Generation, Language, Body Mass Index, and Activity Patterns in Hispanic Children

Abstract

Background

The acculturation hypothesis proposes an overall disadvantage in health outcomes for Hispanic immigrants with more time spent living in the U.S., but little is known about how generational status and language may influence Hispanic children’s relative weight and activity patterns.

Purpose

The association between generation and language was investigated with relative weight (BMI z-scores), physical activity, screen time, and participation in extracurricular activities (e.g., sports, clubs) in a U.S.-based, nationally representative sample of Hispanic children.

Methods

Participants included 2,012 Hispanic children aged 6–11 years from the cross-sectional, 2003 National Survey of Children’s Health. Children were grouped according to generational status (1^st, 2^nd or 3^rd), and the primary language spoken in the home (English vs non-English). Primary analyses included adjusted logistic and multinomial logistic regression to examine the relationships among variables; all analyses were conducted between 2008 and 2009.

Results

Compared to 3rd generation, English speakers, 1^st and 2^nd generation, non-English speakers were over two times more likely to be obese. Moreover, 1^st generation, non-English speakers were half as likely to engage in regular physical activity and sports. Both 1^st and 2^nd generation, non-English speakers were less likely to participate in clubs compared to 2^nd and 3^rd generation, English speakers. Overall, all non–English speaking groups reported less screen time compared to 3^rd generation, English speakers.

Conclusions

The hypothesis that Hispanics lose their health protection with more time spent in the U.S. was not supported in this sample of Hispanic children.

Introduction

Hispanics have become the largest racial/ethnic minority group in the U.S., and are projected to constitute one quarter of the total U.S. population by 2050 (U.S. Census, 2008). Recently, immigrants and their U.S.-born children are making up an increasing proportion of the total U.S. population,¹ the majority of which are of Hispanic or Asian origin.² Children of immigrants have contact with multiple cultural influences, which ultimately can affect their health status and behaviors.³ Research suggests that social, economic, and environmental factors, as well as access to health care, influence the health of populations.^4,5 Studies have shown that adult immigrants tend to have better health, longer life expectancy and lower mortality rates when compared to their U.S.-born counterparts.^6–8 However, studies have also demonstrated that this health protection tends to decrease with more time spent in the U.S.

The acculturation hypothesis, proposes an overall disadvantage in health outcomes for Hispanic immigrants with more time spent living in the U.S.^9,10 Acculturation has been defined as the process by which immigrants adopt the attitudes, values, customs, beliefs and behaviors of a new culture.¹¹ Previous studies have linked acculturation in Hispanic adults with various negative health outcomes including obesity, low fruit and vegetable intake, high fat and sugar intake, smoking, and alcohol consumption;¹² however, the findings relating physical activity to acculturation in this adult population have been inconsistent. Several studies report a positive association between acculturation and physical activity,^13–20 while others have shown a relationship between higher acculturation and lower levels of physical activity in Hispanic adults.^21,22 Therefore, it is not clear whether patterns of physical activity behaviors in Hispanic adults are consistent with the acculturation hypothesis. It is evident that the influence of acculturation on Hispanic health behaviors and outcomes is quite complex and not well understood.²³ While there have been many studies examining the influence of acculturation on health behaviors in adults, little is known about the effects of acculturation on health behaviors in Hispanic children.

Of the few studies that have investigated the role of acculturation on health outcomes in Hispanic youth, most have observed a relationship between acculturation and poorer health and health behaviors.^24–29 Similar to what is found in the adult and adolescent literature, there seems to be a trend for BMI to increase with subsequent generations in the U.S. for Hispanic children.^24,27,29 However, to the author’s knowledge, very few data exist which examine screen time or extracurricular activity prevalence in Hispanic immigrant children. Of the studies the authors identified that examine the association between acculturation and physical activity in Hispanic immigrant children, the results have been conflicting, and include a wide variety of measures to estimate acculturation or physical activity.^{3,25,27,30–33} The purpose of the present study is to investigate associations among generational status and language with relative weight (BMI z-scores), screen time, physical activity patterns and sport and club participation in a nationally representative sample of Hispanic children aged 6–11 years living in the U.S.

