Research | Open | Published:
Contextual influences affecting patterns of overweight and obesity among university students: a 50 universities population-based study in China
International Journal of Health Geographicsvolume 16, Article number: 18 (2017)
Many studies have examined childhood and adolescent obesity, but few have examined young adults and the effect of their home and current living environments on prevalence rates. The present study explores contextual factors affecting overweight and obesity among university students in China and, in particular, focuses on how the SES–obesity relationship varies across different geographical contexts.
Participants were 11,673 students, who were identified through a multistage survey sampling process conducted in 50 universities. Individual data was obtained through a self-administered questionnaire, and contextual variables were retrieved from a national database. Multilevel logistic regression models were used to examine urban and regional variations in overweight and obesity.
Overall the prevalence of overweight and obesity in the study sample was 9.5% (95% CI 7.7, 11.3%). After controlling for individual factors, both attributes of the home location (regional GDP <gross domestic product> per capita and rurality) and the current university location (city population) were found to be important, thus suggesting that the different origins of students affect current levels of obesity. At the individual level, while students with more financial resources were more likely to be obese, the extent of this relationship was highly dependent upon area income and city size.
The results of this study add important insights about the role of contextual factors affecting overweight and obesity among young adults and indicate a need to take into account both past as well as present environmental influences when considering the role of contextual factors in models of the nutrition transition.
The prevalence of overweight and obesity is increasingly evident in both richer and poorer countries . Although greater attention has been paid to environmental determinants of obesity in recent years , this research has largely occurred in western countries and focused on various neighborhood factors affecting obesity prevalence . In low and middle income countries, while the impact of the nutrition revolution has been noticed for some time [4, 5], less attention has been paid to the independent effect of contextual risk factors affecting overweight and obesity and how these may differ from those of richer countries. However, the studies which have occurred have largely focused on factors affecting national variations in obesity  and variations in the impact of socio-economic status (SES)  and other contributing factors . While others have explored urban and regional differences in obesity at the sub-national scale , many of these studies have not always adequately controlled for individual level factors . In some cases regional variations in obesity are largely seen as an outcome of individual level differences  or where independent macro-level effects have been identified, with few exceptions [11, 12], often these have been unspecified [13, 14]. Thus, in view of such trends, it is important to pay greater attention to the significance of various environmental determinants of obesity and why these may be important in countries at earlier stages in the nutrition revolution. The objective of this study, therefore, is to investigate the effects of both the home (region) and current (university city) living contexts on obesity among university students in China.
Because of their high mobility rates, studies of obesity among young adults provide an opportunity to simultaneously examine the effect of a variety of contextual effects, characteristic of their home and current locations, which may contribute to obesity . Current patterns of obesity most likely reflect different cultural and behavioral norms relating to the home locations of students  as well as the socio-economic and other characteristics of the environments where they now reside. A frequent criticism of contextual studies of health is that they are cross-sectional in nature and do not take account of prior environmental conditions that people have been exposed to. While there have been many studies of childhood and adolescent obesity , there has been less focus on the importance of earlier life conditions on current levels of obesity. For example, in the United States, Zheng and Tumin  found that women’s obesity status at older ages was influenced by early childhood conditions and place of residence, while adulthood factors seemed to be more important for males. Among the few studies of younger adults the evidence suggests that, for some groups (e.g. African Americans) neighborhood deprivation clearly plays a role in later patterns of obesity . Similarly in Denmark, birthplace played a role in explaining regional differences in the prevalence of obesity. Young men currently living in provincial rural areas surrounding Copenhagen had a greater risk of obesity, especially if their birthplaces were also rural .
In low and middle income countries attending university may also increase the risks of obesity . Since more affluent students are most likely to attend university, higher rates of obesity are likely to be found among this group , especially among rural dwellers migrating to more urbanised places . However, the strength of the socio-economic status (SES)–obesity relationship is likely to be context dependent. As Jin and Lu  have noted, with the exception of cross-national studies , most of the existing studies on the relationship between SES and obesity have ignored spatial variations in the nature of this relationship. This has been particularly evident in studies within particular countries where the factors producing obesogenic environments, and hence the nature and strength of the SES–obesity relationship, are likely to vary over geographic space. Thus it might be expected that more affluent students originating from higher income regions or who are currently studying in more urban and economically developed environments will be most at risk, because exposure to obesogenic factors is likely to be greater in such places .
