need help
Lamin
Article4..pdf
ORIGINAL PAPER
Socioeconomic inequalities in adult obesity risk in Canada: trends and decomposition analyses
Mohammad Hajizadeh • M. Karen Campbell •
Sisira Sarma
Received: 13 July 2012 / Accepted: 20 February 2013 / Published online: 31 March 2013
� Springer-Verlag Berlin Heidelberg 2013
Abstract This study examines trends in socioeconomic-
related inequalities in obesity risk among Canadian adults
(aged 18–65 years) from 2000 to 2010 using five nationally
representative Canadian Community Health Surveys
(CCHSs). We employed the concentration index (C) to
quantify the socioeconomic inequalities in obesity risk
across different demographic groups and geographic
regions in each survey period. A decomposition analysis of
inequality is performed to determine factors that lie behind
income-related inequality in obesity risk. Although
declining over time, the results show that there exists
income-related inequality in obesity risk in Canada. The
estimated Cs for men indicate that obesity is concentrated
among the rich and its trend is increasing over time. The
findings, however, suggest that obesity is more prevalent
among economically disadvantaged women. While we
found that obesity is mainly concentrated among the poor
in the Atlantic Provinces, the degree of socioeconomic
related inequality in obesity risk is increasing in these
provinces. The results for Alberta showed that obesity is
concentrated among the better-off individuals. The
decomposition analysis suggests that factors such as
demographics, income, immigration, education, drinking
habits, and physical activity are the key factors explaining
income-related inequality in obesity risk in Canada. Our
empirical findings suggest that, in order to combat the
obesity epidemic, health policies should focus on poorer
females and economically well-off males.
Keywords Socioeconomic inequality � Obesity � Decomposition analysis � Canada
JEL Classification I14 � D63 � I18
Introduction
Inequalities in health among different socioeconomic seg-
ments of the population remain one of the main challenges
for public health throughout the world [1]. In all countries,
individuals with higher socioeconomic status (SES) are in
the better state of health compared to their lower SES
counterparts [2]. In Canada, for example, in spite of the
substantial improvements in overall health-related quality
of life, inequality in health remains a public health concern
[3, 4]. Although inequalities in health can result from dif-
ferential access to care and/or differences in health seeking
behaviors of individuals, several studies (e.g., [5–7]) dem-
onstrate that social determinants of health, such as income,
play a major role in the observed health inequalities.
Although inequalities in health have received substantial
attention in the economics and public health literature (e.g.,
[8–13]), the policy implications of these studies may not be
directly applicable to inequality in obesity. To date, studies
on socioeconomic-related inequalities in obesity risk are
sparse despite the dramatic increase in the prevalence of
M. Hajizadeh � M. Karen Campbell � S. Sarma (&)
Department of Epidemiology and Biostatistics, Schulich School
of Medicine and Dentistry, The University of Western Ontario,
London, ON N6A 5C1, Canada
e-mail: [email protected]
M. Hajizadeh
e-mail: [email protected]
M. Karen Campbell
e-mail: [email protected]
M. Karen Campbell
Departments of Obstetrics and Gynecology and Paediatrics,
Schulich School of Medicine and Dentistry, The University
of Western Ontario, London, ON N6A 5C1, Canada
123
Eur J Health Econ (2014) 15:203–221
DOI 10.1007/s10198-013-0469-0
obesity throughout the world. In fact, obesity is now con-
sidered a major public health concern worldwide [14].
According to the World Health Organization [15], obesity
rates have more than doubled since 1980 worldwide, and
currently at least 250 million individuals are considered to
be obese. The percentage of Canadian adults who are
considered obese increased from 10 % in 1970 [16] to
25 % in 2008 [17].
Obesity is a serious public health concern because of its
strong association with many negative physical and mental
health outcomes, including heart disease, diabetes, and
depression as well as mortality [18–21]. The increased
emergence of the obesity epidemic also has serious eco-
nomic consequences [22]. In Canada, for example, it is
estimated that in 2005, the total economic burden of adult
obesity was $3.42 billion; of this, $1.62 billion was due to
direct costs and $1.80 billion was attributable to produc-
tivity losses [23].
Obesity generally results from an energy imbalance,
whereby caloric intake exceeds the energy expenditure
[24]. In spite of this fairly simple equation, obesity is a
complex disease [25]. Although genetic predisposition is
an important consideration in understanding obesity risk, it
cannot explain the observed rapid increase in obesity rates
in recent decades [26, 27]. It is argued by some (e.g., [27,
28]) that psychosocial and technological factors are con-
tributing to the rapidly increasing rate of obesity. Further,
there is a substantial international evidence regarding the
effect of SES on obesity; two systematic reviews by Sobal
and Stunkard [29] and Mclaren [18] demonstrate that
education is negatively associated with obesity risk among
men and women in developed countries, but the effect of
income on obesity risk is unclear.
The existing studies in Canada also illustrate socioeco-
nomic differences in obesity prevalence (e.g., [30–36]).
While these studies provide some evidence on the regional
and socioeconomic differences in obesity prevalence, the
aim of these studies was not to quantify and decompose the
extent of any inequalities in obesity. Understanding the
existence of socioeconomic inequalities in obesity risk will
have considerable policy implications, particularly if a
gradient in obesity attributed to socioeconomic inequalities
in society. In the existing literature, very little is known
regarding how inequalities in obesity vary between dif-
ferent socioeconomic groups, what are the relative contri-
butions of different socioeconomic factors, and how the
role of each factor influencing inequalities in obesity risk
changed over time.
The overall objective of this study is to provide a
comprehensive analysis of socioeconomic-related inequal-
ity in obesity risk among Canadian adults over the past
decade. To date, several health inequality indices have
been proposed in the literature. The most popular measures
of inequality are the Gini coefficient and the concentration
index (C). We employed the C approach to measure and
decompose socioeconomic inequalities in obesity risk
because, as argued by Wagstaff et al. [8], the Gini coeffi-
cient does not measure health inequalities that originate
from the underlying socioeconomic factors. The analysis is
based on five available Canadian Community Health Sur-
veys (CCHSs) representative of the general Canadian
population [37]. This study not only augments the body of
information that is accumulating on health inequality but
also brings out empirical evidence on factors behind socio-
economic related inequalities in obesity risk in Canada.
The decomposition analysis enables us to identify the
contribution of each factor on socioeconomic inequality in
obesity risk, and thus provides valuable information for
policy makers to design effective strategies to reduce
inequality in obesity risk. This issue recently gained
importance not only in the US and Europe (especially in
Central, Southern and Eastern Europe; [38]) but also
throughout the world.
