Abstract
Objective:
To identify levels of moderate-intensity physical activity (MPA) and vigorous-intensity physical activity (VPA) in a representative sample of college students in six East Asian economies and examine their relationship with weight, alcohol consumption and cigarette smoking.
Design:
Cross-sectional survey.
Setting:
College students recruited from 21 different colleges in six East Asian economies.
Method:
Self-reported physical activity, weight, height, alcohol consumption, and smoking were assessed using already-validated instruments. Multiple logistic regression models of MPA and VPA were separately performed for each economy, controlling for age, gender, living area before college, military experience, paid employment, and religion. All the analyses were performed using SAS 9.2 with clustering effects accounted for.
Results:
Being a heavy drinker increased the odds of engaging in VPA in five economies (adjusted odds ratio [AOR] = 1.56 to 2.65). Cigarette smoking was not associated with MPA in any economy and was only associated with VPA in China (AOR = 1.54) and Taiwan (AOR = 1.48), indicating that smokers are more likely to engage in VPA than non-smokers.
Conclusions:
The relationship between college students’ VPA and alcohol consumption and between MPA and cigarette smoking is similar across the East Asian economies, whereas the relationship between VPA and smoking varies substantially.
Introduction
It is well known that regular physical activity can prevent many kinds of chronic diseases, such as heart disease, diabetes, stroke, and osteoporosis.1–4 Many studies have shown that physical inactivity or low levels of physical activity can cause remarkable increases in all-cause mortality rates. 5 In particular, an increase in physical activity during young adulthood can decrease risk of adult mortality. 6 Given the health benefits of regular physical activity, more attention should be paid to physical activity among young adults because individual lifestyle behaviours are often established during young adulthood. 7 It is imperative that health-promotion efforts for young adults include a thorough understanding of how various health-promoting or inhibiting behaviours tend to co-occur.7, 8 Two of the most prevalent health-inhibiting behaviours during young adulthood are smoking and alcohol use. 9 Thus a better understanding of young adults’ smoking and drinking behaviours that may co-occur with physical activity will help practitioners enhance physical activity promotion efforts.
According to problem behaviour theory (PBT), human behaviours are composed of problem behaviours (i.e. delinquent or norm-violative behaviours) and conventional behaviours that are socially approved of. 10 Problem and conventional behaviours can be extended into the health behaviour domain since health behaviours are also likely to be affected by social norms. 11 For example, because of their well-known impact on health, smoking and excessive alcohol consumption can be considered as problem behaviours, and regular physical activity can be considered as a conventional behaviour. PBT also emphasizes that involvement in a specific problem behaviour is associated with more involvement in other problem behaviours and less participation in conventional behaviours due to their socio-ecological linkages across multiple levels of society. 10 It has been suggested that the form problem behaviours take may vary across different ethnic groups, regions, or cultures, which means there is no general set of criteria for defining problem behaviours.12,13
Most studies that have examined relations of physical activity with alcohol consumption and smoking among college students have been conducted in the United States of America (USA)8, 14–18 and European countries.19, 20 The results of previous studies examining the association between physical activity and alcohol consumption vary from country to country. Although most studies conducted in the USA show positive relations between physical activity and alcohol consumption,8, 14, 17, 18 one European study found a negative relation between these two behaviours. 20 Research findings regarding the relation between physical activity and cigarette smoking among college students are relatively consistent.8, 14–17, 19, 20 Given the mixed results in this field of study, it would be meaningful to compare these relationships among young adults living in different cultures. To our knowledge, there are only two studies21, 22 that have investigated cultural differences in associations of physical activity with smoking and alcohol consumption. However, none of these cross-cultural studies considered dose-response effects of alcohol consumption on physical activity. Moreover, there has been no cross-cultural study that has compared these relations among East Asian economies such as Hong Kong, Korea, Malaysia, Singapore, Taiwan, and China.
In addition to understanding the relation of physical activity with alcohol consumption and smoking among East Asian college students, it is also important to determine how physical activity is related to the body mass indexes (BMIs) of these students. Despite a lack of a comprehensive database, an increase in the prevalence of overweight and obesity has been noted in many Asian countries.23, 24 A longitudinal study 25 showed that a substantial proportion of adolescents tend to become and remain obese during the transition into young adulthood. This may be due to a decline in physical activity during this transition. 26 A decrease in energy expenditure, such as a decline in physical activity, may be partially responsible for the increasing obesity in young adulthood. 27 However, there is a paucity of longitudinal data that has prospectively investigated the relation between physical activity and BMI (or BMI groups) among those who have transitioned from adolescence to young adulthood.
