Skip to main content
NCBI home page
PMC Canada Author Manuscripts logoLink to PMC Canada Author Manuscripts
. Author manuscript; available in PMC: 2018 Jan 5.
Published in final edited form as: J Pediatr Psychol. 2012 Feb 24;37(8):843–853. doi: 10.1093/jpepsy/jss037

Generation of Immigration and Body Mass Index in Canadian Youth

Elizabeth C Quon 1, Jennifer J McGrath 1, Marie-Hélène Roy-Gagnon 2
PMCID: PMC5756084  CAMSID: CAMS6752  PMID: 22366576

Abstract

Objective

To investigate the role of acculturation, as measured by generational status, on body mass index (BMI) in a sample of Canadian youth.

Methods

Population-based data from the National Longitudinal Survey of Children and Youth were used. Participants were divided into 2 age ranges: children aged 6–11 years (n = 14,287) and adolescents aged 12–17 years (n = 12,155). Youth were classified into one of five generations of immigration: first-generation, second-generation, mixed-generation, third-generation, and Aboriginal. Parent-and self-report height and weight were used to calculate BMI Z-scores.

Results

Generation of immigration was significantly related to BMI Z-score in both childhood and adolescence. First-generation immigrants had more weight gain compared to other groups during adolescence, but not during childhood.

Conclusions

Acculturation, as measured by generation of immigration, is an important predictor of BMI in Canadian children and adolescents.

Keywords: adolescents, children, culture, obesity, public health, race/ethnicity

Introduction

In line with worldwide trends, Canadian youth are becoming progressively more overweight and obese (Tremblay & Willms, 2000). Recent representative national surveys show the prevalence of overweight to be 18% and the prevalence of obesity to be 8% among children and adolescents aged 2–17 years (Shields, 2006). Canadian children, Aboriginal Canadians, and adult immigrants to Canada have been highlighted as groups at risk of excess weight (Bélanger-Ducharme & Tremblay, 2005). However, there is little known about the rates of excess weight in Canadian immigrant children and adolescents.

Immigration is an important part of Canada’s history and continued development. Since 1990, Canada (population 34 million) has accepted approximately 230,000 immigrants per year, of which about 20% are under the age of 14 years. Immigrants to Canada are classified based on their reason for immigrating: economic class, family class, refugees, and other. In the past 10 years, about 50% of immigrants to Canada have originated from Asia/Pacific, 20% from Africa/Middle East, 17% from Europe, 9% from South/Central America, and 3% from the United States (Citizenship and Immigration Canada, 2010).

The “healthy immigrant effect” refers to the observation that the health of immigrants at the time of immigration is superior to the health of the native-born population, but worsens with time spent in the new country. In Canada, recent immigrants enjoy health advantages over long-term immigrants and the Canadian-born population in terms of overall health status, prevalence of chronic conditions, disability, and life expectancy (Chen, Ng, & Wilkins, 1996; Hyman, 2001; Perez, 2002). The healthy immigrant effect has also been documented in the United States (Singh & Siahpush, 2002), the UK (Kennedy, McDonald, & Biddle, 2006), and Australia (Biddle, Kennedy, & McDonald, 2007). The underlying reasons driving the decline in immigrant health are unclear thus far. Differences may result from a cohort effect, such that immigrants who arrived more recently had a better health profile than immigrants who arrived earlier, due to evolving immigration criteria that selects immigrants on the basis of health screening, education, and job skills (Hyman, 2001). Alternatively, the health status of immigrants may actually deteriorate with time in the host country due to acculturation. Acculturation, or culture change from being in contact with the dominant culture (Redfield, Linton, & Herskovits, 1936), may help to explain why the health of immigrants deteriorates with time since immigration. Acculturation is associated with changes in multiple domains, including the physical environment, one’s cultural identity, social relationships, and psychological and behavioral changes (Berry, Kim, Minde, & Mok, 1987). As immigrants adopt Canadian health-related behaviors, they may become more similar to native-born Canadians and lose their health advantage. Acculturation may be particularly important for the development of obesity, due to North American “obesogenic” lifestyle habits (Swinburn, Egger, & Raza, 1999), such as sedentary lifestyles, large portion sizes, and consumption of high-fat, energy-dense foods. The changes associated with acculturation may also be stressful (Berry & Annis, 1974), which is another risk factor for obesity. Acculturation has been measured in adolescents using a number of proxy variables, including generational status, linguistic variables, and proportion of immigrants in the neighborhood (e.g., Gordon-Larsen, Mullan Harris, Ward, & Popkin, 2003). Generational status, or “generation of immigration,” is based on the youth’s and their parents’ country of birth. Higher generation of immigration reflects a greater number of generations the family has been in the host country and suggests greater acculturation to the dominant culture in the host country (Hsin, La Greca, Valenzuela, Moine, & Delamater, 2010).

Evidence for a healthy immigrant effect for excess weight in youth has been examined in the United States using generational status with mixed results. Popkin and Udry (1998) found that the prevalence of overweight increased from first-to second-generation Hispanic and Asian immigrant adolescents to the United States, but that there was no difference between second- and third-generation immigrants. Similarly, Gordon-Larsen et al., (2003) found the prevalence of overweight was lower among some groups of Hispanic adolescent first-generation compared to second-generation immigrants. In contrast, Haas, Lee, Kaplan, Sonneborn, Phillips, and Liang (2003) found no significant differences between obesity rates in first- and second-generation immigrant children or adolescents (results not stratified by race). A more recent study by Singh, Kogan, and Yu (2009) found that first-generation adolescent immigrants had lower odds of obesity than second-generation, who had lower odds of overweight than third-generation. When stratified by race, only Black and White youth showed this pattern of results, Hispanic and Asian youth did not.