While there have been numerous regional studies of obesity [10, 14, 23, 24] there have been few multi-level approaches [13, 25] which have examined the independent influence of city or regional contextual factors on obesity among young adults. The few studies which have occurred have largely focused on children and adolescents, usually at the local neighborhood level [17, 26]. While neighborhood effects are important, so too are influences which operate at other spatial scales. These may be levels of urbanization or area income differences, both of which are likely to be related to the greater availability of energy dense foods or reduced daily physical activity . In addition the effects of income inequality are likely to be greater in such places and thus should strengthen the relationship between overconsumption among the rich and food insecurity among the poor . Despite the importance of macro-level variables a recent review of contextual determinants of obesity paid little attention to such factors .
In the light of the preceding comments this paper poses two questions:
Independent of individual characteristics, what contextual attributes of a student’s home location are important in explaining current patterns of overweight and obesity and do these remain significant when taking into account attributes of the student’s current university city location?
Given the well-known link between individual SES and increased obesity in low and middle income countries, to what extent is this evident for university students and does the strength of this relationship vary across different geographic contexts?
To answer these questions the rest of the paper is organized as follows. First we outline the methodology of the study. This is followed by the results and a discussion of the most important findings, placing them in a wider context of international research on obesity and other Chinese studies. We conclude by emphasizing some of the wider theoretical and policy implications of our study.
This study reports data from the Global Health Professions Student Survey (GHPSS) on Tobacco Control in China GHPSS (Extended version). The GHPSS is part of the Global Tobacco Surveillance System and is a university-based survey. The GHPSS was initially completed in 31 countries between 2005 and 2007. In China the GHPSS provided a valuable source of information on health and related health behaviours, including obesity.
The study employed a multistage sampling design and collected the sample in 2013. In Stage 1, 50 universities with medical programs were selected from 42 cities across China and differentiated by regional location (see Fig. 1). Stage 2 of the sampling strategy involved the selection of classes within each university, and all students in these classes were eligible as the sampling frame. A more detailed description of the survey and the data can be found in Yang et al. . The study was approved by the Ethics Committee at the Medical Center, Zhejiang University, and verbal consent was obtained from all participants prior to data collection.
Body mass index (BMI) was calculated by dividing body weight (kg) by squared height (m2). Overweight and obesity were defined as recommended by guidelines for the prevention and control of chronic diseases in China (Department of Disease Control and Prevention ): individuals with BMI scores of 24.0–27.9 kg/m2 were categorized as overweight, and those with scores of ≥28.0 kg/m2 were categorized as obese, which is the national standard . As the prevalence of obesity was low in this sample of university students, the categories of overweight and obesity were combined in the analyses.
Height and weight were measured by self-report. Given the potential problems of this measure , we objectively measured height and weight among 170 subjects from two universities in Hangzhou in order to validate the self-reported prevalence of overweight and obesity and for obesity alone. Concordance was 97.1% for the combined prevalence of overweight and obesity, and 98.8% for obesity.
Individual-level independent variables
In order to control for possible individual-level confounders, questions were utilized to determine age, gender, ethnicity, parents’ occupation, monthly expenses and smoking. With few exceptions, most Chinese studies have shown that obesity tends to increase with age  and be higher for male children and adolescents [31, 32]. Given the well-established link between SES and obesity in China, three measures of SES, parents’ occupation, family income and monthly student expenses, were included. Occupation was recorded in three categories: operations and commercial work (operations referring to mainly farmers and workers), staff and administration (which included mainly government jobs and company management jobs); teacher, scientific and technical work) (9, 10). Family income (in RMB Yuan) was measured through the question: “how much was the income of each person in your family over the last year?” Categories ranged from less than ¥1000, ¥1000 to ¥1999, and ¥2000 and over. We also included a variable, monthly student expenses (in RMB Yuan), which was measured though the question: “how much do you spend each month?” In addition, in view of the very high rates of smoking in China, which has been related to overweight and obesity, we also included this as a background factor [13, 33,34,35].
Home location and current contextual factors
Two sets of contextual variables were included, relating to characteristics of the student’s family home location and of the university city where they were studying. In terms of the former, family location was defined in terms of both their home region (Northeast, North, Eastern, Centre, Southwest and Northwest) and whether students came from a city, county ‘town’ or rural area. In addition, gross domestic product (GDP) per capita was included to highlight differences in area income between the students’ home provinces.
