TRANG, NGUYEN H. H. D.1; HONG, TANG K.1; DIBLEY, MICHAEL J.2; SIBBRITT, DAVID W.3
Obesity/overweight is an emerging epidemic in developing countries, in which youth are adopting a more passive lifestyle with less physical activity and changing their dietary habits by increasing the levels of intake of fat and refined sugars. Among the accepted risk factors for excess weight gain in adolescents, decreased physical activity is likely to play an important role (13,17,25).
Many researchers have reported health benefits from physical activity in both the short and the long term (3,32). Knowing the demographic, socioeconomic, psychosocial, environmental, and social determinants of physical inactivity in youth is important to design effective intervention strategies to promote physical activity in this population. Most prior studies of physical activity in adolescents and the associated risk factors for inactivity have been conducted in western populations (12,27). There are few reports from Asia, with a single study from China (21) and two studies from Taiwan (34,35), but none from Vietnam.
Research into gender differences in physical activity of adolescents has consistently found that boys are more active than girls (10,23,24). In Vietnam, gender differences in physical activity might be further exaggerated by the reduced independence of mobility of girls. Thus, a gender-specific approach to the analysis of physical activity data in this population is needed. In studies from western countries, children's activity levels have been shown to be correlated with parental encouragement of physical activity (8,18,27), which was more common with better-educated parents and families of higher socioeconomic status. Furthermore, society in Vietnam places a high value on academic achievement and may result in different parental and school influences on physical activity of adolescents.
There are a wide range of factors that are potentially associated with physical inactivity in children, including the community, the school and home environmental elements, the family and parental characteristics, the child's lifestyle and behaviors, and the child's individual characteristics. To develop valid models to identify which of these factors are most strongly associated with physical inactivity requires a conceptual framework linking these factors at different levels to guide the modeling process (30).
There have been no prior studies of physical activity and factors associated with inactivity in adolescents in Vietnam. This article aims to describe the physical activity patterns across gender, age, school location, and household wealth index of adolescents in Ho Chi Minh City (HCMC), Vietnam.
From November to December 2004, a cross-sectional survey was conducted on a representative sample of students of both genders from grades 6 to 9 (ages 11-16 yr) of junior high schools in urban areas of HCMC. The sample was chosen using a multistage stratified cluster sampling approach. In the first stage, schools were selected within two strata (wealthy and less wealthy districts) by systematic random sampling using a proportionate to population size method. In the second stage, classes were selected by systematic random sampling from a list of classes organized by grade within each school selected in stage one. All the adolescents in the selected classes from stage two were investigated. In total, there were 31 schools with 2718 students. The study was approved by the ethics committee for medical research of Pham Ngoc Thach University of Medicine (PNTUM) and the human research ethics committee of the University of Newcastle, Australia. In addition, the study was approved by the municipal education department, the school health division at the district level, and the headmasters of each of the selected schools. Written consent was obtained from both the students and their parents before data collection.
The Adolescent Physical Activity Recall Questionnaire (APARQ) (4), which was developed and validated among Australian adolescents, was modified for use in Vietnam. It was translated into Vietnamese and assessed in focus group discussions undertaken with children, parents, and physical education teachers to make sure it was appropriate and reliable for use with Vietnamese adolescents. The questionnaire asks students to think about a normal week and write in a table the sports and games they usually did, how many times each week, and the usual amount of time they spent doing them. In HCMC, there are few differences in physical activities between summer and winter school terms. Thus, the original questionnaire (which did differentiate between seasons) was modified to assess usual physical activity during the school terms and the annual summer holidays. The questionnaire recorded information about organized and nonorganized sports and games at schools before and after school hours and on weekends during the school term and the summer holidays. In HCMC, the school term lasts 38 wk and the summer holidays last 14 wk. Additional questions were added to the physical activity recall questionnaire to record "sedentary time," including time spent watching television, playing computer games or video games, and doing homework after school on both weekdays and weekends. These questions were modified for use with Vietnamese youth from a validated 11-item television and computer usage questionnaire (22) and estimate the hours per day in each type of sedentary behavior over the month before interview. Questions on the type of transportation to school and the time spent traveling to school were also included. The questionnaire was administered to the adolescents by interview by the staff of the PNTUM and the Nutrition Center of HCMC and usually took approximately 15 min to complete.