The paper is organised as follows. In the next section,
we review previous literature on socioeconomic related
inequality in obesity. The subsequent two sections describe
data and our methodological approach. Then, we explain
the results followed by discussions. The final section con-
cludes the paper.
Previous literature
In the literature, a very limited number of studies address
inequalities in obesity across different socioeconomic
groups. Using the C index, Zhang and Wang [39] and
Nikolaou and Nikolaou [40] found a strong inverse asso-
ciation between SES and obesity among women in the US
and in 10 European countries: Austria, Belgium, Denmark,
Finland, Greece, Ireland, Italy, Portugal, Spain, and Swe-
den. Although numerous studies documented a higher
prevalence of obesity among minority groups than in
whites in the US, Zhang and Wang [39] found a lower
socioeconomic-related inequality in obesity within minor-
ity populations. Zhang and Wang [41, 42] also investigated
the trend in socioeconomic-related inequality in obesity
between 1971 and 2002 in the US. They found that the
relationship between SES and obesity weakened in both
male and female groups over time. However, the authors
did not apply a correction factor suggested by Wagstaff
[43] or Erreygers [44] to their binary nature of the obesity
outcome variable while calculating the C index over time.
The application of any of these two correction factors does
not change the sign of the C index, but it changes the
absolute value of the C index. Hence, the corrected
C should be employed in order to make precise conclusions
204 M. Hajizadeh et al.
123
about change in socioeconomic-related inequality in
obesity risk over time.
Costa-Font and Gil [45] found evidence of significant
effects of socioeconomic inequalities on the probability of
being obese in Spain. Decomposition of such inequalities
indicates that education attainment and demographic fac-
tors have important influences on observed income-related
inequalities in obesity risk. Similar to Zhang and Wang
[41, 42], this study did not correct the C index either.
Ljungvall and Gerdtham [46] measure and decompose
socioeconomic-related obesity inequality in Sweden using
three waves of longitudinal data from 1980 to 1997. This
study provides evidence on trend in income-related
inequality in adult obesity using the corrected C index
suggested by Wagstaff [43]. The authors find that
inequalities in obesity favor the rich (i.e. obesity concen-
trated among the poor), but there is a decrease in inequality
over time. This study suggests that policies aimed at
reducing income inequality may be the most effective
strategy to address inequality in obesity risk.
Finally, Madden [47] analyzed socioeconomic gradient
of obesity in Ireland using nationally representative data-
sets for 2002 and 2007. He found higher socioeconomic
inequality in obesity prevalence in women than in men.
Decomposition of the corrected C by this study shows that
income and education are the two main factors explaining
the observed inequality in obesity in Ireland.
To sum up, the existing studies that measured socio-
economic inequality in obesity illustrate that obesity is
mainly concentrated among the poor. Nevertheless, the
degree of income-related inequality in obesity has declined
over time. There is also substantial difference in the mag-
nitude of inequality between males and females; obesity is
more highly concentrated among the socioeconomically
disadvantaged women than their male counterparts.
Data and variables
In order to quantify socioeconomic-related inequality in
obesity risk among Canadian adults, we used five confi-
dential master files of the Canadian Community Health
Surveys (CCHSs) conducted by Statistics Canada during
2000/2001–2009/2010. Each CCHS is a large nationally
representative survey of more than 130,000 individuals
aged 12 or older living in all provinces and territories in
Canada, except those living on Crown lands and Indian
reserves, on Canadian Forces bases, in institutions (prisons,
hospitals, universities), and in some remote areas.1
The CCHSs contain information on self-reported weight
and height for all individuals 18 years and older. We
measured self-reported BMI (as weight in kilograms divi-
ded by height in meters squared), based on self-reported
weight and height data. The self-reported BMI is likely to
be biased because individuals have a tendency to over
report their heights and under report their weights. Hence,
the correction factors proposed by Gorber et al. [48] were
applied to the self-reported BMI as Males:
BMIcorrected ¼ �1:08þ 1:08� BMIself�reported; Females:
BMIcorrected ¼ �0:12þ 1:05� BMIself�reported.
In accordance with the definition of the WHO, we
defined obesity as BMI C30 kg/m2. We exclude the
regions of Nunavut, Yukon, and Northwest Territories
from the analysis because these territories are sparsely
populated and culturally distinct compared to the rest of
Canada. Also, we restricted our analysis to those aged
between 18 and 65 years old and excluded pregnant
women to avoid misclassification of our outcome measure
(i.e. obesity).
We used CCHS 2000/2001 and CCHS 2009/2010 to
determine factors explaining income-related inequalities in
obesity risk over time. Based on the current literature, a
variety of demographic (age, gender, and marital status),
socioeconomic and behavioral factors were used as deter-
minants in the decomposition analysis of income-related
inequality in obesity. We included household characteris-
tics (i.e. income and family composition), dwelling own-
ership (as a proxy for wealth), education level, occupation
status, and immigration status in order to capture potential
socio-economic factors influencing obesity. In our analysis,
we used income measured at the household level because it
has been demonstrated to be a better measure of social
status than individual level income [49, 50]; moreover,
individual level income will be endogenous. Household
income has been equivalized to take into account house-
hold size. Similar to recent OECD publications (e.g., [51]),
we employed the square root scale which divides house-
hold income by the square root of household size to
equivalize household annual income. The influence of
behavioral factors on obesity were controlled by using
fruits and vegetables consumption, smoking habits, alcohol
consumption, and physical activity, as some studies (e.g.,
[52–54]) suggest that these variables are linked to obesity.
We controlled the basic characteristics of the residence
areas by using urban/rural variable. Also, the provincial
differences were captured by including a set of provincial
dummy variables, with Ontario as the reference category.
Table 1 presents the definition of all variables used in the
analysis.
Table 2 reports weighted descriptive statistics of all
variables used in the study. The increase in the proportion
1 We utilized restricted micro-level datasets and employed sampling
weights in order to obtain estimates that are representative of the
general Canadian population.