Given that many studies have stressed the link between the penetration of obesogenic environments and a country’s level of urbanisation , we included urban population size and area income (GDP per capita) to describe the university cities where the students were studying. Finally we also determined the characteristics of the universities which the students attended. Given that different universities have different social resources and some are far more prestigious than others, then it was important to include such a measure since, because of large differences in tuition fees, university type is also an indirect measure of family income. University type was determined using the China university ranking system (“high level,” “middle level,” and “low level”) as established by the National Ministry of Education .
All data were entered into a database using Microsoft Excel. The dataset was then imported into statistical analysis system (SAS) (9.3 version) for statistical analyses. Descriptive statistics were calculated to determine the prevalence of overweight and obesity, which was also mapped at the provincial level. For the purposes of mapping prevalence rates were defined according to the home provinces of the students in the sample.
A logistic model was utilized to assess associations between the dependent and each of the independent variables. Both unadjusted and adjusted methods were considered in the data analyses and implemented to examine these associations. SAS survey logistic procedures were applied in the unadjusted analysis, using the university as the clustering unit, in order to account for a within-clustering correlation, attributable to the complex sample for unadjusted analysis. The multilevel analysis was weighted using sampling, subject-level weights, and post-stratification weights, respectively .
In terms of the first study objective, we applied multilevel logistic regression models using the SAS GLIMMIX procedure (Table 2). We started with the Null Model, a three -level (individual, university, and original home province) model with random intercepts. First we constructed an individual model which included variables relating to gender, monthly expenses and smoking. The second (home location) model included the above individual factors but also three variables relating to region of origin, urban–rural background and GDP per capita of the home region. Models were run both including and excluding the student’s home region. The third (university city model) added three new contextual variables to the above individual characteristics, university type, university city GDP and city size. Finally, since we also wished to examine the relative importance of both home location and university city characteristics, the final (combined) model included all of the above variables.
It should be mentioned that family income is an appropriate indicator of family economic resources. However, family income was surveyed in only some universities and obtained for 4902 students due to a printing error in the questionnaire. Consequently we used an indirect measure of family income, monthly student expenses in the multilevel models. This was a valid measure because families tend to fund most students’ expenses. Not surprisingly the latter was significantly associated with family income (r = 0.30; p < 0.001).
With respect to the second objective, we assessed the interaction between our measure of student income, monthly expenses, and four contextual variables: the rural–urban home location of the student, university city population and the GDP per capita of the student’s home region and university city. Thus for each group we were able to assess the strength of the individual SES–obesity relationship as well as the (weighted) prevalence of overweight and obesity. We also examined the relationship between urban–rural origin and obesity by university type.
A total of 12,211 questionnaires were completed. Of those who responded 11,942 students were available for general analysis. BMI was calculated for the 11,673 respondents (97.4%) who provided complete data. There were no significant differences in demographic characteristics between responders and non-responders. Of the sample, 16% were less than 20 years of age and 21% were aged 23 years and over. There were more females (64.2%) than males, most (93%) of the students were of Han ethnicity and the majority (over 75%) came from families where the parents were engaged in operations or commercial work. Almost 38% recorded high levels of monthly expenditures (over 1500 RMB). Most students came from the countryside or townships (67%) and about half attended universities in middle-size cities.
Overall the prevalence of overweight and obesity in the study sample was 9.5% (95% confidence interval (CI) 7.7, 11.3%), and obesity prevalence was 2.2% (95% CI 1.3, 3.1%) there was considerable geographic variation. A higher prevalence occurred in northeast and southwest China (Fig. 2), with the highest rates being recorded in Liaoning (34.9%), Neimenggu (18.7%), Shanxi (15.5%) and Beijing (14.9%).
The unadjusted logistic analysis showed that of the individual-level variables gender (male), higher monthly expenses and smoking were associated with being classified as overweight and obese (Table 1). Parents’ occupation was not significant. However in the limited sample there was a significant relationship between family income and overweight and obesity with higher income families more likely to be overweight or obese. The unadjusted odds ratios (ORs) were 1.61** (95% CI 1.19, 2.17) (for 10,000 and over RMB vs <10,000RMB) and 1.35** (CI 1.14, 1.59) (for 20,000 and more RMB vs <10,000RMB). This relationship remained after adjusting for other individual variables [the respective ORs were 1.61** (CI 1.10. 2.20) and 1.29** (CI 1.08, 1.55)] and all variables [1.38** (CI 1.09, 1.75) and 0.82 (CI 0.57, 1.18)]. In all models there were no significant differences between monthly expenses and overweight and obesity after considering family income.
