Body mass index among immigrant and non-immigrant youth: evidence from the Canadian Community Health Survey.
Wahi, Gita ; Boyle, Michael H. ; Morrison, Katherine M. 等
Over the past two decades, the prevalence of childhood overweight
and obesity has steadily risen around the world. (1-3) In Canada, the
prevalence of childhood overweight/obesity has risen from 15% in 1978 to
26% in 2004. (4) The health consequences of this trend include rising
rates of cardiovascular disease, type 2 diabetes and shorter life
expectancy. (5,6)
Immigrants in Canada are the fastest-growing segment of the
population, contributing to two thirds of its population growth. (7)
Studies contrasting health outcomes among immigrant versus nonimmigrant
youth lend support to a general pattern of findings, termed the healthy
immigrant effect (HIE). The HIE suggests that immigrants arrive in a
host country with a favourable state of health, however over time there
is evidence of declining health and convergence to that of the
native-born population. (8) Key to this experience is a pattern of
findings where, over time, immigrants adopt the host populations'
habits and lifestyle, termed acculturation, putting them at risk for
poor health outcomes. (8) There are several reasons why immigrant status
needs consideration in the epidemic of childhood obesity. First, there
is limited information on obesity among Canadian immigrant children and
youth. Canadian data on adults suggest that immigrants have lower rates
of obesity compared to non-immigrants. (9-11) In contrast, Canadian
immigrant youth may experience lifestyle and socioeconomic risk factors
that place them at risk for obesity and related complications. (8,12-14)
Also, the shift in the national composition of recent immigrants to
Canada, combined with the deteriorating economic circumstances of recent
immigrant families over the past 20 years, calls into question the
applicability of the HIE to more recent cohorts of immigrant children
and adolescents. Foreign-born children are at higher risk for living in
lower socio-economic circumstances and with financial adversity, (13)
which in turn may place them at higher risk for poor health outcomes.
Further, socioeconomic disadvantage during childhood has been associated
with an increased risk of overweight/obesity in adulthood. (15) At
present, we have minimal empirical evidence that quantifies differences
in overweight/obesity between immigrant versus non-immigrant youth, and
the extent to which lifestyle and socio-demographic factors contribute
to these differences, if they exist.
Using a nationally representative sample of youth in Canada, the
objectives of this study are to: i) examine differences in body mass
index (BMI) and prevalence of overweight/obesity between immigrant
versus non-immigrant youth aged 12-19 years and ii) identify the extent
to which lifestyle and socio-demographic factors account for
between-group differences.
METHODS
Participants
Data for analyses were based on four cycles of the Canadian
Community Health Survey (CCHS)--1.1, 2.1, 3.1, 4.1. The CCHS is a
cross-sectional survey conducted biennially by Statistics Canada. A
detailed description of the survey methodology has been described
elsewhere. (16) The reported response rates for each of the four cycles
were 85%, 81%, 79%, and 76% respectively. (17,18) The CCHS interview was
available in 24 different languages and participants could select their
choice of language. Approximately 75% of participants chose to complete
the interview in English. (18) Data analysis was conducted at the
Research Data Centre, operated by Statistics Canada, McMaster University
location. Approval for access to the CCHS restricted data file was
obtained through the joint committee operated by the Social Sciences and
Humanities Research Council (SSHRC) and Statistics Canada.
Sample for analysis
The sample for analysis includes participants aged 12 to 19 years
(n=67,406), with complete data on the body composition measures, i.e.,
height and weight (n=63,509). The average age of respondents in the
sample was 15.5 years with an even distribution between males (51.7%)
and females (48.3%). About 6.4% (n=4,052) of respondents were born
outside of Canada and classified as immigrant youth. The majority of
respondents could converse in English and/or French.
Measurement of overweight/obesity
Respondents were asked to self-report their height and weight. From
that information, we constructed the variable body mass index (BMI) z
score: The BMI for each respondent was calculated by dividing the
self-reported weight in kilograms (kg) by the square of height in metres
([m.sup.2]). The BMI z score (zBMI) was calculated using each
respondent's BMI, age, sex and the external reference of the World
Health Organization (WHO), based on the WHO Reference 2007 for 5-19 year
olds. (19)
Overweight/obese versus normal weight
For the purposes of estimating the proportion of respondents
classified as overweight/obese, BMI was converted to a dichotomous
variable using internationally based cut-off points. (20) These cut-offs
were derived from an international, multicultural sample of
cross-sectional growth surveys from six countries (Brazil, Great
Britain, Hong Kong, Netherlands, United States and Singapore). (20) The
cut-off points for overweight and obese correspond to adult (aged
greater than 18 years) BMI cut-offs of 25 kg/[m.sup.2] (overweight) and
30 kg/[m.sup.2] (obese).