Although there are a number of cross-sectional studies that have investigated such a relation among college-age students, inconsistent results are noted. Some studies in the USA 16 and European countries 22 reported a quadratic association (i.e. as BMI progresses from underweight to overweight, the prevalence of college students who meet the vigorous physical activity guideline increases, but the prevalence declines for those with an obese BMI). Other European studies found that there is a negative association between physical activity and BMI.28, 29 Although these inconsistent results may be due, in part, to the different methods used to measure physical activity or BMI, they may also be due to cultural differences in physical activity. Therefore it would be interesting to see whether such relations differ across different cultures. To our knowledge, there has been no study that has examined the association between physical activity and BMI (or BMI groups) among a representative sample of East Asian college students.
The purpose of the present study therefore was to compare the relation between physical activity and BMI groups and to investigate cultural differences in the relation of physical activity with alcohol consumption and cigarette smoking among college students in six East Asian economies (China, Hong Kong, Korea, Malaysia, Singapore, and Taiwan). To complement previous cross-cultural work, we investigated the dose-response effects of alcohol consumption on physical activity. We also distinguished vigorous-intensity physical activity (VPA) from moderate-intensity physical activity (MPA) because the determinants of VPA and MPA are likely to be different.16, 30 Literature suggests that the form of problem behaviours may vary across different regions or cultures.12, 13 Thus it was hypothesized that the association of physical activity with smoking and alcohol consumption differs across the six East Asian economies because of their different criteria for defining problem behaviours. In addition, despite the inconsistent results, we hypothesized that physical activity is negatively related to BMI (i.e. the more obese, the lesser the MPA and VPA) across the six East Asian economies based on the majority finding in the literature and the notion of energy balance. 27
Method
Design
The present study was conducted in conjunction with another investigation, 31 but the results of this study have neither been presented nor included in other manuscripts. During the 2008–2009 academic year, data were collected from a representative sample of college students (N = 16,558) from 21 institutions in six East Asian economies: three colleges each from Hong Kong, Korea, Malaysia, and Singapore; four colleges from Taiwan; and five colleges from China. After these six economies were chosen, a research consortium was established to develop a questionnaire; implement a sampling plan; collect, manage, and analyse the data; and publish the results of the study. The investigators of the participating economies who were college professors were requested by the lead investigator in the USA to randomly select one college from each of the three to five geographically representative and distant areas in their respective economy. Once participating institutions had been identified, classes were randomly chosen, and instructors of those classes were contacted to obtain permission to administer a survey. Research assistants visited the selected classes to administer the survey. The overall response rate was 78%. The study protocol was approved by the institutional review board of the institutions that collected data, and informed consent was obtained from each participant at the time of the survey.
Sample
The mean age of participants was 20.6 years (SD = 2.73). The sample was composed of 7852 women (47.4%) and 8521 men (51.5%) with 185 non-responses on gender. Women outnumbered men in the samples from Hong Kong, Malaysia, and Singapore, and the reverse was true in the samples from Korea, Taiwan, and China. The majority of the respondents (96%) were 17 to 25 years of age. Of the 16,558 participants, 302 (2%) had missing data on the dependent variables (VPA and MPA). These participants did not significantly differ from those who provided complete data on the dependent variables in terms of age, gender, living area before college, military experience, paid employment, and religion (all ps > .20). No imputation was made on the missing data.
Measures
To assess information about respondents’ physical activity, BMI, alcohol consumption, cigarette smoking, and other potential correlates, a 62-item closed-ended questionnaire was constructed based on pre-validated instruments, including the Behavioural Risk Factor Surveillance System (BRFSS) 32 and the Youth Risk Behaviour Survey (YRBS) 33 questionnaires. After an English version was developed, it was translated separately for the students in Korea, Taiwan, and China. The translation was conducted by two bilingual health behaviour professors using the procedures suggested by Banville et al. 34 A back translation was performed by two bilingual doctoral students majoring in health behaviour. Then the same panel of bilingual experts compared the original version to the back-translated version and edited the questionnaire accordingly. An expert converted the finalized English version of the questionnaire into a computer-scannable format that was administered in English-speaking economies, such as Hong Kong, Malaysia, and Singapore. A pilot test of the questionnaire was conducted with a group of international and American students at a Midwestern US university (n = 30: five Chinese-speaking students, five Korean-speaking students, and 20 English-speaking students) to check against any unexpected problems and to test whether the computer-scannable English-version questionnaire could be properly processed by a machine. No discernible problems were detected in the pilot test.