Of the home contextual factors, students who originated from Northeast China (Fig. 2), from larger towns and cities and from provinces with higher GDP per capita were more likely to be overweight or obese. Of the three contextual factors a more urbanised family home location increased the chances of being overweight and obese to a much greater extent than region of origin or home province GDP. Contextual characteristics of the university cities were also important. Larger and wealthier destination cities were also related to overweight/obesity prevalence levels as was the type of university. Compared to high level universities, students enrolled at lower and middle level universities had a reduced risk of being overweight or obese.
We also performed unadjusted analyses for males and females separately and some gender differences occurred (table not shown). Parental occupation became significant compared to female students who had fathers engaged in operations and commercial work, those whose fathers were teachers or were employed in scientific and technical work were more than twice as likely to be overweight or obese. The effect of daily smoking increased the risk of being overweight or obese for females (unadjusted OR 7.79; 95% CI 1.82, 33.23), but was not significant for males. By contrast, home region GDP was only significant for males (unadjusted OR 1.26; CI 1.11, 1.44), while university city GDP and population size had little effect on male trends in overweight and obesity compared to females (the respective ORs 2.39; CI 1.05, 5.48 and 1.80; CI 1.76, 2.75).
In the multilevel individual and family location models being male, a regular smoker and having higher monthly expenses remained significant (Model 1) as did the student’s urban–rural origin and area income of the home province (Model 2) (Table 2). However, home region became not significant, thus suggesting that regional differences in overweight and obesity were highly related to differences in urbanisation and levels of GDP. For this reason we re-ran Model 2 excluding home region and also omitted this variable in the final model. In the university city model no contextual variables were significant (Model 3). However, if controls were made just for individual factors then university type was significant, with students attending low level universities being less likely to be overweight or obese (OR 0.55; CI 0.40, 0.74). Similarly if controls were made just for individual factors then students attending university in the largest cities were more likely to be overweight or obese (OR 1.55; CI 1.02, 2.36). Finally in the combined model, incorporating both home and university city contextual factors, urban–rural origin, home provincial GDP and population size of the university city remained significant (Model 4).
Controlling for gender, monthly expenses and smoking, students who originated from cities versus rural and township areas were between 60 and 82% more likely to be overweight or obese. In addition we investigated the interaction between home location and individual level expenses since both factors have independent influences on obesity. However, analyses showed no significant interaction between these two factors (Wald Chi Square: 2.74, p: 0.0979).
In terms of the relationship between student monthly expenses and obesity and urban/area income contextual factors the results suggest that association between individual SES and obesity was strongest for students who originated from rural areas and for those who attended universities in smaller cities (Table 3). The results are less clear for home region GDP but in the case of university city GDP higher income students were more likely to be overweight or obese in poorer cities. In terms of the actual prevalence of obesity the highest rates occurred among more affluent students originating from (28.7%) or currently living in (21.2%) larger places and among more affluent students coming from wealthier areas.
This aim of this study was twofold; to assess the importance of home location and university contextual effects on patterns of student overweight and obesity and to investigate the extent to which the SES–obesity relationship varied depending upon the geographic context. While this is not the first Chinese study to investigate obesity among young adults  or university students in particular , it is the first Chinese research to examine these factors using a multi-level framework taking into account both past and current locations. Internationally, the research is also one of the few studies [46, 47] to consider variations in the nature of the SES–obesity relationship within a particular national context.
With respect to the first objective three main conclusions are evident. First, with respect to the student’s home environment, independent of individual characteristics, levels of urbanization and provincial GDP per capita emerged as the key predictors of overweight and obesity. Students who came from county towns or larger cities were twice to almost four times more likely to be obese or overweight compared to students originating from rural areas. Similarly those who were born in more affluent regions were more likely to be overweight or obese, independent of their own individual income status. These patterns suggest the importance of lifestyle and dietary factors on overweight and obesity because students’ basic lifestyles are partly formed during childhood and adolescence. These findings are similar to other research both in China  and elsewhere [13, 18, 26] which points to the significance of early life conditions on patterns of adult obesity.
