Article Text

Download PDFPDF

Cross-national comparisons of socioeconomic differences in the prevalence of leisure-time and occupational physical activity, and active commuting in six Asia-Pacific countries
  1. Adrian Bauman1,
  2. Guansheng Ma2,
  3. Frances Cuevas3,
  4. Zainal Omar4,
  5. Temo Waqanivalu5,
  6. Philayrath Phongsavan6,
  7. Kieren Keke7,
  8. Anjana Bhushan8,
  9. for the Equity and Non-communicable Disease Risk Factors Project Collaborative Group
  1. 1School of Public Health, University of Sydney, Sydney, Australia
  2. 2Institute for Nutrition and Food Safety, Chinese Center for Disease Control and Prevention, China
  3. 3Degenerative Disease Office, Department of Health, Manila, Philippines
  4. 4Disease Control Division, Ministry of Health, Kuala Lumpur, Malaysia
  5. 5Ministry of Health, Fiji
  6. 6School of Public Health, University of Sydney, Sydney, Australia
  7. 7Government of Nauru, Republic of Nauru
  8. 8Division for Health Sector Development, Western Pacific Regional Office World Health Organization, Philippines
  1. Correspondence to Adrian Bauman, School of Public Health, University of Sydney, Medical Foundation Building, New South Wales, Sydney 2006, Australia; adrianb{at}health.usyd.edu.au

Abstract

Background This study describes physical activity patterns and their association with socioeconomic factors in six countries in the Asia-Pacific region, and examines whether physical activity associations with socioeconomic status follow similar patterns across the six countries.

Methods Population-wide representative surveys of non-communicable disease risk factors and socioeconomic factors conducted in Australia, China, Fiji, Malaysia, Nauru and the Philippines between 2002 and 2006 were used. Survey respondents aged 18–64 years who provided information on their socioeconomic status (age, education, income, area of residence) and physical activity level in three domains (leisure-time, occupation, commuting) were included in the study (Australia N=15 786; China N=142 693; Fiji N=6763; Malaysia N=2572; Nauru N=2085; Philippines N=3307).

Results Leisure-time physical activity increased with age in China, showed inverse associations for Fiji and Nauru men, and there were no age relationships in other countries. Individuals in China, Fiji and Malaysia living in urban areas, with higher educational attainment and affluence were physically active during leisure time but less active at work and during commuting compared to those in rural areas, with lower educational attainment and lower income.

Conclusion There is a link between types of physical activity participation and socioeconomic factors in developing countries. Associations with socioeconomic indicators are likely to reflect economic growth. The findings strongly support the need for a comparable non-communicable risk factors surveillance system in developing countries.

  • Chronic disease
  • physical activity
  • surveillance
  • social inequalities
  • social economic
  • chronic DI
  • surveillance SI

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

There is increasing recognition of the role of physical inactivity in the Global Burden of Disease.1 A focus on increasing physical activity, as part of non-communicable disease control, is now established for developed as well as developing countries.2 Rapid change in the social and economic landscapes of developing countries has had profound effects on urbanisation, workforce structure and lifestyle patterns.28 4 5 Changes in the socioeconomic environment have also resulted in physical activity declining as populations shift from active occupations and commuting to sedentary jobs and motorised commuting.3 6–11 Given that leisure-time physical activity participation is not yet common in many developing countries, substantial decline in overall physical activity levels in these settings is expected to continue.7 12–14

The final report of the WHO Commission on Social Determinants of Health15 reinforced the importance of measuring socioeconomic inequality, and called for routine monitoring of the social and economic determinants of health. Although associations between socioeconomic indicators and physical activity have been extensively explored in developed countries for several years,16–18 only recently has a similar body of research emerged from developing regions.12 19–21 Where physical activity levels have been consistently shown to vary by socioeconomic status in developed countries, the same associations may not exist in developing countries. The nature and scope of socioeconomic patterning of physical activity in developing countries must be understood so that appropriate public health responses can be formulated.

