A physically active lifestyle offers significant benefits for an individual's health and is now recognized as one of the most important behaviors for the health and well-being of a population. Regular physical activity helps children and young people to build and maintain healthy bones (17) and muscles and joints (3,22), control body weight, reduce fat, and develop efficient functioning of the heart and lungs (7,14,22). Physical inactivity is recognized as a critical public health issue because of the alarming global trend toward increased physical inactivity in adults and children (35). In addition, physical inactivity is an important preventable factor for noncommunicable diseases (35). Guidelines specific for youth have been put forward to emphasize that regular daily moderate to vigorous physical activity is associated with health benefits (9,24). Patterns of physical activity acquired during childhood and adolescence are more likely to be maintained throughout life, thus providing the basis for an active and healthy lifestyle (15,30).
The National Nutrition and Health Survey in China in 2002 revealed that 32% of adolescents in cities reported seldom doing exercises (not exercising for 30 min per session, three times per week), whereas 17% of boys and 7% of girls reported often doing exercises (exercising for more than 30 min per session, and doing more than three sessions per week) (33). The sedentary behavior in Chinese adolescents is manifested mainly by long hours of doing homework rather than watching TV or playing games, which is observed in American children (2). It is anticipated that with mass-appeal broadcasting, consumer marketing, and increasing purchases of private automobiles, physical activity in Chinese adolescents will diminish (31).
To develop effective physical activity-promotion strategies for youth in China, the key influences and potential mediators of activity levels need to be well understood. Physical activity behavior is so complex that no single factor can explain the level of physical activity in a population. There is a range of factors correlated with physical activity in children and adolescents including biological, psychological, and environmental factors (25). Conducting a cross-sectional study to identify correlates of inactivity is the first step toward identifying mediators, and similar research has been undertaken in Western countries (27,29). But there has been no study in China delineating the factors associated with physical inactivity in adolescents. The purpose of this study was to explore the biological, socioeconomic, and environmental factors associated with physical inactivity in adolescent boys and girls in Xi'an City.
In 2004, the cross-sectional survey was conducted in six districts of Xi'an, a major city in northwest China with a population of 7 million that is undergoing rapid economic change. A representative sample of 1804 adolescents aged 11-17 yr attending junior high schools in Xi'an City were enrolled in the study. A multistage cluster-sampling method was used in which 30 schools were selected proportionate to school population size by systematic random sampling from a frame of all junior high schools in the city. In each selected school, one class out of each grade was randomly selected, and from each class, 20 students were selected, using systematic random sampling. The calculated sample size of 1640 was sufficient to estimate the prevalence of inactivity with a precision of 2.5% and 90% power. Group consent for the study was sought, starting with the municipal education department, the school health division at the district level, and the headmaster at the school level. Individual written informed consent was obtained from the adolescents and their parents. The study was approved by the human research ethics committee at the University of Newcastle and the ethics committee for medical research at Xi'an Jiaotong University.
The level of physical activity was assessed with an adolescent physical activity recall questionnaire that has been validated and found to have moderate reliability and validity in Australian adolescents (7). This questionnaire was adapted for Chinese adolescents to include popular Chinese activities such as jumping rope and parcel shooting, and activities such as cricket and squash were deleted. There were 42 activities listed on the questionnaire, with additional blank spaces to record other activities. The consenting participants estimated the time they spent in the specific activities either as organized or nonorganized activity in an average week in the spring and autumn semesters. This self-administered form took about 20 min to complete.
The intensity of the physical activities were rated with MET values as classified by Ainsworth et al. (1). Equivalent METs were assigned to those Chinese activities by consulting with a physical education instructor who identified a similar activity listed in the compendium. For example, the intensity of jumping elastic string is equivalent to that of jogging, so the MET of jogging was assigned to jumping elastic string.
Sociodemographic and environmental factors at community and household levels were collected by self-administered questionnaires completed on average in less than 30 min by consenting parents of the participants. Socioeconomic information included parental education, occupation, and an inventory of household assets for computing a wealth index as an indicator of the socioeconomic status of the household; this index has been validated in developing countries (19). Environmental factors including 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, parents' involvement with children doing exercise, household facilities for playing games, and family rules for playing games.