Methods

Participants

Data were drawn from participants in the 2003 National Survey of Children’s Health (NSCH), a nationally representative sample of children aged 0–17 years in the U.S. This random-digit-dial survey was conducted by the Health Resources and Services Administration’s Maternal and Child Health Bureau (MCHB) in collaboration with the CDC’s National Center for Health Statistics (NCHS). Details on the survey design are described elsewhere.^34,35 A Spanish-language version of the NSCH questionnaire was created after extensive review and evaluation for accuracy and cultural appropriateness by a team of experienced Spanish-language telephone interviewers and supervisors. In Spanish-speaking households, NSCH interviews were administered in Spanish by a trained bilingual interviewer; families were given the choice to respond in either English or Spanish. In the present study, the full NSCH sample (N =102,353) was limited to Hispanic children aged 6–11 years (n = 3914). Further, children were excluded if they were underweight (n = 210) or had missing data on gender (n = 2), BMI category (n = 1358), language (n = 113), poverty (n = 472), screen time (n = 40), sports (n = 14) and club participation (n = 8), and weekly physical activity (n = 58), resulting in a final sample of 2,022 children. This study was deemed exempt by the Pennsylvania State University IRB.

Measures

Generational Status

Child’s generational status was based on parental report of whether the child and his or her parents were born in the U.S. Generational status was coded according to recommendations from the IOM report on immigrant children: 1st generation—an immigrant to the U.S. not preceded by parents or other family members; 2nd generation—U.S.-born child of a 1st generation immigrant; and 3rd generation—U.S.-born child of a 2nd generation immigrant.³⁶

Language

Child’s language was assessed by asking parents “What is the primary language spoken in your home?” Responses were recoded to 0 (English), and 1 (non-English). See Table 1 for a breakdown of generation groups by language. The 1st generation, English speaking group consisted of only 10 children and thus, was excluded from this study.

Table 1.

Classification of sample participants into generation/language groups by country of birth and primary language

	Country of birth
Groups	n	Mother	Father	Child	Primary language
1st generation, non-English	141	F	F	F	NoE
1st generation, English	10	F	F	F	E
2nd generation, non-English	365	F	F	U.S.	NoE
2nd generation, English	113	F	F	U.S.	E
3rd generation, non-English	78	At least one parent U.S.		U.S.	NoE
3rd generation, English	1315	At least one parent U.S.		U.S.	E

Note: Children were categorized according to generational status (i.e., parent and child place of birth) and primary language spoken in the home (i.e., English, non-English).

E, English speaking; F, foreign born; NoE, non-English speaking; U.S., born in the U.S.

Weight Status

Based on parent-reported child height and weight, age- and gender-specific BMI z-scores were computed using the CDC Growth Charts.³⁷ Child BMI categories were based on the 2007 Expert Committee recommendations³⁸ adopted by the CDC: normal weight (BMI <85th percentile and >5th percentile), overweight (BMI ≥ 85th percentile and < 95th percentile), and obese (BMI ≥ 95th percentile). Underweight children (BMI ≤5th percentile) were not included in the sample.

Physical Activity

Children’s participation in physical activity was determined by asking parents how many days in the past week their child exercised or participated in physical activity for ≥20 minutes that made him or her breathe hard and sweat (i.e, basketball, running, or fast bicycling); response options ranged from 0 to 7. The variables no physical activity (yes=1/no=0) was defined as no reported days of physical activity per week, and regular physical activity (yes=1/no=0) as 3 or more days of physical activity per week.³⁹

Screen time

Parents reported the average number of hours their child spends watching TV, videos, or playing video games on a typical school day; response options ranged from 0 to 24. For the purposes of this study, this variable was recoded to 0 (≤2 hours/day) and 1 (>2 hours/day) based on the American Academy of Pediatrics (AAP) recommendations for children.⁴⁰

Sports participation

Parent’s responded yes or no to whether their child was on a sports team or took sports lessons after school or on the weekends in the past 12 months.

Club participation

Parent’s responded yes or no to whether their child participated in any clubs or organizations after school or on the weekends in the past 12 months (i.e., Scouts, a religious group, or Boy/Girl’s club).