In Canada, there is a lack of research examining the relationship between generation of immigration and excess weight in children and adolescents. However, several cross-sectional studies have demonstrated a healthy immigrant effect for excess weight in Canadian adults. Specifically, recent adult immigrants to Canada have a lower prevalence of overweight than Canadian-born adults, but the prevalence of overweight increases with additional years in Canada (Cairney & Ostbye, 1998; McDonald & Kennedy, 2005; Tremblay, Perez, Ardem, Bryan, & Katzmarzyk, 2005). These results are based on all immigrant groups, with variation in these patterns between races. Analysis of longitudinal data is required to support the idea that immigrant weight gain is a result of additional time spent in Canada.

While most Canadians can trace their ancestry to immigrants, Aboriginal Canadians are those peoples whose ancestors were native to Canada. The Aboriginal population of Canada includes approximately 1 million people, or about 3% of the population (Statistics Canada, 2005). Compared to the general population, Aboriginal Canadians show a larger annual population growth rate, a substantially lower life expectancy, and a higher prevalence of obesity and related morbidities (Katzmarzyk, 2008; Statistics Canada, 2005). Similar health disparities between indigenous and nonindigenous populations have been documented in the United States, Australia, and New Zealand (Ring & Brown, 2003). Like immigrants to Canada, Aboriginal Canadians have distinct cultural characteristics from the general population. Therefore, Aboriginal youth may experience acculturation, although the nature of culture change is expected to be different from that of immigrant youth due to its involuntary nature (Bartlett, 2003). In addition, findings of elevated rates of obesity among Aboriginal youth (First Nations Centre, 2005; Katzmarzyk, 2008) indicate that this group should be considered separately from the highest generation of immigration.

There are other factors that may contribute to overweight in immigrant and Aboriginal children and adolescents, and these may also be related to acculturation and generational status. The prevalence of overweight varies across race in Canadian youth (Shields, 2006). Moreover, the increase in prevalence of overweight from first- to second-generation was larger for Asian immigrants than for Hispanic immigrants in American adolescents (Popkin & Udry, 1998), suggesting a potential interaction between generation of immigration and race in immigrant youth. Socioeconomic status is an important predictor of overweight and obesity in Canadian youth (e.g., Oliver & Hayes, 2005). In addition, poverty levels vary by generation of immigration, with higher rates of poverty in first-generation immigrant families (36%) than in second-(14%) or third-generation immigrant families (13%) in Canada (Beiser, Hou, Hyman, & Tousignant, 1998). Poverty rates for Aboriginal children are significantly higher than other visible minorities (Blackstock, Trocmé, & Bennett, 2004).

The extant research literature suggests that generation of immigration is related to overweight in certain groups of immigrant youth in the United States. However, to date there is a lack of research on the relationship between generational status and excess weight in Canadian children and adolescents. Establishing the association between generational status, as a proxy for acculturation, and excess weight is the first step toward developing policy and interventions to prevent and reduce potential weight gain in Canadian immigrant youth. In addition, establishing this association may aid in understanding broader constructs, such as the healthy immigrant effect and the development of obesity. Research in this area is primarily cross-sectional; therefore, there is a need to examine this association longitudinally.

The primary goal of the current study was to assess the cross-sectional effect of acculturation, as measured by generation of immigration, on body mass index (BMI) in Canadian children and adolescents. The secondary goal was to examine the effect of generation of immigration on change in BMI over time. Using data from the National Longitudinal Survey of Children and Youth, we tested the hypothesis that BMI would be lower overall in more recent generations of immigrants compared to higher generations of immigrants, with the highest BMI in the Aboriginal group. We also tested the hypothesis that change in BMI over time would differ across generation of immigration, with greater increases observed in more recent generations of immigrants.

Methods

Sample

Participants are from the National Longitudinal Survey of Children and Youth (NLSCY), a population-based longitudinal survey of Canadian children and adolescents conducted by Statistics Canada and Human Resources Development Canada. The NLSCY sample is representative of children aged 0–11 years that were living in any Canadian province in 1994/1995, when survey weights are applied. A full description of the NLSCY and sampling design is available elsewhere (Statistics Canada & Human Resources Development Canada, 1995).

The current study was conducted on data from children and adolescents in the original longitudinal cohort of the NLSCY, who were between the ages of 0 and 11 years when they were first identified in Cycle 1 (1994/1995). Data collection occurred every two years, for a total of seven observations per youth over a 12-year time span. Only those observations that occurred when the youth was between the ages of 6 and 17 years were included in the current study. Observations were divided into two age ranges: children (ages 6–11 years) and adolescents (ages 12–17 years). These age ranges were used because height and weight were reported by parents for children until 11 years and by adolescents beginning at 12 years. Children under the age of 6 years were omitted due to the known instability of parent-reported BMI at these ages in the NLSCY (cf. Oliver & Hayes, 2005). Biases in parent-reported BMI for early childhood are largely attributable to underreporting child height at young ages (Akinbami & Ogden, 2009). Youth older than 18 years of age were not included due to changes in the measurement definitions of BMI at this age. Similar age ranges have been previously employed in studies of overweight in Canadian youth (e.g., Oliver & Hayes, 2005; Shields, 2006). Thus, depending on age at initial recruitment, there were between one and three observations per youth for each age range.

Data collection for the NLSCY occurred via a computer-assisted telephone interview with the “person most knowledgeable” about the youth, and his or her spouse, and via a self-completed paper questionnaire for adolescents (aged 12–17 years). The “person most knowledgeable” was the youth’s biological mother (90%) or biological father (8%). The spouse was the youth’s biological father (69%), step-father (7%), or biological mother (5%). There was no spouse for 16% of the observations. The “person most knowledgeable” will subsequently be referred to as “parent.”

The demographics of the current study sample are provided in Table I. Due to over-sampling of rural areas, the NLSCY under-sampled populations of immigrant families and non-White youth, who often settle in large urban areas (Beiser et al., 1998; Statistics Canada, 2001). All analyses are based on this unweighted sample of youth from the NLSCY.

Table I.

