separated from family income. Deaton (2003) shows that equalized family income can well measure the utilization of family resources by family members (4). Therefore, this study re-measures individual income with per capita family income deflator. The regression results are shown in Column (1) in Table 3. The beneficial effect of tea drinking on SRH of older adults remains unchanged even after changing the income variable.
Note: 1. ***, **, and * indicate 1%, 5%, and 10% significance levels. 2. p-value in parentheses.
Second, possible endogenous effects. Considering the effect of tea drinking in the early stage on health in the later stage, a problem in which explanatory variables are related to random error terms may arise. To solve this problem, explanatory variables with a lag of two periods from existing research are analyzed. The analysis results are presented in Column (2) in Table 3. After adding the control variables, the estimated coefficient of variable of tea drinking with a lag of two periods remains significantly negative.
Third, health dynamics. In view of the dynamic status of individual health, which is the past health status, may affect the current health status. Individual SRH with a lagged period is added into the model as a new control variable for regression. The results are shown in Columns (3) in Table 3. The regression results have once again confirmed that the estimated coefficient of tea drinking is significant and negative.
Fourth, change the dependent variable. Etilé and Milcent (2010) study the health samples in France and find that people with high income likely underestimate their health status because they have high health expectations (5). People with low income overestimate their health status for the lack of relevant health information and the low level of education. They think two categories of SRH indicators (“good” and “bad”) make sense. The treatment method is as follows: SRH answers of “very good” and “good” are classified into the same category, and new variable health2 is defined as “good” whose value is 1. The rest is 0. Regression results are displayed in Columns (4) in Table 3. In the regressions, SRH with one lag period is added simultaneously. Per capita household income variable is added into the regression of Column (4). The results reveal a significant positive correlation between the dummy variables of tea drinking and binary SRH variables. This finding is consistent with the basic conclusion because the value of dependent variable is 1 when SRH is good, indicating that drinking tea can help improve SRH. All the robust regression results have further verified the beneficial effects of drinking tea on SRH of older adults.
Note: 1. ***, **, and * indicate 1%, 5%, and 10% significance levels. 2. p-value in parentheses.
Effect of drinking tea frequency
Tea polyphenols have beneficial effects at a certain level in the body. Heavy tea drinking by irregular steps leads to the ups and downs of components such as polyphenols, which offer far fewer healthy benefits than continuous drinking tea. Different ways of drinking tea can directly affect the curative and prevention effect of effective substances in tea, which is also one of the important reasons for the difference in the anti-cancer effect of tea drinking between eastern and western countries. On the basis of large-scale data in China, the study demonstrates whether the sustainability of drinking tea is conducive to the improvement of personal SRH of older adults.
Columns (1) and (2) in Table 4 report the effect of tea drinking frequency on the SRH of old adults. The results show that no matter how the control variables are adjusted, the significant positive correlation exists between the tea drinking frequency and SRH. But when the variables of chronic diseases are added in the model, the quadratic term of tea frequency shows the significant negative correlation with SRH. The results show that tea needs to be consumed frequently enough to reach some cumulative amount in order to contribute to the health of older adults, however, excessive consumption of tea is not healthy for older adults. The increase in tea drinking frequency indicates the stable existence of the healthy components in the body, which is beneficial for improving human body function and immunity.
China has a vast territory, and cities in the north and south regions are divided by the Qinling Mountains-Huaihe River line. Northern regions include Liaoning, Heilongjiang, and Shandong; the rest are southern regions. The geographical location, climate characteristics, historical culture, political economy, and other aspects of the north and south are evidently different. Generally speaking, southerners are graceful and restrained. They pay attention to the process of producing tea and the utensils of making tea. Northerners are bold, unconstrained, are indifferent about the cultural attributes of drinking tea. Most northern regions do not produce tea. The most popular kinds of Chinese tea are from the South. Southerners prefer various kinds of tea, such as green, black, and white tea. Northern climate difference is large that southerners prefer black tea to warm their stomach when the weather is cold. By contrast, green tea is preferred in the summer. The bioavailability of the active components of tea and species differences in their functions may lead to different effects on people, whereas that of the active components of tea leaves may be different in the human body (1). Studies suggest that green tea is better than black tea in preventing certain diseases and improving health (6). The availability of tea polyphenols, catechins, and other components in tea drinking depends on the type of tea and how it is processed. As a result, the study analyzes the heterogeneity effect of tea drinking on SRH between north and south regions. A dummy variable, whose value of south regions are 1, and the rest are 0, is constructed. The interaction term of dummy variable and tea drinking variable is also added. The regression results suggest that interaction items are insignificant, when personal basic information, individual fixed effect and wave fixed effect are added. Therefore, no difference is observed in the effect of the different ways and types of tea drinking on SRH in the north and south of China. The most direct reason for the result is that the data do not involve the type of tea drinking.