Covariates

Income, household education, child age and gender were included as covariates in the analyses. Highest education level in the household was coded as 0 (<12 years), 1 (12 years), 2 (>12 years). Child’s gender was coded as 0 (male), 1 (female). Income was coded as “poor” (below 133% poverty), “near poor” (at 133% poverty and below 185% poverty), and “not poor” (at or above 185% poverty) based on 2002 and 2003 USDHHS Federal Poverty Guidelines for households.

Statistical Analysis

All data analyses were completed between 2008 and 2009 using STATA 9.1 (Stata Corporation, College Station, TX) to account for complex sample design.⁴¹ All proportions and means were weighted to provide estimates for the population of Hispanic children living in the U.S. Associations between generation and language groups for weight status, physical activity, screen time, and club and sports participation were investigated using adjusted logistic and multinomial logistic regression, followed by post hoc multiple comparisons. Logistic regression analyses allowed us to account for unequal sample sizes and each model was adjusted for the covariates: household income and education, and child age and gender. The authors hypothesized that these covariates would have influence on Hispanic children’s physical activity and screen time. In each model, the generation and language groups were dummy coded and the 3^rd generation, English group served as the comparison group. This group was selected as the comparison because they were the group living the longest in the U.S., and whose primary language was English. To examine whether the odds of the dependent variables differed among the noncomparison generation and language groups (i.e., 2^nd generation, English versus 2^nd generation, non-English), post estimation commands in STATA were used to run unadjusted and Bonferroni adjusted post hoc multiple comparisons (p<.05) following each model.

Results

Weighted sample descriptives by generation and language groups are presented in Table 2. The average age of the respondent’s children was 8.7 years, and 50% were women. Over one half of the children resided in poor or near poor households; yet, nearly 50% of the households had education levels exceeding 12 years. The overall proportions of obese and overweight children were 39.5% and 18.8%, respectively. Rates of obesity were highest among 1^st and 2^nd generation, non-English speakers, and over one quarter of 3^rd generation, non-English speakers were overweight. Screen time, physical activity, and sports and club participation were highest among 2^nd and 3^rd generation, English speakers and lowest among 1^st generation, non-English speakers. Of note, the level of occurrence of no physical activity was highest among 1^st generation, non-English speakers, followed by 2^nd and 3^rd generation, non-English speakers. Second generation, English speakers reported the highest club participation rate.

Table 2.

Weighted sample descriptives of children by generation and language groups^a

	Generation/language groups
	1st gen., non-English n = 141	2nd gen., non-English n = 365	2nd gen., English n = 113	3rd gen., non-English n = 78	3rd gen., English n = 1,315	All n = 2012
Age, year, M (SE)	9.1 (0.2)	8.7 (0.1)	9.1 (0.3)	8.2 (0.3)	8.6 (0.1)	8.7 (0.1)
Gender (%)
Men	49.0	49.2	39.7	31.0	53.2	50.0
Women	51.0	50.8	60.3	69.0	46.8	50.0
Weight statusb
BMI z-score, M (SE)	1.64 (0.14)	1.53 (0.09)	1.07 (0.20)	1.09 (0.24)	1.10 (0.06)	1.26 (0.05)
Normal weight (%)	27.1	28.4	49.3	41.5	49.7	41.7
Overweight (%)	16.2	21.4	15.6	27.5	17.5	18.8
Obese (%)	56.7	50.2	35.2	31.1	32.8	39.5
Screen timec, M (SE), hrs/day	1.40 (0.11)	1.64 (0.08)	1.67 (0.14)	1.55 (0.14)	1.90 (0.12)	1.76 (0.07)
> 2 hrs/day (%)	11.9	14.8	16.3	6.7	20.5	17.4
Physical activityd, M (SE), days/wk	3.33 (0.37)	3.93 (0.19)	4.34 (0.32)	4.11 (0.42)	4.31 (0.12)	4.12 (0.10)
No physical activitye (%)	18.1	12.6	3.1	12.5	7.8	10.0
Regular physical activityf (%)	39.4	53.2	59.9	51.5	62.2	57.1
Sports participationg (%)	23.1	34.3	44.3	40.8	59.4	47.9
Club participationh (%)	21.9	21.9	56.0	27.0	47.3	38.0
Household Incomei (%)
Poor	84.7	63.8	38.3	42.7	23.9	46.7
Near poor	9.1	16.8	16.0	24.4	8.5	11.9
Not poor	6.2	19.4	45.7	33.0	67.5	41.3
Household educationj (%)
<12 years	26.3	38.8	12.3	20.8	4.3	16.5
12 years	42.8	37.7	38.2	47.1	29.3	34.0
>12 years	30.9	23.5	49.6	32.2	66.4	49.6

^aChildren were categorized according to generational status (1^st, 2^nd, or 3^rd) and primary language spoken in the home (English or non-English).