Environmental elements at home, school and neighborhoods potentially associated with physical inactivity were also assessed. The environmental questionnaire was constructed using the results of focus group discussions with 10 to 15 community members in different locations of varying socioeconomic status from across HCMC and the results of in-depth interviews with selected local stakeholders. The participants of the group discussions included students, parents, pediatric doctors, nutritionists, physical education teachers, school principals, and sports coaches. The parents and the students as well as selected teachers and school principals were identified from five schools purposively selected to represent the urban areas of HCMC. The doctors and the nutritionists were recruited from the Children's Hospital No 1 and Children's Hospital No 2 and from the Nutrition Center of HCMC, and the sports coaches were identified from "cultural houses" in three urban districts of HCMC. These group discussions identified the key environmental elements of homes, schools, and neighborhoods that needed to be included in the questionnaire. Lack of pathways and dangerous traffic were among the reasons many parents did not allow their children to walk or cycle alone. Some questions were also obtained from the environmental assessment questionnaire used with urban Chinese adolescents (20,21). The home environment questionnaire was piloted in a sample of 50 respondents (adolescents and their parents from different areas in HCMC) and validated by direct observations by the investigators. Information regarding environmental factors at community and household levels as well as sociodemographic factors including parental education, occupation, inventory of household assets for computing a wealth index, and parental weight and height were obtained via a self-administered questionnaire for parents. The questionnaire specifically asked about recreational facilities in the community (open square, playground park, playfield or sports oval, stadium, swimming pool, and sports club), places around the home for children to play, transportation, level of residence, safety concern, fast food restaurant, parents' involvement with children doing exercise, household facilities for playing games, using pocket money, and family rules for playing games. In a confidential setting, the adolescents self-reported their pubertal status using a questionnaire with photographs illustrating five stages of pubertal development for pubic hair, male genitalia, or female breasts, and for female students, the date of their first menstruation was also recorded.
School representatives completed a self-administered questionnaire to record elements of the school environment, including the area of the school yard, the physical education curriculum, and the availability and accessibility of sport facilities at school. The list of environmental factors was based on a questionnaire previously developed in China (20) but which was modified for the situation in HCMC. The items were adjusted using the results of focus group discussions with students, parents, and school teachers. The questionnaire was validated from all schools at the start of the current study by a repeat data collection by an investigator who was blind to the responses from the school and who collected the data with an observational check list and interview of key school staff. Overall, there was greater than 90% agreement for all questions between the schools' responses and the investigators observations; for example, for the question about the "number of physical education classes at the school," there was 94% agreement, whereas for question on "the accessibility of the school's sport room," there was 89% agreement. The disagreements were not clustered in particular schools or questions but were scattered across the data. The details of this validation will be reported elsewhere (14).
Two trained data collectors took anthropometric measurements of the students using standard methods (33). Weight was measured without shoes and heavy clothes using a Tanita electronic scale (Tanita BF 571; Tanita Corporation, Tokyo, Japan) and was recorded to the nearest 100 g. Standing height was measured with a suspended Microtoise tape to the nearest 0.1 cm. Anthropometric standardization exercises were conducted to ensure uniform techniques were used by all data collectors (33).
The "svyset" commands in STATA 9 (StataCorp, College Station, TX, 2005) were used to adjust for the multistage stratified cluster sampling design. In addition, sampling weights were applied for survey parameter estimates to take account of the oversampling of students in the wealthy urban district stratum.
The physical activity data were cleaned based on the APARQ guidelines in which calculations were conducted separately for school term and summer holidays and for organized and nonorganized activities (4). Each activity was assigned a MET value based on the compendium of physical activities, and the sum of these MET values for each adolescent was categorized into low, moderate, or vigorous activity levels (1,2,17). One MET represents the resting metabolic rate equivalent to 3.5 mL·kg−1 body weight·min−1 oxygen consumption. A value greater than one indicates energy expenditure related to activity (11,12).
The level of physical activity was divided into two groups: physically active or insufficiently active (here on termed inactive), according to the classification reported in a study of physical activity in US adolescents (19). The criterion of "active" includes three or more sessions at least 20 min·wk−1 of vigorous activity (METs ≥ 6.0), five or more sessions of at least 30 min·wk−1 of moderate activity (3.5-5.9 METs), three or more sessions per week of strength training, enrolment in physical education, and participation in one or more sport teams. One is considered to be "inactive" if he or she takes part in less than three sessions at least 20 min·wk−1 of vigorous activity (METs ≥ 6.0), less than five sessions of at least 30 min·wk−1 of moderate activity (3.5-5.9 METs), less than three sessions per week of strength training, does not enroll in physical education, or does not participate in any sport teams. To check the appropriateness of this categorization system, a summary measure of physical activity, MET-minutes per week, was calculated using the following formula: physical activity score = MET × total time spent for each activity per week.