It reveal that tea drinking may reduce iron absorption and increase the risk of iron deficiency, especially in vulnerable groups, for tea polyphenols combined with iron (7, 8). The present study discusses whether the conclusion of tea drinking is beneficial to SRH of older adults varies with gender in China. According to Model (1), the sample is divided into male and female groups for analysis. The regression results of the ordered probit model are presented in Columns (3) and (4) in Table 4. The analysis in Table 6 adopts per capita family income to measure income. The analysis results of the male and female groups show the estimated coefficients of the dummy variables whether tea drinking is significantly negative. Therefore, the conclusion of improving the health status of older adults by drinking tea is no significant gender differences. However, the coefficient values were greater in the male geriatric group than in the female geriatric group. This indicates that, excluding other controllable factors, tea consumption promotes the health of elderly men slightly better than that of elderly women.
Existing studies have neither used a wide range of data on tea drinking and older adults’ health in China to verify the correlation between them nor conducted an in-depth analysis. In this context, CHNS panel data, including demographic characteristics, economy, public resources, health, and other indicators, are selected in the present study. An ordered probit model is also utilized to analyze the relationship between tea drinking and older adults’ SRH. The results are as follows. First, a significant negative correlation exists between tea drinking and SRH, indicating that tea drinking is beneficial to the improvement of older adults’ SRH. In addition to the personal basic information, certain external factors can affect human health. Thus, this study adds dummy control variables of chronic diseases, individual smoking or drinking habits, diet preferences, activity preferences, cognitions of healthy exercise and diet, and performs regression. The results adding these dummy control variables respectively do not change the conclusion of tea drinking improving older adults’ SRH.
Second, the robustness test considers the following factors: change the control variables, eliminate the endogenetic effect, eliminate the possible health dynamic influence and the index of SRH is reprocessed. Different robustness test results also confirm that tea drinking is beneficial to older adults’ health. Third, the increase in the frequency of tea drinking is beneficial to the improvement of older adults’ SRH. Such an increase also indicates the stable existence of health components in the body, which are beneficial for improving the function and immunity of the human body. however, excessive consumption of tea is detrimental to the older adults’ health. Fourth, the conclusion of heterogeneity analysis reveals that no difference exists in the effect of different ways and types of tea drinking on SRH in the north and south of China, so as to gender.
Acknowledgement: J. Wang is supported by National Social Science Fund Youth Project of China (Project Number: 18CJY021) and X. Wan is supported by The paper is funded by the Fundamental Research Funds for the Central Universities (NO. 2722020JCG083).
Conflict of interest: The authors declared that they have no conflicts of interest to this work.
Ethical standards: The study used publicly available research data and did not violate ethical standards.
1. Chen L, Lee M J, Li H, Yang CS. Absorption, distribution, elimination of tea polyphenols in rats. Drug Metab Dispos 1997;25(9):1045-1050
2. Courtemanche C, Marton J, Ukert B, Yelowitz A, Zapata D. Early effects of the affordable care act on health care access, risky health behaviors, and self-assessed health. South Econ J 2018;84(3):660-691
3. Lockwood LM. Incidental bequests and the choice to self-insure late-life risks. Am Econ Rev 2018;108(9):2513-50
4. Deaton A. Health, inequality, and economic development. J Econ Lit 2003;41(1):113-158
5. Etilé F, Milcent C. Income-related reporting heterogeneity in self-assessed health: evidence from France. Health Econ 2010; 5(9):965-981
6. Pathy NB, Peeters P, Gils CV, et al. Coffee and tea intake and risk of breast cancer. Breast Cancer Res and Tre 2010;121(2):461-467
7. Zijp IM, Korver O, Tijburg LBM . Effect of Tea and Other Dietary Factors on Iron Absorption. Cri Rev in Food Sci and Nut 2000;40(5):371-398
8. Li W, Yang J, Zhu XS, Li SC, Ho PC. Correlation between tea consumption and prevalence of hypertension among Singaporean Chinese residents aged 40 years. J Hum Hypertens 2016;30: 11-17