Second, there was also evidence of contextual influences in the destination cities of the students. Students who attended university in the larger cities were more likely to be overweight or obese compared to students who lived in smaller cities and this effect remained significant in the final model after controlling for the urban–rural home origins of students. These results are similar to the findings of Ji and Chen  who found that over the period 1985–2010 the rate of increase in overweight and obesity (amongst children and adolescents) was greatest in the largest cities, the authors suggesting that adult prevalence rates in such places were approaching those found in developed countries. They also support the results of other studies of students  and adults  as well as children and adolescents in China  where urban–rural differences in obesity, despite decreasing in recent years , are still very apparent.
Although in the expected direction, university city GDP was not significant. University type was significant when just individual factors were controlled for but not in the final model nor in the university model when the two other university contextual characteristics were included. This most likely reflects the fact that the most prestigious universities are located in the largest cities, places where students will be most exposed to obesogenic environments. Nevertheless the pressures of studying at China’s most prestigious universities should also be taken into account. One can only speculate that sedentary behavior is more likely to be common among such students who have little time for other activities. Li et al. , for example, reported that obese students in Guangdong were more likely to indicate that they never exercised or engaged in daily exercise. Thus it was no surprise that, when university type was considered, the highest prevalence of overweight and obesity was typical of students with urban origins studying at China’s top universities.
Third, the findings suggest that characteristics of the home locations of students are more important than those of the places to which they migrated in influencing overweight and obesity. This should not be surprising given the short length of residence of many of the students in the university cities. It would be tempting to suggest that the selective migration of more affluent students from their homes to university accounted for most of the variation in obesity amongst this student sample. However, other studies have suggested that differences in the prevalence of obesity could not be accounted for by birthplace or later selective migration  and that resident children are more likely to be obese than migrant children . Such findings thus point to the importance of local context effects, which are likely to become more important over time, in affecting obesity prevalence among the student population. Thus the fact that urban size remained significant in the final model suggests the importance of destination factors in modifying patterns of obesity amongst the most at-risk populations.
With respect to the second objective, the study found further evidence of the well known link between individual socio-economic status and obesity, which was evident in all the models. Unlike western countries where obesity is highest amongst the poor and certain ethnic groups , the pattern of higher prevalence of obesity among students from higher income families is typical of countries in the earlier stages of the nutrition transition [5, 6]. However, the fact that income differences in obesity were generally strongest for students originating from rural locations and for those who attended university in smaller cities suggests that the nutrition transition is at an earlier stage than in larger cities where, although obesity rates are higher, individual SES differences are much less pronounced (Fig. 3b). Thus the results suggest a need for more thought to be given to geographical variation in the SES–obesity relationship especially in countries as diverse as China. In other words, models of the nutrition revolution, rather than focusing just on international differences in the relationship between SES differences and obesity (Fig. 3a), need to take greater account of the forces operating within low and middle income countries. For example, at what point in the urbanization process do factors relating to obesogenic environments start to become much more important compared to individual-level determinants of obesity? Do particular thresholds exist? Such considerations are important in all low and middle income countries where the obesogenic epidemic is at an earlier stage. Answering such questions obviously has implications for population programs aiming to target the most at-risk groups.
Thus, in contrast to some other views  the geographic distribution of obesity cannot simply be read off from individual SES variations, but also reflects a range of context effects which will modify the individual SES–obesity relationship. Important here is the level of urbanization and regional affluence which will be highly correlated with the penetration of obesogenic environments. As He et al.  have suggested, these will include the greater availability of energy dense foods at a cheaper cost, the spread of obesity-related health knowledge  or the adoption of higher-SES groups of western cultural norms regarding body shape, all of which may narrow SES differences in obesity. On the other hand, higher rates of income inequality and increased food consumption among the rich [41, 42] or the effects of transportation infrastructure and other labour saving devices which result in reduced daily physical activity , are likely to increase social disparities in obesity. The above results thus suggest an urban–rural diffusion of the obesity epidemic in China. However, while the spread of the obesity epidemic to rural areas, to some extent, has already been identified [30, 38, 47] exactly how the local environment helps shape the social distribution of obesity in different places remains unclear. Thus more research which focuses on the role of contextual factors influencing obesity at smaller spatial scales, such as cities and neighbourhoods, would seem to be a high priority.