Immigrant status and recency
Participants were asked if they were born in Canada. Those who
answered "no" to this question were considered immigrants.
Follow-up questions were asked about country of birth and year of
arrival in Canada. No additional information was collected on
parent's country of birth to further classify participants based on
immigrant generational status. As such, immigrant status is defined as
foreign-born (i.e., 1st generation immigrant) versus Canadian-born
(i.e., non-immigrant). Among immigrants, the length of time living in
Canada (reported in years) was also collected and reported as a
continuous variable.
Measures of socio-demographic covariates
Ethnicity/race of the participant was determined by asking which
racial or ethnic background they belonged to. This variable was dummy
coded (0, Caucasian and 1, non-Caucasian). Language of the participant
was determined by asking if they spoke English and/or French or neither.
This variable was dummy coded (0, either/both English and French and 1,
neither English nor French). Household size was reported by asking the
participant the number of people living in their primary residence.
Source of income was reported by asking the participants the primary
source of their household income. The variable was dummy coded 0,
non-assisted (i.e., salary, wages, self-employed, pension), and 1,
government assisted (i.e., Employment Insurance, government assistance,
no income).
Measurement of lifestyle covariates
Physical activity was a derived measure of energy expenditure (EE)
(kilocalories expended per kilogram of body weight per day) based on
self-reported questionnaire responses. The measure of physical activity
takes into account the following information: the number of times the
respondent reports participating in any activity; the average duration
of the activity (in hours); and the MET, which is the "metabolic
energy cost" of the activity expressed as kilocalories expended per
kilogram of body weight per hours of activity (kcal/kg per hour). To
further illustrate the interpretation of METs: 2 METs would describe an
activity that required twice the amount of energy as compared to a body
at rest. Different activities are pre-assigned METs by the Canadian
Fitness and Lifestyle Research Institute. (21)
Fruit and vegetable intake was measured by the participants'
response to questions of frequency of consumption of fruits and
vegetables per week. Participants reported how many times per week they
ate any fruits or vegetables. Of note, this question denotes frequency
but not amounts of consumption and is not based on Canada's Food
Guide serving sizes.
Statistical analysis
Participants' characteristics are reported using descriptive
statistics, continuous variables reported as means and standard
deviations and categorical variables are reported using percentages.
Differences between immigrant and non-immigrant youth were examined
using chi-square tests for categorical variables and independent t-tests
for continuous variables.
About 33.6% of the sample had one or more missed responses on study
variables. Multiple imputation (MI) (22) using SPSS 19.0 was used to
address missing data in the current study. For MI, SPSS uses an
algorithm based on linear regression, Markov chain Monte Carlo (MCMC)
method. Five imputed datasets were created and subsequently combined
using Rubin's rules for scalar estimands. (22) Given the design of
the CCHS (i.e., individuals clustered within health regions), 2-level,
multilevel regression models (MLM) were used for the analysis. A
multilevel linear model was used for the continuous dependent variable,
zBMI, and a logistic model for the binary response dependent variable,
weight category (i.e., normal weight versus overweight/obese).
Initially, the model is estimated using first-order marginal
quasi-likelihood (MQL). As first-order MQL may underestimate
between-group variation, the final model is estimated using second-order
penalized quasi-likelihood (PQL) and iterative generalized least-squares
estimation. Three models are examined. The first model examines
differences between immigrant and non-immigrant youth after adjusting
for age, sex and CCHS cycle. Lifestyle factors are added in Model 2 and
socio-demographic factors are added in Model 3. Multilevel analyses were
conducted with MLwiN software version 2.24. (23)
RESULTS
Table 1 presents differences between immigrant and nonimmigrant
youth on key study variables. About 88.2% of non-immigrant youth
identified as Caucasian, compared to only 26.9% of immigrant youth.
Immigrant youth consumed fruits and vegetables less frequently compared
to non-immigrant youth. Immigrants have lower zBMI scores and have a
lower prevalence of overweight/obesity. More immigrant youth live in
households that are larger and in households that receive government
income assistance, compared to non-immigrant youth. The mean number of
years (SD) that immigrant youth lived in Canada was 7.1 (4.1) years.