Environmental factors at the school level were obtained by asking school doctors to complete a precoded school environment questionnaire, which took 15 min to complete. Items related to physical activity included the availability of playgrounds, ovals, gyms, sports equipment, schedules for sports meetings, morning exercises, class recess exercise, physical education, and school policies about bicycle riding of students. The items in the environment questionnaires were based on those identified in focus group discussions with students, parents and school doctors, an expert panel meeting at Xi'an Jiaotong University, and from additional factors that were identified in other studies in Western countries (26,28) but not validated for this study.
Parental self-reported weight and height were also recorded (21). Children's height without shoes was measured using a nonstretchable tape suspended from the wall (214 Road Rod) and was recorded to the nearest 0.1 cm. Body weight was measured with the children barefoot and wearing underwear or light clothes, using a digital scale (Tanita HD-305, Japan) to the nearest 0.1 kg.
Duration estimates exceeding 12 h·d−1 for any given activity were considered reporting errors and were excluded from the analysis (18). Time estimates were summed within the same classification of intensity in spring and autumn as either organized or nonorganized in an average week. Then, the average physical activity time per week was computed. The distribution of the time doing physical activity was skewed and categorized based on thresholds of physical activity necessary for health benefits (24). Two definitions of physical inactivity were used in the analysis. In the first definition, active (coded as 0) is equal to or more than 150 min·wk−1 of moderate activity (4-6 METs) or 60 min·wk−1 of vigorous activity (≥ 6 METs); inactive (coded as 1) is less than 150 min·wk−1 of moderate activity (4-6 METs) or 60 min·wk−1 of vigorous activity (≥ 6 METs) (24). The second definition is based on the recent 60-min·d−1 moderate-to-vigorous (> 4 METs) activity cutoff (9). However, the assessment of the distribution among subgroups and the outcome for the logistic regression analysis was based on classifications from the earlier international physical activity guidelines to facilitate cross-country comparisons.
The difference in physical inactivity between subgroups was assessed with chi-square tests. There were five blocks of variables in the analysis. There were four factors at the community level, including recreational facility score, perceived neighborhood safety, local transportation, and residence. Recreational facility scores were computed by assigning 1 to those answering yes and 2 for those answering no, and then summing up the scores of all given facilities (open square, playground park, playfield or sports oval, stadium, swimming pool, and sport club) ranging from 7 to 14 and categorized into tertiles. Higher scores indicated more difficult access to recreational facilities.
There were 14 factors at the school level, including availability of playgrounds, ovals, gyms, and sports equipment; organized morning exercise and recess activity; extracurricular sports (yes, no), sports meetings (twice per year, once per year, once in 2 yr) and duration (1 d, 2 d, 3 d, > 3 d); physical education and health education sessions; and school bike-riding policy (allowed, forbidden). There were eight factors at the household level, including availability of a yard, vacant fields, sidewalks, or lanes (yes, no); floor of residence (low: 1-3 floor, medium 4-7 floor, high > 8 floor); means of getting into the house (by stairs, by elevator, no need); video game machine or shops around the home; and number of TVs at home (none, one, more than one). There were nine factors at the family level, including family size, wealth index, family rules for playing games or watching TV (yes, no), parental involvement in exercise with children, parental education, occupation, and BMI status. There were six individual factors, including age, gender, birth weight (low, < 2500 g; normal, 2500-3500 g; high, > 3500 g), ethnicity (Han is the major ethnic group, Hui is the main minority ethnic group in Xian), BMI status (normal, overweight/obese), and pubertal stage with validated confidential pictograms (20). Pubertal stage was classified into prepubescence, pubescence, or postpubescence according to WHO categorization based on pictogram questionnaire (36).
All candidate variables (P < 0.30) from univariate logistic regressions were included in logistic regression models for the sample and separately by gender. A hierarchical approach based on a conceptual framework describing the order of relationships between risk factors and physical activity was applied in identifying risk factors associated with physical inactivity (10,32). As seen in Figure 1, which illustrates this conceptual model, the factors at the community level may directly or indirectly determine all variables being studied; the next hierarchical level comprises school factors, which are partly determined by community factors; the third level includes household and the fourth level includes parental and family factors and, finally, the adolescents' personal characteristics.