^bNormal weight is defined as having a BMI above the 5^th percentile but below the 85^th percentile; overweight at or above the 85^th percentile; and, obese at or above the 95^th percentile.

^cScreen time includes TV, videos, and video games.

^dPhysical activity was measured in days per week and defined as exercising or participating in a physical activity ≥20 minutes that made the child breathe hard and sweat.

^eNo physical activity is defined as exercising or participating in physical activity 0 days per week.

^fRegular physical activity is defined as participating in ≥3 days of exercise or physical activity per week.

^gSports participation is defined as belonging to a sports team or taking part in sports lessons in the past 12 months.

^hClub participation is defined as participating in any clubs or organizations in the past 12 months.

ⁱIncome was grouped into 3 categories based on USDHHS Federal Poverty Guidelines for households: “poor”, below 133% poverty; “near poor”, at or above 133% poverty but below 185% poverty; and, “not poor”, at or above 185% poverty.

^jHighest level of education in the household.

Gen., generation;

For the 1st and 2^nd generation, non–English speaking children, the adjusted odds of being overweight or obese were significantly greater compared to the 3^rd generation, English speakers (Table 3). For example, 2^nd generation, non-English speakers were 2.8 and 2.2 times more likely to be overweight and obese, respectively, and 1^st generation, non-English speakers were 2.8 times more likely to be obese relative to the comparison group. In contrast, the odds of exceeding 2 hours of daily screen time was lower among the non–English speaking groups compared to 3rd generation, English speakers. The odds of having >2 hours of daily screen time were 60% less in 1st generation, non-English speakers, 47% less in 2^nd generation, non-English speakers, and 76% less in 3^rd generation, non-English speakers compared to 3^rd generation, English speakers.

Table 3.

AORs of dependent variables by generation status and language.^a

	Generation/Language Groups
	1st Generation, non-English n=141	2nd Generation, non-English n=365	2nd Generation, English n=113	3rd Generation, non-English n=78	3rd Generation, English n=1315
Weight Statusb
Overweight	1.61 (0.65, 3.95)	2.81 (1.02, 3.19)*	0.84 (0.38, 1.89)	1.69 (0.70, 4.09)	1.00
Obese	2.75 (1.35, 5.62)**	2.16 (1.32, 3.53)**	1.02 (0.51, 2.03)	1.06 (0.37, 3.02)	1.00
Screen timec
>2 hours/day	0.40 (0.18, 0.91)*	0.53 (0.31, 0.92)*	0.71 (0.32, 1.58)	0.24 (0.08, 0.70)**	1.00
Regular physical activityd	0.49 (0.25, 0.96)*	0.85 (0.54, 1.33)	0.98 (.52, 1.87)	0.79 (0.36, 1.76)	1.00
Sport Participatione	0.44 (0.23, 0.86)*	0.86 (0.54, 1.37)	0.79 (0.41, 1.54)	1.04 (0.48, 2.27)	1.00
Club Participationf	0.43 (0.20, 0.96)*	0.44 (0.27, 0.72)**	1.66 (0.83, 3.30)	0.55 (0.23, 1.31)	1.00

Note: Regression analyses were adjusted for child age and gender, household income and highest education level; The comparison group was 3^rd gen., English speakers; numbers in parentheses represent the 95% CI.

^aChildren were categorized according to generational status (i.e., 1^st, 2^nd, 3^rd) and primary language spoken in the home (i.e., English or non-English).

^bNormal weight is defined as having a BMI above the 5^th percentile but below the 85^th percentile, overweight at or above the 85^th percentile, and, obese at or above the 95^th percentile. Reference category is normal weight.

^bScreen time includes TV, videos, and video games. Reference category is ≤2 hours of screen time.

^cRegular physical activity is defined as participating in ≥3 days/week of exercise or vigorous physical activity. Reference category is <3 days/week of vigorous physical activity.