Table 2 presents the results for zBMI. Model 1 demonstrates a
significant, negative association between immigrant status and zBMI. The
zBMI is 0.44 lower among immigrant compared to nonimmigrant youth. Also
zBMI increased by 0.02 for every year an immigrant respondent resided in
Canada (p<0.05). Measures of diet (fruit and vegetable consumption)
and activity level (energy expenditure) were added in Model 2. This did
not modify the association between zBMI and immigrant status. However,
there was a negative association between frequency of fruit/vegetable
consumption and zBMI (b=-0.01, se=0.002), and a positive association
with energy expenditure (b=0.002, SE=0.001). The addition of
socio-demographic factors in Model 3 did not affect the strength of
association between immigrant status and zBMI. Speaking neither English
nor French and more family members residing in a household were
significantly associated with lower zBMI scores. Having an income source
that was government assisted was associated with a higher zBMI score.
Table 3 presents results from the binary weight variable. The
trends had many similarities to the zBMI models. The association between
immigrant status and weight category is significant and negative (OR
0.66, 95% CI: 0.45-0.86). Neither the direction nor the magnitude of the
association between weight category and immigrant status changed in
Models 2 or 3. Of note, energy expenditure demonstrated a significant,
negative association with weight category, and having more family
members residing in a household was negatively associated with
overweight/obesity. Again, having a government-assisted income source
was associated with overweight/obesity.
DISCUSSION
Using data from a series of cross-sectional Canadian Community
Health Surveys, the current study compares body mass index among
immigrant and non-immigrant Canadian youth. A secondary goal was to
identify the extent to which lifestyle and socio-demographic factors
account for between-group differences. We found that Canadian immigrant
youth have lower zBMI and a lower prevalence of overweight/obesity,
relative to Canadian-born youth. Also, for immigrant youth, length of
time in Canada (i.e., recency) was associated with higher zBMI scores
and increased odds of overweight/obesity. Further, there were
differences in lifestyle factors including less frequent consumption of
fruits and vegetables among immigrant youth compared to Canadian-born
youth.
Previously published work has consistently demonstrated that
immigrant adults in Canada have lower rates of overweight/obesity
relative to Canadian-born adults. (9-11) Furthermore, evidence from
longitudinal data from the National Population Health Survey (NPHS)
demonstrated lower BMI among immigrants compared to non-immigrants in
Canada and that BMI converged to nonimmigrant levels over a 12-year
period among Caucasian male immigrants, however the sample did not
include anyone under age 18 years. (10) The observation of these
findings in Canadian youth is important as it gives valuable insight to
an important developmental period that could be pursued for targeted
health promotion and prevention efforts.
Despite some differences in lifestyle factors between immigrant and
non-immigrant youth, the addition of lifestyle factors, including diet
and activity, did not modify the association between immigrant status
and zBMI or proportion of overweight/obesity (Tables 2 and 3). However,
there were some contrary results between the continuous and binary
outcome variables. For energy expenditure, the odds of
overweight/obesity decreased by 2.0% (OR 0.98, 95% CI: 0.97-0.99) with
each 1-unit increase of energy expenditure. This finding is in line with
previously reported literature that reports a negative association
between physical activity and rates of overweight/obesity. (24-26)
However we also found a significant association between zBMI and energy
expenditure, in an opposite direction: zBMI increased with an increase
in energy expenditures. These contradictory results may be due to a lack
of robustness in the measure, or non-linearity in the association
between the measures due to reporting biases or adjustments to the
measurement of zBMI. Albeit statistically significant, the magnitude of
this association with overweight/obesity is small compared to previously
reported estimates. It has been shown that children who engage in
vigorous physical activity are 33% less likely to be obese compared to
those who do not. (26) Further, there was no significant difference in
physical activity, as measured by energy expenditure (kkd), between
immigrant and non-immigrant youth. That being said, it has been
demonstrated that Canadian immigrant adults and children in the United
States have lower levels of physical activity (27,28) than their
non-immigrant counterparts, and this differing lifestyle routine is one
hypothesis with regard to the increasing risk for obesity with time
spent living in North America. Also of note is the negative association
between frequency of fruits and vegetables consumption and zBMI score.