The distribution of activity levels before categorization was highly skewed toward high MET scores with an overall mean value of 1017.5 and a median of 817.6. The mean physical activity MET score in the active category was 1121.2 and in the inactive category was 693.5. These findings indicate that the physical activity categorization system adequately identified children of different overall levels of physical activity.
Prevalence with 95% confidence intervals (CI) was calculated for inactivity. Using two-tailed significance tests, categorical data were tested with Pearson chi-square, whereas normally distributed continuous data were tested with Student's t-test.
In this article, we examined five groups of study factors:
1) The community environmental risk factors included the location of residence, parental reported neighborhood safety, and accessibility to surrounding recreational facilities and the children's transportation to school. Recreational facility scores were computed by summing up the scores of each facility (1 for yes and 0 for no) with the totals ranging from 7 to 14, and these scores were categorized into quartiles. The highest scores indicated the most difficult access to recreational facilities, whereas the lowest scores represented the easiest access. Children's transportation was divided into two types: active and passive transportation, in which active transportation was defined as children who in a typical week walked or cycled to school on at least 1 d.
2) The school environmental risk factors such as school location, accessibility to school yard and sports rooms (no restrictions for using, could be used during school days, or could be used during school days with limited time), having morning recess exercises, physical activity after class, and school sport meetings. School sport meeting was a combined variable having three values: 1) once in 2 yr, lasting more than a day; 2) once in 2 yr, lasting 1 d or less; and 3) one to two school sport meetings/year. This variable was derived from two questions, the frequency and the duration of school sport meetings.
3) The family and household environmental factors covered the household physical environment, family interaction, parenting style, and parental characteristics. The household physical environmental factors included number of televisions at home (0, 1, >1), location of television in the house, computer ownership, availability of game machines or game shops around the home, and household wealth index. To assess the economic status of the household, the ownership of 14 different assets (telephone, radio, video cassette player, CD system, DVD player, air conditioner, refrigerator, computer, gas stove, microwave, bicycle, motorbike, car, and television) were used to construct a household wealth index using the principal components method to assign a weight for each asset (7). The index was ranked and divided into four categories (quartiles). The family interaction factors examined included adults participating in playing sport with their children, whereas the parenting style factors included supporting the child's activity and monitoring of television viewing or game playing. The individual parental characteristics consisted of parental body mass index (BMI) status, education, and occupation.
4) The child's behavioral risk factors included time spent in sedentary activities consisting of watching television, playing video/computer games, and studying. Time spent watching television per weekday was categorized into <2 h and ≥2 h. Time spent playing video/computer games was categorized into groups (playing games <2h and ≥2 h·d−1) based on a review of studies from 34 countries (15).
5) The individual child characteristics of gender, age, pubertal stage, and overweight status. The age of the surveyed children was divided into four groups: less than 12 yr, 12 yr, 13 yr, and 14 yr or older. There have been no prior studies of physical activity and factors associated with inactivity in adolescents in Vietnam. Overweight and obesity were assessed using body mass index (BMI) [i.e., weight/height2 (kg·m−2)] and were defined using International Obesity Task Force (IOTF) cutoffs (5). The analyses examined combined overweight and obesity (equivalent to adult BMI ≥25 kg·m−2), which is termed "overweight."
Statistical models were constructed separately by gender. A hierarchical approach was used for model building (30), which was based on a conceptual framework that describes the hierarchical relationship between community-, school-, and family-level factors (6) (Fig. 1). We chose this approach rather than a standard stepwise model-building approach based solely on statistical significance because we wished to also incorporate into the model-building knowledge on known interrelationships between the factors involved.
FIGURE 1-Conceptual ...Image Tools
The first step in the multivariate analyses examined the effects of the community-level variables of residency, accessibility of recreational facilities, and accessibility of fast food shops. Only those variable(s) significantly associated with inactivity in adolescents (P value <0.05) in this first model was retained for subsequent steps in the model building.
In the second step, all school environment variables were added to the model with the significant community-level variables. Only those school environment variable(s) that was significantly associated with inactivity in adolescents (P value <0.05) in this first model was retained for subsequent steps in the model building. Similar steps were used for the household environment, the family and parental characteristics, and lastly the individual child characteristics.
