Ⅰ. Introduction

Globally, a large body of evidence has established that health status is highly associated with behavioral risk factors that can lead to chronic diseases and mortality (Frasure-Smith et al., 1993) as well as posing a public health burden. Four main behavioral factors linked to chronic diseases are tobacco use, unhealthy diet, insufficient physical activity, and harmful alcohol use (Alwan, 2011), and understanding the magnitude of these risks to health is a key to preventing chronic diseases and mortality. Identifying risk factors is also important for health policy planning and evaluation, particularly for disease prevention and promotion (World Health Organization, 2009).

Health status can be measured with self-rated health (SRH) scale, which has been recognized as a simple and relatively inexpensive general health measurement tool that is easy to administer a survey (Bombak, 2013). Furthermore, SRH is a widely used health indicator in survey research on general health assessments in countries around the world (Bardage et al., 2005; Kaleta, Makowiec-Dabrowska, & Jegier, 2008; Zarini et al., 2014); it was found to be a valid measure of health status and mortality prediction in populations without cardiovascular diseases and functional disability (Bardage et al., 2005; DeSalvo, Bloser, Reynolds, He, & Muntner, 2006; Idler & Benyamini, 1997; Wu et al., 2013).

However, the relationship between SRH and behavioral risks appeared to be a complex association (Bombak, 2013); a number of studies have shown that behavioral risk factors are associated with SRH (Bodde, Seo, & Frey, 2009; Fylkesnes & Førde, 1991; Schulz et al., 1994; Zarini et al., 2014), but there are conflicting findings concerning the relationship between the two variables (Bombak, 2013; Manderbacka, Lundberg, & Martikainen, 1999). Some have argued that SRH varies by socioeconomic status (SES) (Delpierre, Lauwers-Cances, Datta, Berkman, & Lang, 2009; Kaleta, Makowiec-Dąbrowska, & Jegier, 2008).

Consequently, it remains unclear from the literature whether behavioral risk factors are associated with how people assess their health. In addition, SRH can have different implications for diverse social and cultural backgrounds (Bardage et al., 2005; Jerković, Sauliūnė, Šumskas, Birt, & Kersnik, 2015; Prus, 2011; Zimmer, Natividad, Hui-Sheng, & Chayovan, 2000). Bardage et al. (2005) examined SRH among older adults in Finland, Sweden, the Netherlands, Spain, and Israel; Jerković et al. (2015) compared SRH among elderly populations in Slovenia, Lithuania, and the United Kingdom; Prus (2011) compared SRH between general population in the United States and Canada; and Zimmer, Natividad, Hui-Sheng, and Chayovan (2000) compared SRH among older adults in the Philippines, Taiwan, and Thailand.

Although SRH has been widely used and compared across European countries, research on the subject has been scarce in Asian countries (Page & Suwanteerangkul, 2009; Zimmer et al., 2000). South Korea (Korea), a high-income country, and Thailand, a middle-income country, are geographically located in Eastern Asia. They share the three behavioral risk factors that account for the highest public health burdens in both countries: diet, alcohol consumption, and smoking (Institute for Health Metrics and Evaluation, 2010), all of which contribute to chronic disease and mortality. However, the two countries do also differ from each other in meaningful ways, including economic development levels, health systems, social welfare programs, and income inequality, which are likely to yield different implications for health status in the two countries. To date, no study has compared SRH and the associated behavioral risk factors between these two countries.

This study, thus, aimed to examine the associations of diet, alcohol consumption, smoking, and physical activity with self-rated health across Korean and Thai adult populations.