Demographics of the Sample by Age Range

Predictors
6–11 years
n (%)		 Predictors
12–17 years
n(%)
Gender
 Male 7,214 (50.5) 6,122 (50.4)
 Female 7,073 (49.5) 6,033 (49.6)
Generation
 First 209 (1.5) 175 (1.4)
 Second 601 (4.2) 477 (3.9)
 Mixed 975 (6.8) 825 (6.8)
 Third 11,730 (82.1) 10,025 (82.5
 Aboriginal 772 (5.4) 653 (5.4)
Race
 White 13,322 (93.2) 11,365 (93.5)
 East Asian 136 (1.0) 109 (0.9)
 South Asian 99 (0.7) 89 (0.7)
 South East Asian 109 (0.8) 91 (0.7)
 Black 108 (0.8) 87 (0.7)
 Aboriginal 325 (2.3) 267 (2.2)
 Other 188 (1.3) 147 (1.2)
M (SD) M (SD)
Child age (years) 8.67 (1.70) 14.07 (1.67)
Household income (CAD) 56,856 (35,996) 72,991 (47,969)
Parental education (years) 12.67 (2.05) 13.20 (2.20)
Open in a new tab

Measures

Acculturation

Youth were classified into one of five groups based on generation of immigration or Aboriginal heritage. Generation classifications were made based on the parent’s country of birth, the spouse’s country of birth, and the child’s country of birth, as reported by parents. Higher generation of immigration reflects a greater number of generations removed from immigration and a potential for greater acculturation. “First-generation” was defined as youth born in a country other than Canada; “second-generation” was defined as youth born in Canada, with both parent and spouse born in a country other than Canada; “mixed-generation” was defined as youth born in Canada, with one parent born in Canada and one parent born in another country; third-generation or higher (hereafter referred to as “third-generation”) was defined as youth born in Canada, with both parent and spouse born in Canada (or parent born in Canada and no spouse listed). Previous studies have similarly used country of birth to define generation of immigration in youth (Gordon-Larsen et al., 2003; Haas et al., 2003; Popkin & Udry, 1998). The nonimmigrant Aboriginal group is a distinct category and is not assumed to be an ordinal ranking relative to the other generations. “Aboriginal” was defined as youth born in Canada and identified by their parents as “an Aboriginal person who is North American Indian, Inuit, or Métis.”

Ethnicity/Race

Race was determined by the parent’s response to the question, “How would you best describe this child’s race or color?” The 12 forced choice response categories in the NLSCY were aggregated according to geographic and cultural proximity to avoid small group sizes. The ethnoracial groups were: White, East Asian, South East Asian, South Asian, Black, Aboriginal, and Other.

Socioeconomic Status

Socioeconomic status was measured using two variables at every observation: household income during the previous twelve months and number of years of education the parent had completed. Household income was derived from open-ended questions answered by the parent and his or her spouse about various sources of income (e.g., wages and salary, self-employment, social assistance). Parental years of education was derived from two forced-choice questions: years of elementary and high school, and highest level of education attained beyond high school. For our analyses, we used mean household income and mean years of parental education aggregated across observations for each participant.

BMI

BMI was calculated as weight (kg)/height (m)2. For children aged 6–11 years, the parent reported the child’s height and weight. Adolescents aged 12–17 years self-reported their height and weight. Height and weight were reported at each observation. Parent- and self-reports are subject to bias, with parents likely to over-report BMI and adolescents likely to under-report BMI (e.g., Shields, 2006). BMI Z-scores and percentiles were derived using age- and sex-specific values from the Centers for Disease Control and Prevention (Kuczmarski et al., 2000). Overweight status was defined as 85th ≤ BMI < 95th percentile; Obese status was defined as BMI ≥ 95th percentile.

Analytical Strategy

Missing Data

Youth with missing information on generation of immigration or race were excluded using listwise deletion (4,203 children and 463 adolescents). Multiple imputation (n = 5) was performed using AmeliaView II to impute missing outcome data for nonresponse (i.e., youth who missed one or more observations of data collection) and partial nonresponse (i.e., youth with missing/incomplete information on outcome data). BMI Z-scores from other observations were included in the imputation model, along with height, weight, age, gender, race, generation of immigration, household income, and parental education. After multiple imputation, the final sample yielded 33,853 observations from 14,287 children (M = 2.4 observations per child) and 27,059 observations from 12,155 adolescents (M = 2.3 observations per adolescent). Results were largely identical when analyses were run on the original versus imputed data set; therefore, only results based on the imputed dataset are presented.

Hypothesis Testing

Hierarchical linear modeling techniques (Bryk & Raudenbush, 1987) were used to fit regression models to the longitudinal data. This type of the modeling is recommended for the analysis of longitudinal pediatric psychology data (DeLucia & Pitts, 2006). To assess the influence of generation of immigration on youth’s BMI, a two-level model was specified in which observations (level 1) were nested within individuals (level 2). The level 1 model describes within-individual BMI change over time and the level 2 model describes the effect of between-individual predictor variables (generation of immigration, ethnicity/race, socioeconomic status, gender). Hierarchical linear models were specified using HLM 6.2 software. Separate models were constructed for child and for adolescent age ranges.

Given the statistical advantages to using age- and sex-specific BMI Z-scores (Must & Anderson, 2006), we chose to test hypotheses using BMI Z-scores as the dependent variable. For ease of interpretation, descriptive statistics are presented in BMI percentiles and weight status categories. For each age range, six a priori defined models were tested. The cross-sectional hypothesis that BMI would be lower in more recent generations of immigrants compared to higher generations, and highest in the Aboriginal group, was tested with three intercept-only models, using only level 2 (between-individual) predictors. First, an unconditional model with no predictors determined baseline intercept estimates and provided a comparison for subsequent models. Next, a generation model was specified with generation of immigration entered as a categorical level 2 predictor (third-generation is reference group). Finally, age at initial data collection, gender (male is reference group), household income (grand mean centered), and parental education (grand mean centered) were entered as level 2 predictors in the full model. By controlling for these covariates, we were able to determine whether generation of immigration independently predicted BMI Z-score.