Socioeconomic inequalities in adult obesity risk in Canada 205
123
Table 1 The description of the variables
Variables Description
Outcome variable
Obesity 1 = if BMI C30, 0 otherwise
Demographic variables
Sex and age (years)
Male 18–34 1 = if male aged 18–34, 0 otherwise
Male 35–49 1 = if male aged 35–49, 0 otherwise
Male 50–65 1 = if male aged 50–65, 0 otherwise
Female 18–34 1 = if female aged 18–34, 0 otherwise
Female 35–49 1 = if female aged 35–49, 0 otherwise
Female 50–65 1 = if female aged 50–65, 0 otherwise
Marital status
Married 1 = if married or de facto married, 0 otherwise
Divorced or widowed 1 = if divorced or widowed, 0 otherwise
Single 1 = if single, 0 otherwise
Socioeconomic variables
Equivalized household income Household income divided by the square root of household size
Household arrangements
Single 1 = if household composition is single, 0 otherwise
Couple without a child 1 = if household composition is couple without a child, 0 otherwise
With child(ren) aged less than 6 1 = if household has child(ren) aged less than 6, 0 otherwise
With child(ren) aged between 6 and 12 1 = if household has child(ren) aged between 6 and 12, 0 otherwise
With children aged under 6 and 12 1 = if household has children aged under 6 and 12, 0 otherwise
Other 1 = if household composition is other, 0 otherwise
Home ownership 1 = if individual lives in a house owned by a household member, 0 otherwise
Education level
Less than secondary school 1 = if individual has less than secondary education, 0 otherwise
Secondary school 1 = if individual has secondary education, 0 otherwise
Some post-secondary 1 = if individual has some post-secondary education, 0 otherwise
Post-secondary degree/diploma 1 = if individual has post-secondary degree/diploma, 0 otherwise
Occupation status
White-collara 1 = if individual is employed in white-collar occupations, 0 otherwise
Blue-collarb 1 = if individual is employed in blue-collar occupations, 0 otherwise
Sales and services 1 = if individual is employed in sales and services occupations, 0 otherwise
Other occupations 1 = if individual is employed in other types of occupation, 0 otherwise
Unable to work 1 = if individual is unable to work, 0 otherwise
Student 1 = if individual is a full time student, 0 otherwise
Unemployed 1 = if individual is unemployed, 0 otherwise
Immigration status
Canadian birthplace 1 = if individual is Canadian born 0 otherwise
[10 years 1 = if individual is migrated to Canada more than 10 years ago, 0 otherwise
B10 years 1 = if individual is migrated to Canada within last 10 years, 0 otherwise
Behavioral variables
Fruit and vegetables consumption Total daily servings of fruit and vegetables
Physical activity
Active 1 = if individual’s leisure-time energy expenditure is C3 kcal/kg per day, 0 otherwise
Moderately active 1 = if individual’s leisure-time energy expenditure is between
1.5 and 2.9 kcal/kg per day, 0 otherwise
Inactive 1 = if individual’s leisure-time energy expenditure is below 1.5 kcal/kg per day, 0 otherwise
206 M. Hajizadeh et al.
123
of obese individuals in Canada is quite clear: the preva-
lence of obesity increased from 21 % in 2000/2001 to
25 % in 2009/2010. As expected, the data reflect Canada’s
aging population: the proportion of older male and female
populations (i.e. aged 50–65 years) increased between
2000/2001 and 2009/2010. The overall marital status dis-
tribution indicates a 3 % increase in the proportion of
singles over time.
With regard to socioeconomic variables, it is evident
that equivalized household annual income increased
markedly from $34,985 CAD in 2000/2001 to $49,538
CAD in 2009/2010 (unadjusted for inflation). Household
characteristics showed that couples without a child were
the most common household arrangements. Single-member
households accounted for 23 % of total households in
Canada in 2000/01, and this figure increased to 28 % in
2009/2010. Over the last 10 years, the proportion of home
ownership increased from 0.71 to 0.76.
According to Table 2, the proportion of individuals with
less than secondary school education declined during the
past decade, while the proportion of individuals with a
post-secondary degree increased significantly. Occupa-
tional status of employed individuals demonstrates that
white- and blue-collar categories account for the majority
of occupations. The proportion of immigrants, including
recent (those in Canada less than 10 years) and long-term
(those in Canada 10 years or more) immigrants increased
from 0.13 in 2000/2001 to 0.15 in 2009/2010.
With regard to behavioral variables, the average number
of daily servings of fruit and vegetables stayed constant at
approximately five servings. Over the course of 10 years,
the proportions of different levels of leisure time physical
activity indicate that Canadian adults are becoming more
physically active. The proportion of individuals who con-
sumed alcohol regularly increased from 0.25 in 2000/2001
to 0.31 in 2009/2010, whereas it declined for the three
Table 1 continued
Variables Description
Drinking habit
Abstainer 1 = if individual is never drinkers, 0 otherwise
Rarely 1 = if individual is rarely drinkers (\once/month and once/month), 0 otherwise
Occasional 1 = if individual is occasional drinkers (2–3 times/month, once/week), 0 otherwise
Regular 1 = if individual is regular drinkers (2–3 times/week, 4–6 times/week
and every day), 0 otherwise
Smoking habit
Never 1 = if individual is never smoked, 0 otherwise
Former 1 = if individual is former smokers, 0 otherwise
Occasional 1 = if individual is occasional smokers, 0 otherwise
Daily 1 = if individual is daily smokers, 0 otherwise
Ecologic variables
Geographical area
Rural 1 = if individual resides in rural area, 0 otherwise
Urban 1 = if individual resides in urban area, 0 otherwise
Province
Newfoundland and Labrador (NL) 1 = if individual resides in Newfoundland, 0 otherwise
New Brunswick (NB) 1 = if individual resides in New Brunswick, 0 otherwise
Nova Scotia (NS) 1 = if individual resides in Nova Scotia, 0 otherwise
Prince Edward Island (PE) 1 = if individual resides in Prince Edward, 0 otherwise
Saskatchewan (SK) 1 = if individual resides in Saskatchewan; 0, otherwise
Manitoba (MB) 1 = if individual resides in Manitoba, 0 otherwise
Alberta (AB) 1 = if individual resides in Alberta, 0 otherwise
Ontario (ON) 1 = if individual resides in Ontario, 0 otherwise
Quebec (QC) 1 = if individual resides in Quebec, 0 otherwise
British Columbia (BC) 1 = if individual resides in British Columbia, 0 otherwise
a This category includes occupations related to management, business, finance, administration, natural and applied sciences, health, social
sciences, education, culture and recreation, religion, and art b This category includes occupations related to trades, transport and equipment operator, and occupations unique to primary industry, processing,
manufacturing, and utilities
Socioeconomic inequalities in adult obesity risk in Canada 207
123
T a
b le
2 D
es cr
ip ti
v e
st at
is ti
cs o
f v
ar ia
b le
s (f
u ll
sa m
p le
) M ea
n
2 0
0 0
/2 0
0 1
2 0
0 3
/2 0
0 4
2 0
0 5
/2 0
0 6
2 0
0 7
/2 0
0 8
2 0
0 9
/2 0
1 0
O u
tc o
m e
v ar
ia b
le
O b
es it
y 0
.2 1
0 8
0 .2
1 7
4 0
.2 2
6 6
0 .2
3 6
5 0
.2 4
7 3
D em
o g
ra p
h ic
v ar
ia b
le s
S ex
an d
ag e
(y ea
rs )
M al
e 1
8 –
3 4
0 .1
6 9
0 0
.1 6
5 8
0 .1
6 9
9 0
.1 5
3 3
0 .1
5 7
3
M al
e 3
5 –
4 9
0 .2
1 2
6 0
.1 8
8 0
0 .1
8 2
7 0
.1 7
4 4
0 .1
6 1
7
M al
e 5
0 –
6 5
0 .1
5 6
7 0
.1 8
1 1
0 .1
7 9
3 0
.1 9
7 0
0 .2
0 7
0
F em
al e
1 8
– 3
4 0
.1 4
3 5
0 .1
4 2
7 0
.1 4
6 9
0 .1
3 3
6 0
.1 3
7 6
F em
al e
3 5
– 4
9 0
.1 7
7 5
0 .1
5 0
3 0
.1 4
8 5
0 .1
4 7
9 0
.1 3
5 7
F em
al e
5 0
– 6
5 0
.1 4
0 7
0 .1
7 4
8 0
.1 7
2 5
0 .1
9 3
8 0
.2 0
0 7
M ar
it al
S ta
tu s
M ar
ri ed
0 .6
1 4
6 0
.5 9
8 9
0 .5
8 2
8 0
.5 9
8 2
0 .5
8 5
1
D iv
o rc
ed o
r w
id o
w ed
0 .1
3 9
4 0
.1 4
2 6
0 .1
3 9
6 0
.1 4
1 1
0 .1
3 9
3
S in
g le
0 .2
4 5
9 0
.2 5
8 5
0 .2
7 7
6 0
.2 6
0 8
0 .2
7 5
6
S o
ci o
ec o
n o
m ic
v ar
ia b
le s
E q
u iv
al iz
ed h
o u
se h
o ld
an n
u al
in co
m e
($ )
3 4
,9 8
5 (2
6 ,2
5 1
) 3
9 ,5
7 6
(2 9
,4 2
6 )
4 1
,8 5
3 (3
0 ,2
7 0
) 4
7 ,1
5 2
(3 4
,7 0
2 )
4 9
,5 3
8 (3
5 ,7
9 9
)
