Abstract
Objective
There has been an increase in overweight among women in low- and middle-income countries, but whether these trends differ for women in different occupations is unknown. We examined trends by occupational class among women from 33 low- and middle-income countries in four regions.
Design
Cross-national study with repeated cross-sectional demographic health surveys (DHS).
Subjects
Height and weight were assessed at least twice between 1992 and 2009 in 248,925 women aged 25–49 years. Interviews were conducted to assess occupational class, age, place of residence, educational level, household wealth index, parity, and age at first birth and breastfeeding. We used logistic and linear regression analyses to assess the annual percent change (APC) in overweight (BMI > 25 kg/m2) by occupational class.
Results
The prevalence of overweight ranged from 2.2% in Nepal in 1992–1997 to 75% in Egypt in 2004–2009. In all four regions, women working in agriculture had consistently lower prevalence of overweight, while women from professional, technical, managerial as well as clerical occupational classes had higher prevalence. Although the prevalence of overweight increased in all occupational classes in most regions, women working in agriculture and production experienced the largest increase in overweight over the study period, while women in higher occupational classes experienced smaller increases. To illustrate, overweight increased annually by 0.5% in Latin America and the Caribbean and by 0.7% in Sub-Saharan Africa among women from professional, technical, and managerial classes, as compared to 2.8% and 3.7%, respectively, among women in agriculture.
Conclusion
The prevalence of overweight has increased in most low and middle income countries, but women working in agriculture and production have experienced larger increases than women in higher occupational classes.
Keywords: overweight, occupation, labour force, developing countries
Introduction
In low and middle income countries, the prevalence of overweight (BMI >25 kg/m2) among adults aged 20 years and older ranges from 18% to 59% in 2008, while the prevalence of obesity (BMI >30 kg/m2) ranges from 7% to 24% in 2008 1. These cross-country differences are strongly clustered by region, with the prevalence of overweight ranging from 14% in South East Asia to 62% in the Americas2. Most low and middle income countries have experienced large and faster increases in overweight and obesity over the last decade3, 4, paired with a decline in underweight 5, 6. To some extent, these trends may be attributable to a shift towards more sedentary lifestyles and less healthy dietary habits as a result of rapid economic growth and industrialization7, 8. In addition, trends in overweight may also be the result of trends in labour force participation. For example, in most countries, there has been a decrease in the number of women working in the agricultural sector and an increase in women working in more sedentary occupations9–11. In addition, the nature of agriculture and production work has evolved towards less physically demanding and more sedentary routines9. Until now, no studies have examined how these changes have affected women working in different occupational sectors.
Understanding trends in overweight for women from different occupational groups may provide important insights into the nature of the obesity epidemic in low- and middle-income countries. During the second half of the 20th century, female labour force participation rates in many low- and middle-income countries rose sharply, while fertility rates dropped dramatically 12, 13. On average, it is estimated that approximately 50% of women in low- and middle-income countries are currently in the labour force 14. During recent decades, women in most countries have also experienced major changes in their occupation. For example, women working in agriculture have benefited from technological developments, which may have changed the level of physical activity for agricultural tasks15. This transition is likely to have pushed women into better paid jobs that may have enabled them to change their diet and increase their leisure time physical activity, while at the same time adopting more sedentary occupations. As a result of these changes, women from different occupations may have experienced different trends in overweight and obesity. In addition, cross-national differences in the pace of these changes 16 may have led to different patterns across countries.
Few studies have examined variations in overweight and obesity according to occupation in low and middle income countries. Existing studies have focused primarily on high-income countries, and suggested that the pattern of overweight and obesity by occupation differs by gender17, 18. Some studies found that women in professional, management, clerical, and service occupations have lower overweight prevalence and have experienced smaller increases in overweight than women in production and agriculture 19, 20. However, whether this applies to women in low - and middle -income countries is uncertain. One study found that the prevalence of overweight is lower only among men in agricultural occupations compared with men in non-agricultural occupations, while there is no difference by occupation among women21. A comparative study in Asian countries found that the population in rural areas tends to be more physically active during work, but they are less active during leisure time 16. This illustrates a lack of sufficient understanding of these trends and how they vary in women across countries.
In this study, we examine trends in overweight by occupational class among women in 33 low- and middle-income countries over the last two decades. We hypothesized that in low- and middle-income countries women in agricultural and manual occupations have lower risk of overweight as compared to women working in service and professional occupations. However, we expect a shift towards more rapid increases in overweight in agricultural and manual occupations. Our motivation for this hypothesis comes from previous evidence of more rapid declines in occupational physical activity associated with an increasing role of technology in the agricultural and production sectors 22.
Materials and Methods
Study population
Data were obtained from the Demographic Health Survey (DHS), a survey of nationally representative samples of women in reproductive ages (15–49 years) living in 33 middle- and low income countries interviewed between 1992 and 2009. Sampling was based on a stratified two-stage cluster design divided by residence areas (urban and rural), geographic or administrative regions within each country. Responses were high and ranged between 90% and 96%. The survey provides information on maternal and child health for women and their children < 5 years. In most countries, measurements were taken every 5 years starting in the early 1990s 23. All DHS surveys were approved by the Opinion Research Corporation Company (ORC) Macro Institutional Review Board 23, 24, and informed consent was obtained from participants at time of interview. This study is based on the publicly available version of the data available online (http://www.measuredhs.com) with no identifiable information on survey participants.
We used data for 33 countries (Supplementary table 2). For analytical purposes we grouped the countries into four regions based on the World Bank classification by region: North Africa and West and Central Asia, South and Southeast Asia, Sub-Saharan Africa, and Latin America and the Caribbean 25. These countries were selected because they had more than one survey and included information on height and weight from women aged between 15 and 49 years between 1992 and 2009. In most waves, weight and height were only measured in women who had had at least one child during the five years prior to the survey, so that results are generalizable to this population only. We restricted our sample to eligible non-pregnant women aged 25–49 year who had valid data for weight and height in all waves (n= 270,202). We excluded women who had biologically implausible values for height or weight (n= 5,732 observations), and women with missing data for relevant covariates (n=15,545). The final sample included 248,925 women.