Similar to the case with energy expenditure, the size of this effect in
the current study is small, however in comparison to other reported
studies, similar in magnitude. For example, in this study, with every
1-unit increase in frequency of fruit and vegetable consumption, we
observed a corresponding decrease in zBMI score of 0.01 (SE 0.002,
p<0.05). In comparison, a large prospective cohort of American
children showed that with every 1 serving of vegetable consumption, boys
aged 9-14 had a 0.003 decrease in zBMI (95% CI: -0.004, -0.001). (29)
Although of similar direction, the unit of sampling, frequency versus
serving size, makes direct comparisons difficult. Nonimmigrant youth
reported greater frequency of consumption of fruits and vegetables than
immigrant youth. Although the difference was quite small, the difference
in diet may contribute to the observation that the odds of
overweight/obesity among immigrant youth increased with time spent in
Canada. One factor that may influence the ability to access a healthy
diet may be socioeconomic status. As previous studies have demonstrated,
families with lower socio-economic status consume less fruits and
vegetables compared to those of higher socio-economic status, (30) and
immigrant youth in this study were more likely to have household income
in the form of government support, a proxy measure for socio-economic
status.
The addition of socio-demographic factors did not attenuate the
association between immigrant status and weight measures, suggesting
that the observed group differences between immigrant and non-immigrant
youth are not attributable to socio-demographic factors.
Government-assisted income source was significantly, positively
associated with BMI.
Strengths of the study
At present, we have not been able to identify any other studies
that address the issue of overweight and obesity among Canadian
immigrant youth, which makes this study unique in describing a growing
segment of the Canadian youth population. The strengths of this study
include adequate statistical power, with a robust sample size of over
60,000 youth. We also had the opportunity to combine both
socio-demographic and lifestyle factors in the models, which allowed for
a comprehensive examination of possible contributing factors in the
association between overweight/obesity and immigration status.
Limitations of the study
As mentioned previously, the data for this secondary data analysis
were cross-sectional in nature. Although four cycles of the CCHS were
combined for the analysis, there was only one data point per respondent,
ruling out a longitudinal analysis and any opportunity to examine
temporal associations among the study variables. Specifically, we were
not able to address how differences in lifestyle factors may have
affected BMI over time for each group. There was a relatively large
amount of missing information, including ~16% missed responses for
frequency of fruit and vegetable consumption. To address loss of data,
multiple imputation was used to account for variables that were missing.
Additionally, the dependent variables were reliant on self-reported
height and weight, which may contribute to reporting bias. In a previous
study, Shields and colleagues compared self-reported height and weight
to measured height and weight in a subsample of the 2005 CCHS survey.
(31) They demonstrated that, in general, respondents over-estimated
their height compared to their actual measured height. (31) Furthermore,
respondents generally under-reported their weight, with a larger
difference among female respondents than males. (31) These reporting
biases were similar for immigrants and non-immigrants. Although the
dependent variable may be affected by self-report, it is likely that the
two groups, immigrant and non-immigrant youth, are affected in similar
fashion. The majority of other included variables were also
self-reported, such as lifestyle factors, including frequency of fruit
and vegetable consumption and energy expenditure. A systematic review of
assessments of direct versus self-reported physical activity showed low
to moderate correlation between measures. (32) Further a systematic
review examining the reliability and validity of sedentary behaviours,
found self-reported methods to be reliable but their validity untested.
(33)
CONCLUSION
This study adds to the growing body of evidence in North America
that adult and youth immigrants compared to native-born peers have
different levels of overweight and obesity. (9-11,34) We demonstrate
that immigrant youth in Canada are less often overweight/obese and have
lower BMI z scores compared to nonimmigrant youth. Also, an important
observation to highlight for future primary prevention strategies is the
impact of recency on risk of becoming overweight and obese among
Canadian immigrant youth. Further, immigrant youth may have modifiable
risk factors, including less frequent consumption of fruits and
vegetables, although the magnitude of the difference between groups is
modest and conclusions should be drawn with caution. By prospectively
studying this subgroup of Canadian youth, we will be able to explore
facilitators and barriers to healthy lifestyle behaviours and to design
effective intervention strategies.
Conflict of Interest: None to declare.
REFERENCES
(1.) Tremblay MS, Willms JD. Secular trends in the body mass index
of Canadian children. CMAJ 2000;163(11):1429-33.
(2.) World Health Organization.
http://www.who.int/dietphysicalactivity/childhood (Accessed December 21,
2010).