The longitudinal hypothesis that change in BMI over time would differ across generation of immigration was tested with three slope or “change” models, using level 1 (within-individual) and level 2 (between individual) predictors. First, an unconditional change model included age (with 0 indicating 6 years old for child analyses and 12 years old for adolescent analyses) as a level 1 predictor to determine baseline change over time estimates. Next, a generation change model was specified with generation of immigration (third-generation is reference group) as a level 2 predictor of BMI Z-score change. Finally, in the full change model, we also entered age at initial data collection, gender (male is reference group), and parental education (grand mean centered) as level 2 predictors of BMI Z-score change. Household income was not included in this model due to multicollinearity with parental education.

Results

We examined rates of overweight and obesity, as well as mean BMI percentiles by gender and generation of immigration. Results are based on means and percentages of all observations in each age range. Descriptive results are provided in Tables II and III.

Table II.

BMI Percentile and Weight Status by Gender and Generation of Immigration: Child

Generation Girls
Boys
BMI percentile, M (SD) Overweight (%) Obese (%) BMI percentile M (SD) %Overweight (%) Obese (%)
First 56.05 (33.66) 14.4 12.6 66.31 (32.23) 17.1 18.9
Second 56.58 (32.97) 11.1 15.4 62.23 (32.32) 16.1 17.8
Mixed 58.74 (31.51) 14.0 13.1 59.86 (32.03) 13.1 16.4
Third 60.20 (31.76) 15.2 14.8 63.71 (31.89) 16.8 19.1
Aboriginal 66.34 (30.53) 18.1 19.9 68.10 (31.23) 18.4 23.1
Open in a new tab

Note. BMI percentiles were derived using CDC age- and sex-specific values (Kuczmarski et al., 2000). Overweight: 85th ≤ BMI < 95th percentile; Obese: BMI ≥ 95th percentile.

Table III.

BMI Percentile and Weight Status by Gender and Generation of Immigration: Adolescent

Generation Girls
Boys
BMI percentile, M (SD) Overweight (%) Obese (%) BMI percentile, M (SD) Overweight (%) Obese (%)
First 55.46 (28.21) 18.6a 58.93 (30.12) 11.1 12.0
Second 53.55 (27.97) 7.6 6.0 59.28 (28.66) 11.6 10.4
Mixed 54.62 (27.56) 10.4 6.0 58.26 (27.97) 12.4 7.7
Third 56.95 (27.27) 10.8 6.7 61.72 (28.49) 13.8 11.2
Aboriginal 61.91 (27.19) 13.6 8.5 63.87 (28.68) 13.2 13.7
Open in a new tab

Note. BMI percentiles were derived using CDC age- and sex-specific values (Kuczmarski et al., 2000). Overweight: 85th ≤ BMI < 95th percentile; Obese: BMI ≥ 95th percentile.

a

Rates of obesity were suppressed due to small cell counts within the NLSCY; therefore, a combined percentage for overweight and obesity (BMI ≥ 85th percentile) is reported. NLSCY restricts release of small cell count data to protect anonymity.

Hypothesis testing results using hierarchical linear modeling and BMI Z-score as the dependent variable are provided separately for children (Table IV) and adolescents (Table V). Note that all analyses were also run with BMI percentile as the dependent variable and results were largely identical. For children, the unconditional model demonstrated that 33.0% of the total variance was between-individual and 67.0% of the total variance was within-individual. The generation model tested the cross-sectional hypothesis that one’s generation of immigration would predict BMI, irrespective of time. Results showed a significant effect of generation of immigration on BMI Z-score, accounting for 0.5% of between-individual variance. Specifically, mixed-generation had significantly lower mean BMI Z-scores than third-generation, while Aboriginal children had significantly higher mean BMI Z-scores. First- and second-generation were not significantly different from third-generation. In the full model, when gender, cohort, race, income, and education were included as covariates, the effect of generation of immigration was no longer significant. Results showed that being a girl, having higher household income, and higher parental education were significantly associated with lower BMI Z-scores. The effect of race was also significant, as East Asian was related to lower BMI Z-scores, and Aboriginal was related to higher BMI Z-scores, compared to White. These additional covariates explained an additional 2.9% of between-individual variance.

Table IV.

Hierarchical Linear Modeling Results for BMI Z-scores: Child

Predictors Intercept-onlya
Changeb
Generation
B (95% CI)		 Full
B (95% CI)		 Generation
B (95% CI)		 Full
B (95% CI)
Intercept 0.386* (0.366 to 0.406) 0.444* (0.412 to 0.477) −0.03* (−0.046 to−0.028) −0.036* (−0.051 to −0.021)
Generation (ref. = Third)
 First −0.018 (−0.172 to 0.136) −0.032 (−0.194 to 0.130) −0.044 (−0.116 to 0.028) −0.030 (−0.105 to 0.045)
 Second −0.075 (−0.166 to 0.016) −0.073 (−0.181 to 0.035) −0.017 (−0.057 to 0.023) −0.010 (−0.059 to 0.039)
 Mixed −0.087* (−0.158 to −0.016) −0.054 (−0.126 to 0.018) 0.007 (−0.024 to 0.038) 0.009 (−0.022 to 0.040)
 Aboriginal 0.209* (0.129 to 0.289) 0.073 (−0.019 to 0.165) −0.013 (−0.049 to 0.024) 0.005 (−0.036 to 0.046)
Gender (ref. = male)
 Female −0.134* (−0.170 to −0.098) −0.032* (−0.047 to −0.016)
 Age at recruitment 0.007 (−0.008 to 0.022) 0.010* (0.0031 to 0.017)
 Income (10K, CAD) −0.001* (−0.001 to −0.0008) –
 Education −0.033*(−0.043 to −0.023) 0.001 (−0.003 to 0.005)
Race (ref. = White)
 East Asian −0.228* (−0.424 to −0.032) −0.042 (−0.131 to 0.047)
 Southeast Asian −0.203 (−0.421 to 0.015) 0.055 (−0.040 to 0.150)
 South Asian 0.182 (−0.041 to 0.405) −0.050 (−0.146 to 0.046)
 Black 0.140 (−0.071 to 0.351) −0.075 (−0.168 to 0.018)
 Aboriginal 0.294* (0.155 to 0.433) −0.051 (−0.113 to 0.011)
 Other 0.170 (−0.001 to 0.341) −0.011 (−0.088 to 0.067)
Open in a new tab

Note.

a

The unconditional main effects model yielded an intercept estimate of 0.388 (0.209 to 0.567).

b

The age-only change model yielded an intercept estimate of −0.038 (−0.046 to −0.030).