Outcome measures
Anthropometric measures (height and weight) were taken as part of the survey using standardized procedures by trained personnel. Weight was measured using a solar powered scale (UNICEF electronic scale or Uniscale) with a precision of 0.1 kg. Height was measured using a wooden support board to place under the scale with a precision of 0.1 cm 26. Body Mass Index (BMI) was defined as body weight in kilograms divided by the square of height in meters. The prevalence of obesity (≥ 30 kg/m2) was very low (less than 5–10%) in many of the countries included in the study. We therefore focused on overweight (cut-off of ≥ 25 kg/m2)27, as a better indicator of changes in BMI trends and a relevant outcome from a prevention perspective 28.
Independent variables
Occupational classification
All respondents were asked whether they were currently in paid employment (yes or no), and women who reported being in the labour force were asked to describe their job, which was subsequently classified based on the one-digit occupational codes from the International Standard Classification of Occupations (ISCO- 88) (http://www.ilo.org/public/english/bureau/stat/isco/index.htm) 29. We collapsed codes into the following broad categories: 1) Managerial, professional, and technical, 2) clerical workers, 3) sale workers, 4) service (domestic and household/no domestic) workers, 5) production (skilled and unskilled) workers, 6) agricultural workers and 7) not working.
Covariates
Education was measured by asking respondents their highest level of education completed, and the highest year completed at that level. We created 5 categories: level 1 (no education), level 2 (incomplete primary), level 3 (complete primary), level 4 (secondary), and level 5 (post-secondary education).
Wealth was measured based on the DHS wealth index, which combines information on household assets such as televisions and bicycles, materials used for housing construction, water access and sanitation facilities. Each household asset was assigned a weight or factor score generated through principal component analysis (PCA) of assets. The resulting asset score was standardized and collapsed into quintiles as: 1) Lowest, 2) second, 3) middle, 4) fourth and 5) highest (http://www.measuredhs.com/topics/Wealth-Index.cfm) 30. We incorporated age and urban/rural residence, parity, age at first birth, and duration of breastfeeding as potential confounders of the association between occupational class and overweight. In the majority of surveys parity was measured as total number of children ever born. We categorized this variable into three categories: One child, 2–3 children, and four or more children. Age at first birth was measured as a continuous variable in years. Months of breastfeeding was based on the most recent child born. About 56% of women reported that they were breastfeeding their last child at the time of interview, but they did not report for how many months. For these women, we assumed that they had started breastfeeding this child at the time of birth, and imputed duration of breastfeeding based on the current age of the child. We categorized values for all women into 0, 1–12, 13–24, and ≥ 25 months.
Statistical Analyses
We started by describing mean BMI and the proportion of overweight (BMI ≥ 25 kg/m2) in each country by period and region. Logistic regression analysis was then used to assess the association between overweight and occupational class. These models did not always converge due to the high prevalence of overweight in several countries. Therefore, we used binary linear regression analysis models (or Poisson regression analysis if the latter failed to converge) to obtain prevalence ratios of overweight by occupational class 31. All models controlled for age, residence, country, parity, age at first birth, breastfeeding, and year of interview. In subsequent models, we adjusted for educational level (model 2) and the household wealth index (model 3). Educational level was considered a potential confounder because it is typically completed prior to occupational achievement. In contrast, household wealth index was considered both a confounder and a potential mediator, as part of wealth is the result of occupational achievements. As expected, occupational class, educational level, and household wealth were correlated but correlations were relatively weak (Polychoric correlation coefficients between 0.27 and 0.34). We therefore placed particular emphasis on models that controlled for all covariates.
We conducted the analyses in the following steps. To provide an overall assessment of the association between occupational class and overweight, we first collapsed data across three time periods (1992–1997, 1998–2003 and 2004–2009) and presented odds ratios for each region. Subsequently, to assess trends in overweight, we incorporated year as a continuous variable in the logistic regression models. The coefficient for year in these models was interpreted as the Annual percent change (APC) in overweight. To assess whether trends differed by occupational class, we then incorporated a set of interaction terms between occupational class and year and parameterised the model to obtain APC estimates for each occupational class in each region. In addition, we predicted probabilities of overweight for each year, occupational class, and region combinations based on these models.
Analyses were carried out using SAS 9.2 (SAS, Cary, NC, USA). Significance was considered at p < 0.05, and all analyses incorporated appropriate survey sample weights.
Results
Average age was around 32 years in most regions, except in South & Southeast Asia where it was 30.5 (table 1). The proportion of women who reported to be working was highest in Sub- Saharan Africa (75.4%), followed by Latin America and the Caribbean (63.0%), South and Southeast Asia (47.1%), and North Africa/West and Central Asia (22.2%). The distribution of occupational class differed across regions. A majority of women worked in agriculture, except in Latin America and Caribbean where there was also a large proportion of women working in services and sales. The distribution of education varied largely across countries. Around half of women in Sub-Saharan Africa and South/Southeast Asia had no education, this proportion was 30.8% in North Africa/West and Central Asia, and 16% in Latin America and the Caribbean. In total, 72.7% of women in Sub-Saharan Africa and 69.1% of women in South and Southeast Asia lived in rural areas as compared with 48.8% in North Africa/West and Central Asia and 46.6% in Latin America and the Caribbean.
Table 1.
Characteristics of 248,925 women aged 25 – 49 years by region across 33 low and middle income countries participating in Demographic Health Survey in the period of 1992 to 2009.