H o
u se
h o
ld ar
ra n
g em
en ts
S in
g le
0 .2
2 9
7 0
.2 5
6 7
0 .2
7 3
8 0
.2 7
0 8
0 .2
7 4
8
C o
u p
le w
it h
o u
t a
ch il
d 0
.2 5
1 2
0 .2
9 6
1 0
.3 0
1 6
0 .3
1 4
2 0
.3 2
6 9
W it
h ch
il d
(r en
)a g
ed le
ss th
an 6
0 .0
9 5
9 0
.0 9
5 0
0 .0
9 1
5 0
.0 9
2 3
0 .0
9 1
6
W it
h ch
il d
(r en
) ag
ed b
et w
ee n
6 an
d 1
2 0
.1 2
2 3
0 .0
9 4
3 0
.0 9
0 5
0 .0
8 7
6 0
.0 7
7 0
W it
h ch
il d
re n
ag ed
u n
d er
6 an
d 1
2 0
.0 6
2 7
0 .0
6 0
0 0
.0 5
4 0
0 .0
5 4
0 0
.0 5
2 8
O th
er 0
.2 3
8 2
0 .1
9 8
0 0
.1 8
8 7
0 .1
8 1
1 0
.1 7
6 8
H o
m e
o w
n er
sh ip
0 .7
1 4
3 0
.7 3
7 4
0 .7
3 7
6 0
.7 5
8 6
0 .7
6 1
5
E d
u ca
ti o
n le
v el
L es
s th
an se
co n
d ar
y sc
h o
o l
0 .2
0 7
2 0
.1 8
0 3
0 .1
5 2
8 0
.1 3
8 5
0 .1
2 6
4
S ec
o n
d ar
y sc
h o
o l
0 .2
0 8
0 0
.2 0
2 9
0 .1
7 2
0 0
.1 7
6 9
0 .1
8 0
2
S o
m e
p o
st -s
ec o
n d
ar y
0 .0
8 8
9 0
.0 8
1 1
0 .0
9 0
8 0
.0 8
4 4
0 .0
8 5
5
P o
st -s
ec o
n d
ar y
d eg
re e/
d ip
lo m
a 0
.4 9
5 8
0 .5
3 5
7 0
.5 8
4 4
0 .6
0 0
2 0
.6 0
7 9
O cc
u p
at io
n st
at u
s
W h
it e-
co ll
ar 0
.3 5
0 0
0 .3
5 5
8 –
0 .3
6 6
1 0
.3 6
0 1
B lu
e- co
ll ar
0 .2
2 3
0 0
.2 2
2 9
– 0
.1 9
2 6
0 .1
8 0
5
S al
es an
d se
rv ic
es 0
.1 5
0 2
0 .1
6 0
3 –
0 .1
3 6
7 0
.1 3
7 0
208 M. Hajizadeh et al.
123
T a
b le
2 co
n ti
n u
ed
M ea
n
2 0
0 0
/2 0
0 1
2 0
0 3
/2 0
0 4
2 0
0 5
/2 0
0 6
2 0
0 7
/2 0
0 8
2 0
0 9
/2 0
1 0
O th
er o
cc u
p at
io n
s 0
.0 4
5 0
0 .0
3 2
2 –
0 .0
2 5
6 0
.0 2
3 1
U n
ab le
to w
o rk
0 .0
2 8
0 0
.0 2
3 6
– 0
.0 3
3 7
0 .0
3 3
7
S tu
d en
t 0
.0 5
6 5
0 .0
5 7
1 –
0 .0
5 1
5 0
.0 5
8 7
U n
em p
lo y
ed 0
.1 4
7 2
0 .1
4 8
0 –
0 .1
9 3
8 0
.2 0
6 9
Im m
ig ra
ti o
n st
at u
s
C an
ad ia
n b
ir th
p la
ce
[ 1
0 y
ea rs
0 .0
8 9
5 0
.0 9
0 1
0 .0
9 1
1 0
.1 1
0 9
0 .1
1 1
2
B 1
0 y
ea rs
0 .0
3 0
7 0
.0 3
0 2
0 .0
3 4
4 0
.0 3
4 0
0 .0
3 6
8
B eh
av io
ra l
v ar
ia b
le s
F ru
it an
d v
eg et
ab le
s co
n su
m p
ti o
n 4
.8 4
9 9
(4 .8
0 6
2 )
4 .6
6 6
2 (2
.5 5
9 5
) –
4 .7
9 5
9 (2
.6 7
2 6
) 4
.7 6
1 8
(2 .6
3 0
1 )
P h
y si
ca l
ac ti
v it
y
A ct
iv e
0 .2
2 4
8 0
.2 5
8 3
0 .2
5 4
6 0
.2 4
9 0
0 .2
6 7
4
M o
d er
at el
y ac
ti v
e 0
.2 4
5 8
0 .2
5 6
2 0
.2 5
7 6
0 .2
5 4
7 0
.2 5
7 1
In ac
ti v
e 0
.5 2
9 4
0 .4
8 5
5 0
.4 8
7 8
0 .4
9 6
3 0
.4 7
5 5
D ri
n k
in g
h ab
it
A b
st ai
n er
0 .1
6 1
7 0
.1 6
1 8
0 .1
5 7
3 0
.1 5
7 3
0 .1
5 9
5
R ar
el y
0 .2
9 4
4 0
.2 7
0 4
0 .2
6 6
0 0
.2 5
5 2
0 .2
5 5
0
O cc
as io
n al
0 .2
9 7
1 0
.2 8
7 2
0 .2
8 3
6 0
.2 6
9 4
0 .2
7 4
3
R eg
u la
r 0
.2 4
6 7
0 .2
8 0
6 0
.2 9
3 1
0 .3
1 8
0 0
.3 1
1 2
S m
o k
in g
h ab
it
N ev
er 0
.2 7
1 8
0 .2
8 5
2 0
.3 0
3 6
0 .3
1 5
2 0
.3 2
8 6
F o
rm er
0 .4
0 2
3 0
.4 2
8 8
0 .4
1 6
2 0
.4 1
6 0
0 .4
1 2
9
O cc
as io
n al
0 .0
4 7
2 0
.0 5
5 5
0 .0
5 6
5 0
.0 4
9 7
0 .0
5 3
4
D ai
ly 0
.2 7
8 7
0 .2
3 0
5 0
.2 2
3 7
0 .2
1 9
0 0
.2 0
5 1
G eo
g ra
p h
ic v
ar ia
b le
s
U rb
an /r
u ra
l
U rb
an 0
.7 3
3 6
0 .7
3 2
8 0
.7 4
4 3
0 .7
2 9
9 0
.7 2
7 8
R u
ra l
0 .2
6 6
4 0
.2 6
7 2
0 .2
5 5
7 0
.2 7
0 1
0 .2
7 2
2
P ro
v in
ce s
N L
0 .0
3 1
8 0
.0 3
1 6
0 .0
3 3
3 0
.0 3
3 4
0 .0
3 3
0
N B
0 .0
3 9
8 0
.0 3
6 8
0 .0
3 9
6 0
.0 4
3 4
0 .0
4 1
4
N S
0 .0
4 1
1 0
.0 3
6 5
0 .0
3 9
4 0
.0 3
9 8
0 .0
3 8
4
P E
0 .0
2 8
4 0
.0 1
4 9
0 .0
1 5
2 0
.0 1
8 2
0 .0
1 4
6