Physical activities were assessed by two questions drawn from the literature. 21 To assess VPA, participants were asked for the number of days they exercised or participated in physical activity for at least 20 minutes that made them sweat or breathe hard, such as basketball, soccer, running, swimming laps, fast bicycling, or similar aerobic activities, in the past seven days. Participants who reported three or more days on this question were defined as vigorously active. Participants were also asked for the number of days they exercised or participated in physical activity for at least 30 minutes that did not make them sweat or breathe hard, such as fast walking, slow bicycling, skating, or mopping floors, in the past seven days. Participants who reported five or more days on this question were defined as moderately active. These physical activity criteria were drawn from physical activity guidelines recommended by the American College of Sports Medicine. 35 BMI was calculated by using self-reported weight and height (kg/m2). Participants’ BMIs were categorized into four groups (underweight: < 18.5; normal weight: 18.5–22.9; overweight: 23.0–27.4; and obese: ≥ 27.5) using the World Health Organization (WHO) recommendations for Asians. 36
Alcohol consumption was assessed using three items from the BRFSS. Participants were asked the number of occasions that they had at least one alcoholic drink (i.e. a can or bottle of beer, a glass of wine, or a shot of liquor) and the average number of drinks they consumed on each occasion in the past 30 days. By multiplying the responses to these two questions and dividing it by 30, average daily alcohol consumption in the past 30 days was computed. Binge drinking was assessed by an item drawn from the BRFSS: ‘How many times in the last two weeks have you had five or more (if you are female, four or more) alcoholic drinks at one sitting?’. Based on the average daily alcohol consumption and binge drinking experience, alcohol consumption was grouped into three categories (abstainer, moderate drinker, and heavy drinker). Participants whose average daily alcohol consumption was zero were defined as abstainers. Using the criteria suggested from the CDC, 37 participants whose average daily alcohol consumption was two drinks or less (one or less for female) were defined as moderate drinkers, and those whose average daily alcohol consumption was more than two drinks (more than one for female) were defined as heavy drinkers. Those who engaged in binge drinking at least once in the last two weeks were grouped into the heavy drinker category. In fact, 96.6% of heavy drinkers, defined by average daily alcohol consumption, were also binge drinkers. Current smoking status was determined using two items from the BRFSS. Participants who reported having smoked at least 100 cigarettes in their lifetime and who currently smoke every day or some days were defined as current smokers.
Analysis
To examine how BMI group, current cigarette smoking, and alcohol consumption were associated with physical activity, multiple logistic regression models of MPA and VPA were separately performed for each economy, controlling for age, gender, living area before college, military experience, paid employment, and religion. Living area before college was included as a covariate since rural residence is often correlated with lower socioeconomic status (SES) in comparison to urban residence38, 39 and SES has been shown to be associated with physical activity levels.40, 41 All the analyses were performed using SAS 9.2, accounting for clustering effects within the same institution.
Results
Descriptive analysis
The mean BMI was 21.3 (SD = 3.6). Table 1 shows the levels of MPA and VPA by correlates among college students in the six Eastern Asian economies. The percentage of students who were moderately active (i.e. engaging in MPA at least 30 minutes a day on five days or more in the past seven days) varied from 7% (Taiwan) to 22% (Malaysia). The percentage of vigorously active students (i.e. engaging in VPA at least 20 minutes a day on three days or more in the past seven days) varied from 24% (Korea) to 52% (Malaysia). Men were more physically active than women in both MPA and VPA across all the six economies except Singapore and Taiwan where the MPA level was the same between the sexes. As expected, those with military experience were more vigorously active than their counterparts in all the economies except Taiwan. In terms of BMI group, underweight students were least vigorously active and overweight students were most vigorously active in each economy. Interestingly, in terms of substance use behaviour, cigarette smokers and heavy alcohol drinkers were more physically active than non-smokers and abstainers, respectively.