Interaction terms were examined after determining the full model. Crude and adjusted odds ratios (OR) from the hierarchical approach were also presented with 95% CI. However, the OR presented in the final multivariate analyses were from the equation corresponding to the level in which the risk factor of interest was first entered to avoid the possibility that mediating variables would reduce the explanatory power of more the distant determinants. For variables that have more than two categories, a multi-degrees-of-freedom test was used to provide an overall P value for such variables. OR were based on sequential rather than simultaneous adjustment.
There were 2718 junior high school students who participated in this study from November to December 2004. Information was collected from all of them about their level of physical activity in a usual week during school time and summer time. Anthropometry measurements and self-assessed pubertal status were gathered from 2715 students (three with missing data). "Family forms" were self-administered by 2687 of the children's parents (99%). There were 31 children for whom these data were not collected because either the children lost the forms or did not give them to their parents or their parents were not willing to answer the questionnaire. Furthermore, 24 of the participants were removed from the analysis due to missing dietary and physical behaviors data leaving 2660 subjects (98%) for the analysis, including 1332 boys and 1328 girls.
Of these 2660 participants, 645 (24.3%) were classified as inactive. Among the boys, 235 (16.7%) were categorized as inactive, and among the girls, 410 (31.8%) were inactive. Table 1 shows the percentage of physical inactivity of the adolescents by child, parental, and family characteristics, with statistically significant findings for the entire sample notated with T, for males with M, and for females with F. Overweight children were more inactive than nonoverweight children (P = 0.0003), although most of this difference was from overweight boys (P < 0.0001). Time spent studying after class was strongly associated with inactivity, and there was a higher percentage of inactive children among those who had the highest level of time spent studying after class (P = 0.01). The same results were found among students who were spending more time playing video games (P < 0.0001).
Table 2 shows the distribution of physical inactivity according to the environmental factors at the community, school, and household levels. Students who were transported to school were less active in comparison to those who walked or cycled to school (P < 0.0001). Among boys, 29.0% of those traveling by automobile were inactive whereas only 5.4% of those going to school on foot or by bike were inactive (P < 0.0001). Among girls, the percentages were 36.7% and 21.7%, respectively (P=0.0003). At the school level, an increased number of physical education sessions per week was inversely associated with percentages of inactivity among students overall (17.4%) (P = 0.04) and girls separately (22.4%) (P = 0.02). At the household level, availability of video games shops near the home was associated with an increase in the percentage of adolescent inactivity, especially in males (P < 0.01).
Table 3 shows the results of the hierarchical regression model building conducted separately for each gender, in which variables were tested in sequence based on a conceptual framework of factors. In boys, transportation was strongly associated with physical inactivity: those taken to schools by parents were seven times more likely to be inactive than those went to schools by foot or by bicycles (OR = 7.4, 95% CI = 5.1-10.8). Boys from schools having only one sports meeting per 2 yr and where the sports meeting lasted longer than a day were less likely to be inactive (OR = 0.5, 95% CI = 0.3-1.1) than those from schools having only one sports meeting per 2 yr and where the sports meeting lasted less than a day. The availability of game shops nearby strongly increased the odds of inactivity (OR = 5.3, 95% CI = 3.8-7.4). There was also a strong association between the family's economic status and physical inactivity, and there was an apparent dose-response relationship with higher odds of inactivity as the family became wealthier (OR = 2.6, 95% CI = 1.5-4.5; wealthiest households compared with poorest). Those who had both overweight parents were more likely to be inactive than those who did not (OR = 2.7, 95% CI = 1.2-6.5). In boys, time spent playing video games were strongly associated with inactivity (OR = 2.6, 95% CI = 1.4-4.7). Also, overweight increased the odds of being inactive 4.2 times (OR = 4.2, 95% CI = 2.7-6.7). The levels of inactivity decreased as the age increased (≥14 yr old: OR= 0.1, 95% CI = 0.0-0.2).
In girls, the odds of being inactive increased when students had passive transportation (OR = 2.6, 95% CI = 1.4-3.9). Students from wealthy urban schools were more likely to be inactive than those from less wealthy urban schools (OR = 3.0, 95% CI = 1.4-6.5), whereas students from schools having two or more physical education sessions per week were less likely to be inactive than students from schools having only one physical education session per week (OR = 0.3, 95% CI = 0.2-0.4). Those who spent more than 2 h·d−1 playing video games (OR = 2.1, 95% CI = 1.5-3.1) had increased odds of physical inactivity. Also, the levels of inactivity decreased as age increased (≥14yr old: OR = 0.3, 95% CI = 0.2-0.5; Table 3).