This study has a number of limitations. First, and most important, is that our range of contextual factors was relatively limited. Although it was beyond the scope of this study greater attention to cultural factors affecting urban and regional differences in diet, the food environment and to the effects of income inequality on patterns of food consumption is necessary. More attention also needs to be paid to the nature of obesogenic environmental factors and they affect different SES groups. Second, the study was based on cross sectional design, which precludes causal inference and calls for cohort and other longitudinal studies on overweight and obesity in China to advance etiologic understanding, and to inform the design and evaluation of generic and tailored interventions. Third, the sampling frame was university students, most of which were medical students. Hence, the findings are not generalizable to the whole of China. Fourth, because not all universities recorded data for the urban–rural origins of the students, this was available for only 41.9% of the total sample. As a result the sample appears to under represent urban students (those who came from cities and county towns).
This study adds important insights about the impact of home and university environments on the prevalence of overweight and obesity among university students. By emphasizing the importance of contextual effects we suggest that more attention needs to be paid to the environmental conditions, such as urbanization and rapid economic development. These have helped change health beliefs, patterns of food consumption, health behavior and lifestyles and have contributed to the growing obesity epidemic in China and other low and middle income nations. Many of these factors have influenced the growing obesity of children and adolescents, but, as would be expected, have become particularly apparent among young adults. As Dutton and McLaren  have suggested modeling these population-level factors is an important avenue for future research. Given that not all places or population groups are equally affected then more attention to how local environments are changing and the effects of such changes on the diet, lifestyles, as well as the social norms and perceptions of obesityon the part of different groups is necessary.
Effective strategies that take into account contextual influences are also needed to implement policy and public health interventions to prevent overweight and obesity, especially since that there is minimal evidence that the obesity epidemic is slowing down. However, central and provincial governments in China have largely ignored food and other policies which may go some way towards slowing the growth of obesity  although the recent policy shift encouraging more healthy cities suggests that this view is changing . This is also true among university environments and thus this study provides preliminary evidence for public health policymakers and educators and suggests the need for an approach for intervening to avert or reduce overweight and obesity among college students. Such an approach needs to take account of the diverse origins of students, since such differences will be important in affecting the social distribution of obesity and the attitudes of students to any policy interventions. Thus we recommend that there is a need for strategies which address both environmental and individual factors as such a multi-faceted approach is essential to help curb the growing obesity epidemic.
gross domestic product
Global Health Professions Student Survey
body mass index
statistical analysis system
Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, Mullany EC. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2014;384:766–81.
Pearce J, Witten K, editors. Geographies of obesity: environmental understandings of the obesity epidemic. Burlinton: Ashgate Publishing Ltd; 2010.
Kim D, Kawachi I. Contextual determinants of obesity: an overview. In: Pearce J, Witten K, editors. Geographies of obesity: environmental understandings of the obesity epidemic. Burlinton: Ashgate Publishing Ltd; 2010. p. 39–54.
Popkin BM. The nutrition transition and obesity in the developing world. J Nutr. 2001;131:871S–3S.
Popkin BM, Adair LS, Ng SW. Global nutrition transition and the pandemic of obesity in developing countries. Nutr Rev. 2012;70:3–21.
Monteiro CA, Moura EC, Conde WL, Popkin BM. Socioeconomic status and obesity in adult populations of developing countries: a review. Bull World Health Organ. 2004;82:940–6.
Peltzer K, Pengpid S, Samuels T, Özcan NK, Mantilla C, Rahamefy OH, et al. Prevalence of overweight/obesity and its associated factors among university students from 22 countries. Int J Environ Res Public Health. 2014;11:7425–41.
Halkjaer J, Sørenson TI. Psychosocial and demographic determinants of regional differences in the prevalence of obesity. J Biosoc Sci. 2004;36:141–52.
Wan Y, Jiang X, He Y, Zhang Y, Liang Y, Pan F, Xu Y, Shang L. Body mass index of young men in China: results from four national surveys conducted between 1955 and 2012. Medicine. 2016;95:e2829.
Ji CY, Cheng TO. Prevalence and geographic distribution of childhood obesity in China in 2005. Int J Cardiol. 2008;131:1–8.
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;64:858–70.
He W, James SA, Merli MG, Zheng H. An increasing socioeconomic gap in childhood overweight and obesity in China. Am J Public Health. 2014;104:e14–22.