^eSports participation is defined as belonging to a sports team or taking part in sports lessons in the past 12 months. Reference category is no sports participation.

^fClub participation is defined as participating in any clubs or organizations in the past 12 months. Reference category is no club participation.

^*p<.05;

^**p<.01.

Overall, the non–English speaking groups were less likely to report regular physical activity, and participation in clubs and sports than the 3^rd generation, English speakers. Specifically, 1^st generation, non-English speakers were 2.0 times less likely to report regular physical activity and ~2.3 times less likely to participate in sports or clubs relative to 3^rd generation, English speakers. In addition, 2^nd generation, non-English speakers were 2.3 times less likely to participate in clubs compared to the comparison group. Unadjusted post hoc multiple comparisons revealed that all non–English speaking groups were less likely to participate in clubs than 2^nd generation, English speakers (p<.05). After adjusting for multiple tests, only findings for the 1^st and 2^nd generation, non-English speakers remained significant. No significant unadjusted or adjusted post hoc comparisons were found for overweight, obesity, weight status, screen time, regular physical activity, and sports participation.

Discussion

These findings suggest that 1st and 2^nd generation, non–English speaking children are more likely to be obese, and 1^st generation, non-English speakers have significantly lower physical activity than 3^rd generation, English speaking children. This finding is distinct from is observed in the Hispanic adult and adolescent population where obesity and overweight tend to increase steadily across generations. These results also indicate that children who speak a language other than English in the home are less likely to engage in 2 or more hours of daily screen time. Finally, 1^st and 2^nd generation, non–English speaking children were less likely to participate in clubs, and 1^st generation, non-English speakers were less likely to participate in sports after school or on the weekends. To the authors’ knowledge, this is the first study to address the association between generational status and language with relative weight, physical and extracurricular activities, and screen time in a U.S.-based, nationally representative sample of Hispanic children.

Ethnic minority and other low-income populations are disproportionately affected by overweight and obesity.⁴² Based on a 2006 report, almost one half of Latino children aged 6–11 years were overweight, exceeding rates of overweight in non-Hispanic white boys/men and girls/women, and black men/boys.⁴³ In this sample, children were more likely to be obese if they were 1^st and 2^nd generation, non-English speakers; this finding corroborates other studies with Hispanic children.^44,45 Generational status and language use are considered to be indirect indicators of the complex acculturation phenomenon, and are highly correlated to each other;⁴⁶ furthermore, they have been used as proxy measures of acculturation in previous studies with Hispanics.^{1,14,25,27,29,30,47–49} According to the acculturation hypothesis, one would assume that children’s relative weight would be lowest in the 1^st generation and increase with subsequent generations; however, this was not demonstrated by the findings. This could imply that acculturation has a different association with relative weight in Hispanic children than what previous literature has shown. Although parent report of height and weight to estimate child BMI is not ideal, it was the only estimate of relative weight provided by this secondary data set. This limitation may explain why it appears as though relative weight was lower in 2^nd and 3^rd generation children. It is possible that parent’s acculturation level, child age, or an interaction between the two introduced some self-report biases into the data.

The odds of having 2 or more hours of screen time was significantly lower for all non–English speaking groups regardless of generational status, compared the 3^rd generation, English speakers. This finding supports previous studies which examined the relationship between screen time and acculturation.^30,50,51 These results would suggest that speaking a language other than English in the home appears to be a protective factor for watching less TV, videos and playing fewer video games on a typical school day. Future studies should investigate after school time use in Hispanic children to discern the activities these non–English speaking children are engaging in, apart from screen time, and how they may affect health.

Descriptively, weekly physical activity level was the lowest among the non–English speaking generation groups. Further, 1^st generation, non-English speakers were significantly less physically active than 3^rd generation, English speakers. These findings are consistent with two recent studies of Hispanic children and adolescents,^30,33 and add to the literature by providing a picture of generational shifts in physical activity levels for Hispanic immigrant children by language. Despite evidence for the acculturation hypothesis which suggests a decline in health behaviors over time with greater acculturation, this study implies that this may not necessarily apply to Hispanic children’s physical activity behaviors. Due to the preliminary nature of the present study, analyses were not performed to explore gender differences in physical activity by language and generation. Future studies should include analyses which look at the gender differences in physical activity for Hispanic children by acculturation level. Also, this study is limited by the fact that the physical activity measure was based on parental report in response to one question. However, at this age, parental reports are appropriate given the inability of young children to correctly recall the duration or frequency of physical activities.^52,53 Future studies should use more reliable and valid measures of child physical activity.