In the hierarchical approach, the first step was to examine the association of physical inactivity with factors at the community level. All candidate variables at the community level were included in the model. Only those significantly associated with inactivity were retained for subsequent model building. In the second step, school variables were added into the first model, and only those significantly associated with inactivity (including dose-response patterns in variables with three or more categories) were retained. The household-level variables were then added to this new model, and a similar procedure was repeated for all levels.
The adjusted OR and the 95% CI presented in the results section were not derived from the final model of this procedure but from the equation corresponding to the level where the factor of interest was first entered and became significant before the next level of variables were added. This was intended to avoid the possibility that mediating variables might take away some of the explanatory power of more distant determinants. For example, part of the effect of community on physical inactivity may be mediated through parental education. The overall effect of community should therefore be examined in a model without parental education; otherwise, the role of community would be underestimated. This approach is superior to purely statistical associations (multivariate model) in that it allows an assessment of whether an effect is direct or mediated through other factors (31).
Data were analyzed using the statistical package STATA 9 (STATA Corporation, College Station, TX), and adjustment was made for the multistage cluster-sampling design, using the svy commands in which similarities within clusters were adjusted by Taylor linearized variance estimation.
Of the 1804 selected participants, 1787 completed the adolescent physical activity recall questionnaire; 889 of these respondents were girls and 898 were boys. Two records were excluded from analysis based on the exclusion criteria. Overall, using 150 min·wk−1 of moderate or 60 min·wk−1 of vigorous activity as the cutoff, 1505 (84.2%) were classified as physically active. Among the girls, 707 (79.5%) were categorized as physically active, and among the boys, 798 (88.9%) were physically active. In contrast, using 60 min·d−1 of moderate to vigorous activity as the cutoff, 44% of adolescents were classified as inactive. There were 1768 parents of participants who completed the household information questionnaires. All school doctors completed the school environment questionnaires.
Description of Physical Inactivity
Table 1 shows the prevalence of physical inactivity in adolescents based on the cutoff of 150 min·wk−1 of moderate or 60 min·d−1 of vigorous activity by characteristics of the adolescents, their parents, and family factors. Boys were significantly more active than girls (P < 0.001). The higher the paternal education, the lower the percentage of inactivity (P = 0.01). More inactive adolescents were found with overweight/obese than with normal-weight parents (P = 0.05). Parents frequently exercising with children or sparing time to exercise with children was inversely related with inactivity in adolescents (P = 0.01). Different patterns of inactivity were found for boys and girls depending on their ethnicity and BMI status. Overweight/obese ethnic minority (e.g., Hui and others except Han) boys were more inactive, whereas overweight/obese ethnic minority girls were more active than their counterparts. Among boys, activity levels decreased at the age of 13, then increased at 14, and decreased after 15 yr old (P = 0.01). The percentage of inactivity was higher in boys younger than 13 yr than in those older than 13 yr (16 vs 10%, P = 0.01). Prepubertal girls were more inactive than pubertal or postpubertal girls (P = 0.04). Girls from extended families were more likely to be active (P = 0.04).
Table 2 shows physical inactivity and its relationship with factors at the household, school, and community levels. At the community level, lack of recreational facilities was associated with a higher percentage of inactivity in the sample and in girls (P < 0.001). Boys traveling by automobile were more inactive compared with those on foot or by bike (18.8 vs 16.1 or 15.1%, P < 0.001). Perceived unsafe neighborhoods were associated with a higher percentage of inactive adolescents, but the difference was not statistically significant (P = 0.08). At the school level, lack of an oval was associated with inactivity (P = 0.04); lack of recess exercise or extracurricular sport or sports meetings was associated with higher percentages of inactivity, especially in boys (P < 0.01). Girls attending schools where bike riding was forbidden were more inactive (P < 0.01); surprisingly, for boys, this pattern was reversed. Only 1 of 30 schools in our survey forbade bicycle riding to school. At the household level, lack of vacant fields (P = 0.02) or sidewalks was related to a higher percentage of inactivity in adolescents (P = 0.01). Lack of video game shops around the house was associated with a higher percentage of inactive boys (P = 0.02).