Ⅱ. Literature Review

1. Country profiles

Korea and Thailand provided universal health coverage with public health insurance in 1989 and 2002, respectively, and now all citizens of both countries have the right to health services (Kwon, 2009; Evans et al., 2012). However, the countries differ in their sources of financing, health service delivery systems, and benefits packages. Korea administers a single-payer scheme, the National Health Insurance, which mainly relies on social health insurance, and medical aid is provided to the poor through general taxation (Kwon, 2009). In contrast, Thailand adopts multi-payer schemes: the Social Security Scheme (SSS) for private-sector employees, excluding dependents, the Civil Servant Medical Benefit Scheme (CSMBS) for government employees and their dependents, and the Universal Coverage Scheme (UCS) for the remaining population, the majority of which are self-employed or informal sector workers. These three schemes cover approximately 16%, 9%, and 75% of the Thai population, respectively (Evans et al., 2012). The schemes have differing financing sources. The SSS is financed by social health insurance, whereas the CSMBS and the UCS are financed by general taxation; in short, Thai public health insurance mainly relies on taxation. The health service delivery system in Korea relies mainly on private providers, whereas the major provider in Thailand is the public sector. In addition, Korea requires co-payment at the point of service, but Thailand does not. Although both countries provide comprehensive benefits, Thailand’s benefits packages vary: the benefits under the CSMBS are slightly greater than those from the SSS or UCS (Evans et al., 2012).

The overall demographic structures of the two countries are similar. In 2010, life expectancy at birth was 79.7 years in Korea and 74.1 in Thailand, and age-standardized year life with disability rate (per 100,000) were 9,575 and 10,369, respectively (Institute for Health Metrics and Evaluation, 2010). However, in 2015, Korea’s mortality rate for children under 5 years per 1,000 live births was 3%, whereas the rate was 12% in Thailand. In addition, in 2014, survival to age 65 for females in Korea and Thailand were 94% and 84% and for males, 86% and 72%, both respectively (http://worldbank.org).

The Gini coefficient reflects inequality of income distribution in a country. In 2013, the Gini coefficients for Korea and Thailand were 0.302 (Organisation for Economic and Co-operative Development, 2016) and 0.379 (http://worldbank.org), respectively, indicating that Korea had less inequality than did Thailand.

As with other high- and middle-income countries, the most important risk factors in these two countries are those associated with chronic diseases (World Health Organization, 2009); they shared 3 of the top 15 behavioral risk factors that account for their disease burdens as measured in disability-adjusted life years (DALYs), including dietary risks, alcohol consumption, and smoking. Total alcohol consumption per capita in 2015 (liters of pure alcohol, projected estimates for more than 15 years) was similar in both countries (Korea: 10.9%; Thailand: 8.3%), and smoking prevalence in 2012 was also similar (Korea: males, 51.7%, females, 4.4%; Thailand: males, 42.3%, females, 2.4%) (http://worldbank.org). These risk factors are the leading causes of chronic diseases such as cardiovascular and circulatory diseases, cancer, diabetes, and injuries (see Figure 1; Institute for Health Metrics and Evaluation, 2010).

새창으로 보기

Figure 1

Percentages of disability-adjusted life years lost due to 15 leading risk factors in Korea and Thailand, 2010

Source: GBD (2010).

Ⅲ. Methods

The data came from the Korea National Health and Nutritional Examination Survey (KNHANES) and the Thai National Health Examination Survey (NHES). KNHANES has been conducted by the Center for Disease Control and Prevention (CDC) from 1998 to the present, aiming for examining the general health, nutrition status, health behavior, the status of chronic diseases, and so on, while NHES has been conducted by the National Health Examination Survey Office, Health Systems Research Institute from 1991 to the present, aiming for examining the general health, disease prevalence, and health risk factors. Additionally, the questionnaires that have been adopted in KNHANES and NHES, particular on behavior risk factors, are relatively similar.

The 2009 NHES, the latest version available, is used here to compare the cases of Thailand and Korea due to the NHES is only the Thai National Health Survey that includes behavioral risk factors of Thai adult population. In this sense, it would be more meaningful to compare behavioral risk factors between the two national surveys that had been conducted in the same year. Noted that a strict approval procedure is required for the data access in Thailand.

According to the 2009 KNHANES and 2009 NHES, the response rates were 82.8% and 93%, respectively. The face-to-face survey interviews were conducted in both countries based on sample representatives of the Korean and Thai populations. However, this study included only data concerning those aged between 18 and 59 years, a total of 5,649 and 7,753 respondents, respectively.