Lastly, the authors examined whether participation in sports and clubs differed by generation and language groups. The results showed that the odds of participating in sports were significantly lower for 1^st generation, non-English speakers; this finding confirms previous research with immigrant children.^30,54 Additionally, 1^st and 2^nd generation, non–English speaking children were 2.3 times less likely to participate in clubs after school or on the weekends when compared to 3rd generation, English speakers. The tendency for 1^st generation children to be less physically active and less likely to participate in sports and clubs could partly reflect cultural preferences; for example, immigrant families may not place high value on extracurricular activities, and may encourage the child to spend more time in academic studies, language classes, or with the family.^30,55

It is also important to recognize the role of social context, specifically those barriers which may influence non–English speaking children’s physical activity levels. Language barriers, exclusion/discrimination from social organization, and socioeconomic disadvantage could all contribute to the observed physical activity disparity.⁵⁶ Lack of English-speaking abilities, coupled with a lack of knowledge of physical activity opportunities and outlets in the community may prevent less acculturated children from participating in physical activities, particularly if parents are primarily responsible for fostering children’s participation in these activities. While the authors attempted to control for SES by adjusting for household income and education, the role of context, culture and other socioeconomic factors in influencing physical activity access and behavior cannot be denied. More research is needed to explore the complex relationships among culture, SES, acculturation and physical activity participation among Hispanic children, as well as sports and club participation in immigrant children. It is possible that the findings of this study vary by Hispanic subgroup, but due to the limitations of using a secondary data set, the authors were unable to explore this possibility. It is evident that Hispanic subgroups differ substantially in historic, sociocultural, and economic backgrounds, and thus in their health and disease experience.⁹ Future population-based studies should gather additional demographic data, such as country of origin, as well as other family and neighborhood data, to give a more complete picture of the heterogeneity that exists within these populations.^57,58

While few researchers would discount the importance of considering acculturation when examining health behaviors and outcomes in Hispanic immigrants, many would agree that the way it is currently measured can be problematic. Proxy measures, while widely used, are limited in scope and sensitivity because they do not directly measure acculturative change, that is, changes in attitudes, values or behaviors.⁵⁹ Thomson and Hoffman-Goetz (2009) in their review of measures of acculturation for Hispanic populations found that only two acculturation scales were theory-based.^59–61 Additional theoretic models are needed to paint a more complete picture of the associations among specific components of acculturation and health outcomes.¹³ The measurement of acculturation, however, is complex. Originally, acculturation was thought to be a linear process whereby immigrants moved from exclusive involvement in the native culture, to exclusive involvement in the new host culture, de-emphasizing their own culture.^12,61–64 Proxy measures are often used in this unidimensional model.¹³ More recently, a bilinear model of acculturation has been proposed. This model assumes a monocultural–bicultural continuum where individuals can endorse the values and beliefs of both their native and host culture simultaneously.^64–68 One such scale measures involvement in the native and host culture separately, allowing for four types of adaptation: traditional, integrated, assimilated, and marginal.⁶⁹ Instruments created to reflect this multidimensional and multidirectional view of acculturation measure various dimensions including changes in behaviors, attitudes, norms and values as well as language preferences.^13,70,71 However, these multidimensional scales are not without criticism, including the inherent assumption that there is a well-defined mainstream culture and ethnic culture, and that cultural contact is occurring between two historically distinct groups.⁷² Future population-based studies of Hispanics would generate more-useful results if they include theory-based, multidimensional acculturation scales, or measures of change in attitudes, behaviors, and beliefs over time with adaptation to a new culture.

According to 2000 Census data, one of every five children in the U.S is from an immigrant family where one or both parents is foreign-born.⁷³ This reality advocates the need for further study on the health of immigrant families and their children. Future interventions which target Hispanic children and their parents should take into account the child’s generation status and language use at home, as these factors appear to play a role in activity levels and relative weight.