| North Africa/West & Central Asia1 | South & Asia2 Southeast | Sub-Saharan Africa3 | Latin America & Caribbean4 | |
|---|---|---|---|---|
|
| ||||
| Characteristics | n/mean (%/SD) | n/mean (%/SD) | n/mean (%/SD) | n/mean (%/SD) |
| Age (years) | ||||
| Age at first birth | 32.00 (5.3) | 30.53 (4.9) | 32.61 (5.9) | 32.71 (5.7) |
| Currently employed | 21.63 (4.2) | 20.63 (4.1) | 19.48 (3.9) | 20.96 (4.5) |
| No | 20,058 (77.8) | 28,401 (52.9) | 28,401 (24.6) | 21,798 (37.0) |
| Yes | 5720 (22.2) | 22,864 (47.1) | 87,224 (75.4) | 37,192 (63.0) |
| Occupational class | ||||
| Professional. technical & managerial | 2381 (9.2) | 1391 (2.9) | 2997 (2.6) | 4241 (7.2) |
| Clerical | 774 (3.0) | 326 (0.7) | 924 (0.8) | 1810 (3.1) |
| Sales | 310 (1.2) | 1469 (3.0) | 24,340 (21.1) | 9576 (16.2) |
| Services | 396 (1.5) | 1245 (2.6) | 2979 (2.6) | 6214 (10.5) |
| Production | 490 (1.9) | 3254 (6.7) | 8293 (7.2) | 5987 (10.2) |
| Agriculture | 1369 (5.3) | 15,179 (31.3) | 47,691 (41.3) | 9364 (15.9) |
| Not working | 20,058 (77.8) | 25,668 (52.9) | 28,401 (24.6) | 21,798 (37.0) |
| Education | ||||
| No education | 7935 (30.8) | 23,006 (47.4) | 63,514 (54.9) | 9696 (16.4) |
| Incomplete primary | 2721 (10.6) | 6166 (12.7) | 24,224 (21.0) | 17,990 (30.5) |
| Primary | 5628 (21.8) | 12,320 (25.4) | 21,672 (18.7) | 15,776 (26.7) |
| Secondary | 6258 (24.3) | 2534 (5.2) | 4030 (3.5) | 8091 (13.7) |
| Higher | 3236 (12.6) | 4506 (9.3) | 2185 (1.9) | 7437 (12.6) |
| Wealth Index | ||||
| Lowest | 5825 (22.6) | 9433 (19.4) | 24,944 (21.6) | 14,861 (25.2) |
| Second | 5081 (19.7) | 8709 (17.9) | 23,035 (19.9) | 13,881 (23.5) |
| Middle | 5084 (19.7) | 8988 (18.5) | 22,436 (19.4) | 12,369 (21.0) |
| Fourth | 5000 (19.4) | 9876 (20.4) | 22,091 (19.1) | 10,202 (17.3) |
| Highest | 4788 (18.6) | 11,526 (23.8) | 23,119 (20.0) | 7677 (13.0) |
| Residence | ||||
| Urban | 13,197 (51.2) | 14,999 (30.9) | 31,598 (27.3) | 31,488 (53.4) |
| Rural | 12,581 (48.8) | 33,533 (69.1) | 84,027 (72.7) | 27,502 (46.6) |
| Parity (number of live births) | 4.06 (2.4) | 3.73 (2.1) | 5.11 (2.6) | 4.14 (2.6) |
| Breastfeeding (number of months) | 13.22 (8.7) | 17.76 (12.3) | 14.68 (9.4) | 13.56 (9.5) |
North Africa/West & Central Asia countries included (Egypt, Jordan, Kazakhstan, Turkey)
South & Southeast Asia countries included (Bangladesh, Cambodia, India, Nepal)
Sub-Saharan Africa countries included (Benin, Burkina Faso, Cameroon, Chad, Cote D’Ivoire, Ethiopia, Ghana, Guinea, Kenya, Madagascar, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Rwanda, Tanzania, Uganda)
Latin America & Caribbean countries included (Bolivia, Colombia, Guatemala, Haiti, Nicaragua, Peru)
Supplementary Table 1 shows trends in labour force participation and occupational distributions by region and period. Two patterns are noteworthy. First, labour force participation among women has not clearly increased in all regions. The proportion of women in the labour force in South and Southeast Asia decreased from 64.3% in 1992–97 to 46.2% in 2004–09, and from 24.3% to 19.7% in North Africa/West and Central Asia. In Sub-Saharan Africa, labour force participation increased marginally, while in Latin America there has been a large increase from 50.7% in 1992–97 to 79.6% in 2004–09. Second, the proportion of women working in agriculture has not changed in the same way in all regions. In South and Southeast Asia, the proportion of women working in agriculture declined from 52% in 1992–97 to 27.3% in 2004–09, while no consistent changes occurred in North Africa/West and Central Asia and Sub-Saharan Africa. In Latin America and the Caribbean, the proportion of women working in agriculture increased from 12.0 per cent in 1992–1997 to 18.9 per cent in 2004–2009.
The proportion of overweight ranged from 2.2% in Nepal to 75% in Egypt in the period 1992–2009 (figure 1). Overall, women in North Africa/West and Central Asia had the highest prevalence of overweight (73.7%) in the period 2004–2009, while women in South and Southeast Asia had the lowest prevalence (13.1%) in this period. In all regions the prevalence of overweight increased substantially. For example, in South/Southeast Asia it increased from 3.4% in 1992–1997 to 13.1% in 2004–2009, whereas in North Africa/West and Central Asia this increase ranged from 56.4% in 1992–1997 to 73.7% in 2004–2009. During the same period, the prevalence of underweight decreased in all regions, albeit at substantially slower pace than the increase in overweight. The prevalence of underweight changed in North Africa/West and Central Asia from 2.1% to 0.7% and in South and Southeast Asia from 30.7% to 24.6%.
Figure 1.
Prevalence of overweight (BMI ≥ 25.0 kg/m2) among women aged 25 – 49 years across 33 low and middle income countries participating in Demographic Health Survey (1992 – 2009).
Overweight prevalence increased most dramatically among women in North Africa/West in all occupational classes, and to a lesser extent in South/Southeast Asia, particularly among women working in professional, sales, services and production jobs (figure 2). Trends in Latin America and the Caribbean showed diverging patterns with increasing overweight trends among women working in agriculture, but less clear changes for women in other occupational classes. In contrast, in Sub-Saharan Africa, the prevalence of overweight has remained rather stable at relatively lower levels than other regions in all occupational groups, except for modest increases in women working in Clerical, Sales and Production occupations.
Figure 2. Trends in the prevalence of Overweight (BMI ≥25.0 kg/m2) among women aged 25 – 49 years by occupational class and region across 33 low and middle income countries participating in Demographic Health Survey (1992 – 2009).
Variables included in the model: age, region, period, occupational class, interaction term region* time period *occupational class.