Siddiqui MZ, Donato R. Overweight and obesity in India: policy issues from an exploratory multi-level analysis. Health Policy Plan. 2016;31:582–91.
Willms JD, Tremblay MS, Katzmarzyk PT. Geographic and demographic variation in the prevalence of overweight Canadian children. Obes Res. 2003;11:668–73.
Poobalan A, Aucott L. Obesity among young adults in developing countries: a systematic overview. Curr Obes Rep. 2016;5:2–13.
Kumanyika SK. Environmental influences on childhood obesity: ethnic and cultural influences in context. Physiol Behav. 2008;94:61–70.
Alvarado SE. Neighborhood disadvantage and obesity across childhood and adolescence: evidence from the NLSY children and young adults cohort (1986–2010). Soc Sci Res. 2016;57:80–98.
Zheng H, Tumin D. Variation in the effects of family background and birth region on adult obesity: results of a prospective cohort study of a Great Depression-era American cohort. BMC Public Health. 2015;15:1.
Curtis DS, Fuller-Rowell TE, Doan SN, Zgierska AE, Ryff CD. Racial and socioeconomic disparities in body mass index among college students: understanding the role of early life adversity. J Behav Med. 2016;39:866–75.
Pengpid S, Peltzer K. Prevalence of overweight/obesity and central obesity and its associated factors among a sample of university students in India. Obes Res Clin Pract. 2014;8:e558–70.
Li Y-q, Li J, Ze W-x, Li X-y, Bin L, Zo H-q. Investigation of obesity in undergraduate students of an university in Guangdong province and its influence factors. J Southeast Univ (Med Sci Edit). 2013;4:012.
Antiporta DA, Smeeth L, Gilman RH, Miranda JJ. Length of urban residence and obesity among within-country rural-to-urban Andean migrants. Public Health Nutr. 2016;19:1270–8.
Li R, Wang D, Chen J, Chai J, Tang M. Regional differences in smoking, drinking, and physical activities of Chinese residents. Asia Pac J Public Health. 2015;27:NP230–9.
Myers CA, Slack T, Martin CK, Broyles ST, Heymsfield SB. Change in obesity prevalence across the United States is influenced by recreational and healthcare contexts, food environments, and Hispanic populations. PLoS ONE. 2016;11:e0148394.
Matozinhos FP, Gomes CS, de Souza Andrade AC, Mendes LL, Pessoa MC, de Lima Friche AA. Neighbourhood environments and obesity among adults: a multilevel analysis of an urban Brazilian context. Prev Med Rep. 2015;2:337–41.
Nau C, Schwartz BS, Bandeen-Roche K, Liu A, Pollak J, Hirsch A. Community socioeconomic deprivation and obesity trajectories in children using electronic health records. Obesity. 2015;23:207–12.
Yang T, Yu L, Bottorff JL, Wu D, Jiang S, Peng S, Young KJ. Global Health Professions Student Survey (GHPSS) in tobacco control in China. Am J Health Behav. 2015;39:732–41.
Department of Disease Control and Prevention, National Ministry of Health. Guidelines for the prevention and control of overweight and obesity in China. Beijing: National Ministry of Health; 2003.
Le A, Judd SE, Allison DB, Oza-Frank R, Affuso O, Safford MM, Howard VJ, Howard G. The geographic distribution of obesity in the US and the potential regional differences in misreporting of obesity. Obesity. 2014;22:300–6.
Lao XQ, Ma WJ, Sobko T, Zhang YH, Xu YJ, Xu XJ, Yu DM, Nie SP, Cai QM, Xia L, Thomas GN, Griffiths SM. Overall obesity is leveling-off while abdominal obesity continues to rise in a Chinese population experiencing rapid economic development: analysis of serial cross-sectional health survey data 2002–2010. Int J Obes. 2015;39:288–94.
Andegiorgish AK, Wang J, Zhang X, Liu X, Zhu H. Prevalence of overweight, obesity, and associated risk factors among school children and adolescents in Tianjin, China. Eur J Pediatr. 2012;171:697–703.
Liu W, Lin R, Li B, Pallan M, Cheng KK, Adab P. Socioeconomic determinants of childhood obesity among primary school children in Guangzhou, China. BMC Public Health. 2016;16:1.