Using population-wide survey data from six Asia-Pacific countries at various stages of socioeconomic development, the aim was to describe how physical activity - at work, during commuting and leisure time - is distributed by age, education, income, and rural/urban areas within a country, and to compare these associations across countries at different stages of socioeconomic development. Age was chosen because of evidence of age-related decline in physical activity22–24; educational attainment because of its known association with physical activity22 24–26; and income and rural/urban indicators reflect affluence and urbanisation, respectively, with both factors also shown to be related to physical activity levels.3 27

Methods

Data sources and participants

This study was a series of cross-sectional analyses of data from China, Fiji, Malaysia, Nauru, the Philippines and Australia. These countries were selected as they were part of the WHO study to examine the relationships between non-communicable disease risk factors and indices of socioeconomic status. To be eligible for the study, countries also needed to have national population-level data with broadly comparable demographic and socioeconomic measures, using comparable sampling methods that would yield a representative sample, as well as clearly identified self-report measures of physical activity.

Data sources were derived from representative population surveys of individuals aged 15 years and older. Surveys were conducted between 2002 and 2006. A summary of the survey designs and response rates are presented in table 1. The survey populations in all countries comprised non-institutionalised individuals living in private dwellings. Institutional ethics committees approved the study in each country. As this was a secondary analysis of existing data, the only exclusion/inclusion criterion for this analyses was that participants were aged 18–64 years.

Table 1

Characteristics of surveys in countries that participated in the study and economic performance

Settings

To examine socioeconomic differences in physical activity patterns cross-nationally, this study analysed data from countries at varying stages of socioeconomic development and population sizes (see table 1). For the purpose of this study, developing countries are those in the low-income to middle-income categories,28 although it is acknowledged that countries with apparently high income may also still be developing countries. In this study, Australia with a population of 21 million people is classified as a high-income economy (a developed country); China with approximately 1.3 billion people is regarded as a lower middle-income economy (a developing country); whereas Nauru with 13 528 people is considered the least affluent. Overall, Nauru represents the extreme comparison to China in terms of population size and to Australia in terms of economic development. China has the fastest growing economy of all six countries with the Gross Domestic Product increasing by around 8–10% per year over the past 20 years.28 31

Defining education, income and area

Because this was a comparative analysis, the objective was to profile physical activity by comparable socioeconomic and demographic measures. As countries used diverse socioeconomic status measures, only education and income were examined. These were considered to yield more comparable data than occupation-based measures. Education was categorised into three broad qualitative categories: completed primary education (low), completed secondary education (medium) and completed tertiary education (high). Four countries had income data: measures of individual income per annum (China), household income per month (Malaysia) or per week (Australia) were grouped into low, medium, high or very high income. For the Philippines, household income per annum was grouped into quartiles according to the country's own income distribution. These classifications were not based on exactly the same level, because the focus was on comparing relative rather than absolute differences in socioeconomic status. Country-specific cut points were used because they were deemed to be more meaningful to the country context.

Only the surveys from China, Fiji, Malaysia and Australia included data stratified by urban and rural residence. For China, the levels of economic development were used to group areas into either urban or rural. Large, medium and small cities were defined as urban areas and the remaining areas as rural. In Malaysia, areas with a population of 10 000 or more were classified as urban areas and those with less than 10 000 were classified as rural areas, according to the 2000 population census. For Australia, metropolitan and outer metropolitan areas were classified as urban, whereas rural and remote areas were defined as rural.

Defining physical activity

Although countries used diverse measures of physical activity, most collected information on time spent in moderate-intensity and vigorous-intensity activity at work, commuting and leisure time. Additional items asked about walking and time spent in sedentary activity. Three physical activity domains were examined separately: leisure, occupational and commuting. A standard protocol was used to define high leisure-time physical activity, high work-related activity and high commuting-related physical activity conferring health-enhancing benefits. Table 2 provides an overview of the physical activity measures and definitions used in each country.

Table 2

Definition of high levels of physical activity according to domain

Statistical analysis

Analyses were conducted using either SPSS (Statistical Package for the Social Sciences) or SAS V9.1 (Statistical Analysis Software) for Windows. Multiple logistic regression analyses were performed to model the probability of having moderate to high levels of physical activity, adjusting for age and all other socioeconomic measures. Different models were built for each country, and separately for men and women. All analyses were adjusted for cluster sampling effects (Australia, China, Fiji, Malaysia, the Philippines) and a Finite Population Correction Factor adjustment was computed and applied to the variance estimates to compensate for the large sampling fraction (45%) (Nauru). Adjusted ORs were produced with 95% CIs.

Results

Gender distributions were broadly comparable among countries (table 3). Fiji, Nauru, Malaysia and the Philippines had higher proportions of surveyed populations aged less than 50 years. The proportion of rural populations was much higher in China than in Malaysia and Fiji. Some differences in income distributions were found: China had a higher proportion of respondents with medium income compared with Malaysia (76% vs 43%). The distributions of education levels differed among countries, with China and Nauru reporting lower tertiary education compared with Malaysia and Fiji. The Philippines and Australia had the highest proportions of respondents with tertiary education.