North Africa/West & Central Asia countries included (Egypt, Jordan, Kazakhstan, Turkey)
South & Southeast Asia countries included (Bangladesh, Cambodia, India, Nepal)
Sub-Saharan Africa countries included (Benin, Burkina Faso, Cameroon, Chad, Cote D’Ivoire, Ethiopia, Ghana, Guinea, Kenya, Madagascar, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Rwanda, Tanzania, Uganda)
Latin America & Caribbean countries included (Bolivia, Colombia, Guatemala, Haiti, Nicaragua, Peru)
In all regions, living in urban areas was associated with significantly higher risk of overweight as compared to living in rural areas (table 2). Having more children was generally associated with overweight, except in South & Southeast Asia where women with ≥4 children had the lowest risk. Age at first birth, education, wealth, and breastfeeding were associated with overweight, but associations differed across regions. Changes in breastfeeding patterns over time did not influence the observed trends in overweight, and do not explain occupational differences in trends. In South and Southeast Asia and Sub-Saharan Africa, women with lower education and wealth had considerable less overweight than women with higher education and wealth.
Table 2.
Odds ratio estimates of the association of overweight (BMI ≥ 25.0 kg/m2) with age, residence, parity, age at first birth, breastfeeding, education and wealth among women aged 25 – 49 years by region across 33 low and middle income countries participating in Demographic Health Survey (1992 – 2009).
| Models | North Africa/West & Central Asia1 OR* (95% CI) | South & Southeast Asia2 OR* (95% CI) | Sub-Saharan Africa3 OR* (95% CI) | Latin America & Caribbean4 OR* (95% CI) |
|---|---|---|---|---|
| Age | 1.07 (1.07– 1.08) | 1.06 (1.05– 1.07) | 1.04 (1.04– 1.04) | 1.05 (1.04– 1.05) |
| Residence | ||||
| Urban | 1.48 (1.39– 1.58) | 2.90 (2.69– 3.12) | 3.00 (2.90– 3.12) | 1.85 (1.78– 1.93) |
| Rural | Reference | |||
| Parity | ||||
| 1 child | Reference | |||
| 2–3 Children | 1.14 (1.03– 1.25) | 1.03 (0.93– 1.14) | 1.21 (1.12– 1.31) | 1.39 (1.31– 1.47) |
| > 4 Children | 1.09 (0.98– 1.22) | 0.62 (0.55– 0.71) | 1.25 (1.16– 1.36) | 1.34 (1.25– 1.43) |
| Age at first birth | 0.99 (0.98– 0.99) | 1.02 (1.01– 1.02) | 0.98 (0.98– 0.98) | 0.96 (0.96– 0.97) |
| Breastfeeding | ||||
| Never breastfeeding | 1.03 (0.92– 1.16) | 1.56 (1.33– 1.82) | 1.23 (1.13– 1.33) | 1.33 (1.22– 1.44) |
| 1– 12 months | 0.93 (0.88– 0.99) | 1.05 (0.97– 1.14) | 1.10 (1.06– 1.14) | 1.11 (1.07– 1.15) |
| 13– 24 months** | Reference | |||
| > 25 months | 0.86 (0.76– 0.98) | 1.01 (0.92– 1.11) | 0.83 (0.78– 0.89) | 1.23 (1.16– 1.31) |
| Education | ||||
| No education | 0.87 (0.79– 0.96) | 0.13 (0.12– 0.15) | 0.20 (0.18– 0.22) | 0.63 (0.59– 0.68) |
| Incomplete Primary | 1.13 (1.01– 1.27) | 0.25 (0.22– 0.29) | 0.31 (0.28– 0.34) | 1.06 (1.00– 1.13) |
| Primary | 1.23 (1.12– 1.35) | 0.49 (0.45– 0.54) | 0.52 (0.47– 0.55) | 1.25 (1.18– 1.32) |
| Secondary | 1.28 (1.17– 1.41) | 0.84 (0.74– 0.97) | 0.74 (0.66– 0.83) | 1.20 (1.13– 1.28) |
| Higher | Reference | |||
| Wealth | ||||
| Lowest | 0.31 (0.29– 0.35) | 0.10 (0.08– 0.12) | 0.29 (0.27– 0.31) | 0.48 (0.45– 0.52) |
| Second | 0.50 (0.45– 0.55) | 0.15 (0.13– 0.17) | 0.34 (0.32– 0.37) | 0.67 (0.63– 0.72) |
| Middle | 0.62 (0.56– 0.68) | 0.27 (0.24– 0.31) | 0.47 (0.44– 0.50) | 0.82 (0.78– 0.87) |
| Fourth | 0.84 (0.76– 0.92) | 0.47 (0.42– 0.51) | 0.62 (0.59– 0.65) | 0.96 (0.91– 1.02) |
| Highest | Reference | |||
All models were adjusted for age, residence, educational level and time period.
Since most women were breastfeeding between 13 to 24 months, this category was selected as reference in order to obtain precise estimates in the model.
North Africa/West & Central Asia countries included (Egypt, Jordan, Kazakhstan, Turkey)
South & Southeast Asia countries included (Bangladesh, Cambodia, India, Nepal)
Sub-Saharan Africa countries included (Benin, Burkina Faso, Cameroon, Chad, Cote D’Ivoire, Ethiopia, Ghana, Guinea, Kenya, Madagascar, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Rwanda, Tanzania, Uganda)
Latin America & Caribbean countries included (Bolivia, Colombia, Guatemala, Haiti, Nicaragua, Peru)
Table 3 shows odds ratios of the associations between overweight and occupational class. Controlling for basic demographics (model 1), in all regions, women working in agriculture had less often overweight than women working in other occupations, followed by women working in production. In contrast, women who worked in non-manual occupational classes tended to have the highest overweight prevalence. This association was substantially attenuated after controlling for education, but remained statistically significant in all regions except North Africa/West and Central Asia. The association was further attenuated but also remained significant after controlling for household wealth.
Table 3.