Levels of moderate- and vigorous-intensity physical activity by correlates (N = 16,558)
Notes: MPA = percentage of those who engaged in moderate-intensity physical activity at least 30 minutes a day on five days or more in the past seven days; VPA = percentage of those who engaged in vigorous-intensity physical activity at least 20 minutes a day on three days or more in the past seven days.
Body mass indexes were categorized into four groups (underweight: < 18.5; normal weight: 18.5–22.9; overweight: 23.0–27.4; and obese: ≥ 27.5) using the WHO recommendations for Asians.
Multivariate analysis
Table 2 shows the results of the logistic regression analyses of MPA and VPA after adjusting for age, gender, living area before college, military experience, paid employment, and religion. Smoking and BMI group showed little relation with MPA in the multivariate model although overweight students in Korea were more likely (adjusted odds ratio [AOR] = 1.35) and underweight students in China were less likely (AOR = 0.68) than those with normal weight to be moderately active. Alcohol consumption was independently associated with MPA in three economies. Moderate drinkers were less likely to engage in MPA in Hong Kong (AOR = 0.65) and Malaysia (AOR = 0.63), and heavy drinkers were more likely than abstainers to engage in MPA in China (AOR = 1.39).
Logistic regression of moderate- and vigorous-intensity physical activity (N = 16,558)
Notes: AOR = adjusted odds ratio; PA = physical activity; CI = confidence interval; ref = reference category.
Odds ratio adjusted for age, gender, living area before college, military experience, paid employment, and religion.
Engaging in moderate-intensity physical activity at least 30 minutes a day on five days or more in the past seven days.
Engaging in vigorous-intensity physical activity at least 20 minutes a day on three days or more in the past seven days.
P < .05; **P < .01; ***P < .001.
The three variables of interest (BMI group, cigarette smoking, and alcohol consumption) showed a stronger independent association with VPA than with MPA. Students who were in the underweight category were significantly less likely to engage in VPA compared to normal-weight students in four economies (AOR ranges from 0.57 to 0.77) with the other two economies showing the same directionality. Being a heavy drinker significantly increased the odds of engaging in VPA in every economy (AOR ranges from 1.56 to 2.65) except Singapore. By contrast, being a smoker significantly increased the odds of engaging in VPA only in Taiwan (AOR = 1.48) and China (AOR = 1.54).
Discussion
This is the first study that has compared the association of BMI group, cigarette smoking, and alcohol consumption with physical activity among a representative sample of college students from six East Asian economies. The percentage of students who were moderately active varied from 7% in Taiwan to 22% in Malaysia, and the percentage of vigorously active varied from 24% in Korea to 52% in Malaysia. This substantial variation in the levels of physical activity might lead to the a priori hypothesis of this study in that the association of physical activity with smoking and alcohol consumption differs across the six economies. However, findings from this study only partially supported this hypothesis. Alcohol consumption was independently (significant even after controlling for covariates) and consistently (showing the same directionality) associated with VPA across all the six economies except Singapore. The association between BMI group and VPA was also consistent in four economies. By contrast, the association between cigarette smoking and VPA was inconsistent. Unlike VPA, MPA showed a weak and inconsistent relation, if any, with BMI group, smoking, and alcohol consumption. Little difference appears to exist in the relation between college students’ VPA and alcohol consumption, but substantial difference exists in the relation between VPA and smoking behaviour across the six economies.
The finding that college students who are heavy drinkers are more likely than abstainers to engage in VPA is in line with previous research with college students in the USA 8, 14, 17, 18 and European countries.20, 22 However, these results are contrary to PBT, which supports the notion that involvement in a specific problem behaviour is associated with less participation in conventional behaviours. 10 This might imply that alcohol consumption is not necessarily considered a problematic behaviour by college students 20 across the globe. In fact, research has found that many young adults believe that alcohol consumption has more positive than negative effects. 42 For example, college students report that alcohol consumption leads them to have better relations with their friends and have more fun, relaxation, and smoother socialization. 43 Another way of interpreting this result, especially with regard to the insignificant relation between MPA and alcohol consumption, is that many college students who are prone to being adventurous and extroverted are likely to consume alcohol to meet their conviviality and express their social competencies. 44 These types of students may seek interesting and exciting activities, such as VPA or sporting events, rather than MPA. Research indicates that college students who engage in regular VPA are more likely to participate in social gatherings and show a higher acceptance for alcohol consumption. 18 Findings of this study affirm that heavy-drinking college students are more likely than abstainers to engage in VPA rather than MPA and suggest that this phenomenon might be ubiquitous.