Table 3

Survey characteristics of participating countries by demographic and socioeconomic status for men and women

Associations between age and physical activity

Table 4 shows that in China, the odds of engaging in leisure-time physical activity and active commuting increased markedly with age, for both sexes. Being older in China, however, reduced the odds of being physically active at work. In contrast, leisure-time physical activity decreased with age among men in Fiji and Nauru, whereas only Australian women aged 50 years and older were more likely to participate in leisure-time activity. In Fiji, although men aged 50 years and older were more likely to report active commuting (OR 1.2, 95% CI 1.0 to 1.6), women in the same age group were less likely to engage in active commuting (OR 0.7, 95% CI 0.6 to 0.9). There were no clear associations between age and physical activity in the Philippines or Malaysia.

Table 4

Adjusted association* (OR and 95% CI) between physical activity (PA) domains and age, by sex and country

Associations between education and physical activity

Table 5 shows that after adjusting for other socioeconomic and demographic variables, respondents with higher education in China, Fiji and Australia were significantly more likely to do leisure-time physical activity compared to those with lower education. However, highly educated respondents in China and Fiji (men only) were much less likely to engage in physical activity as part of their work compared to those with lower education. Educated men and women in China were also less likely to engage in active commuting. There were no significant associations between education and physical activity in the Philippines, Malaysia and Nauru.

Table 5

Adjusted association* (OR and 95% CI) between physical activity (PA) domains and education, by sex and country

Associations between income and physical activity

Table 6 shows the adjusted association between income and physical activity for Australia, China, Malaysia and the Philippines. Wealthy respondents in China were twice as likely to participate in leisure-time physical activity compared with the less affluent (men OR 2.1, 95% CI 1.8 to 2.4; women OR 2.7, 95% CI 2.4 to 3.1). Similarly, Filipino women in the highest income group were more likely to be physically active during leisure time (OR 2.9, 95% CI 1.2 to 7.2) than women in other countries except China. Wealthier men and women in China were less likely to be physically active at work or engaging in active commuting for 30 min or more each day. There were no significant associations between income and occupational physical activity in the Philippines or Malaysia.

Table 6

Adjusted association* (OR and 95% CI) between physical activity (PA) domains and income, by sex and country

Associations between area and physical activity

Table 7 shows a strong and significant association between area and leisure-time physical activity in China, with men (OR 5.7,95% CI 5.1 to 6.4) and women (OR 8.7, 95% CI 7.8 to 9.7) living in urban areas significantly more likely to be physically active during their leisure time than rural populations. In contrast, urban men and women were significantly less likely to report being active as part of their work compared with rural residents. A similar pattern of associations was also observed for men and women in Fiji and Malaysia with regard to occupational physical activity. In Fiji and China (men only), living in urban areas significantly reduced the probability of residents engaging in walking or cycling for commuting purposes compared with rural residents.

Table 7

Adjusted association*(OR and 95% CI) between physical activity (PA) domains and area, by sex and country

Discussion

This study contrasts different physical activity domains in relation to socioeconomic status in six Asia-Pacific countries. The relationship of leisure-time physical activity and socioeconomic status has been systematically examined in developed countries18; the present study shows these established patterns do not hold for developing countries, and further, that they become more complex when non-leisure domains of physical activity are considered. These additional domains, such as work and active transport, are important in maintaining energy expenditure and preventing obesity in developing countries.

This study found the pattern of associations between age, education, income and area and the different domains of physical activity vary among countries. Specifically, leisure-time physical activity increased with age in China and to a certain extent in Australia (women), but showed inverse associations among Fijian and Nauruan men, and no age relationships in other countries. In China, Fiji and Malaysia as in Australia, urban, educated and wealthy populations were more physically active during leisure time but less active as part of their work and commuting compared with populations in rural areas, with lower education and income.

Some limitations in comparing the six national surveys should be noted. First, different socioeconomic measures were used across countries. However, meaningful categorisation of education, income and areas were used where possible to ensure comparability. For example, achieved level of education rather than number of years of education completed was used to overcome the heterogeneity of the various educational systems. Second, the reliance on self-report physical activity measures coupled with different cultural nuances in the six countries could result in reporting biases, although broad categorisation is likely to classify ‘active’ and inactive adults reasonably well. Third, due to the different physical activity measures and differential categorisation of physical activity level, precise comparisons of prevalence were difficult. It is therefore possible that these factors combined may explain the differences in the relationship observed between countries. Additionally, important differences in survey implementation due to cultural approaches, adaptation and translation of questionnaires could result in differential interpretations of the survey and contribute to the challenges of contextualising the findings of this multicountry study.