Socioeconomic inequalities in adult obesity risk in Canada 209
123
other categories of drinking habits (abstainer, rarely, and
occasional) during the same period. As expected, the
smoking habits of Canadians showed a declining trend: the
proportion of daily smokers decreased from 0.28 in 2000/
2001 to 0.21 in 2009/2010.
Methods
To date, several indices have been proposed by health
economists to quantify inequalities in health. The most
distinctive of these indices are the index of dissimilarity,
the range, the relative index of inequality, the Gini coef-
ficient, and the concentration (C) index [40]. As argued by
Wagstaff et al. [8], only the relative index of inequality and
the C index are the most complete indices which accurately
capture socioeconomic inequality in health. A socioeco-
nomic inequality index needs to satisfy three minimum
requirements (Wagstaff et al. [8]): (1) the index should
reflect the health inequalities that originate from the
socioeconomic factors; (2) it should be representative of
the entire population; and (3) it should be ‘‘sensitive to
changes in the distribution of the population across
socioeconomic groups.’’ The two most commonly used
indices in health inequality literature are the Gini coeffi-
cient and the C [40]. In this study, we employed the C
approach to measure the degree of socioeconomic-related
inequality in obesity risk because, as Wagstaff et al. [8]
argued, the Gini coefficient does not satisfy the first
requirement. Our approach involves two steps: calculation
of the C and decomposition of the C.
Calculation of the C
The C index approach requires the plotting of the enu-
merated population of individuals, to be ranked in
ascending order of SES against the cumulative percentage
of the health or health-related outcome variable of interest.
The C index can be computed using the following ‘‘con-
venient regression’’ approach suggested by Kakwani et al.
[69]:
2r2 r
yi
l
� � ¼ aþ bri þ ei; ð1Þ
where r2 r is the variance of fractional rank, yi is the out-
come variable of interest (i.e. obesity) for individual i, l is
the mean outcome variable for the entire population, ri ¼ i=N is the fractional rank of individual ii in the distribution,
with i = 1 for the lowest and i = N for the highest SES.
The ordinary least squares (OLS) estimate of b represents
the C [12]. Note that the standard error of b provides an
estimate of the standard error of C, which is inaccurate.
This is because the nature of the fractional rank variableT a
b le
2 co
n ti
n u
ed
M ea
n
2 0
0 0
/2 0
0 1
2 0
0 3
/2 0
0 4
2 0
0 5
/2 0
0 6
2 0
0 7
/2 0
0 8
2 0
0 9
/2 0
1 0
S K
0 .0
5 9
6 0
.0 5
3 5
0 .0
5 6
7 0
.0 5
9 2
0 .0
6 0
2
M B
0 .0
6 4
4 0
.0 5
6 6
0 .0
5 5
4 0
.0 5
8 6
0 .0
5 6
4
A B
0 .1
1 7
8 0
.1 0
8 9
0 .0
9 4
9 0
.0 9
8 7
0 .1
0 2
3
Q C
0 .1
7 3
9 0
.2 2
0 4
0 .2
2 8
1 0
.1 8
8 5
0 .1
8 9
4
B C
0 .1
4 0
4 0
.1 1
9 6
0 .1
1 6
6 0
.1 1
9 4
0 .1
2 1
1
O N
0 .3
0 2
8 0
.3 2
1 3
0 .3
2 0
9 0
.3 4
0 8
0 .3
4 3
2
N u
m b
er o
f o
b se
rv at
io n
s 9
0 ,4
7 3
8 9
,1 7
4 8
8 ,9
7 4
8 5
,4 3
9 7
9 ,0
4 6
S D
in p
ar en
th es
es .
In fo
rm at
io n
o n
d ai
ly se
rv in
g s
o f
fr u
it an
d v
eg et
ab le
s w
as n
o t
co ll
ec te
d in
P ri
n ce
E d
w ar
d Is
la n
d ,
O n
ta ri
o ,
A lb
er ta
an d
B ri
ti sh
C o
lo m
b ia
p ro
v in
ce s
in th
e C
C H
S 2
0 0
5 .