Factors Associated with Physical Inactivity in Adolescents
Table 3 shows the results of the hierarchic multiple logistic regression models and highlights the factors associated with physical inactivity in the sample. At the community level, the access to public facilities (OR: 1.4, 95% CI: 1.0-1.9 for moderate and OR: 1.7, 95% CI: 1.2-2.4 for difficult) and concerns about neighborhood safety (OR: 2.1, 95% CI: 1.1-4.1) were positively associated with inactivity. At the school level, lack of extracurricular sports (OR: 1.3, 95% CI: 1.1-1.6) and sports meetings (OR: 2.0, 95% CI: 1.4-2.9) were significantly associated with physical inactivity, but physical education was inversely associated with inactivity (OR: 3.1, 95% CI: 1.6-6.0 for twice a week and OR: 2.6, 95% CI: 1.3-5.1 for three times a week). At the household level, adolescents living in a house without sidewalks were 30% more likely to be inactive (OR: 1.3, 95% CI: 1.0-1.6). At the family level, those adolescents whose fathers had secondary education were 40% less likely to be inactive (OR: 0.6, 95% CI: 0.4-0.9); boys were 50% less likely to be inactive than girls (OR: 0.5, 95% CI: 0.3-0.6), and those aged 14 yr were 30% less likely to be inactive than those less than 13 yr (OR: 0.7, 95% CI: 0.5-0.9), but there was no difference in physical activity level among other age groups.
Table 4 shows the result of model building conducted separately for each gender. In boys, those aged 13-13.9 yr were 60% less likely to be inactive than those younger than 13 yr (OR: 0.4, 95% CI: 0.2-0.97). When boys were classified into two groups (younger than 13 yr and older than 13 yr), those older than 13 yr were 40% less likely to be inactive than those younger than 13 yr (OR: 0.6, 95% CI: 0.4-0.9). After adjustment by age, boys passively transported to school were 3.2 (95% CI: 1.7-6.0) times more likely to be inactive than those traveling to school on foot. Adolescent boys living in surroundings without vacant fields were 1.7 times (95% CI: 1.2-2.5) more likely to be inactive. Boys were 40% less likely to be inactive if an adult in the family spent time exercising with them than those whose parents or other adults did not commit time to exercise (OR: 0.6, 95% CI: 0.4-0.95). Lack of class recess sports (OR: 2.2, 95% CI: 1.2-4.0) and sports meetings (OR: 1.5, 95% CI: 1.0-2.2) were associated with low levels of physical activity, and adolescents at schools forbidding bike riding to school were 60% less likely to be inactive (OR: 0.4, 95% CI: 0.2-0.8). Unavailability of video game shops near their home was associated with inactivity in boys by 50% (OR: 1.5, 95% CI: 1.1-2.1). Also, overweight/obese boys were 40% more likely to be inactive than those with normal BMI, but this difference was not statistically significant (OR: 1.4, 95% CI: 0.9-2.1).
Nine factors were significantly associated with inactivity in girls. As it became more difficult to access public facilities for exercise in the community, the risk of physical inactivity progressively increased (moderate OR: 1.7, 95% CI: 1.04-2.6; difficult OR: 2.4, 95% CI: 1.6-3.5). Fewer sports meetings (OR: 1.7, 95% CI: 1.03-2.8) or sports meetings at schools lasting more than 3 d (OR: 0.5, 95% CI: 0.2-0.97) were associated with inactivity. More sessions of physical education was associated with physical inactivity in girls (OR: 0.5, 95% CI: 0.2-0.97). Lack of sidewalks around the house was associated with physical inactivity (OR: 1.5, 95% CI: 1.04-2.0). In female adolescents, mothers staying at home (OR: 0.6, 95% CI: 0.4-0.9) or fathers attaining higher education (OR: 0.5, 95% CI: 0.3-0.9) was associated with higher levels of activity compared with mothers holding a nonprofessional job or fathers with lower education. Girls from an extended family were 60% less likely to be inactive (OR: 0.4, 95% CI: 0.2-0.8) than those from single-parent family. Those aged 14 yr were 40% less likely to be inactive than those aged younger than 13 yr (OR: 0.6, 95% CI: 0.4-0.97).