Odds ratios estimates of the association of overweight (BMI ≥ 25.0 kg/m2) with occupational classes among women aged 25 – 49 years by region across 33 low and middle income countries participating in Demographic Health Survey (1992– 2009).
| Model | North Africa/West & Central Asia1. OR (95%CI) | South & Southeast Asia2. OR (95%CI) | Sub-Saharan Africa3. OR (95%CI) | Latin America & Caribbean4. OR (95%CI) |
|---|---|---|---|---|
| Model 1 | ||||
| Professional, technical & managerial | 1.10 (0.94– 1.29) | 4.21 (3.46– 5.11) | 3.65 (3.34– 3.99) | 1.54 (1.41– 1.68) |
| Clerical | 1.44 (1.15– 1.79) | 6.22 (4.44– 8.73) | 3.26 (2.82– 3.78) | 1.26 (1.13– 1.41) |
| Sales | 1.26 (0.97– 1.64) | 3.37 (2.76– 4.12) | 2.15 (2.04– 2.26) | 1.83 (1.71– 1.97) |
| Services | 1.24 (0.96– 1.61) | 2.47 (1.97– 3.09) | 2.20 (1.99– 2.43) | 1.61 (1.49– 1.74) |
| Production | 0.94 (0.75– 1.18) | 1.75 (1.46– 2.10) | 1.60 (1.49– 1.71) | 1.59 (1.47– 1.72) |
| Not working | 0.84 (0.65– 1.07) | 3.81 (3.05– 4.76) | 1.63 (1.47– 1.82) | 1.66 (1.53– 1.80) |
| Agriculture | Reference | |||
| Model 2 | ||||
| Professional, technical & managerial | 0.99 (0.84– 1.17) | 2.10 (1.71– 2.58) | 2.18 (1.98– 2.13) | 1.66 (1.51– 1.83) |
| Clerical | 1.11 (0.89– 1.40) | 3.56 (2.52– 5.01) | 2.06 (1.77– 2.39) | 1.33 (1.19– 1.50) |
| Sales | 1.23 (0.94– 1.61) | 2.83 (2.31– 3.46) | 2.03 (1.93– 2.13) | 1.82 (1.70– 1.96) |
| Services | 1.15 (0.89– 1.49) | 2.32 (1.85– 2.91) | 1.90 (1.72– 2.10) | 1.59 (1.47– 1.72) |
| Production | 0.88 (0.70– 1.11) | 1.66 (1.38– 1.99) | 1.49 (1.39– 1.60) | 1.58 (1.46– 1.70) |
| Not working | 0.77 (0.61– 0.99) | 3.01 (2.40– 3.78) | 1.56 (1.40– 1.74) | 1.67 (1.54– 1.81) |
| Agriculture | Reference | |||
| Model 3 | ||||
| Professional, technical & managerial | 0.84 (0.71– 1.00) | 1.50 (1.22– 1.84) | 1.75 (1.58– 1.94) | 1.50 (1.36– 1.65) |
| Clerical | 0.95 (0.75– 1.19) | 2.47 (1.75– 3.48) | 1.58 (1.36– 1.84) | 1.14 (1.02– 1.29) |
| Sales | 1.13 (0.87– 1.48) | 1.94 (1.58– 2.39) | 1.62 (1.54– 1.71) | 1.67 (1.55– 1.79) |
| Services | 1.06 (0.82– 1.37) | 1.75 (1.39– 2.20) | 1.47 (1.33– 1.63) | 1.51 (1.40– 1.64) |
| Production | 0.80 (0.64– 1.00) | 1.33 (1.11– 1.61) | 1.23 (1.15– 1.32) | 1.44 (1.33– 1.56) |
| Not working | 0.69 (0.54– 0.88) | 2.06 (1.64– 2.60) | 1.29 (1.16– 1.44) | 1.54 (1.42– 1.67) |
| Agriculture | Reference | |||
Variables included in the model 1: age, residence, country, parity, age at first birth, breastfeeding, time period, occupational class.
Variables included in the model 2: variables of model 1 + educational level.
Variables included in the model 3: variables of model 2 + household wealth.
North Africa/West & Central Asia countries included (Egypt, Jordan, Kazakhstan, Turkey)
South & Southeast Asia countries included (Bangladesh, Cambodia, India, Nepal)
Sub-Saharan Africa countries included (Benin, Burkina Faso, Cameroon, Chad, Cote D’Ivoire, Ethiopia, Ghana, Guinea, Kenya, Madagascar, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Rwanda, Tanzania, Uganda)
Latin America & Caribbean countries included (Bolivia, Colombia, Guatemala, Haiti, Nicaragua, Peru)
Figure 3 shows the annual percent change in the prevalence of overweight for each occupational class. In all regions, overweight has increased across most occupational classes. However, in Latin American and the Caribbean and Sub-Saharan Africa, increases in overweight have been significantly larger among women working in agriculture and production as well as women out of the labour force as compared to women working in professional, technical, managerial, and clerical occupations.
Figure 3. Annual percent change in the odds of overweight (BMI ≥25.0 kg/m2) among women aged 25– 49 years by occupational class and region across 33 low and middle income countries participating in Demographic Health Survey (1992 – 2009).
Variables included in the model: age, household wealth, residence, country, educational level, parity, occupational class, age at first birth, months of breastfeeding, interaction term year of survey*occupational class.
North Africa/West & Central Asia countries included (Egypt, Jordan, Kazakhstan, Turkey)
South & Southeast Asia countries included (Bangladesh, Cambodia, India, Nepal)
Sub-Saharan Africa countries included (Benin, Burkina Faso, Cameroon, Chad, Cote D’Ivoire, Ethiopia, Ghana, Guinea, Kenya, Madagascar, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Rwanda, Tanzania, Uganda)
Latin America & Caribbean countries included (Bolivia, Colombia, Guatemala, Haiti, Nicaragua, Peru)
Discussion
Our results indicate that women working in agricultural occupations have a lower overweight prevalence than women working in higher occupational classes. However, there is a tendency towards larger increases in overweight in the agricultural group than among women working in services, clerical, and professional groups. The results point at the possible role of changes in industrialization, technology and the nature of agricultural work, and raise questions about the prevention strategies that may be required to prevent rising overweight and obesity trends among working women.
Comparisons with previous findings
To our knowledge, this is the first comparative study to assess trends in overweight by occupational class in a large number of low - and middle -income countries. Consistent with our results, previous studies in developing countries suggested that women classified as agricultural workers have a lower prevalence of overweight than non-agricultural workers16, 21, 32, 33. Some studies have found that agriculture activities have significantly higher energy expenditures and are associated with lower weight gain32, 33. One study in six countries in the Asian- Pacific region found that the lowest educated and wealthy populations had the highest levels of occupational physical activity and active commuting16. For example, farmers from Jamaica walk 60% more than their counterparts in urban areas 33. Consistently, studies in developed countries suggest that jobs characterized by spending more than 50% of the working day sitting can increase weight gain substantially34, 35. A study showed that sedentary employees often have an energy imbalance, characterized by increased levels of glucocorticoids such as insulin and cortisol, which may be a predisposing factor for abdominal obesity 36.