A ls
o ,
th e
o cc
u p
at io
n al
st at
u s
o f
ad u
lt s
is n
o t
av ai
la b
le in
th is
cy cl
e
210 M. Hajizadeh et al.
123
induces a certain pattern of autocorrelation in the data. A
solution for this problem is to use the Newey-West esti-
mator proposed in the literature [55], which corrects for
autocorrelation and heteroskedasticity [56]. The index is
positive if better outcome is concentrated on the higher
SES, and negative if it is concentrated on the lower SES.
The range of the index is from -1 to ?1, with a value of
zero indicating ‘‘perfect equality’’ [56].2
Wagstaff [43] illustrated that, when the outcome variable
of interest is binary, the minimum and maximum of the
concentration index are not -1 and ?1, and it depends on
the mean of the variable (l). For large samples, the lower
and the upper bounds are equal to l-1 and 1-l, respec-
tively. So, as the mean increases, the feasible interval of the
index shrinks. Therefore, one should be cautious in using
the C index to compare socioeconomic-related inequality
across time, geographic regions, and specific socioeco-
nomic groups when the outcome variable of interest is
binary [12]. One response to this problem, as suggested by
Wagstaff [43], is to normalize the concentration index by
multiplying the calculated index by 1=1� l. As the out-
come variable in our study is binary, we employed the
corrected C suggested by Wagstaff [43] to quantify the
socioeconomic-related inequalities in obesity risk.3
We measured income-related inequalities in obesity risk
by gender, age groups, urbanization level, and provinces.
The stratified analysis by region is due to the fact that some
of the empirical studies (e.g., [57, 58]) indicated significant
provincial and rural–urban differences in the prevalence of
obesity in Canada. A separate analysis by gender was also
undertaken because previous studies have established
substantial gender differences in the association between
socioeconomic factors such as income and obesity preva-
lence in Canada (e.g., [18, 59–62]). In order to measure
socioeconomic inequality in obesity risk at different life
stages, we calculated the C among three age groups: 18–34
(young adult), 35–49 (early middle-age) and 50–64 (late
middle-age).
Decomposition of the C
In this step, we used a decomposition technique to estimate
and break down inequalities, and thus quantify and com-
pare the income-related effect to that of other factors such
as education, age, gender, different health behaviors, and
socioeconomic status. Wagstaff et al. [63] demonstrated
that the C index of outcome variable y can be decomposed
into the contribution of the k factors (x) which determine
obesity risk. They showed that, if we have a linear
regression model linking our outcome variable of interest,
y; to a set of k determinants, xk such as:
y ¼ X
k
bkxk þ e: ð2Þ
The C index for y, can be written as:
C ¼ X
k
bkxk
l
� � Ck þ
GCe
l ð3Þ
In Eq. (3), l is the mean of y, xk is the mean of xk, CK is
the concentration index for xk, and GCe is the generalised
concentration index for the error term defined as
GCe ¼ 2 n
Pn i¼1
eiri, where ri is the fractional rank of the ith
person in the relevant distribution [63]. According to Eq.
(3), C is equal to a weighted sum of the concentration
indices of the regressors (i.e. total sum of contributions
made by all determinants), in which the weight for each xk
is the elasticity of y with respect to xk, gk ¼ bKxk=l. The
error term in Eq. (3) reflects the income-related inequality
in obesity risk that is not explained by systematic
differences in income [12]. Normalizing the C index by
multiplying it by 1=1� l yields :
Cnormalized ¼ C
1� l ¼ P
k bkxk
l
� � Ck
1� l þ
GCe l
1� l ð4Þ
Although it is preferable to estimate binary models using
a probit or logit regression technique, we followed the
standard practice in the existing literature [46, 47] and
employed a linear probability model (LPM) because the
decomposition works with a linear model [47]. The
estimated parameters of LPM are generally consistent.
We performed the decomposition analysis for the total
population as well as for males and females separately.
Results
In this section, first we illustrate the distribution of obesity
across different demographic groups and geographic areas
in Canada between 2000/01 and 2009/10, then we report
income-related inequalities in obesity risk in Canada, and
finally, the results of decomposition analysis are presented.
Prevalence and trends in obesity
The descriptive statistics showed that the prevalence of
obesity increased from 21 % in 2000/2001 to 25 % in
2009/10. The data also showed variation by gender, age,
2 For more explanation of the C index, see Wagstaff et al. [8],
Kakwani et al. [69], Lambert [70], and O’Donnell et al. [12]. 3 Erreygers [44] also proposed a corrected version of the concentra-
tion index as: 4l b�a
C, where, b and a are upper and lower limits of the
health variable, respectively. We obtained similar results using the
Erreygers’ correction in the calculation of Cs.
Socioeconomic inequalities in adult obesity risk in Canada 211
123
region, and province in the prevalence of obesity. As
illustrated in Fig. 1a, the prevalence of obesity increased
for both male and female population in Canada over the
study period. The prevalence of obesity is higher in males
compared to their female counterparts in all age groups.
According to Fig. 1b, men aged 50–64 have the highest
prevalence of obesity.
As can be seen from Fig. 2a, there are differences in the
obesity rates between rural and urban areas of Canada.
While 25 % of individuals in rural areas of Canada were
categorized as obese in 2000/2001, this figure was 20 % in
urban areas. The obesity rates among men and women
increased by 5 percentage points from 2000/2001 to
20009/2010. The variation is even greater between prov-
inces in Canada. For example, while 38 % of the adult
population of Newfoundland and Labrador (NL) and New
Brunswick (NB) were obese in 2009/10, the obesity rates in
Quebec (QC) and British Columbia (BC) were 23 %.
Trends in socioeconomic-related inequities in obesity
risk
Table 3 reports the corrected C indices as a measure of
income-related inequalities in obesity risk by sex, age
group, and province in Canada.4 The results showed that
obesity is concentrated among the poor in Canada. The
inequality, however, decreased over the study period. The
stratified analysis of inequality for men and women dem-
onstrates that obesity is concentrated among better-off
men, whereas the findings for women suggest that obesity
is concentrated among the poor. Further, the results
revealed that the concentration of adult obesity among the
rich increased between 200/01 and 2009/10 in all male age
groups, especially in the 35–49 years age group. Income-
related inequality in obesity risk favors the rich in all
female groups. The inequality is weakened for young
women aged 18–34 years, whereas it widened for adults
aged 35–49.
