High Levels of Physical Inactivity and Sedentary Behaviors Among US Immigrant Children and Adolescents

Objective  To examine the prevalence and correlates of physical inactivity and sedentary behavior among immigrant and US-born children.

Design  Cross-sectional analysis using data from the 2003 National Survey of Children's Health, a telephone survey conducted between January 29, 2003, and July 1, 2004.

Setting  United States.

Participants  Multivariate logistic and least squares regression models were used to analyze immigrant differentials among 68 288 children aged 6 through 17 years.

Main Exposure  Ethnic-immigrant status.

Main Outcome Measures  Prevalence and odds of regular physical activity, inactivity, television watching, and lack of sports participation.

Results  Physical inactivity and sedentary behaviors varied widely among children in various ethnic-immigrant groups. For example, 22.5% of immigrant Hispanic children were physically inactive compared with 9.5% of US-born white children with US-born parents. Approximately 67% of immigrant Hispanic children did not participate in sports compared with 30.2% of native Asian children. Overall, immigrant children were significantly more likely to be physically inactive and less likely to participate in sports than native children; they were, however, less likely to watch television 3 or more hours per day than native children, although the nativity gap narrowed with increasing acculturation levels. Compared with native white children, the adjusted odds of physical inactivity and lack of sports participation were both 2 times higher for immigrant Hispanic children with foreign-born parents, and the odds of television watching were 1.5 and 2.3 times higher for native Hispanic and black children, respectively.

Conclusions  Immigrant children in each ethnic minority group generally had higher physical inactivity and lower sports participation levels than native children. To reduce disparities, health education programs need to promote physical activity among children in immigrant families.

Because of a dramatic increase in the prevalence of childhood obesity and diabetes mellitus during the past 2 decades, physical activity (PA) has assumed an increasingly prominent role in disease prevention and health promotion efforts in the United States and is considered 1 of the 10 leading health indicators for the nation.^1-5 Consequently, monitoring levels of regular PA and sedentary behavior among both children and adults has become an important public health surveillance activity in the United States.^2-4

Although data on the prevalence of childhood physical inactivity and sedentary behavior are routinely available by sex and major racial/ethnic groups in the United States,^2,3,6,7 such estimates for immigrant children are less well known.⁸ The immigrant population in the United States has increased considerably in the last 36 years. In 2006, there were 37.5 million immigrants, an increase of 27.9 million since 1970. Immigrants now account for 12.6% of the total US population.^9-11 The increase in the immigrant child population has also been substantial. The proportion of US children living with at least 1 foreign-born parent increased from 12.1% in 1990 to 18.1% in 2000. Of these children, 4.4% were foreign born in 2000.¹²

Given such a rapid increase in the immigrant population, it is important to know how patterns of PA, inactivity, and sedentary behaviors for this increasing segment of the population differ from those of the majority native population. Health, social, and behavioral profiles of immigrants vary substantially from those of the US-born population.^11,13,14 There is also evidence that PA and other obesity-related behaviors differ between immigrants and the US-born population and that acculturation modifies the health and behavioral risks of immigrants.^8,11,13,15 The purposes of this study were, therefore, to estimate the prevalence of PA, inactivity, lack of sports participation, and television watching among immigrant and US-born children aged 6 through 17 years after adjusting for age, sex, race/ethnicity, socioeconomic status (SES), place of residence, and neighborhood safety using a large, nationally representative sample of US children and to examine the extent to which immigrant PA patterns vary by ethnicity and level of acculturation.

The data for the present analysis are from the National Survey of Children's Health (NSCH).^16-18 The survey was conducted by the National Center for Health Statistics with funding and direction from the Maternal and Child Health Bureau of the Health Resources and Services Administration.^17,18 The survey included an extensive array of questions about the family, including parental health, stress and coping behaviors, family activities, and parental concerns about their children.^17,18 Interviews were conducted with parents, and special emphasis was placed on factors related to children's well-being.