(3.) Krebs NF, Jacobson MS. Prevention of pediatric overweight and
obesity. Pediatrics 2003;112(2):424-30.
(4.) Shields M. Overweight and obesity among children and youth.
Health Rep 2006;17(3):27-42.
(5.) Reilly JJ, Methven E, McDowell ZC, Hacking B, Alexander D,
Stewart L, et al. Health consequences of obesity. Arch Dis Child
2003;88:748-52.
(6.) Olshansky SJ, Passaro DJ, Hershow RC, Layden J, Carnes BA,
Brody J, et al. A potential decline in life expectancy in the United
States in the 21st century. N Engl J Med 2005;352(11):1138-45.
(7.) Statistics Canada. Immigration in Canada: A portrait of the
Foreign-born population. Available at:
http://www.census2006.ca/census-recensement/2006/assa/97-557/index-eng.cfm (Accessed October 27, 2009).
(8.) Newbold KB. Chronic conditions and the healthy immigrant
effect: Evidence from Canadian immigrants. J Ethn Migr Stud
2006;32(5):765-84.
(9.) Cairney J, Ostbye T. Time since immigration and excess body
weight. Can J Public Health 1999;90(2):120-24.
(10.) Setia MS, Quesnel-Vallee A, Abrahamowicz M, Tousignant M,
Lynch J. Convergence of body mass index of immigrants to the
Canadian-born population: Evidence from the National Population Health
Survey (1994-2006). Eur J Epidemiol 2009;24:611-23.
(11.) McDonald JT, Kennedy S. Is migration to Canada associated
with unhealthy weight gain? Overweight and obesity among Canada's
immigrants. Soc Sci Med 2005;61:2469-81.
(12.) Newbold B. The short-term health of Canada's new
immigrant arrivals: Evidence from LSIC. Ethn Health 2009;14(3):315-36.
(13.) Beiser M, Hou F, Hyman I, Tousignant M. Poverty, family
process, and the mental health of immigrant children in Canada. Am J
Public Health 2002;92(2):220-27.
(14.) Beiser M. The health of immigrants and refugees in Canada.
Can J Public Health 2005;96(Suppl 2):S30-S44.
(15.) Parsons TJ, Power C, Logan S, Summerbell CD. Childhood
predictors of adult obesity: A systematic review. Int J Obes Relat Metab
Disord 1999;23(Suppl 8):S1-S107.
(16.) Statistics Canada. Canadian Community Health Survey--Annual
component. Available at:
http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&
SDDS=3226&lang=en&db=imdb&adm=8&dis=2 - b3 (Accessed
March 26, 2012).
(17.) Statistics Canada. The impact of considering birthplace in
analyses of immigrant health. Available at:
http://www.statcan.gc.ca/pub/82-003 x/2011004/article/11589-eng.htm
(Accessed March 26, 2012).
(18.) Statistics Canada. Quality measures - Canadian Community
Health Survey. Available at:
http://www.statcan.gc.ca/imdb-bmdi/document/3226_D9_T9_ V1-eng.pdf
(Accessed March 26, 2012).
(19.) World Health Organization. Child growth standards. Available
at: http://www.who.int/growthref/tools/en/ (Accessed November 1, 2009).
(20.) Cole TJ, Bellizi MC, Flegal KM, Dietz WH. Establishing a
standard definition for child overweight and obesity worldwide:
International survey. BMJ2000;320(7244):1240-43.
(21.) Canadian Fitness and Lifestyle Research Institute. Available
at: http://www.cflri.ca (Accessed December 10, 2010).
(22.) Rubin DB. Multiple Imputation for Nonresponse in Surveys. New
York, NY: John Wiley & Sons, Inc., 1987.
(23.) Rasbash J, Steele F, Browne WJ, Goldstein H. A user's
guide to MLwiN version 2.10: Centre for Multilevel Modelling, University
of Bristol, 2009.
(24.) Warburton DER, Nicol CW, Bredin SSD. Health benefits of
physical activity: The evidence. CMAJ 2006;174(6):801-9.
(25.) Patrick K, Norman GJ, Calfas KJ, Sallis JF, Zabinski MF, Rupp
J, et al. Diet, physical activity, and sedentary behaviors as risk
factors for overweight in adolescence. Arch Pediatr Adolesc Med
2004;158(4):385-90.