The Cs suggested variations in income-related distri-
bution of obesity across rural/urban and between different
provinces. While obesity is concentrated among the poor in
rural areas, the positive association found between income
and obesity in urban areas weakened over time. The
(a)By gender
(b)By age groups
Fig. 1 Trends in the prevalence (%) of obesity among Canadian
adults (aged 18–64), by gender and age groups: 2000/2001–2009/
2010
(a) By rural/urban
(b) By provinces
Fig. 2 Trends in the prevalence (%) of obesity among Canadian
adults (aged 18–64), by rural/urban and provinces: 2000/2001–2009/
2010
4 Application of the correction factor to the calculation of the C index
increased the absolute value of the C, but it did not change the overall
trends in socioeconomic inequalities in obesity risk that we found
without applying the correction factor.
212 M. Hajizadeh et al.
123
calculated Cs demonstrated that obesity is concentrated
mainly among the poor in the Atlantic Provinces (i.e.
Newfoundland and Labrador, New Brunswick, Nova Sco-
tia, Prince Edward Island) and the degree of income-related
inequality in obesity risk increased over time. In contrast,
obesity was concentrated among the poor in Manitoba,
Alberta, Ontario, and Quebec in 2000/01, but the observed
inequality in obesity risk in these provinces declined over
time. The obesity has become more concentrated among
the rich in Alberta in 2010.
Decomposition of socioeconomic-related inequalities
in obesity risk
As discussed above, the calculated Cs indicate that socio-
economic inequality in obesity exists among adults in
Canada. Therefore, we decompose income-related
inequalities in order to determine the contributions of
observable variables to the inequalities in obesity risk. The
decomposition analysis enables us to identify how much a
certain observed factor contributes to the inequality. A
factor that is distributed unequally by income and has an
effect on the probability of obesity can contribute to
inequality in obesity risk. Table 4 presents the results of the
decomposition analysis of income-related inequalities in
obesity risk for Canada in 2000/01 and 2009/10. The
table reports coefficient results ðbKÞ,5 the elasticities
gk ¼ bKxk=lð Þ, the concentration index of the explanatory
variables ðCKÞ, and the contribution of explanatory vari-
ables to the C index.
Based on the regression coefficients for demographic
variables reported in the table, older male age groups have
higher probability of being obese compared to the male 18–
34 group. The probability of being obese is lower among
all female age groups as compared to the male 18–34
Table 3 Income-related inequalities in obesity by sex, age group, and province in Canada: 2000/2001–2009/2010
Corrected concentration index (C)
2000/01 2003/04 2005/06 2007/08 2009/10
Canada
-0.0368*** (0.0049) -0.0424*** (0.0050) -0.0349*** (0.0049) -0.0287*** (0.0049) -0.0145*** (0.0051)
Sex and age (years)
Male 0.0151* (0.0068) 0.0075 (0.0069) 0.0230*** (0.0067) 0.0339*** (0.0069) 0.0492*** (0.0072)
Male 18–34 0.0198 (0.0134) 0.0232* (0.0135) 0.0285** (0.0129) 0.0209 (0.0141) 0.0307** (0.0149)
Male 35–49 0.0040 (0.0106) 0.0065 (0.0113) 0.0101 (0.0111) 0.0141 (0.0116) 0.0901*** (0.0124)
Male 50–64 0.0212 (0.0082) 0.0181 (0.0085) 0.0256 (0.0082) 0.0313* (0.0088) 0.0712*** (0.0093)
Female -0.1079*** (0.0069) -0.1176*** (0.0071) -0.1228*** (0.0070) -0.1163*** (0.0070) -0.1060*** (0.0073)
Female 18–34 -0.1269*** (0.0139) -0.1180*** (0.0142) -0.1431*** (0.0140) -0.1255*** (0.0147) -0.0909*** (0.0154)
Female 35–49 -0.1175*** (0.0109) -0.1257*** (0.0124) -0.1430*** (0.0123) -0.1434*** (0.0122) -0.1646*** (0.0132)
Female 50–64 -0.1070*** (0.0084) -0.1111*** (0.0091) -0.1358*** (0.0090) -0.1165*** (0.0092) -0.1147*** (0.0098)
Rural/urban
Urban -0.0344*** (0.0058) -0.0409*** (0.0058) -0.0245*** (0.0057) -0.0222*** (0.0058) -0.0068*** (0.0061)
Rural -0.0329*** (0.0092) -0.0309*** (0.0093) -0.0649*** (0.0093) -0.0431*** (0.0091) -0.0285*** (0.0093)
Province
NL -0.0449*** (0.0247) -0.0850*** (0.0257) -0.0793*** (0.0241) -0.0950*** (0.0248) -0.0735*** (0.0251)
NB -0.0707*** (0.0225) -0.0592** (0.0237) -0.0551** (0.0218) -0.1237*** (0.0216) -0.0852*** (0.0222)
NS -0.0626*** (0.0222) -0.0126 (0.0242) -0.0962*** (0.0224) -0.0800*** (0.0223) -0.0790*** (0.0236)
PE -0.0414 (0.0275) -0.0782** (0.0376) 0.0245 (0.0362) -0.1116*** (0.0339) -0.1159*** (0.0415)
SK -0.0083 (0.0190) -0.0472** (0.0204) -0.0016 (0.0194) 0.0120 (0.0189) -0.0281 (0.0204)
MB -0.0382** (0.0181) -0.0592*** (0.0205) -0.0152 (0.0204) -0.0173 (0.0200) -0.0139 (0.0205)
AB -0.0162 (0.0143) -0.0299** (0.0150) -0.0183 (0.0160) 0.0265* (0.0155) 0.0541*** (0.0158)
QC -0.0501*** (0.0120) -0.0856*** (0.0108) -0.0203* (0.0104) -0.0609*** (0.0114) -0.0368*** (0.0119)
BC 0.0148 (0.0138) -0.0067 (0.0156) 0.0030 (0.0152) 0.0095 (0.0158) 0.0236 (0.0164)
ON -0.0416*** (0.0089) -0.0247*** (0.0086) -0.0420*** (0.0086) -0.0269*** (0.0084) -0.0184** (0.0088)
SE in parentheses
*** p \ 0.01, ** p \ 0.05, * p \ 0.1
5 Similar to other studies in the literature (e.g., [45, 46, 71]), based on
large population-based data, the overall explanatory power of our
model was low, with R-squared values of 0.0501 (2000/2001) and
0.0616 (2009/2010). This indicates that individual level factors
generally explain very small variation in obesity prevalence.
Socioeconomic inequalities in adult obesity risk in Canada 213
123
group, with the exception of the older female age group
(age 50–65).6 Also, married people are more likely to be
obese than their single counterpart.