The NSCH was a telephone survey conducted between January 29, 2003, and July 1, 2004. It had a sample size of 102 353, including a sample of approximately 2000 children per state. A random-digit-dial sample of households with children under 18 years of age was selected from each of the 50 states and the District of Columbia. One child was randomly selected from all children in each identified household to be the subject of the survey. The respondent was the parent or guardian who knew most about the child's health status and health care. Consequently, all NSCH data were based on parental reports. The interview completion rate, measuring the percentage of completed interviews among known households with children, was 68.8%.¹⁸ Substantive and methodologic details of the survey are described elsewhere.¹⁸ The National Center for Health Statistics Research Ethics Review Board approved all data collection procedures.

Analysis of immigrant differentials was conducted for 68 288 children aged 6 through 17 years for whom complete information on PA was available. Physical activity, the dependent variable, was derived from the question, “During the past week, on how many days did the child exercise or participate in physical activity for at least 20 minutes that made him/her sweat and breathe hard, such as basketball, soccer, running, swimming laps, fast bicycling, fast dancing, or similar aerobic activities?” The physical inactivity variable was defined as the percentage or likelihood of children with no days of vigorous PA in the past week. Regular, vigorous PA was defined as the percentage or likelihood of children engaging in 3 or more days of PA in the past week, following the Centers for Disease Control and Prevention recommendations and in accordance with the Healthy People 2010 guidelines.^2,5-7 A third measure of PA was derived that indicated the number of days of physical inactivity among children in the past month.

We considered 2 additional outcome measures of sedentary behaviors among children.⁶ Lack of sport participation was a dichotomous variable derived from the question, “During the past 12 months, was the child on a sports team or did s/he take sports lessons after school or on weekends?” Television viewing was based on the question, “On an average school day, how many hours does the child usually watch television, watch videos, or play video games?” The number of hours spent per day watching television was coded into 2 categories: less than 3 and 3 or more. This classification is based on the current Centers for Disease Control and Prevention recommendations for tracking this indicator and on prior research showing increased obesity risks associated with 3 or more hours per day of television viewing.^6,8,19

Nativity/immigrant status, the main covariate of interest, was defined on the basis of children's own nativity and that of their parents.^8,15,20-23 It consisted of 4 categories: foreign-born children with both immigrant parents (first generation), US-born children with both immigrant parents (second generation), US-born children with 1 immigrant parent (second generation), and US-born children with both US-born parents (third or higher generation). The first 3 groups together make up the overall immigrant category that consisted of children born to 1 or both immigrant parents, whereas the fourth category consists of third-generation or higher, native-born children (also referred to as native children). Race/ethnicity was classified into 4 categories: non-Hispanic white, non-Hispanic black, Hispanic, and all other ethnic groups, including Asians. The joint variable of ethnic-immigrant status included 12 categories. Although white and black children were classified into 2 broad immigrant groups because of small numbers, Hispanic and “all other ethnic groups” each consisted of 4 immigrant groups.

Although in the study of adult health differentials, immigrant status is defined solely on the basis of an individual's country of birth information,^11,13,14 defining immigrant status based on both the child's and parental nativity is the preferred approach in studies of child health and well-being.^8,15,20-23 This is mainly because parental characteristics are considered critical in shaping and influencing physical and psychological health and behavioral risks of children and adolescents, particularly those of young children. In addition, children who are born abroad make up less than one-third of all children born in immigrant families.^12,21 Furthermore, taking into account the generational status of children allows us to measure acculturation, estimating the effect of which on PA is 1 of the aims of our study.^8,15,20,21

In addition to immigrant status and race/ethnicity, we considered the following sociodemographic factors that are known to influence the likelihood of inactivity and sedentary behaviors in children: age (6-9, 10-11, 12-14, and 15-17 years), sex, household composition (2-parent biological or stepfamilies, single mother, and other), metropolitan/nonmetropolitan residence, household or parental educational level (<12, 12, and ≥13 years), household poverty status as a ratio of annual family income to poverty threshold (<100%, 100%-199%, 200%-399%, and ≥400%), and perceived neighborhood safety.^1-3,7,24-28

Primary language spoken at home (English or any other language) was strongly associated with immigrant status, our primary independent variable. Language spoken at home was, thus, used as a defining characteristic of immigrant status and a dimension of acculturation rather than a covariate.^13,21,29 The neighborhood safety variable was based on the question, “How often do you feel the child is safe in your community or neighborhood: never, sometimes, usually, or always?” Neighborhood safety was dichotomized by grouping the responses usually and always as “safe” and never and sometimes as “unsafe.”