*

p < .05

Table V.

Hierarchical Linear Modeling Results for BMI Z-scores: Adolescent

Predictors Intercept-onlya
Changeb
Generation
B (95% CI)		 Full
B (95% CI)		 Generation
B (95% CI)		 Full
B (95% CI)
Intercept 0.296*(0.277 to 0.315) 0.371 (0.336 to 0.406) 0.008* (0.001 to 0.015) 0.001 (−0.012 to 0.014)
Generation (ref. = Third)
 First −0.091 (−0.160 to 0.122) −0.098 (−0.247 to 0.051) 0.058* (0.008 to 0.108) 0.074* (0.020 to 0.128)
 Second −0.093* (−0.182 to −0.004) −0.075 (−0.179 to 0.029) −0.007 (−0.040 to 0.026) 0.008 (−0.030 to 0.046)
 Mixed −0.075* (−0.143 to −0.007) −0.044 (−0.113 to 0.025) −0.006 (−0.031 to 0.019) −0.004 (−0.029 to 0.021)
 Aboriginal 0.147* (0.071 to 0.223) 0.092* (0.005 to 0.179) −0.0004 (−0.028 to 0.028) 0.009 (−0.024 to 0.042)
Gender (ref. = male)
 Female −0.017(−0.051 to 0.017) −0.015* (−0.028 to −0.002)
 Age at recruitment 0.003 (−0.007 to 0.013) 0.008* (0.003 to 0.013)
 Income (10K, CAD) −0.001* (−0.001 to −0.0009) –
 Education −0.023* (−0.032 to −0.014) −0.0006 (−0.004 to 0.002)
Race (ref. = White)
 East Asian −0.393* (−0.585 to −0.201) −0.083* (−0.158 to −0.008)
 Southeast Asian −0.014 (−0.228 to 0.200 −0.003 (−0.085 to 0.079)
 South Asian 0.154 (−0.060 to 0.368) −0.009 (−0.089 to 0.071)
 Black 0.068 (−0.142 to 0.278) −0.017 (−0.102 to 0.068)
 Aboriginal 0.071 (−0.063 to 0.205) −0.026 (−0.076 to 0.024)
 Other 0.111 (−0.056 to 0.278) −0.041 (−0.105 to 0.023)
Open in a new tab

Note.

a

The unconditional main effects model yielded an intercept estimate of 0.294 (0.277 to 0.311).

b

The age-only change model yielded an intercept estimate of 0.008 (0.002 to 0.014).

*

p < .05.

For children, the unconditional change model indicated that age was related to an overall decline in BMI Z-score, accounting for 6.3% of within-person variance. The generation change model tested the hypothesis that one’s generation of immigration would predict BMI change over time. Results showed that generation of immigration was not significantly associated with change in BMI Z-score. In the full change model, gender and age at recruitment were significantly associated with BMI Z-score change, with boys and children who were younger at recruitment demonstrating an increase in BMI Z-score over time.

For adolescents, the unconditional model demonstrated that 51.5% of the total variance was between-individual and 48.5% of the total variance was within-individual. The generation model showed a significant effect of generation of immigration on BMI Z-score, accounting for 0.3% of between-individual variance. The pattern of results was similar to childhood, with second-and mixed-generation associated with lower BMI Z-scores, Aboriginal associated with higher BMI Z-scores, and first-generation showing no difference, compared to third-generation. When additional covariates were added in the full model, the main effect of generation of immigration was attenuated for all generations; only Aboriginal remained significantly different from third-generation. In adolescents, gender was not associated with BMI Z-score, while higher household income and higher parental education were associated with lower BMI Z-scores. Race showed a significant effect on BMI Z-score, specifically, East Asian was associated with lower BMI Z-scores compared to all other ethnoracial groups.

For adolescents, the unconditional change model indicated that age was related to an overall increase in BMI Z-score, accounting for 12.9% of within-person variance. The generation change model showed that first-generation was associated with an increase in BMI Z-score over time, compared to all other generations. This effect remained significant in the full change model. Similar to the findings in children, younger age at initial recruitment and male gender were related to an increase in BMI Z-score over time. East Asian race was associated with a decrease in BMI Z-score over time, compared to other races.

Discussion

This study assessed the cross-sectional and longitudinal effects of acculturation, as measured by generation of immigration, on BMI in Canadian youth. We used a population-based sample of children and adolescents to establish an association between generational status and excess weight in this population.

Our first hypothesis was that BMI would be elevated in higher generations of immigrants compared to more recent generations of immigrants, and would be the most elevated in the Aboriginal group. We demonstrated that generation of immigration was related to BMI in both child and adolescent age ranges. As expected, second-, and mixed-generations had lower overall BMI than third-generation (second-generation was statistically significant in adolescents only); however, no differences were observed between first-generation and higher generations. Aboriginal group was associated with higher overall BMI than all other groups, as expected. Therefore, our results partially support our hypothesis in that children and adolescents in more recent generations of immigrants would be less overweight. The finding that first-generation youth were more overweight than higher generations was contrary to our hypothesis, which was based on previous research and the healthy immigrant effect.