Due to a lack of relevant information in some countries, it was not possible to calculate total physical activity and the contribution of the various activity domains to total physical activity. Physical activity related to housework was not assessed by participating countries, and therefore not considered in the analyses. It was also not possible to look at physical activity patterns across income and area for all countries because of missing data. However, data comparability was facilitated by reference to international recommended definitions of health-enhancing physical activity.33

The surveys have a number of strengths. The high response rates, the national samples and the multistage clustered random selection of samples help to ensure that the data were nationally representative. The large-scale samples derived from several national surveys are also a strength of the study. Another important advantage is that with the exception of the Australian data, all surveys examined a range of physical activity domains, particularly physical activity derived from commuting and occupation. These settings are important sources of energy expenditure, and not just leisure-time physical activity, in many developing countries.

The findings of this study clearly demonstrate a link between physical activity types and socioeconomic factors. In both established and developing countries (Australia and China respectively), leisure-time physical activity levels are generally higher in educated and affluent populations. The present findings are similar to socioeconomic differentials in physical activity patterns seen in developed countries.18 34 In new and emerging economies of Eastern Europe, cross-national data showed that higher education is a strong and consistent predictor of leisure-time activity in men and women.19 These findings could be explained in part by previous studies from developed countries showing that individuals of higher socioeconomic status are usually the first to respond to public health programmes.35 It is expected the same will occur in developing countries, with urbanised and educated populations also more likely to respond to public health campaigns promoting leisure-time physical activity.

However, in Malaysia, Nauru and the Philippines, the pattern of association was different; highly educated groups did not indicate a physically active lifestyle during leisure time and Nauru even showed an inverse relationship, although the associations did not reach statistical significance in any of the three countries. In view of the low variations in the prevalence of physical activity levels across socioeconomic groupings in these countries, the absence of significant associations is not surprising. Another explanation is that the lack of association between education and physical activity could reflect the limited understanding of the health benefits of physical activity. Participating in physical activity across the various domains is generally a novel concept in many developing countries, and there is a need to improve understanding of incidental physical activity as a health-enhancing behaviour.

Consistent with previous studies,3 36 it was found that high levels of occupational and active commuting were more common in low income and low education groups. The patterns are pronounced in China and to a lesser degree in Fiji. However, it is not expected that these populations will be protected from chronic diseases because they are enjoying relatively higher levels of physical activity. Two notable issues are worthy of consideration. First, assuming that sustained economic development is likely to affect all levels of society through significant changes in infrastructure, nature of employment, urban planning and transportation - all of which could significantly influence physical activity behaviour of the population - it is likely that those in lower socioeconomic groups will also experience a gradual decline in total energy expenditure traditionally gained through work and active commuting. These changes will render those with lower educational or economic status more vulnerable to the disease burden associated with physical inactivity.6 35 37–39

Second, urbanisation in China, Fiji and Malaysia was significantly associated with lower occupational activity and lower active commuting. The present results also showed a highly significant association between location of residence (ie, urbanisation) and leisure-time physical activity (high), but no relationship in Australia. It is postulated that the trend observed in this study will hold in other developing countries where the rates of walking and cycling for commuting are expected to decline in parallel with increasing urbanisation and automobile ownership.6

However, the argument that physical activity decline is generally associated with economic affluence does not hold true in all cases. An interesting result to note from this study is that Nauru, as the smallest country (in size and population) with the least developed economy, reported relatively high prevalence of physical inactivity during leisure (men 84.7%, women 95.2%), work (men 60.5%, women 70.1%) and commuting (men 77.8%, women 81.6%) compared to Australia, Fiji and Malaysia where similar types of data are available (data not shown here). As indicated in tables 4 and 5 age and education were not associated with any domains of physical activity. This could be explained by generally similar prevalence of physical activity levels across age and educational groupings, with the exception of leisure-time and occupational physical activity among young men. Thus, broad socioeconomic factors (such as education, age) arguably are less relevant to the likelihood of Nauruans engaging in health-enhancing physical activity. It could be that physical activity participation in Nauru and most likely in other participating countries is determined by other factors, such as cultural influences and climatic conditions. For example, poverty can also reduce opportunities for physical activity as primary care providers are more concerned with the challenges of daily survival.40 Societal expectations that women should not engage in activities that take them away from family responsibilities,41 or religious beliefs about modest clothing by women may be barriers to leisure-time physical activity participation.42 The influence of sociocultural factors points to the need to consider and assess domestic activities as an important source of energy expenditure for women in developing countries, which might confer health-enhancing benefits.43 Identifying a wide range of correlates of physical activity participation across different domains (ie, work, active community, leisure time, household and domestic chores) in cross-national studies, therefore, would add to knowledge about the determinants of physical activity participation in these domains and the overall directions for public health efforts.