This is the first report assessing the association between physical inactivity and factors at community, school, household, family, and individual levels among adolescents in Xi'an City. There has been no study in China detecting the environmental factors correlated with physical inactivity in adolescents. This study identified environmental factors and individual factors of physical inactivity, and the factors varied in boys and girls. These findings are essential for the development of appropriate strategies to promote physical activity and to contribute to the prevention of overweight/obesity, an emerging public health problem in adolescents in Xi'an City (16).
Consistent with other studies (26,27), we found that boys were more physically active than girls, and this difference may be the result of perceived societal gender roles. Studies have indicated that girls report lower physical activity self-efficacy, less perceived benefits, and more perceived barriers to being physically active than boys, both in Chinese and U.S. adolescents (26,37).
A review of studies of physical activity in adolescents has concluded that physical activity declines with age (25). In this urban adolescent sample, especially in boys, physical activity increased at 13-14 yr and then decreased at 15-17 yr. One possible explanation is that children younger than 13 yr are transported to school and other events by adults, because, in China, it is not usual for children younger than 12 yr to ride a bicycle by themselves (31). Fourteen percent of children younger than 13 yr were chauffeured by adults with automobiles, compared with 8-10% of children in older age groups. Therefore, the younger groups tended to have fewer chances to be physically active. Another possible reason was systematic overreporting by older adolescents than by younger ones because of more consciousness about socially desirable behavior. In a society where tertiary education is highly valued, older adolescents would be likely to allocate more time to school courses to prepare for the competitive senior high school entrance examination. In addition, pubertal development, combined with an increase in self-awareness, may make adolescents reluctant to put themselves in situations where physical changes may be noticeable (11).
Among the indices of socioeconomic status, only paternal education was significantly associated with levels of physical activity in adolescents after adjustment for age and gender. Mother's occupation was dropped in the model-building process. Adolescents whose fathers had attained higher education tended to be more active. This reflected that among SES factors included in the analysis, only paternal education showed a strong association with level of physical activity.
The results of this study also showed that higher parental BMI was associated with adolescents' physical inactivity. The univariate analysis showed that if parents spent time exercising with children, adolescents had significantly higher levels of physical activity. This may be explained with the social cognitive theory proposed by Bandura (5) that describes physical activity as a modeling process, with the behavior of children and adolescents influenced by family members. There are studies showing that physically active parents tend to have physically active children and that physically inactive parents tend to have physically inactive children (12). Parents may influence the physical activity of children by their encouragement, assistance in organizing exercise sessions, and in providing home sports equipment. The results of our study indicated that effective promotion of physical activity in adolescents in Xi'an City should involve families because parents could potentially set models and establish a supportive atmosphere for adolescents.
School activities such as sports meetings, physical education sessions, and extracurricular sports were independently associated with physical inactivity. Schools can take advantage of sports meetings and physical education sessions to promote an active lifestyle for adolescents, either through modeling by school teachers, peers, and coaches or by training students to acquire positive knowledge and attitudes about physical activity and sport and activity skills to keep healthy. An intervention study in elementary schools conducted in the United States established that physical education curricula could provide students with substantially more physical activity during physical education classes (23). Based on these findings, targeted school approaches should be developed by combining the existing strategies from Western countries to promote physical activity of students. The unexpected direction between more sessions of physical education and physical inactivity in girls poses challenges for interpretation and needs further investigation.
Concerns about neighborhood safety were found to be positively associated with physical inactivity in adolescents in univariate and multivariate analysis. Similar results have been reported from a study in Australia where concerns about traffic and road safety, the absence of lights or crossings to help children cope with traffic, and the necessity to cross multiple roads to reach play areas were negatively associated with regular walking or cycling to local destinations (29).