The pattern of results based on generational status was different from Singh et al. (2009), who demonstrated lower odds of overweight and obesity in first-generation compared to second- and third-generation adolescents, while controlling for race and other covariates. Similarly, Popkin and Udry (1998) and Gordon-Larsen et al. (2003) found lower prevalence of overweight in first- compared to second-generation adolescents in Hispanic and Asian American adolescents. In fact, our results may be more in line with Haas et al. (2003), who found no difference between first-generation and second-generation or higher children and adolescents in a multicultural sample. Therefore, it is possible that lower BMI in first-generation immigrant children and adolescents may be more evident when samples are stratified by race. Although, small non-White sample sizes prevented us from running our analyses this way, visual inspection of means suggest patterns consistent with higher BMI in Black, South Asian and Other first-generation immigrants and lower BMI in East Asian, Southeast Asian, and White first-generation immigrants. Therefore, the expected patterns of results may be more likely to emerge in East Asian, Southeast Asian, and White youth. Another possible explanation for our findings was that first-generation immigrant youth originated from countries with similar or higher prevalence of childhood obesity to Canada (such as United States, England), which would suggest that these youth would not be greatly affected by acculturation to an obesogenic environment. Again, visual inspection of means suggest patterns consistent with the idea that youth from the United States or England (about 18% of the first-generation group) did not have higher BMI than youth originating from other countries, such as Germany, China, India, and Philippines. Finally, an overall small sample of first-generation immigrants may have reduced our power to detect differences in this group. Despite these possibilities, it is important to consider that these results are reflective of actual patterns of overweight in Canadian immigrant youth. Specifically, first-generation immigrant children and adolescents may be at increased risk for excess weight compared to children of immigrant parents. This could be due to stress associated with immigration (Young, Spitzer, & Pang, 1999) and acculturation (Berry & Annis, 1974).

Our second hypothesis was that change in BMI over time would differ across generation of immigration. This was the first study to examine longitudinal patterns of overweight across generational status in children and adolescents. We found that generation of immigration was not associated with BMI change within childhood, but that first-generation was associated with significantly more weight gain over time compared to all other generations within adolescence. Given that first-generation immigrants did not have lower BMI cross-sectionally in childhood or adolescence, the observed weight gain relative to other generations is especially concerning. It is possible that a greater disparity between original and dominant cultures in first-generation immigrants is related to more acculturative stress, especially during the transitional time of adolescence.

Current findings for both child and adolescent age groups are congruent with previous study findings that Aboriginal youth have a higher prevalence of obesity than non-Aboriginal youth (Katzmarzyk, 2008). This study demonstrated that overall mean BMI is higher in this population, but did not show more increases over time. While research is ongoing in this area, it important to continue to uncover the factors underlying for Canadian Aboriginal peoples’ increased risk for obesity and related morbitities.

One limitation of this study is the use of parent- and self-reported heights and weights to calculate BMI. Although parent- and self-reports have been used previously to document important changes in overweight and obesity in Canadian youth (Tremblay & Willms, 2000), they may also be less accurate and reliable than measured height and weight. Parent-reported height and weight have been shown to lead to overestimation of child overweight and obesity (Shields, 2006), which may be primarily driven by underestimation of young children’s height (Akinbami & Ogden, 2009). Although adolescent self-reported height and weight correlates highly with measured height and weight (Goodman, Hinden, & Khandelwal, 2000), they tend to lead to underestimation of overweight and obesity (Shields, 2006; Strauss, 1999). Gender and race may be related to degree of reporting errors in parent-reports (O’Connor & Gugenheim, 2011) and adolescent self-reports (Goodman et al., 2000; Strauss, 1999). However, the validity of parent- and self-reports of height and weight by generation of immigration is largely unknown (Bates, Acevedo-Garcia, Alegria, & Krieger, 2007). The use of a continuous measure of BMI as the dependent variable in the current study may help to minimize errors from misclassification into weight categories (Strauss, 1999). We have previously referred to limitations in the NLSCY sample that led to low numbers of first-generation immigrants and non-White ethnoracial groups. Small sample sizes may have limited statistical power to observe significant differences between these groups. In addition, the current study grouped immigrants, refugees, and Canadian children born abroad, although their risk for obesity may vary. Given these limitations, it will be important to replicate the findings of this study using measured height and weight in a Canadian sample with a larger number of first-generation immigrants and non-White youth. This could be done by oversampling large urban centers, such as Toronto, Vancouver, or Montreal.

Future researchers should aim to uncover the mechanisms by which generational status is related to the development of overweight in children and adolescents. One avenue of research is to explicitly measure acculturation, rather than infer through generational status, to understand how culture change is related to weight gain. Acculturation may be assessed using cultural identity questionnaires, language proficiency and/or preference, or neighborhood ethnic or immigrant composition. Examination of time since immigration is another important variable to examine in this age group. A longitudinal design should be emphasized to understand the temporal pattern and begin to infer causality. Another area of future research is to measure important related risk factors, such as diet/food, physical activity/inactivity, stress, and sleep patterns, in relation to immigration, acculturation, and weight gain. Findings in these areas will promote further understanding of the healthy immigrant effect, including whether it is driven by lifestyle factors and whether excess weight may mediate the relationship between time since immigration and deterioration of health.

As researchers begin to understand how overweight and obesity develop in immigrant populations, clinicians and public health officials can design and implement strategies to prevent immigration-related weight gain and interventions targeted at reducing existing obesity in immigrant youth. Pediatricians and other health professionals should be cognizant of the risk for obesity in immigrant children, and may begin educating immigrant families about achieving healthy lifestyles in their new country. Researchers and policy makers may learn from the prevention and intervention strategies that have been implemented in Aboriginal children and adolescents, as these have often integrated the community and culture (e.g., Potvin, Cargo, McComber, Delormier, & Macaulay, 2003). It may be possible to build upon these interventions and change them to fit with the specific needs of immigrant youth. It will also be important to continue to adapt the existing interventions to address acculturation issues in Aboriginal youth.