(26.) Janssen I, LeBlanc AG. Systematic review of the health
benefits of physical activity and fitness in school-aged children and
youth. Int J Behav Nutr Phys Act 2010;7(40).
(27.) Tremblay MS, Bryan SN, Perez CE, Ardern CI, Katzmarzyk PT.
Physical activity and immigrant status. Can J Public Health
2006;97(4):277-82.
(28.) Singh GK, Yu SM, Siahpush M, Kogan MD. High levels of
physical inactivity and sedentary behaviors among US immigrant children
and adolescents. Arch Pediatr Adolesc Med 2008;162(8):756-63.
(29.) Field AE, Gillman MW, Rosner B, Rockett HR, Colditz G.
Association between fruit and vegetable intake and change in body mass
index among a large sample of children and adolescents in the United
States. Int J Obes (Land) 2003;27:821-26.
(30.) Azagba S, Sharaf MF. Disparities in the frequency of fruit
and vegetable consumption by socio-demographic and lifestyle
characteristics in Canada. Nutr J 2011;10.
(31.) Shields M, Gorber SC, Tremblay MS. Estimates of obesity based
on self-report versus direct measures. Health Rep 2008;19(2):61-76.
(32.) Prince SA, Adamo KB, Hamel ME, Hardt J, Gorber SC, Tremblay
M. A comparison of direct versus self-report measures for assessing
physical activity in adults: A systematic review. Int J Behav Nutr Phys
Act 2008;5(56).
(33.) Lubans DR, Hesketh K, Cliff DP, Barnett LM, Salmon J, Dollman
J, et al. A systematic review of the validity and reliability of
sedentary behaviour measures used with children and adolscents. Obes Rev
2011;12(10):781-99.
(34.) Singh GK, Kogan MD, Yu SM. Disparities in obesity and
overweight prevalence among US immigrant children and adolescents by
generational status. J Community Health 2009;34(4):271-81.
Received: October 10, 2013
Accepted: April 27, 2014
Gita Wahi, MD, MSc, [1] Michael H. Boyle, PhD, [2] Katherine M.
Morrison, MD, [1] Katholiki Georgiades, PhD [2]
[1.] Department of Pediatrics, McMaster University, Hamilton, ON
[2.] Offord Centre for Child Studies, Department of Psychiatry and
Behavioural Neurosciences, McMaster University, Hamilton, ON
Correspondence: Dr. Gita Wahi, Department of Pediatrics, McMaster
University, 1280 Main Street West, Hamilton, ON L8S 4K1, E-mail:
wahig@mcmaster.ca
Table 1. Descriptive characteristics of respondents by immigrant
group (n=63,509)
Characteristic Non- Immigrant p-value
immigrant (n=4,052)
(n=59,457)
Age, years 15.51 (0.01) 15.92 (0.03) <0.001
(mean, sd)
Sex, female (%) 48.4 46.9 0.022
Race, Caucasian (%) 88.2 26.9 <0.001
Language, English 99.5 96.7 <0.001
or French (%)
Weekly consumption 5.22 (0.02) 5.08 (0.04) 0.001
of fruits and
vegetables
(mean, sd)
Energy expenditure 3.91 (0.02) 3.83 (0.05) 0.118
(mean, sd)
Number living in 4.17 (0.01) 4.38 (0.02) <0.001
household
(mean, sd)
Income source 3.5 5 <0.001
(% assisted)
Overweight or 21.8 18 <0.001
obese (%)
BMI z score 0.34 (0.004) 0.09 (0.001) <0.001
(mean, sd)
Table 2. Multilevel linear model of zBMI, beta coefficients
([beta]) and standard errors (SE)
Model 1 Model 2
[beta] SE [beta]
Fixed effect intercept 1.02 0.03 * 1.04
Immigrant -0.44 0.04 * -0.44
(ref = non-immigrant)
Time since immigration 0.02 0.004 * 0.02
Age -0.03 0.002 * -0.03
Female (ref = male) -0.33 0.01 * -0.32
CCHS cycle 2 0.04 0.01 * 0.04
(ref = cycle 1)
CCHS cycle 3 0.04 0.01 * 0.05
CCHS cycle 4 0.04 0.01 * 0.04
Fruit/vegetable -0.01
consumption
Energy expenditure 0.002