The multiple logistic regression analysis indicated that lack of side walks and facilities such as playgrounds, open squares, sports ovals, and sports clubs in the community were significantly associated with physical inactivity. The association was consistent with another study in which environmental factors were reported either by parents or children themselves (29). We found that 90% of adolescents used active transportation (48% walked and 42% rode bicycles) to travel to school; thus, having safe sidewalks or footpaths is of great importance to maintain the current level of physical activity and to further promote activity through walking and cycling. These findings suggest that in China, physical activity-promotion programs should partly focus on urban planning to ensure the expansion of urban environments that are safe and convenient for communities and adolescents. In the United States and Australia, where automobiles dominate transportation, research has identified that environmental factors such as safe and convenient pedestrian paths and bicycle trails are associated with walking and cycling and that it is cost-effective to build safe pathways to promote physical activity (11,34).
Our gender-specific analyses revealed different factors associated with physical inactivity. Girls' physical activity was more influenced by family and socioeconomic status than boys, whereas boys were more influenced by family models. This result can be explained by studies indicating that boys seemed to receive more parental support and encouragement to be physically active than girls (4). It is postulated that boys are allowed more freedom to display aggressive behavior and to engage in more vigorous activities, whereas girls seem to be encouraged to be more dependent and less exploratory in their behavior (6). In Xi'an City, the environmental factors impacted boys mainly in transportation, recess exercise, spaces around house, school policies for bike riding, and parental involvement in exercising with them; for girls, the environmental factors were school physical education and facilities in the community and around the house. Research findings have shown significant differences in independent mobility between boys and girls. Boys enjoy far more independence than girls (4). However, in our study, for both boys and girls, school organized sports meetings were associated with increased physical activity.
Our study showed that having video game shops in the community was associated with increased physical activity in boys. In this study, we found that boys played activities such as snooker, skateboarding, basketball, and volleyball on the streets or close to streets where video games shops were located. These active boys would have had more opportunities to play video games away from home. This may explain the paradox of why the availability of video game shops seemed to be associated with increased physical activity. However, there is a need to explore how the availability of video game shops in these urban settings and playing games affect physical activity behavior. Watching TV and playing games differ in that children tend to have snacks and may watch attractive food commercials while watching TV, but when children play video games, they tend to concentrate on the games and have less food or even skip meals, although both take up time that could be spent being physically active.
It is impossible to compare the prevalence of physical activity level in our study with the national survey because of the different measurement instruments used. In addition, different definition criteria were used to classify active or inactive children (33). In our study, the percentage of physically inactive children was similar to the levels reported in studies in the United States (18) and an Australian survey (8). All of these studies used similar or the same physical activity recall questionnaire and the same definition for active and inactive children. Our results indicate the need for programs to promote physical activity in Chinese adolescents, considering that there have been none to date, in contrast to Western countries where various programs to promote physical activity in children have already been implemented (13,23). Using the recent guidelines, we found that only 54% of adolescents met the recommendations, and similar factors were identified at the levels of community, school, household, and personal characteristics. This implies the ample opportunity for and the great importance of launching physical activity-promotion programs among adolescents in Xi'an City from a public health perspective.
The limitations of this study include some of the instruments used and the study design. The environmental factors were measured with a questionnaire, with items identified from focus group discussions with the local community. However, this instrument was not specifically validated for use in a Chinese city. Future investigations should validate inventories of environmental factors relevant to urban settings in China. Also, the cross-sectional design limits the exploration of causal relationships between physical inactivity and the associated factors because of the lack of a time dimension to the relationships between the potential risks factors and levels of physical activity. Nonetheless, the study has provided useful information as a first stage in the identification of the determinants of physical inactivity in adolescents in Xi'an City. The strengths of the study include the representative sample of urban adolescents and the high response rate to all the types of data collected, which suggests that the findings should be of relevance to other cities in western China. Another strength of the study is the use of appropriate analytical methods, which took account of the cluster-sampling design and appropriately assessed the association of a variety of factors at different levels, including environmental factors, with physical inactivity in adolescents.