In summary, the current study demonstrated that there are both cross-sectional and longitudinal differences in BMI between generations of immigration. Specifically, children of immigrant parents tend to have lower BMI than children of Canadian-born parents across childhood and adolescent age groups. However, contrary to most previous findings, first-generation immigrant youth did not have lower BMI than higher generations. In addition, first-generation immigrants were shown to gain weight relative to other groups across the adolescent age group. These findings suggest that the healthy immigrant effect for excess weight may not be straightforward in Canadian children and adolescents. First-generation immigrant youth appear to be at risk for developing obesity, which has important implications for future research as well as health policy and practices targeted at this population.

Acknowledgments

The authors thank Dr William Bukowski and Sivan Rotenberg for their statistical advice and helpful comments.

Funding

This work was supported by Canadian Institutes of Health Research (CGM 89256), Fonds de la recherche en santéau Quebec (Dossier 16843), and Quebec Inter-University Centre for Social Statistics.

Footnotes

Conflicts of interest: This analysis was based on Statistics Canada master files for NLSCY Cycles 1–7, which contain anonymized data collected from 1994 to 2007. The responsibility for the use and interpretation of these data is solely that of the authors. The opinions expressed in this article are those of the authors and do not represent the view of Statistics Canada.

References

  1. Akinbami LJ, Ogden CL. Childhood overweight prevalence in the United States: The impact of parent-reported height and weight. Obesity. 2009;17:1574–1580. doi: 10.1038/oby.2009.1. [DOI] [PubMed] [Google Scholar]
  2. Bates LM, Acevedo-Garcia D, Alegria N, Krieger N. Immigration and generational trends in body mass index and obesity in the United States: Results of the National Latino and Asian American Survey, 2002–2003. American Journal of Public Health. 2007;97:70–77. doi: 10.2105/AJPH.2006.102814. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bartlett JG. Involuntary culture change, stress phenomenon, and Aboriginal health status. Canadian Journal of Public Health. 2003;94:165–166. doi: 10.1007/BF03405058. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bélanger-Ducharme F, Tremblay A. Prevalence of obesity in Canada. Obesity Reviews. 2005;6:183–186. doi: 10.1111/j.1467-789X.2005.00179.x. [DOI] [PubMed] [Google Scholar]
  5. Beiser M, Hou F, Hyman I, Tousignant M. Growing up Canadian: A study of new immigrant children. Ottawa, ON: Human Resources Development Canada; 1998. [Google Scholar]
  6. Berry JW, Annis RC. Acculturative stress: The role of ecology, culture and differentiation. Journal of Cross-Cultural Psychology. 1974;5:382–405. [Google Scholar]
  7. Berry JW, Kim U, Minde T, Mok D. Comparative studies of acculturative stress. International Migration Review. 1987;21:491–511. [Google Scholar]
  8. Biddle N, Kennedy S, McDonald JT. Health assimilation patterns amongst Australian immigrants. Economic Record. 2007;83:16–30. [Google Scholar]
  9. Blackstock C, Trocmé N, Bennett M. Child maltreatment investigations among Aboriginal and non-Aboriginal families in Canada. Violence Against Women. 2004;10:901–916. [Google Scholar]
  10. Bryk AS, Raudenbush SW. Application of hierarchical linear models to assessing change. Psychological Bulletin. 1987;101:147–158. [Google Scholar]
  11. Cairney J, Ostbye T. Time since immigration and excess body weight. Canadian Journal of Public Health. 1998;90:120–124. doi: 10.1007/BF03404114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chen J, Ng E, Wilkins R. The health of Canada’s immigrants in 1994–1995. Health Reports. 1996;7(4):33–45. [PubMed] [Google Scholar]
  13. Citizenship and Immigration Canada. Facts and figures 2010: Immigration overview - Permanent and temporary residents. Ottawa, ON: Citizenship and Immigration Canada; 2010. [Google Scholar]
  14. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard definition for child overweight and obesity worldwide: International survey. British Medical Journal. 2000;320:1240–1245. doi: 10.1136/bmj.320.7244.1240. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. DeLucia C, Pitts SC. Applications of individual growth curve modeling for pediatric psychology research. Journal of Pediatric Psychology. 2006;31:1002–1023. doi: 10.1093/jpepsy/jsj074. [DOI] [PubMed] [Google Scholar]
  16. First Nations Centre. First Nations Regional Longitudinal Health Survey 2002/03: Results for adults, youth, and children living in First Nations communities. Ottawa, ON: First Nations Information Governance Committee; 2005. [Google Scholar]
  17. Goel MS, McCarthy EP, Phillips RS, Wee CC. Obesity among US immigrant subgroups by duration of residence. Journal of the American Medical Association. 2004;292:2960–2867. doi: 10.1001/jama.292.23.2860. [DOI] [PubMed] [Google Scholar]
  18. Goodman E, Hinden BR, Khandelwal S. Accuracy of teen and parental reports of obesity and body mass index. Pediatrics. 2000;106:52–58. doi: 10.1542/peds.106.1.52. [DOI] [PubMed] [Google Scholar]
  19. Gordon-Larsen P, Mullan Harris K, Ward DS, Popkin BM. Acculturation and overweight-related behaviors among Hispanic immigrants to the US: The National Longitudinal Study of Adolescent Health. Social Science & Medicine. 2003;57:2023–2034. doi: 10.1016/s0277-9536(03)00072-8. [DOI] [PubMed] [Google Scholar]