Previous studies suggest that overweight and obesity have increased more among women with lower education and occupational class17, 18, 20. Consistent with these findings, our results show that in Latin American and the Caribbean and Sub-Saharan Africa, women working in agriculture and production have had larger increases in overweight than women working in higher occupational classes. This pattern may partly be explained by a faster adoption of healthy behaviours such as greater intake of fruits and vegetables and increased leisure physical activity17, 18, 20 among higher educated individuals.
Explanation of results
Our results show that overweight is higher among professionals and white collar occupations, while production and agricultural workers generally have lower overweight. In developing countries, agriculture is considered as a vulnerable employment characterized by low pay and strenuous working conditions 14. In many countries women have less education and work experience than men, making them more vulnerable for less favorable forms of employment15. Women in agricultural occupations may enjoy less disposable income is for food consumption 37, which may paradoxically contribute to maintain relatively low levels of fat consumption and overweight. Furthermore, evidence suggests that working in agriculture, production, and other blue collar occupations involves more physical activity and energy expenditure than other jobs 35, 38, 39. For example, a study reported that the average number of daily steps was significantly higher (8,757 steps per day) among blue collar workers than professionals and other white collar workers (2,835 steps per day) 39. In addition, agriculture workers living in rural areas may have more access to fruits and vegetables, whole grains and non-hydrogenated fat than white-collar workers generally concentrated in urban areas37.
Despite the general increase in overweight in most occupations, we found larger increases in overweight prevalence over time in production and agricultural groups. This pattern does not seem to be entirely explained by selective transition of agricultural workers into other occupations, as the proportion of women working in agriculture changed little over the study period in most regions. A possible explanation is that physical activity associated with agricultural work has declined due to industrialization and mechanization in the agricultural sector. Improvements in community infrastructure and technological changes are associated with declines in physical activity22, 36, 40, 41. For example, in China improvements in community infrastructures and services explained 40% of the decline in total physical activity among women41. Furthermore, a study showed that from the 1960s onwards India has experienced a major increase of modern farm machines such as tractors and pump sets, which have grown by a factor of 56 and 22 in the period 1962–2006, respectively. This mechanization of agriculture may partly explain the increasing prevalence of overweight and obesity among agricultural workers22.
Why has overweight increased across all occupational classes in most regions? A possible explanation for the increase in overweight is that the prevalence of underweight has decreased. In sensitivity analyses, we found that parallel to increases in overweight, underweight prevalence has decreased, but the magnitude of this decrease is much smaller and does not account for the increase in overweight. In addition, most countries in Asia, Latin America, Northern Africa, Middle East and in urban areas in sub-Saharan Africa have experienced a shift in their dietary patterns. This includes a large increase in the consumption of fat and sugars10, 40, paralleled by a decline in total cereal intake and fiber. For example, in India, the traditional coarse grains such as sorghum, millet, and corn, have been replaced by simple carbohydrates such as rice, as they are easier to cook than other foods 22. Furthermore, agriculture policies have increased food production and exports, which eventually have contributed to higher consumption of edible oil and animal foods such as red meat 4, 42. To illustrate, in Egypt, average per capita consumption of protein from meat increased from 16.3 g/person/day in 1981 to 25.5g/person/day in 2000 43. These patterns are likely to have influenced dietary patterns of most women in low and middle income countries, although our study shows that the magnitude of these effects differ across occupations.
Limitations
To our knowledge, this is the first study to assess trends in overweight by occupational class in several low and middle income countries. However, several limitations should be considered. Although DHS surveys provide comparable data for many countries, our results are limited to women aged 25 to 49 who had recently had children, and are not generalizable to all women of reproductive age. However, among all women aged 25–49 in countries participating in DHS, the proportion having at least one fertile pregnancy ranged from 93.6 in the Central African Republic to 99% in Bolivia, suggesting that our results cover most women in this age range 44. Similarly in our data 56% of women were breastfeeding and this might have a potential direct impact on women’s weight. Nevertheless, in sensitivity analysis, we found that the pattern among women who were not breastfeeding (excluding the 56% who was breastfeeding at interview time) was almost identical to that for the entire sample. In addition, DHS is a repeated cross-sectional study, and it does not follow participants over time. Therefore, assessments of individual weight change and their associated determinants were not possible.
Some studies have reported a weaker association between BMI, percentage of body fat, and health risks in Asian countries as compared to other populations, suggesting that different cut-off points for overweight and obesity should be used in this region45–47. Nonetheless, we followed the WHO expert consultation that advises that BMI cut-off points of 25.0 kg/m2 to 29.9 kg/m2 for overweight and ≥ 30.0 kg/m2 for obesity be used in international comparative studies including Asian countries 48. BMI cut-off points in Asian populations have not been well defined and vary from 22 kg/m2 to 25 kg/m2 for overweight, 26 kg/m2 to 31 kg/m2 for obesity 48, 49. Nevertheless, in a sensitivity analysis, we found that trends in overweight in the Asian region based on these cut-off points followed a very similar pattern as those described in our study.
Misclassification of educational level and occupational class is another potential source of bias. Individuals may report education differently across countries, and reclassification of national levels into an internationally comparable classification may not be accurate. Nevertheless, we used broad categories instead of very specific levels to diminish the potential bias associated with misclassification. Occupational class may also suffer from misreporting, and the same occupational category may be interpreted differently among women in different countries. Direct comparisons of estimates across regions and countries were therefore not possible. However, our study focused on generic patterns of prevalence and trends in overweight within countries across occupational classes, diminishing any bias introduced by cross-national differences in occupational classifications.
Conclusion
In conclusion, our study showed that women working in agricultural occupations have less often overweight than women working in higher occupational classes. The prevalence of overweight has increased in all regions, but women working in agriculture and production have experienced larger increases than women in higher occupational classes. These trends may result from a combination of reduced physical activity at work due to changes in mechanisation and technology affecting women working in agriculture and production, together with changes in dietary habits that have affected women from all occupational classes. Findings highlight the need to target overweight and obesity prevention efforts towards women from all occupational classes. While women in higher occupations have a higher prevalence of overweight, the epidemic is rapidly extending towards women working in agricultural and production sectors. Overweight and obesity prevention efforts are therefore likely to benefit women from all occupational groups in low- and middle-income countries.