Our findings confirm the usefulness of the conceptual framework presented in Figure 1. We found factors at each level significantly associated with physical inactivity, indicating the need to consider factors at all levels when identifying determinants of physical inactivity. Some of the more proximal factors seemed to reduce the effects of more distal factors (e.g., paternal education reduced the effect of neighborhood safety concerns), implying the need for a hierarchical conceptual framework and appropriate analysis. Many of the factors associated with physical inactivity in our study were the same as those reported in studies from Western countries, such as age, gender, parental education (25), parental involvement in exercise with children (12), school curriculum (23), neighborhood safety concerns (29), and the availability of recreational facilities (29). The strength of association for these factors was different from our study because our analysis included environmental factors and individual factors at different levels, whereas the analyses of other studies were unadjusted bivariate correlation models in which only a limited number of factors were examined. The availability of video game shops and school rules for bicycle riding to school were also identified in our study but not reported in studies from Western countries. However, the factors we identified as associated with physical inactivity were similar overall to those already reported, suggesting that effective programs to promote physical activity developed in the West (13,23) could be modified and applied in China, although new interventions might be needed for those factors unique to China.
In summary, physical activity in adolescents from Xi'an City varied by gender, age, and paternal education. Conveniently located sports facilities in neighborhoods and communities were associated with physical activity independently of age and gender. School activities such as recess exercise, extracurricular sports, schedules of sports meetings, and physical education were associated with levels of physical of activity in boys and girls. Parents also influenced their children's activity behavior. Effective physical activity promotion should aim at personal behavior change through integrated family involvement, modification of school curricula, and supportive community strategies.
This research was sponsored by a research training fellowship (066971/2/02/A) from the Health Consequences of Population Change Program of The Wellcome Trust. We thank the Municipal Education Department in Xi'an City for coordination and all the field workers for their contributions in data collection. The authors have no professional relationships with any companies or institutions that would benefit from these research findings. The results of the present study have no direct relationship with any specific health care product and imply no endorsement of any particular product.
1. Ainsworth, B. E., W. L. Haskell, M. C. Whitt, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med. Sci. Sports Exerc.
2. Andersen, R., C. Crespo, S. Bartlett, L. Cheskin, and M. Pratt. Relationship of physical activity and television watching with body weight and level of fatness among children: results from the Third National Health and Nutrition Examination Survey. JAMA
3. Armstrong, N., and J. Welsman. Physical activity and muscular strength. In: Young People and Physical Activity
, N. Armstrong and J. Welsman (Eds.). New York: Oxford University Press, pp. 137-151, 1997.
4. Armstrong, N., and J. Welsman. Promoting physical activity. In: Young People and Physical Activity
, N. Armstrong and J. Welsman (Eds.). New York: Oxford University Press, pp. 245-260, 1997.
5. Bandura, A. Social Foundations of Thought and Action
. Englewood Cliff: Prentice-Hall, 1986.
6. Bar-Or, O., and T. Baranowski. Physical activity, adiposity, and obesity among adolescents. Pediatr. Exerc. Sci.
7. Booth, M. L., A. D. Okely, T. Chey, and A. Bauman. The reliability and validity of the adolescent physical activity recall questionnaire. Med. Sci. Sports Exerc.
8. Booth, M. L., A. D. Okely, T. Chey, A. Bauman, and P. Macaskill. Epidemiology of physical activity participation among New South Wales school students. Aust. N. Z. Pub. Health
9. Cavill, N., S. Biddle, and J. F. Sallis. Health enhancing physical activity for young people. Pediatr. Exerc. Sci.
10. Davison, K., and L. Birch. Childhood overweight: a contextual model and recommendations for future research. Obes. Res.
11. Duncan, M., and K. Mummery. Psychosocial and environmental factors associated with physical activity among city dwellers in regional Queensland. Pre. Med.
12. Fogelholm, M., O. Nuutinen, M. Pasanen, E. Myohanen, and T. Saatela. Parent-child relationship of physical activity patterns and obesity. Int. J. Obes.
13. Gortmaker, S. L., K. Peterson, J. Wiecha, et al. Reducing obesity via a school-based interdisciplinary intervention among youth
(Planet Health). Arch. Pediatr. Adolesc. Med.
14. Gutin, B., Z. Yin, M. Humphries, and P. Barbeau. Relations of moderate and vigorous physical activity to fitness and fatness in adolescents. Am. J. Clin. Nutr.
15. Kemper, H., J. Twisk, L. Koppes, W. V. Mechelen, and G. Post. A 15-year physical activity pattern is positively related to aerobic fitness in young males and females (13-17 years). Eur. J. Appl. Physiol.