For this study, we used the same questions for the data analysis that were used in the surveys in both countries. Health status was defined based on SRH, specifically, the single question, “How do you rate your general state of health? Excellence, good, fair, poor, or very poor.” We analyzed self-rated health as a dichotomous measure, coding both “Excellence” and “good” as “good health,” and “fair”, “poor,” and “very poor” as “poor health.” Diet was assessed based on fruit and vegetable intake per day (grams) and coded smoking as (1) never smokers, (2) former smokers, and (3) current smokers. Alcohol consumption was measured as the amount of drinking per year and was coded into (1) high risk and (2) low risk. Physical activity was measured based on duration and frequency of moderate exercise, more than 30 minutes of exercise more than 5 days per week, was coded into yes and no.

Confounding variables included gender, age, education, family income, employment status, residence area, national health system, BMI, and depression. Education was classifed into five levels: (1) no education, (2) elementary school, (3) middle school, (4) high school, and (5) vocational and university. Family income was classified into five quintiles and employment status was classified into employed and unemployed, and residence area was classified into rural and urban areas. BMI was calculated as weight/height2. Depression was measured by using a yes/no question: “Did you have feelings (sad, do not interest in anything, or loss energy) more than two weeks?” The national health insurance system was categorized based on the system in each country. For Korea, the categories were (1) employee health insurance, (2) self-employed health insurance, and (3) medical aid beneficiaries and for Thailand, they were (1) Social Security Scheme (SSS), (2) Universal Coverage Scheme (UCS), and (3) Civil Servant Medical Benefit Scheme (CSMBS).

Multiple logistic regressions were conducted to analyze the factors associated with SRH, analyzing the data separately by country, controlling for gender, age, education, family income, employment status, residence area, BMI, depression, and national health insurance system. The results are presented as odds ratios (ORs) and confidence intervals. For all statistical analyses, we used STATA software, version 13, all tests were two-sided, and considered p values less than 0.05 statistically significant.

Ⅳ. Results

Table 1 shows the results of comparing the general characteristics of the Korean and Thai samples; the two countries had similar proportions of gender and age among their respondents. More than half of the Thai sample had an elementary school education, whereas one-third of the Korean sample had completed high school. The mean family income in Korea was approximately seven times higher than that in Thailand, and compared with Korea, Thailand had a greater proportion of employed respondents. In both samples, more people lived in urban areas, but Korea had a greater proportion of people in its urban areas than did Thailand. The mean BMIs were similar in the two countries, and the proportions of respondents with depression were low in both samples. Most respondents were covered under the health insurance for self-employed or informal sector workers. A higher proportion of Thai respondents reported good health than Korean respondents, but the Korean sample consumed greater portions of fruits and vegetables per day. Both samples had low-risk alcohol consumption and were non-smokers, but both countries also had large proportions of respondents who did not exercise much.

Table 2 shows the odds ratios (OR) for each of the status characteristics and behavioral risk factors in relation to SRH, within each country, when controlling for gender, age, education, family income, employment status, residence area, BMI, depression, and national health insurance system. Controlling for other variables in the model, only smoking and physical activity showed a statistically significant association with SRH in both countries. Fruit and vegetable intake had a significant association with SRH in Korea, but not in Thailand.

Korean adults who consumed few servings of fruits and vegetables per day were more likely to report poor SRH, which similar to previous study (Choe, Kwon, & Paik, 2006). In contrast, fruit and vegetable intake in Thailand did not show a significant association. According to Satheannoppakao et al. (2009), 36.5% and 68.0% of Thai adults consumed fruits and vegetables on a daily basis; compared with adults in Korea (fruits: 65.5%, vegetables: 99.1%; http://stats.oecd.org, 2014), Thai adults consumed considerably fewer fruits and vegetables, a possible reason for the insignificant association with SRH in Thai adults.