Supplementary Material
Acknowledgments
Sandra Lopez-Arana was supported by the European Union Erasmus Mundus Partnerships program Erasmus-Columbus (ERACOL) at Erasmus MC in the Netherlands. Mauricio Avendano was supported by a Starting Researcher grant from the European Research Council (ERC) (grant No 263684), a fellowship from Erasmus University Rotterdam and a grant from the National Institute of Ageing (R01AG037398).
Footnotes
Conflict of interest
The authors declare no conflict of interest.
References
- 1.WHO. Global Status Report on Noncommunicable Diseases 2010. World Health Organization; Geneva: 2011. [Google Scholar]
- 2.WHO. Global Health Risks. Mortality and burden of disease attributable to selected major risks. World Health Organization; Geneva: 2009. [Google Scholar]
- 3.Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ, et al. National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country- years and 9. 1 million participants. Lancet. 2011;377:557–567. doi: 10.1016/S0140-6736(10)62037-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Popkin BM, Adair LS, Ng SW. Global nutrition transition and the pandemic of obesity in developing countries. Nutr Rev. 2012;70:3–21. doi: 10.1111/j.1753-4887.2011.00456.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mendez MA, Monteiro CA, Popkin BM. Overweight exceeds underweight among women in most developing countries. Am J Clin Nutr. 2005;81:714–721. doi: 10.1093/ajcn/81.3.714. [DOI] [PubMed] [Google Scholar]
- 6.Razak F, Corsi DJ, Sv S. Change in the body mass index distribution for women: analysis of surveys from 37 low- and middle-income countries. PLoS medicine. 2013;10 (1):e1001367. doi: 10.1371/journal.pmed.1001367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Balarajan Y, Villamor E. Nationally Representative Surveys Show Recent Increases in the Prevalence of Overweight and Obesity among Women of Reproductive Age in Bangladesh, Nepal, and India. J Nutr. 2009;139:2139–2144. doi: 10.3945/jn.109.112029. [DOI] [PubMed] [Google Scholar]
- 8.Filozof C, Gonzalez C, Sereday M, Mazza C, Braguinsky J. Obesity prevalence and trends in Latin-American countries. Obes Rev. 2001;2:99–106. doi: 10.1046/j.1467-789x.2001.00029.x. [DOI] [PubMed] [Google Scholar]
- 9.Dodzin S, Vamvakidis A. Trade and industrialization in developing economies. J Dev Econ. 2004;75:319–328. [Google Scholar]
- 10.Popkin BM. The Nutrition Transition and Obesity in the Developing World. J Nutr. 2001;131:871S–873S. doi: 10.1093/jn/131.3.871S. [DOI] [PubMed] [Google Scholar]
- 11.Treiman DJ. Industrialization and Social Stratification. Sociol Inq. 1970;40:207–234. [Google Scholar]
- 12.UN. Completing the Fertility Transition. 48/49. UN/PB; 2009. pp. 1–507. [Google Scholar]
- 13.Cáceres-Delpiano J. Can We Still Learn Something From the Relationship Between Fertility and Mother’s Employment? Evidence From Developing Countries. Demography. 2012;49:151–174. doi: 10.1007/s13524-011-0076-6. [DOI] [PubMed] [Google Scholar]
- 14.ILO. Global Employment Trends 2012. Preventing a deeper job crisis. International Labour Office; Geneva: 2012. [Google Scholar]
- 15.FAO. Closing the Gender Gap for Development. Food and Agriculture Organization of the United Nations; Rome: 2011. The State of Food and Agriculture. Women in Agriculture. [Google Scholar]
- 16.Bauman A, Ma G, Cuevas F, Omar Z, Waqanivalu T, Phongsavan P, et al. Cross- national comparisons of socioeconomic differences in the prevalence of leisure-time and occupational physical activity, and active commuting in six Asia-Pacific countries. J Epidemiol Community Health. 2011;65:35–43. doi: 10.1136/jech.2008.086710. [DOI] [PubMed] [Google Scholar]
- 17.Allman-Farinelli MA, Chey T, Merom D, Bauman AE. Occupational risk of overweight and obesity: an analysis of the Australian Health Survey. J Occup Med Toxicol. 2010;5:14–23. doi: 10.1186/1745-6673-5-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Galobardes B, Morabia A, Bernstein MS. The Differential Effect of Education and Occupation on Body Mass and Overweight in a Sample of Working People of the General Population. Ann Epidemiol. 2000;10:532–537. doi: 10.1016/s1047-2797(00)00075-2. [DOI] [PubMed] [Google Scholar]
- 19.Church TS, Thomas DM, Tudor-Locke C, Katzmarzyk PT, Earnest CP, Rodarte RQ, et al. Trends over 5 decades in U.S. occupation-related physical activity and their associations with obesity. PLoS One. 2011;6:1–7. doi: 10.1371/journal.pone.0019657. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pigeyre M, Dauchet L, Simon C, Bongard V, Bingham A, Arveiler D, et al. Effects of occupational and educational changes on obesity trends in France: The results of the MONICA-France survey 1986–2006. Prev Med. 2011;52:305–309. doi: 10.1016/j.ypmed.2011.02.004. [DOI] [PubMed] [Google Scholar]
- 21.Gregory CO, Dai J, Ramirez-Zea M, Stein AD. Occupation Is More Important than Rural or Urban Residence in Explaining the Prevalence of Metabolic and Cardiovascular Disease Risk in Guatemalan Adults. J Nutr. 2007;137:1314–1319. doi: 10.1093/jn/137.5.1314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Sarkar A, Aronson KJ, Patil S, Hugar LB, vanLoon GW. Emerging health risks associated with modern agriculture practices: A comprehensive study in India. Environ Res. 2012;115:37–50. doi: 10.1016/j.envres.2012.03.005. [DOI] [PubMed] [Google Scholar]
- 23.Rutstein SO, Rojas G. Guide to DHS statistics. Calverton, MD: ORC Macro; 2006. [Google Scholar]
- 24.ICF I. Survey Organization Manual for Demographic and Health Surveys. ICF International; Calverton. Maryland: 2012. [Google Scholar]
- 25.World Bank. World Trade Indicators (July 2009-July 2010) 2012 accessed at http://data.worldbank.org/about/country-classifications.