The contribution of the various hierarchies to the final model (shown in Table 3) is demonstrated via pseudo-R2 values. Community-level variables had a pseudo-R2 of 0.121 for males and 0.029 for females. With the addition of the school-level variables, the pseudo-R2 increased to 0.130 for males and 0.067 for females. When the household/familial level was added, the R2 was 0.224 for males and 0.079 for females. The pseudo-R2 increased to 0.389 for males and 0.108 for females after having level of child's characteristics.
This is the first study to provide a detailed documentation of factors associated with inactivity among adolescents in HCMC and in Vietnam. The results show a striking difference in patterns of inactivity between genders and reveal the importance of environmental factors at community, school, home, and family in promoting or decreasing adolescent physical activity.
Consistent with other studies (21,28), we found that the type of transportation to school was significantly associated with inactivity. School environments such as exercise at recess times and frequency and duration of sports meetings were significantly related to the inactivity status of adolescents. Although the recess exercises included only moderate activities, they played a role in promoting a healthy lifestyle for adolescents. Environmental factors have been noted to be strongly associated with the level of physical activity in adolescents (12), and our findings confirm the importance for physical activity in adolescents of providing opportunities at school for participation in physical education.
Among the neighborhood environmental factors, we found that having home play yards significantly reduced the likelihood of physical inactivity of adolescents, but the presence of video game shops around the home had the reverse effect. These associations were mostly consistent with other studies (21,26), but our findings did differ from a study of adolescents in China, which reported that the presence of local game shops was associated with higher levels of physical activity. Local reports indicate that many Vietnamese children are likely to spend many hours playing online and video games. The ready availability of a video game shop near their home may help promote an inactive lifestyle and reduce physical activity among adolescents.
Our study also found that the time spent attending courses after class (approximately 135 min·d−1) was even longer than the time spent watching television or playing video games and that this sedentary behavior was significantly associated with inactivity. These findings highlight the level of the school-work burden for high school students in Vietnam. This is motivated by the pressure on students to obtain good marks and high grades to optimize their future study opportunities. To save time on transportation, parents take their children by motorbike or by car to these after-school classes rather than let the students walk or use bicycles by themselves. This again decreases the amount of time spent on physical activity and increases the inactivity of these children.
Findings from other studies have indicated that overweight children have lower levels of physical activity (9,27). In HCMC, overweight status was strongly associated with inactivity in adolescent boys but not in adolescent girls. These findings indicate that the factors associated with overweight and obesity are gender specific, and this has been observed in other studies of adolescents in Asia (20). The pseudo-R2 values for the gender-specific models were much larger for the male models compared with the female models. This suggests that the variables collected and used in the analyses for this study were much more salient for males than for females in predicting physical inactivity in adolescents. Therefore, there is need for future research to identify better predictors of physical inactivity in adolescent females. A higher frequency of physical activity among boys than girls has been reported from several studies of physical activity in adolescents (21,24,29), including a study of Vietnamese adolescents in California (16), and the results of all these studies were similar to our findings in adolescents in urban Vietnam. The gender differences in levels of physical activity could be due to girls perceiving more barriers to and/or disliking physical activity more than boys (24). In addition, there may be a decrease in accessibility to structured activity for adolescent girls and a decrease in the social desirability of physical activity for girls at puberty undergoing breast development (10). Adolescent girls may avoid putting themselves in situations where physical changes in their bodies are noticeable.
Many studies have mentioned differences in adolescent activity across age groups. Physical activity has been reported to decline with younger age groups (21,23). Sallis et al. (23) stated that the decline in physical activity with age is antithetical to public health goals. In our study, the younger groups had fewer chances to be active because they were taken to schools by parents. This again decreases the amount of time spent on physical activity and increases the inactivity of these children.