Fewer than 0.5% of the observations had missing data for the PA and other sedentary outcomes, which were excluded from the analysis. For each relevant covariate, the missing or unknown responses were used as a separate category in regression models instead of excluding them from the multivariate analyses, which would have resulted in a significant decrease in the effective sample size available for analysis. There were no missing data for age and sex in the study sample.

The χ² statistic was used to test the overall association between each covariate and the prevalence of each outcome. The t statistic was used to test the difference in prevalence between any 2 groups. Multivariate logistic regression was used to examine the association between each binary PA outcome and selected sociodemographic factors, including immigrant status. Least squares regression was used to model the mean number of days of physical inactivity in the past month, results of which are discussed but are not shown in tabular form for the sake of brevity. Because household educational level and annual income were strongly correlated (γ [the magnitude of the ordinal association] = 0.65), to avoid estimation problems due to colinearity, either household poverty or educational level along with other covariates were used as predictors in the multivariate models. Three sets of multivariate models were estimated. The first set of regression models yielded the adjusted effect of the overall 4-category immigrant status after adjusting for race/ethnicity, age, sex, household composition, metropolitan/nonmetropolitan residence, household poverty status, and neighborhood safety. The second set of models included the 12-catergory ethnic-immigrant status along with age, sex, household composition, metropolitan/nonmetropolitan residence, household poverty status, and neighborhood safety. The third set of models was similar to the second set, with parental educational level substituted for household poverty status. Sex-specific interactions were examined by estimating separate regression models of inactivity and sedentary behaviors for boys and girls. Adjusted prevalence estimates were derived from the fitted logistic and least squares models. To account for the complex sample design of the NSCH, SUDAAN software was used to conduct all statistical analyses.³⁰

The 4 broad immigrant groups varied substantially in their household composition, SES, language use, perceived neighborhood safety, and place of residence (Table 1). Although 79.4% of children with at least 1 foreign-born parent lived in traditional 2-parent households, 28.1% of US-born children with both US-born parents lived in single-mother households. At least 35% of children with both immigrant parents lived below the poverty line compared with 14.7% of US-born children with both US-born parents and 9.6% of children with at least 1 immigrant parent. Of children with both immigrant parents, 67.6% lived in non–English-speaking households compared with 3.5% of US-born children with both US-born parents. More than 27% of children with both immigrant parents reported living in unsafe neighborhoods compared with 14.3% of US-born children with both US-born parents.

Table 2 gives the prevalence of PA, inactivity, and sedentary behaviors by immigrant status and selected sociodemographic covariates. For all US children 11.4% were physically inactive, and 73.5% of children engaged in regular PA 3 or more days per week. There was 42.1% of children who did not participate in sports, whereas 17.0% of children watched television 3 or more hours per day. Differentials by immigrant status were substantial. Of immigrant Hispanic children with both immigrant parents, 22.5% were physically inactive compared with 9.5% of US-born children with both US-born parents. Immigrant Hispanic children were, on average, inactive for 16 days in the past month, 4 days more than US-born white children with both US-born parents. Of immigrant Hispanic children, 66.6% did not participate in sports compared with 30.2% of US-born Asian children with both US-born parents. Of US-born black children with both US-born parents, 31.9% watched television 3 or more hours per day compared with only 5.4% of US-born Asian children with both US-born parents. Overall, immigrant children were significantly more likely to be physically inactive and less likely to engage in regular PA and participate in sports than US-born children with both US-born parents; they were, however, less likely to watch television 3 or more hours per day than US-born children with both US-born parents, although the nativity gap narrowed with increasing levels of acculturation.