Current smokers were more likely to report poor SRH than were non-smokers in both countries. Similar to the findings from previous studies in both countries, Choi and Kim (2008) reported a relationship between smoking and frequency of health and medical facility visits in Korea. Similar research also showed that average number of visits to medical facilities had a negative relationship with average amount of smoking per year, and in Thailand, Zimmer, Natividad, Hui-Sheng, and Chayovan (2000) found that current smokers were more likely to report poor SRH than were non-smokers.

Both Korean and Thai adults who did not perform more than 30 minutes of exercise more than five days per week were more likely to report poor SRH than those who performed, and physical activity had the highest ORs in both countries (Korea: OR=1.46 and Thailand: OR=1.21). Therefore, physical activity has a powerful impact on SRH. These results are similar to findings from previous studies in both Korea (Choi, 2014; Han et al., 2009; Kim, 2012) and Thailand (Kaewpan & Kalampakorn, 2012; Mota et al., 2012; Page & Suwanteerangkul, 2009).

Gender, education, family income, and depression effects were statistically significant in both countries; female, less education, lower family incomes, and depressed respondents were more likely to report poor SRH. Although higher age, being unemployed, and living in urban areas were significantly associated with poor SRH in Thailand, they were not in Korea. Separately, Thai adults on the SSS were more likely to report poor SRH than were those on the CSBMS.

From the results, it can be seen that the control variables affected SRH differently in Korea and Thailand. While age, employment status, residence area, and national health insurance system had significant effects on SRH in Thailand, in Korea, they did not. The findings showed that SRH in Thailand varied by demographic characteristics, SES, residence area, and health insurance scheme, and they were similar to findings from previous studies in Thailand that demographic and SES variables were significantly associated with SRH (Seubsman, Kelly, Yiengprugsawan, Sleigh, & the Thai Cohort Study, 2011; Tangcharoensathien et al., 2006; Yiengprugsawan, Lim, Carmichael, Sidorenko, & Sleigh, 2007; Zimmer & Amornsirisomboon, 2001). However, this study’s findings differed from other Thai studies with regard to residence area and health insurance scheme (Tangcharoensathien et al., 2006; Yiengprugsawan, Lim, Carmichael, Sidorenko, & Sleigh, 2007): we found that respondents who lived in urban areas were more likely to report poor SRH than were those in rural areas. One explanation for this is that rural residents are less-exposed to pollution than are those who live in urban areas. From the health insurance system perspective, residents of rural areas have better access to primary health care from health care centers, which heavily emphasize on disease prevention and health promotion, than did urban residents, and this is because primary care is a gatekeeper in the district health system in Thailand. In contrast to rural areas, urban Thai residents can choose either hospitals or private clinics - which mainly focus on curative care - as their primary contacts. Thus, they might have less access to health education, which in turn might lead to poor SRH. Because of the differences in the benefits across schemes, CSMBS members receive greater benefits than SSS and UCS members (Evans et al., 2012), which could contribute to poorer SRH among SSS members, compared to the CSMBS, and SRH did not significantly differ between SSS and UCS members.

In contrast to Thailand, Korea was characterized by relationships between demographic and SES variables such as gender and education and SRH. Thailand also showed significant associations between these variables, but Korea had higher ORs. As reported in previous studies, the relationship between SES and SRH is an important issue in Korea (Kim, 2007). According to Kim (2007), SES factors such as gender and education not only affect SRH but also suggest social inequality. Therefore, persons of lower SES need to be taken under consideration when designing and implementing social interventions in Korea (Kim, 2007). As found in Kim and other studies in Korea (Kim et al., 2010; Choi, 2014; Shin, Shin, & Rhee, 2012), our analyses showed that females and those with less education tended to report poor SRH.

Another control variable that affected SRH in this study in both countries was depression, and this result is similar to findings from previous studies in Korea that those with poor SRH were more likely to be depressed (Lee, 2005; Mui, 2000). Other studies in Thailand also found that those with poor psychosocial scores such as stress, loneliness, hopelessness, and unhappiness tended to report poor SRH (Haseen, Adhikari, & Soonthorndhada, 2010; Page & Suwanteerangkul, 2009).