The multivariate analysis revealed that different factors were associated with physical inactivity between genders. Apart from the common variables, which were significantly associated with physical inactivity in both males and females, such as transportation, time spent playing video games, and age, males were more affected by family factors whereas females were more influenced by community environmental factors. The family's economic status and parental overweight status strongly associated with physical activity in boys. Many studies have found parental influences on children's activity (8,18). It has been reported that students from low-income families in Western countries have limited access to public resources that support physical activity, and so they became more inactive (12,24). However, in Vietnam, children from the wealthiest families were least likely to be active. This is probably because parents from wealthier households usually provide their children with a "modern" life, including up-to-date recreational facilities such as television, computers, and other "labor-saving" household devices. However, these two factors were found to be significantly related to inactivity only in the model for male adolescents. Historically, in Vietnamese society, females play a greater role in house work than males (31). Furthermore, boys are more independent in their mobility compared with girls. Hence, the availability of game shops or play yards around the home was significantly associated with the activity status of boys whereas these factors were not associated with girls' activity.
The accessibility to public recreation facilities and the school location were significantly associated with physical inactivity in girls. Many studies have reported that perceptions of the local neighborhood may influence children's physical activity. The findings from these studies showed that where there were no parks or sports grounds near the home, there was a lower likelihood of walking or cycling and that unsafe neighborhoods were associated with increased inactivity (12,26). Schools in wealthy areas are usually located in the central districts of HCMC, where the traffic flow is heavy. The danger from traffic is probably the reason why students from wealthy district schools were more inactive than those from less wealthy district schools, especially for girls who tend to have less independence of mobility than boys. In addition, the play yards for students in schools located in wealthy urban districts were significantly smaller that those in schools located in less wealthy urban districts (1520 m2 compared with 1732 m2 in average, P value <0.001). Thus, it can be expected that students from wealthy district schools were less likely to be active.
There were several limitations in our study. Firstly, the environmental factors were measured in this study using a questionnaire that was not specifically validated for use in Vietnamese adolescents. Secondly, given the cross-sectional nature of this study, it is not possible to determine whether a cause-and-effect relationship exists between the determinants and the inactivity behaviors. The risk factors may be either the causes or the results of physical inactivity. Another limitation of the study is the incomplete assessment of pubertal status for male students, where the attainment of adult voice was not assessed. These limitations in the method of pubertal assessment in males may have reduced the strength of the association between pubertal status and inactivity.
Despite these limitations, the findings from this study highlight the importance of environmental factors in physical inactivity of adolescents in HCMC. Identification of these factors allows for the development of better-targeted policies and programs to promote physical activity in adolescents. If students can easily access public recreational facilities and if neighborhoods were made safe and easy to walk in, then children would be more active. National recommendations are needed about an adequate level of physical activity for Vietnamese adolescents and should be a part of population-wide promotion of physical activity for youth.
The authors gratefully acknowledge Dr. Tran Thi Minh Hanh, the head of the Department of Community Nutrition, Nutrition Centre, Ho Chi Minh City, for her support during the nutrition survey in the year 2004. The authors would like to thank Dr. Phan Nguyen Thanh Binh and Dr. Ho Thi Kim Lien for their enormous help in data collection, entry, and cleaning of the survey. The 2004 survey was conducted with a grant from the Nestle Foundation (Switzerland) and additional funds from the Nutrition Centre, Ho Chi Minh City Health Department. The authors have no professional relationships with any companies or institutions that would benefit from these research findings. The results of the present study do not constitute endorsement of any particular product by the American College of Sports Medicine.
1. Ainsworth BE, Haskell WL, Leon AS, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc
2. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc
. 2000;32(9 suppl):S498-516.
3. Andersen LB, Harro M, Sardinha LB, et al. Physical activity and clustered cardiovascular risk in children: a cross-sectional study (The European Youth Heart Study). Lancet
4. Booth ML, Okely AD, Chey TN, Bauman A. The reliability and validity of the Adolescent Physical Activity Recall Questionnaire. Med Sci Sports Exerc
5. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ
6. Davison KK, Birch LL. Childhood overweight: a contextual model and recommendations for future research. Obes Rev
7. Filmer D, Pritchett LH. Estimating wealth effects without expenditure data-or tears: an application to educational enrollments in states of India. Demography
8. Fogelholm M, Nuutinen O, Pasanen M, Myöhänen E, Säätelä T. Parent-child relationship of physical activity patterns and obesity. Int J Obes Relat Metab Disord
9. Fonseca H, Gaspar de Matos M. Perception of overweight and obesity among Portuguese adolescents: an overview of associated factors. Eur J Public Health
10. Goran MI, Gower BA, Nagy TR, Johnson RK. Developmental changes in energy expenditure and physical activity in children: evidence for a decline in physical activity in girls before puberty. Pediatrics
11. Gordon-Larsen P, McMurray RG, Popkin BM. Adolescent physical activity and inactivity vary by ethnicity: The National Longitudinal Study of Adolescent Health. J Pediatr
12. Gordon-Larsen P, McMurray RG, Popkin BM. Determinants of adolescent physical activity and inactivity patterns. Pediatrics
13. Gortmaker SL, Peterson K, Wiecha J, et al. Reducing obesity via a school-based interdisciplinary intervention among youth: Planet Health. Arch Pediatr Adolesc Med
14. Hong KT. School Environment: A Comparison between Two Data Collection Methods
. Ho Chi Minh City: Pham Ngoc Thach University of Medicine; 2004. p. 10.