  20. Haas JS, Lee LB, Kaplan CP, Sonneborn D, Phillips KA, Liang S. The association of race, socioeconomic status and health insurance status with the prevalence of overweight among children and adolescents. American Journal of Public Health. 2003;93:2105–2110. doi: 10.2105/ajph.93.12.2105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hsin O, La Greca AM, Valenzuela J, Moine CT, Delamater A. Adherence and glycemic control among Hispanic youth with Type 1 diabetes: Role of family involvement and acculturation. Journal of Pediatric Psychology. 2010;35I:156–166. doi: 10.1093/jpepsy/jsp045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hyman I. Immigration and health. Health Policy Working Paper 01–05. Ottawa, ON: Health Canada; 2001. [Google Scholar]
  23. Janssen I, Katzmarzyk PT, Boyce WF, King MA, Pickett W. Overweight and obesity in Canadian adolescents and their associations with dietary habits and physical activity patterns. Journal of Adolescent Health. 2004;35:360–367. doi: 10.1016/j.jadohealth.2003.11.095. [DOI] [PubMed] [Google Scholar]
  24. Katzmarzyk PT. Obesity and physical activity among Aboriginal Canadians. Obesity. 2008;16:184–190. doi: 10.1038/oby.2007.51. [DOI] [PubMed] [Google Scholar]
  25. Kennedy S, McDonald JT, Biddle N. The healthy immigrant effect and immigrant selection: Evidence from four countries. SEDAP Research Paper. 2006:164. [Google Scholar]
  26. Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z, … Johnson CL. CDC growth charts for the United States: Methods and development. National Center for Health Statistics. Vital Health Statistics. 2000;11:246. [PubMed] [Google Scholar]
  27. McDonald JT, Kennedy S. Is migration to Canada associated with unhealthy weight gain? Overweight and obesity among Canada’s immigrants. Social Science and Medicine. 2005;61:2469–2481. doi: 10.1016/j.socscimed.2005.05.004. [DOI] [PubMed] [Google Scholar]
  28. Must A, Anderson SE. Body mass index in children and adolescents: Considerations for population-based applications. International Journal of Obesity. 2006;30:590–594. doi: 10.1038/sj.ijo.0803300. [DOI] [PubMed] [Google Scholar]
  29. O’Connor DP, Gugenheim JJ. Comparison of measured and parents’ reported height and weight in children and adolescents. Obesity. 2011;19:1040–1046. doi: 10.1038/oby.2010.278. [DOI] [PubMed] [Google Scholar]
  30. Oliver LN, Hayes MV. Neighbourhood socio-economic status and the prevalence of overweight Canadian children and youth. Canadian Journal of Public Health. 2005;96:415–420. doi: 10.1007/BF03405180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Perez CE. Health status and health behaviour among immigrants. Health Reports. 2002;13(Suppl):1–13. [Google Scholar]
  32. Popkin BM, Udry JR. Adolescent obesity increases significantly in second and third generation U.S. immigrants: The National Longitudinal Study of Adolescent Health. Journal of Nutrition. 1998;128:701–706. doi: 10.1093/jn/128.4.701. [DOI] [PubMed] [Google Scholar]
  33. Potvin L, Cargo M, McComber AM, Delormier T, Macaulay AC. Implementing participatory intervention and research in communities: Lessons from the Kahnawake Schools Diabetes Prevention Project in Canada. Social Science & Medicine. 2003;56:1295–1305. doi: 10.1016/s0277-9536(02)00129-6. [DOI] [PubMed] [Google Scholar]
  34. Redfield R, Linton R, Herskovits MJ. Memorandum for the study of acculturation. American Anthropologist. 1936;38:149–152. [Google Scholar]
  35. Ring I, Brown N. The health status of indigenous peoples and others. British Medical Journal. 2003;327:404. doi: 10.1136/bmj.327.7412.404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Shields M. Overweight and obesity among children and youth. Health Reports. 2006;17:27–42. [PubMed] [Google Scholar]
  37. Singh GK, Kogan M, Yu SM. Disparities in obesity and overweight prevalence among US immigrant children and adolescents by generational status. Journal of Community Health. 2009;34:271–281. doi: 10.1007/s10900-009-9148-6. [DOI] [PubMed] [Google Scholar]
  38. Singh G, Siahpush M. Ethnic-immigrant differentials in health behaviors, morbidity, and cause-specific mortality in the United States: An analysis of two national data bases. Human Biology. 2002;74:83–109. doi: 10.1353/hub.2002.0011. [DOI] [PubMed] [Google Scholar]
  39. Statistics Canada, & Human Resources Development Canada. National Longitudinal Survey of Children: Overview of survey instruments for 1994–95 data collection, Cycle 1. Catalogue no. 89F0078XIE. Ottawa, ON: Statistics Canada; 1995. [Google Scholar]
  40. Statistics Canada. Projections of the Aboriginal populations, Canada, Provinces and Territories, 2001 to 2017. Catalogue no. 91547XIE. Ottawa, ON: Statistics Canada; 2005. [Google Scholar]
  41. Statistics Canada. Visible minorities in Canada. Catalogue No. 85F0033MIE. Ottawa, ON: Statistics Canada; 2001. [Google Scholar]
  42. Strauss RS. Comparison of measured and self-reported weight and height in a cross-sectional sample of young adolescents. International Journal of Obesity. 1999;23:904–908. doi: 10.1038/sj.ijo.0800971. [DOI] [PubMed] [Google Scholar]
  43. Swinburn B, Egger G, Raza F. Dissecting obesogenic environments: The development and application of a framework for identifying and prioritizing environmental interventions for obesity. Preventive Medicine. 1999;29:563–570. doi: 10.1006/pmed.1999.0585. [DOI] [PubMed] [Google Scholar]
  44. Tremblay MS, Perez CE, Ardem CI, Bryan SI, Katzmarzyk PT. Obesity, overweight, and ethnicity. Health Reports. 2005;16:23–34. [PubMed] [Google Scholar]
  45. Tremblay MS, Willms JD. Secular trends in the body mass index of Canadian children. Canadian Medical Association Journal. 2000;163:1429–1433. [PMC free article] [PubMed] [Google Scholar]
  46. Young DE, Spitzer DL, Pang F. Understanding the health care needs of Canadian immigrants. Edmonton, AB: Report for the Prairie Centre for Excellence on Immigration and Integration; 1999. [Google Scholar]

RESOURCES