To attain meaningful findings, this study compared Thailand’s 2009 data to Korea’s data of the same year despite the fact that Korea National Health and Nutritional Examination Survey data is currently available until 2015. It is also important to look at the latest health risks in Korea. Therefore, the analysis results of Korea in 2009 and 2015 are also presented as follows.

Comparison between the data of 2009 and 2015 versions yielded the main similar findings; fruit and vegetable intake, current smoking and exercise were significantly associated to SRH. Noted that a physical activity variable was defined differently between 2009 and 2015; it was measured as appropriate exercise in 2009, while in 2015, it was measured as aerobic physical activity practice. Among the control variables, age and BMI were significantly associated to SRH, which were different from the 2009 results.

Overall, the results of data analysis did not show a significant difference despite the 6-year gap, which means that the health risk behaviors of Koreans have not changed much, and the policy and intervention for health behaviors are still needed.

Ⅴ. Discussion and Conclusion

Cross-country comparisons provide a unique opportunity to examine how differences in different countries’ social and cultural contexts shape the relationships between behavioral risk factors and health status. Using multiple logistic regression to predict self-reported good health, we found similar patterns of behavioral risk factors in the two countries, including physical activity and smoking, were significantly associated with SRH; in addition, the sizes of their impacts tended to be similar between the two. Unlike in Thailand, Korean adults who consumed few fruits and vegetables were more likely to report poor SRH.

However, it can be seen that social determinants of health include micro-level social factors such as SES and other demographic characteristics (gender, education, family income, depression as a psychosocial factor) that were related to SRH in both countries. Interestingly, age, employment status, residence area, and national health insurance were significantly associated with SRH in Thailand, but not in Korea.

Findings from this study showed that lower socioeconomic status contributed to poor SRH in both Korea and Thailand, and according to Kim (2007), these social and health disparities increase with age. Although Korean adults have longer life expectancies and lower mortality than Thai adults, health promotion interventions still need to be introduced among adults of lower socioeconomic status. Both the Korean and Thai governments should establish fundamental policies to improve health status particularly among these vulnerable groups.

In contrast to Korea, residence area and health insurance variables contributed to poor SRH in Thailand; as we observed earlier, there were differences in Thailand in these two variables. Thus, the Thai government should consider how to improve disease prevention and health promotion programs in urban areas as well as to improve the equality of benefits package across three schemes.

In addition, psychosocial factors such as depression are also important to consider when designing and implementing health promotion policies because these affected SRH in both countries. This factor appears to be more complex than others in that seeking care for depression is perceived as a stigma, which in turn might lead to suicide if patients do not receive treatment properly.

The implications of this study are as follows. First, the effects of behavioral risk factors on SRH in Thailand and Korea are similar, and thus, it can be inferred that universal risk factors could affect health similarly.

Second, although Korea and Thailand showed similar results for the behavioral risk factors, some control variables differed based on social and cultural contexts; it is important to understand and analyze the social and cultural backgrounds of each country. With this study, we aimed not only to simply analyze behavioral risk factors but also to take into account the context of each country, including demographic characteristics, SES, and health insurance systems as control variables.

Finally, the last analysis between 2009 and 2015 Korean among adult population emphasized the fact that the policy and intervention for health behaviors in Korea are strongly needed to be modified.

Limitations of this study are as follows. First, we limited the study to comparing behavioral risk factors between two countries, and thus, the results cannot be generalized to other countries; additional studies should expand to other countries in Asia to establish whether the results are similar. Second, we only analyzed Korean and Thai adults, and it is necessary to conduct subgroup analysis of SRH by country, especially the demographic and SRH variables that showed significant associations with SRH, for example, gender, family income, and education. Finally, the study’s cross-sectional design has the limitation of assuming cause-effect relationships, and these apparent relationships should be examined in longitudinal studies.