- 26.ORC M. Interviewer’s Manual for Use with Model “A” Questionnaire for High Contraceptive Prevalence Countries. ORC Macro Calverton; Maryland, USA: 2002. [Google Scholar]
- 27.WHO. Physical status: the use and interpretation of anthropometry. World Health Organization; Geneva: 1995. [PubMed] [Google Scholar]
- 28.Neuman M, Finlay JE, Davey Smith G, Subramanian SV. The poor stay thinner: stable socioeconomic gradients in BMI among women in lower- and middle-income countries. Am J Clin Nutr. 2011;94:1348–1357. doi: 10.3945/ajcn.111.018127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.ILO. International Standard Classification of Occupations (ISCO- 88) International Labor Organization; Geneva: 1990. [Google Scholar]
- 30.Rutstein SO, Johnson K. The DHS Wealth Index. Calverton, MD: ORC Macro; 2004. [Google Scholar]
- 31.Zou G. A Modified Poisson Regression Approach to Prospective Studies with Binary Data. Am J Epidemiol. 2004;159:702–706. doi: 10.1093/aje/kwh090. [DOI] [PubMed] [Google Scholar]
- 32.Keighley ED, McGarvey ST, Turituri P, Viali S. Farming and adiposity in Samoan adults. Am J Hum Biol. 2006;18:112–122. doi: 10.1002/ajhb.20469. [DOI] [PubMed] [Google Scholar]
- 33.Levine JA, McCrady SK, Boyne S, Smith J, Cargill K, Forrester T. Non-exercise physical activity in agricultural and urban people. Urban Stud. 2011;48:2417–2427. doi: 10.1177/0042098010379273. [DOI] [PubMed] [Google Scholar]
- 34.Bockerman P, Johansson E, Jousilahti P, Uutela A. The physical strenuousness of work is slightly associated with an upward trend in the BMI. Soc Sci Med. 2008;66:1346–1355. doi: 10.1016/j.socscimed.2007.12.015. [DOI] [PubMed] [Google Scholar]
- 35.Choi B, Schnall PL, Yang H, Dobson M, Landsbergis P, Israel L, et al. Sedentary work, low physical job demand, and obesity in US workers. Am J Ind Med. 2010;53:1088–1101. doi: 10.1002/ajim.20886. [DOI] [PubMed] [Google Scholar]
- 36.Huneault L, Mathieu MÈ, Tremblay A. Globalization and modernization: an obesogenic combination. Obes Rev. 2011;12:e64–e72. doi: 10.1111/j.1467-789X.2010.00817.x. [DOI] [PubMed] [Google Scholar]
- 37.Mishra CP. Nexus of poverty, energy balance and health. Indian J Community Med. 2012;37:71–78. doi: 10.4103/0970-0218.96083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Brun TA, Geissler CA, Mirbagheri I, Hormozdiary H, Bastani J, Hedayat H. The energy expenditure of Iranian agricultural workers. Am J Clin Nutr. 1979;32:2154–2161. doi: 10.1093/ajcn/32.10.2154. [DOI] [PubMed] [Google Scholar]
- 39.Steele R, Mummery K. Occupational physical activity across occupational categories. J Sci Med Sport. 2003;6:398–407. doi: 10.1016/s1440-2440(03)80266-9. [DOI] [PubMed] [Google Scholar]
- 40.Monda KL, Gordon-Larsen P, Stevens J, Popkin BM. China’s transition: The effect of rapid urbanization on adult occupational physical activity. Soc Sci Med. 2007;64:858–870. doi: 10.1016/j.socscimed.2006.10.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Ng SW, Norton EC, Popkin BM. Why have physical activity levels declined among Chinese adults? Findings from the 1991–2006 China health and nutrition surveys. Soc Sci Med. 2009;68:1305–1314. doi: 10.1016/j.socscimed.2009.01.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Hawkes C, Ruel M. The links between agriculture and health: an intersectoral opportunity to improve the health and livelihoods of the poor. Bull World Health Organ. 2006;84:984–990. doi: 10.2471/blt.05.025650. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Mehio Sibai A, Nasreddine L, Mokdad AH, Adra N, Tabet M, Hwalla N. Nutrition Transition and Cardiovascular Disease Risk Factors in Middle East and North Africa Countries: Reviewing the Evidence. Ann Nutr Metab. 2010;57:193–203. doi: 10.1159/000321527. [DOI] [PubMed] [Google Scholar]
- 44.Rutstein SO, Shah IH. Infecundity, Infertility, and Childlessness in Developing Countries. ORC Macro and the World Health Organization; Calverton, Maryland: 2004. [Google Scholar]
- 45.Pan WH, Flegal KM, Chang HY, Yeh WT, Yeh CJ, Lee WC. Body mass index and obesity-related metabolic disorders in Taiwanese and US whites and blacks: implications for definitions of overweight and obesity for Asians. Am J Clin Nutr. 2004;79(1):31–9. doi: 10.1093/ajcn/79.1.31. [DOI] [PubMed] [Google Scholar]
- 46.Rush EC, Freitas I, Plank LD. Body size, body composition and fat distribution: comparative analysis of European, Maori, Pacific Island and Asian Indian adults. Br J Nutr. 2009;102(4):632–41. doi: 10.1017/S0007114508207221. [DOI] [PubMed] [Google Scholar]
- 47.Wen CP, David Cheng TY, Tsai SP, Chan HT, Hsu HL, Hsu CC, et al. Are Asians at greater mortality risks for being overweight than Caucasians? Redefining obesity for Asians. Public Health Nutr. 2009;12(4):497–506. doi: 10.1017/S1368980008002802. [DOI] [PubMed] [Google Scholar]
- 48.WHO EC. Appropiate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363:157–163. doi: 10.1016/S0140-6736(03)15268-3. [DOI] [PubMed] [Google Scholar]
- 49.Low S, Chin MC, Ma S, Heng D, Deurenberg-Yap M. Rationale for redefining obesity in Asians. Ann Acad Med Singapore. 2009;38(1):66–9. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.