Table 3 and Table 4 give the odds and likelihood of inactivity and sedentary behaviors among various ethnic-immigrant groups after adjusting for all of the socioeconomic and demographic differences outlined in Table 1 except household language use. Compared with US-born white children with both US-born parents, the adjusted odds of physical inactivity were at least 113% higher for Hispanic children with both immigrant parents, 28% higher for US-born Hispanic children with both US-born parents, and 18% higher for US-born black children with both US-born parents. Compared with US-born white children with both US-born parents, the odds of regular PA were approximately 50% lower for immigrant Hispanic children, 21% lower for US-born Hispanic children with both US-born parents, and 40% lower for immigrant black children. Compared with US-born white children with both US-born parents, the odds of not participating in sports were at least 99% higher for Hispanic children with both immigrant parents, 67% higher for immigrant black children, and 25% higher for US-born black children with both US-born parents. Compared with US-born white children with both US-born parents, the odds of excess television watching were 51% higher for US-born Hispanic children with both US-born parents, 31% higher for immigrant Hispanic children with both immigrant parents, and 129% higher for US-born black children with both US-born parents. Adjusted immigrant differentials in physical inactivity and PA levels generally did not differ by sex (data not shown).

The effects of other covariates were as expected (Table 3 and Table 4). Compared with children aged 6 through 9 years, adolescents aged 15 through 17 years were 4 days more inactive in the past month and had 291% higher odds of inactivity, 54% lower odds of PA, and 53% higher odds of watching television 3 or more hours per day. Although girls were substantially more likely to be inactive than boys, they had 24% lower odds of watching television 3 or more hours per day than boys. Compared with children with annual family incomes that exceeded 400% of the poverty threshold, those below the poverty threshold had 83% higher odds of inactivity, 340% higher odds of not participating in sports, and 152% higher odds of watching television 3 or more hours per day. Compared with children of college-educated parents, children with parents without a high school diploma had 71% higher odds of physical inactivity, 217% higher odds of not participating in sports, and 102% higher odds of watching television 3 or more hours per day. Children in single-mother households had 50% higher odds of physical inactivity, 23% lower odds of PA, 36% higher odds of not participating in sports, and 20% higher odds of watching television 3 or more hours per day than children in 2-parent households. Children in nonmetropolitan areas had 12% higher odds of inactivity and 15% lower odds of PA than children in metropolitan areas. Compared with children in safe neighborhoods, children living in unsafe neighborhoods had 18% higher odds of inactivity, 23% lower odds of PA, 26% higher odds of not participating in sports, and 28% higher odds of watching television 3 or more hours per day.

To our knowledge, this is the largest US study that has examined the prevalence and sociodemographic determinants of physical inactivity and sedentary behaviors among US-born and immigrant children, using a nationally representative sample of 68 288 children. This study defined immigrant status on the basis of both parental nativity and nativity status of children,^8,15,20,21 which allowed us to also assess the impact of acculturation, albeit indirectly, on PA and sedentary behaviors. Another unique aspect of the study involved assessing differential effects of immigrant status on several sedentary outcomes in Hispanic children and in children of black, white, and Asian ethnicities. Although the role of acculturation and other social correlates in PA and obesity-related behaviors has previously been examined among Hispanic immigrants,⁸ our study is the first attempt, to our knowledge, to examine these patterns in non-Hispanic immigrant groups at the national level.

Substantial immigrant differentials in PA and physical inactivity levels remained even after controlling for several socioeconomic and demographic characteristics. Lower PA and higher inactivity levels among immigrant children may, at least, partly reflect cultural and normative influences that may not necessarily promote participation in leisure-time PA and organized sports.³ Immigrant families may not be fully aware of the many physical and psychological health benefits of PA and sports participation.^3,24,25 In addition, immigrant parents may place higher value on children's time devoted to reading and learning activities, language lessons, academic performance, and participation in family activities rather than on PA.²³ Immigrant parents may also discourage their children's participation in sports because of fear of bullying, linguistic barriers, and household employment situation.²⁹ Higher inactivity levels in immigrants may also partly reflect immigrant differences in individual characteristics (such as obesity and health status), familial support, socioeconomic characteristics other than annual household income and educational level, parental behaviors such as parental PA levels, and differences in the physical or built aspects of the neighborhood environment, such as access to recreational facilities, outdoor parks, playgrounds, bike trails, walking paths or sidewalks, modes of transportation and vehicular traffic congestion, and media advertising promoting PA.^{1,3,19,24,25,28,31}

Future research is needed to examine the role of these specific cultural and parental influences and social environmental factors in explaining ethnic-immigrant differentials in PA shown herein. Although ethnic-immigrant patterns in PA were similar for boys and girls, they may vary by age and SES. From a policy standpoint, it would be useful to identify if children of immigrant parents in specific age cohorts or SES strata are at particularly high risks of physical inactivity and sedentary behavior compared with children of US-born parents.