16. Li, M., M. Dibley, D. Sibbritt, and H. Yan. An assessment of adolescent overweight and obesity in Xi'an City, China. Int. J. Pediatric. Obes.
17. MacKelvie, K. Physical activity and bone in childhood and adolescence. In: Physical Activity and Bone Health
, K. Khan, H. McKay, P. Kannus, D. Bailey, J. Wark, and K. Bennell (Eds.). Champaign, IL: Human Kinetics, pp. 111-128, 2001.
18. Marshall, S., S. Biddle, J. Sallis, T. McKenzie, and T. Conway. Clustering of sedentary behaviors and physical activity among youth
: a cross-sectional study. Pediatr. Exerc. Sci.
19. MICS. Assessing the economic status of household. End Decade Assessment. UNICEF Statistics, 2003.
20. Morris, N., and J. Udry. Validation of a self-administered instrument to assess stage of adolescent development. J. Youth Adolesc.
21. Niedhammer, I., I. Bugel, S. Bonefant, M. Goldberg, and A. Leclerc. Validity of self-report weight and height in French GAZEL cohort. Int. J. Obes.
22. Riddoch, C. Relationships between physical activity and physical health in young people. In: Young and Active? Young People and Health-Enhancing Physical Activity-Evidence and Implications
, S. Biddle, J. Sallis, and N. Cavill (Eds.). London: Health Education Authority, pp. 17-48, 1998.
23. Sallis, J., T. McKenzie, J. Alcaraz, N. Faucette, and M. Hovell. The effect of a 2-year physical education program (SPARK) on physical activity and fitness in elementary school students. Am. J. Public Health.
24. Sallis, J., and K. Patrick. Physical activity guidelines for adolescents: consensus statement. Pediatr. Exerc. Sci.
25. Sallis, J., J. Prochaska, and W. Taylor. A review of correlated of physical activity of children and adolescents. Med. Sci. Sports Exerc.
26. Sallis, J., J. Zakarian, M. Hovell, and C. Hofstetter. Ethnic, socioeconomic, and sex differences in physical activity among adolescents. J. Clin. Epidemiol.
27. Sallis, J., J. J. Prockaska, W. C. Taylor, J. O. Hill, and J. C. Geraci. Correlates of physical activity in a national sample of girls and boys in grades 4 through 12. Health Psych.
28. Santos, M., C. Esculcas, and J. Mota. The relationship between socioeconomic status and adolescents' organized and non-organized physical activities. Pediatr. Exerc. Sci.
29. Timperio, A., D. Crawford, A. Telford, and J. Salmon. Perceptions about the local neighborhood and walking and cycling among children. Pre. Med.
30. Trudeau, F., L. Laurencelle, and R. Shephard. Tracking of physical activity from childhood to adulthood. Med. Sci. Sports Exerc.
31. Tudor-Locke, C., B. E. Ainsworth, L. S. Adair, S. Du, and B. M. Popkin. Physical activity and inactivity in Chinese school-aged youth
: the China Health and Nutrition Survey. Int. J. Obes.
32. Victora, C., S. Huttly, S. Fuchs, and M. Olinto. The role of conceptual frameworks in epidemiological analysis: a hierarchical approach. Int. J. Epidemiol.
33. Wang, D. Behavior
and lifestyle of Chinese population. In: Summary Report of China National Nutrition Survey 2002
, X. Qi, C. Chen, J. Ma, L. Li, and K. Rao (Eds.). Beijing: People's Health Publication House, pp. 45-48, 2004.
34. Wang, G., C. S. Macera, B. Scudder-Soucie, T. Schmid, M. Pratt, and D. Buchner. Cost effectiveness of a bicycle/pedestrian trail development in health promotion. Pre. Med.
35. WHO. Health and Development Through Physical Activity and Sport
, Geneva: World Health Organization, p. 1, 2004.
36. WHO. Physical Status: The Use and Interpretation of Anthropometry. Report of a WHO Expert Committee. World Health Organ Tech. Rep. Ser.
37. Wu, T., N. Pender, and S. Noureddine. Gender differences in the psychosocial and cognitive correlates of physical activity among Taiwanese Adolescents: a structural equation modeling approach. Int. J. Behav. Med.