15. Janssen I, Katzmarzyk PT, Boyce WF, et al. Comparison of overweight and obesity prevalence in school-aged youth from 34 countries and their relationships with physical activity and dietary patterns. Obes Rev
16. Kaplan CP, Zabkiewicz D, McPhee SJ, et al. Health-compromising behaviors among Vietnamese adolescents: the role of education and extracurricular activities. J Adolesc Health
17. Kimm SY, Glynn NW, Obarzanek E, et al. Relation between the changes in physical activity and body-mass index during adolescence: a multicentre longitudinal study. Lancet
18. Klesges RC, Malott JM, Boschee PF, et al. The effects of parental influences on children's food intake, physical activity, and relative weight. Int J Eat Disord
19. Levin S, Lowry R, Brown DR, Dietz WH. Physical activity and body mass index among US adolescents: youth risk behavior survey, 1999. Arch Pediatr Adolesc Med
20. Li M, Dibley MJ, Sibbritt D, Yan H. Factors associated with adolescents' overweight and obesity at community, school and household levels in Xi'an City, China: results of hierarchical analysis. Eur J Clin Nutr
21. Li M, Dibley MJ, Sibbritt D, Yan H. Factors associated with adolescents' physical inactivity in Xi'an City, China. Med Sci Sports Exerc
22. Peterson KE, Field AE, Fox MK. Validation of YRBSS Questions on Dietary Behaviors and Physical Activity among Adolescents in Grades 9-12
. Atlanta (GA): Division of Adolescent and School Health, Centers for Disease Control and Prevention; 1996.
23. Sallis JF, Prochaska JJ, Taylor WC. A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc
24. Sallis JF, Zakarian JM, Hovell MF, Hofstetter CR. Ethnic, socioeconomic, and sex differences in physical activity among adolescents. J Clin Epidemiol
25. Steinbeck KS. The importance of physical activity in the prevention of overweight and obesity in childhood: a review and an opinion. Obes Rev
26. Timperio A, Crawford D, Telford A, Salmon J. Perceptions about the local neighborhood and walking and cycling among children. Prev Med
27. Trost SG, Kerr LM, Ward DS, Pate RR. Physical activity and determinants of physical activity in obese and non-obese children. Int J Obes Relat Metab Disord
28. Tudor-Locke C, Ainsworth BE, Adair LS, Du S, Popkin BM. Physical activity and inactivity in Chinese school-aged youth: the China Health and Nutrition Survey. Int J Obes Relat Metab Disord
29. Tudor-Locke C, Ainsworth BE, Adair LS, Popkin BM. Physical activity in Filipino youth: the Cebu Longitudinal Health and Nutrition Survey. Int J Obes Relat Metab Disord
30. Victora C, Huttly S, Fuchs S, Olinto MT. The role of conceptual frameworks in epidemiological analysis: a hierarchical approach. Int J Epidemiol
31. Wells M. Viet Nam: Gender Situation Analysis
. Hanoi: Asian Development Bank; 2005. p. 315.
32. WHO. CDC/WHO Collaborative Center Workshop on Global Advocacy for National Physical Activity Plans
. San Diego: World Health Organization; 2007. p. 90.
33. WHO Expert Committee. Physical Status: The Use and Interpretation of Anthropometry: Report of a WHO Expert Committee (WHO Technical Report Series; 854)
. Geneva: World Health Organization; 1995. p. 454.
34. Wu SY, Pender N, Noureddine S. Gender differences in the psychosocial and cognitive correlates of physical activity among Taiwanese adolescents: a structural equation modeling approach. Int J Behav Med
35. Wu TY, Pender N. Determinants of physical activity among Taiwanese adolescents: an application of the health promotion model. Res Nurs Health