In most cases, effective public health actions will need to be quite different to developed countries, and consider the socioeconomic status and economic development of the populations so that appropriate health promotion strategies can be tailored to meet the specific needs of population groups. For example, for the wealthy and the more educated the focus may be on promoting leisure-time physical activity and increasing awareness about the health-enhancing benefits of regular physical activity. Among the lower socioeconomic status groups, health communications promoting social norms that encourage energy expenditure maintenance on work and retention of an active lifestyle through active commuting may be required. Although with industrialisation, only the latter may remain as a potential strategy. However, halting the decline of physical activity will require a comprehensive range of strategies targeting both individual-level change and the creation of supportive infrastructure. Regardless of economic development, a national focus on developing or maintaining more active societies through developing infrastructure that favours cyclists and pedestrians, making available cycling lanes and public recreational spaces are some of the strategies that deserve more attention.

Conclusion

This study shows that in order to understand how a country experiences change in physical activity levels, assessment of individual and contextual factors are needed to better understand the determinants of such changes. A comparable physical activity surveillance system and measures implemented within and between countries would add greatly to improving the evidence base in this area. The WHO Stepwise approach to risk factor surveillance (Stepwise surveys),30 which has already been implemented in at least 39 developing countries, documents the extent of non-communicable disease risk factors using standardised questions and protocols. The Stepwise surveys use the Global Physical Activity Questionnaire (GPAQ) to assess physical activity separately for the domains of work, transport and leisure-time activities.32 Such a surveillance system will allow for meaningful monitoring and comparisons to be made within and between countries.

The findings of this study have relevance for countries undergoing economic and social transitions. For many countries, although improved economic resources and urbanisation are important for improving living standards, increasing life expectancy and combating infectious diseases, the same resources can also result in increased risk for chronic non-communicable diseases. The challenge for these countries is how best to ameliorate the negative consequences of economic development in maintaining low levels of physical inactivity in order to reduce the imminent burden of chronic non-communicable diseases.

What is already known on this subject

  • Individuals with high socioeconomic status are more likely to be physically active than their counterparts.

  • However, the same association may not hold in developing countries.

  • The scope and distribution of socioeconomic patterning of physical activity in these countries needs to be understood to inform the development of appropriate public health strategies.

What this study adds

  • Measuring all domains of physical activity is an important component of international surveillance studies, especially when involving economically developing countries.

  • Although economic development plays a key role in physical activity decline, broader cultural and environmental factors are equally important in determining this behaviour.

Acknowledgments

Thank you to the following coinvestigators and country-level authors for statistical and specific advice on the manuscript and management of the overall study: Azah N, Chunming Chen, Tien Chey, Zhaohui Cui, Charmaine Duante, Gauden Galea, Xiaoqi Hu, Lingzhi Kong, Dante D. Morales, Mustapha F, Anis Salwa, Ismail Samad, Stephanie Schoeppe, Ruby Thoma, Si Thu Win Tin, Timaima Tuiketei, Godfrey Waidubu. We thank the following institutions which allowed access, analysis and reporting of the survey data: Australia - Australian Bureau of Statistics; China - Department of Disease Control and Prevention, Ministry of Health, National Institute for Nutrition and Food Safety, Chinese Center for Disease Control and Prevention; Fiji - Ministry of Health; Malaysia - Diseases Control Division, Ministry of Health; Republic of Nauru - Ministry of Health; Philippines - Degenerative Disease Office, Department of Health; Food and Nutrition Research Institute; WHO Western Pacific Region; WHO South Pacific Office.

References

Footnotes

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the University of Sydney, Ref: 03-2009/11569.

  • Provenance and peer review Not commissioned; externally peer reviewed.