Immigrant patterns in PA outcomes shown herein differ markedly from immigrant differentials observed for other health behaviors and health outcomes in the United States, including breastfeeding, smoking, obesity, infant mortality, low birth weight, morbidity, mortality, and life expectancy.^11,13-15,20 Immigrants have been shown to have a significant advantage over their US-born counterparts in these health and behavioral outcomes, which tends to decrease with increasing acculturation levels or length of residence in the United States.^11,13,14 Acculturation is a process by which immigrants adopt the values, attitudes, beliefs, practices, and lifestyle characteristics of the native-born population.^11,13,14 Higher inactivity levels in immigrants and increased PA associated with acculturation are consistent with similar immigrant patterns observed for preventive health services and health care use in the United States.^14,22,32 The positive association between acculturation and PA has also been noted for adults.³³ However, acculturation has been associated with increased risks of other obesity-related behaviors, such as fast food consumption and television watching in adolescents,³⁴ a finding consistent with our study.

Household language use is just 1 measure of acculturation that we used to differentiate the 4 broad immigrant groups.^13,21,29 To examine more fully the impact of acculturation on immigrants' risks of physical inactivity and sedentary behaviors, other direct measures of acculturation are needed. These measures may include the extent to which one identifies with one's own culture, celebrates one's own holidays or cultural events, eats one's own ethnic foods, participates in one's own ethnic-cultural networks, and is accepted by members (friends, relatives, or coworkers) of one's own ethnic group or the majority group.^13,21,35

This study has some limitations. First, the childhood PA and inactivity, sports participation, and television viewing measures in our study were based on parental reports and may not accurately reflect the true prevalence. However, the NSCH-based estimates are consistent with the self-reported estimates from the Youth Risk Behavior Survey,^2,6,7 which validates our analysis. Second, our analysis of immigrant and acculturation patterns is limited because the NSCH lacked data on such variables as the length of immigration, citizenship, naturalization, and legal status.^{9,11,13,14,21} Third, Asian Americans account for more than a quarter of the US immigrant population. However, we have presented a limited PA analysis for Asian children because NSCH data for them were available for only 5 states. Furthermore, the survey did not identify specific Hispanic and Asian subgroups, who are extremely heterogeneous in their socioeconomic, behavioral, and health characteristics and who are, therefore, expected to also differ in their PA patterns.^{9,11,13,21,32}

In conclusion, immigrant children in each ethnic minority group generally had higher levels of physical inactivity and lower levels of sports participation than children of US-born parents. Acculturation, as measured in this study, appears to have a beneficial impact in terms of increased PA and sports participation in children, particularly among Hispanic individuals. However, the evidence presented herein reveals that more than one-third of all Hispanic, Asian, and black immigrant children and adolescents fail to meet the recommended levels of PA. Lack of participation in organized sports remains a major concern in immigrant children, with more than half of Hispanic and black immigrant children not participating in sports. Given the health benefits of PA,^3,24,25 continued higher physical inactivity and lower activity levels in immigrant children are likely to reduce their overall health advantage over US-born populations during adulthood. To reduce disparities in childhood PA, health education programs designed to promote PA should target not only children from socially disadvantaged households and neighborhoods but also children in immigrant families.

Correspondence: Gopal K. Singh, PhD, Maternal and Child Health Bureau, Health Resources and Services Administration, US Department of Health and Human Services, 5600 Fishers Ln, Room 18-41, Rockville, MD 20857 (gsingh@hrsa.gov).

Accepted for Publication: January 24, 2008.

Author Contributions: Dr Singh had full access to all of the data and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Singh. Acquisition of data: Singh. Analysis and interpretation of data: Singh, Yu, Siahpush, and Kogan. Drafting of the manuscript: Singh. Critical revision of the manuscript for important intellectual content: Singh, Yu, Siahpush, and Kogan. Statistical analysis: Singh, Yu, and Siahpush. Study supervision: Singh and Kogan.

Financial Disclosure: None reported.