Language neither
English nor
French (ref = English
and/or
French language)
Income source
government
assisted (ref = Income
source not
government assisted)
Non-Caucasian
(ref = Caucasian)
Household size
Random effects
variance
Level 2 (HR) 0.02 0.002 0.02
Level 1 (respondent) 1.25 0.01 1.25
-2*log likelihood 195094.91 195081.28
Model 2 Model 3
SE [beta] SE
Fixed effect intercept 0.03 * 1.22 0.04 *
Immigrant 0.04 * -0.42 0.04 *
(ref = non-immigrant)
Time since immigration 0.004 * 0.02 0.004 *
Age 0.002 * -0.03 0.002 *
Female (ref = male) 0.01 * -0.32 0.01 *
CCHS cycle 2 0.01 * 0.03 0.01 *
(ref = cycle 1)
CCHS cycle 3 0.01 * 0.04 0.01 *
CCHS cycle 4 0.01 * 0.03 0.01 *
Fruit/vegetable 0.002 * -0.01 0.002 *
consumption
Energy expenditure 0.001 * 0.002 0.001 *
Language neither
English nor
French (ref = English
and/or
French language) -0.16 0.06 *
Income source
government
assisted (ref = Income
source not
government assisted) 0.07 0.03 *
Non-Caucasian -0.02 0.02
(ref = Caucasian)
Household size -0.03 0.004 *
Random effects
variance
Level 2 (HR) 0.002 0.02 0.002
Level 1 (respondent) 0.01 1.25 0.01
-2*log likelihood 194990.49
ref = reference category; HR = health region.
* p<0.05.
Table 3. Multilevel logistic model of overweight/obesity,
Odds Ratio (OR) and 95% Confidence Interval (CI)
Model 1
OR 95% CI
Fixed effect intercept 0.41 0.26, 0.56
Immigrant (ref = non-immigrant) 0.66 0.45, 0.86
Time since immigration 1.02 1.00, 1.04
Age 1.00 0.99, 1.01
Female (ref = male) 0.57 0.53, 0.61
CCHS cycle 2 (ref = cycle 1) 1.04 0.98, 1.09
CCHS cycle 3 1.06 1.01, 1.11
CCHS cycle 4 1.08 1.03, 1.14
Fruit/vegetable consumption
Energy expenditure
Language neither English nor
French (ref = English and/or
French language)
Income source government
assisted (ref = Income source
not government assisted)
Non-Caucasian (ref = Caucasian)
Household size
Random effects variance
Level 2 (HR) 0.06 0.01
Level 1 (respondent) 1 0
Model 2
OR 95% CI
Fixed effect intercept 0.47 0.32, 0.63
Immigrant (ref = non-immigrant) 0.65 0.45, 0.86
Time since immigration 1.02 1.00, 1.04
Age 1.00 0.99, 1.01
Female (ref = male) 0.56 0.52, 0.60
CCHS cycle 2 (ref = cycle 1) 1.04 0.99, 1.09
CCHS cycle 3 1.07 1.02,1.12
CCHS cycle 4 1.09 1.04, 1.15
Fruit/vegetable consumption 1.00 0.99, 1.01
Energy expenditure 0.98 0.97, 0.99
Language neither English nor
French (ref = English and/or
French language)
Income source government
assisted (ref = Income source
not government assisted)
Non-Caucasian (ref = Caucasian)
Household size
Random effects variance
Level 2 (HR) 0.06 0.01
Level 1 (respondent) 1 0
Model 3
OR SE
Fixed effect intercept 0.73 0.54, 0.92
Immigrant (ref = non-immigrant) 0.64 0.43, 0.85
Time since immigration 1.02 1.01, 1.04
Age 0.99 0.98, 1.00
Female (ref = male) 0.56 0.51, 0.60
CCHS cycle 2 (ref = cycle 1) 1.02 0.97, 1.08
CCHS cycle 3 1.05 1.00, 1.10
CCHS cycle 4 1.07 1.02, 1.13
Fruit/vegetable consumption 1.00 0.99, 1.01
Energy expenditure 0.98 0.97, 0.99
Language neither English nor
French (ref = English and/or
French language) 0.76 0.45, 1.07
Income source government
assisted (ref = Income source
not government assisted) 1.13 1.03, 1.23
Non-Caucasian (ref = Caucasian) 1.07 0.97, 1.16
Household size 0.92 0.90, 0.94
Random effects variance
Level 2 (HR) 0.06 0.01
Level 1 (respondent) 1 0
ref = reference category; HR = health region.
* p<0.05.