Lv J, Chen W, Sun D, Li S, Millwood IY, Smith M, Guo Y, Bian Z, Yu C, Zhou H, Tan Y, Chen C, Chen Z, Li L. Gender-specific association between tobacco smoking and central obesity among 0.5 million Chinese people: the China Kadoorie Biobank Study. PLoS ONE. 2015;10:e0124586.
Pei L, Cheng Y, Kang Y, Yuan S, Yan H. Association of obesity with socioeconomic status among adults of ages 18 to 80 years in rural Northwest China. BMC Public Health. 2015;15:1.
John U, Hanke M, Rumpf HJ, Thyrian JR. Smoking status, cigarettes per day, and their relationship to overweight and obesity among former and current smokers in a national adult general population sample. Int J Obes. 2005;29:1289–94.
National Ministry of Education. Annual report on university graduates’ employment in 2015. Beijing: National Ministry of Education; 2016.
Grilli L, Pratesi M. Weighted estimation in multilevel ordinal and binary models in the presence of informative sampling designs. Surv Methodol. 2004;30:93–104.
Ji CY, Chen TJ, Working Group on Obesity in China (WGOC). Empirical changes in the prevalence of overweight and obesity among Chinese students from 1985 to 2010 and corresponding preventive strategies. Biomed Environ Sci. 2013;26:1–2.
Zou Y, Zhang R, Zhou B, Huang L, Chen J, Gu F, Zhang H, Fang Y, Ding G. A comparison study on the prevalence of obesity and its associated factors among city, township and rural area adults in China. BMJ Open. 2015;5:e008417.
Song Y, Ma J, Wang HJ, Wang Z, Hu P, Zhang B, Agard A. Secular trends of obesity prevalence in Chinese children from 1985 to 2010: urban–rural disparity. Obesity. 2015;23:448–53.
Jones-Smith JC, Gordon-Larsen P, Siddiqi A, Popkin BM. Cross-national comparisons of time trends in overweight inequality by socioeconomic status among women using repeated cross-sectional surveys from 37 developing countries, 1989–2007. Am J Epidemiol. 2011;173:667–75.
Subramanian SV, Perkins JM, Khan KT. Do burdens of underweight and overweight coexist among lower socioeconomic groups in India? Am J Clin Nutr. 2009;90:369–76.
Dutton DJ, McLaren L. How important are determinants of obesity measured at the individual level for explaining geographic variation in body mass index distributions? Observational evidence from Canada using quantile regression and Blinder–Oaxaca decomposition. J Epidemiol Commun Health. 2016;70:367–73.
Wang H, Zhai F. Programme and policy options for preventing obesity in China. Obes Rev. 2013;14:134–40.
Xu B, Yang J, Zhang Y, Gong P. Healthy cities in China: a Lancet Commission. Lancet. 2016;388:1863–4.
Jin H, Lu Y. The relationship between obesity and socioeconomic status among Texas school children and its spatial variation. Appl Geogr. 2017;79:143–52.
Gao Y, Huang Y, Song F, Dai D, Wang P, Li H, Zheng H, et al. Urban-rural disparity of overweight/obesity distribution and its potential trend with breast cancer among Chinese women. Oncotarget. 2016;7:56608–18.
TY conceived the study design, conceptualized the ideas, and supervised the data management and analyses. YL, SJ, SP and YF organized the data collection. TY wrote the initial manuscript. RB and LL provided the revision of the manuscript. All authors reviewed previous drafts. All authors read and approved the final manuscript.
We thank local teams from the “Building advocacy capacity for tobacco control in medical universities in China (Bloomberg global control project)” project organizing the data collection.
The authors declare that they have no competing interests.
Availability of data and materials
Data sharing not applicable to this article because the datasets we used belongs to our Centre for Tobacco Control Research Zhejiang University School of Medicine. Please contact corresponding author for data requests.
Ethics approval and consent to participate
The study was approved by the Ethics Committee at the Medical Center, Zhejiang University, and verbal consent was obtained from all participants prior to data collection.
This study was partly funded by the National Nature Science Foundation of China (Major Project, 71490733), the National Nature Science Foundation of China (71473221), Center for Tobacco Control Research, Zhejiang University School of Medicine. This project universities, local government and Center for Disease Control and Prevention also partly fund this survey. These funding was used for the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Tingzhong Yang and Lingwei Yu are co-first authors