It is interesting that cigarette smoking was not associated with MPA in any economy and was only associated with VPA in China and Taiwan. Among college students in China and Taiwan, smokers were more likely to engage in VPA than non-smokers. This contrasts sharply with studies14, 16, 17 which found that, among US college students, smokers are less likely to engage in both VPA and MPA than non-smokers. This means that, collectively, the MPA levels of East Asian college students are not different between smokers and non-smokers, whereas the levels are different among their US counterparts.
Research indicates that college students have different risk perceptions for different substances in terms of their effect on physical capabilities and health. 14 East Asian college students may have a lower risk perception of smoking or generally be more accepting of it than their US counterparts, which might have led to the non-differentiation between the MPA of smokers and non-smokers among East Asian college students and a higher level of VPA among smokers in China and Taiwan. It deserves mention that young adults frequently quote social reasons for their participation in VPA, including that they want to be with friends or to be part of a team. 45 Previous research has found that smoking is socially accepted and encouraged among Chinese and Taiwanese adults 46 as smoking is believed to help reinforce social relationships or friendships. 47 Therefore smoking and VPA may have a positive relation in China and Taiwan because these behaviours are likely to occur for the same reason. Yet given the heterogeneity in the relation between college students’ smoking and VPA within the East Asian region as well as between continents, smoking, especially in the context of predicting college students’ physical activity, should be tackled with a transversal rather than universal perspective.48
As for the relationship between BMI group and physical activity, the hypothesis that physical activity is negatively related to BMI was not tenable. Rather, as indicated by some researchers,16, 22 the relationship is more quadratic than linear. As BMI progressed from underweight to overweight, the prevalence of college students who engaged in MPA or VPA also increased, but this prevalence declined for those with an obese BMI across the six economies. In the multivariate logistic model where covariates were controlled, underweight students were less likely to engage in VPA than those with normal weight across all the economies except Korea and Singapore where the AORs with the same directionality failed to reach statistical significance. Previous research asserts that underweight students are less likely to participate in VPA than students of other weight categories because underweight students are less likely to be involved in team sports than others. 22
Limitations
This study has several limitations. First, it is not possible to investigate causal relations between BMI group, cigarette smoking, alcohol consumption, and physical activity because this study used cross-sectional data. Future research that employs a longitudinal study design would be desirable to establish causality. Second, this study is based on self-reported measures that may have introduced bias. Third, due to the use of Chinese and Korean versions of the questionnaire, the exact meanings of some original question items may have been lost in the translation. However, this possibility was minimized by re-translation (i.e. backward and forward translation), which helps increase the likelihood of equivalence across the instruments. Fourth, a small number of questions that were used to assess VPA, MPA, smoking, and alcohol consumption in this study might not have captured all the aspects of these variables, although these question items are pre-validated ones that are commonly used in this type of survey research.
Conclusions
Despite these limitations, our study provides important insight into cross-cultural differences and similarities in the relation between BMI group, cigarette smoking, alcohol consumption, and physical activity among college students in six East Asian economies. The relation between college students’ VPA and alcohol consumption and between MPA and cigarette smoking was similar across the six economies, whereas a substantial difference was found in the relation between VPA and smoking. Findings from this study indicate that smoking, especially in the context of predicting college students’ physical activity, should be tackled with a transversal rather than universal perspective. In addition, the present study affirms that heavy-drinking college students are more likely than abstainers to engage in VPA rather than MPA and that this phenomenon might be true across many cultures. This result does not support PBT. According to PBT, heavy-drinking college students should engage in less VPA than abstainers. The opposite was observed in this study. Future cross-cultural studies are needed to examine if college students consider alcohol consumption a problematic behaviour. 20 Across the East Asian region, aggressive interventions, such as anti-alcohol campaigns or education, need to be implemented to change the positive perception of alcohol consumption by college students, especially those who engage in VPA on a regular basis. Findings of this study indicate that, unlike alcohol consumption, smoking should be tackled with a transversal rather than universal perspective in the East Asian economies.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.