While equivalized household income negatively asso-
ciated with the probability of obesity in 2000/2001, the
estimated coefficient for this variable was not statistically
significant in 2009/10. The findings suggest an inverse
relationship between home ownership and obesity in 2000/
2001. Also, higher levels of education are significantly
protective against obesity. The coefficients on occupation
variables suggest that adults working in white collar
occupations are comparatively less obese than all other
occupation categories with the exception of students.
As expected, immigration status is negatively associated
with obesity—the coefficient of both short-term
(B10 years) and long-term ([10 years) immigrants are
statistically significant. According to the results for 2000/
2001 and 2009/2010, all immigrants have lower chances of
obesity (4–5 %) than the native born.
Although the coefficient on the consumption of fruits
and vegetables is not statistically significant in the year
2000/2001, it was negatively associated with obesity in
2009/2010. Relative to physically active adults, moderately
active and inactive adults have higher probability of being
obese by 5 and 20 %, respectively. Based on the coeffi-
cients on drinking habit, it is evident that regular drinkers
are 9 % less likely to be obese than abstainers. Compared
to never smokers, the probability of being obese is higher
among former smokers; however, daily smokers tend to
have lower likelihood of being obese.
The coefficients on geographic variables showed that
obesity is significantly higher among rural adults than
urban adults. Whereas individuals residing in the provinces
of Quebec and British Columbia have a significantly lower
likelihood of obesity as compared to Ontario, the proba-
bility of being obese is generally higher in the rest of the
provinces.
Table 4 also reports the concentration indices of all
explanatory variables, CK . The positive (negative) sign of
the CK for a certain variable demonstrates that the factor
concentrated among rich (poor) individuals. For example,
as reported in the table, higher educational attainment (i.e.
holding post-secondary degree/diploma), home ownership
and fruits and vegetables consumption are concentrated
among high-income earners, whereas being a daily smoker
and engaging in no physical activity are more concentrated
among the poor.
Based on the elasticities and concentration index of each
explanatory variable, we calculated the contribution of
each factor to the C index for obesity, gkCk=1� l. The
term ‘‘contribution’’ in Table 4 shows how much the var-
iation in a given explanatory factor among different income
groups (i.e. Ck=1� lÞ) can explain the relationship
between obesity and income in Canada via its partial
association with obesity (i.e. gkÞ [64]. The negative con-
tribution of a certain factor to the C reveals that the
socioeconomic distribution of the factor and the association
of the relevant factor and obesity leads to an increase in the
probability of being obese among the poor and vice versa.
According to Table 4, while income makes a negative
contribution to the inequality in obesity risk in Canada in
2000/2001, it is not a significant contributor to the
inequality in 2009/2010. Factors such as occupation status,
drinking habits and educational attainment are other
determinants that contribute negatively to the observed
inequality in obesity risk in Canada. Education makes a
negative contribution to the inequality because this factor
has a negative elasticity (indicating that an increase in
educational attainment decreases the probability of being
obese) and higher education attainments are mainly con-
centrated among the better offs (see the positive signs of
the CK for post-secondary degree/diploma in Table 4). In
contrast, factors such as demographic characteristics,
immigration, and smoking factors contribute positively to
the inequality in obesity risk. The positive contribution of
the factor ‘‘smoking habit’’ is due to the fact that smoking
has a negative elasticity (i.e. a decrease in smoking
increases the probability of being obese) and smoking
prevalence is higher in low-income groups than in high-
income groups.
Turning our attention to the geographic factors, it is
apparent that Alberta and Quebec contribute positively to
the inequality in obesity risk in Canada, whereas New-
foundland and Labrador, New Brunswick, Nova Scotia,
and Manitoba make negative contributions. The positive
contribution of Alberta to the inequality is due to the fact
that, compared to Ontario the probability of being obese is
higher in this province (therefore, it has a positive elas-
ticity) and the people of this province are generally richer
compared to other provinces (see CK for this province).
Thus, the product of these two effects leads to the positive
contribution of Alberta to the overall C index.
Figure 3 shows the results of the decomposition analysis
for the total population as well as for men and women
separately. As can be seen, similar to the results obtained
for the total population, factors such as drinking habits,
physical activity, occupation status, and education are the
main factors that contribute to the concentration of obesity
among the poor in male and female populations. On the
other hand, demographic variables, immigration status, and
6 We also performed the decomposition analysis using separate age
(continuous variable) and gender (male dummy) variables in the
regression. As expected, the results showed that age and being male
are both associated with higher probability of being obese in Canada.
Also, the total contribution of age and gender factors to income-
related inequality in obesity risk was identical in both models.
214 M. Hajizadeh et al.
123
T a
b le
4 D
ec o
m p
o si
ti o
n o
f in
co m
e- re
la te
d in
eq u
al it
ie s
fo r
o b
es it
y in
th e
C an
ad ia
n ad
u lt
p o
p u
la ti
o n
: 2
0 0
0 /2
0 0
1 an
d 2
0 0
9 /2
0 1
0
2 0
0 0
/2 0
0 1
2 0
0 9
/2 0
1 0
C o
ef fi
ci en
ts E
la st
ic it
ie s
C k
C k = 1 �
l C
o n
tr ib
u ti
o n
C o
ef fi
ci en
ts E
la st
ic it
ie s
C k
C k =
1 �
l C
o n
tr ib
u ti
o n
A b
so lu
te
v al
u e
S u
m m
ed A
b so
lu te
v al
u e
S u
m m
ed
D em
o g
ra p
h ic
v ar
ia b
le s
A g
e an
d se
x (y
ea rs
)
M al
e 1
8 –
3 4
(R ef
.)
M al
e 3
5 –
4 9
0 .0
4 6
0 *
* *
0 .0
4 0
4 0
.0 7
3 6
0 .0
9 3
7 0
.0 0
3 8
* *
* 0
.0 8
9 4
* *
* 0
.0 5
4 4
0 .0
6 8
2 0
.0 9
0 4
0 .0
0 4
9 *
* *
M al
e 5
0 –
6 5
0 .0
5 1
0 *
* *
0 .0
3 2
1 0
.0 9
8 0
0 .1
2 4
7 0
.0 0
4 0
* *
* 0
.0 7
7 3
* *
* 0
.0 5
9 2
0 .0
9 2
1 0
.1 2
2 1
0 .0
0 7
2 *
* *
F em
al e
1 8
– 3
4 -
0 .0
5 4
9 *
* *
- 0
.0 4
2 9
- 0
.1 1
8 2
- 0
.1 5
0 4
0 .0
0 6
4 *
* *
- 0
.0 5
4 5
* *
* -
0 .0
3 3
8 -
0 .1
3 2
4 -
0 .1
7 5
5 0
.0 0
5 9
* *
*
F em
al e
3 5
– 4
9 -
0 .0
1 5
1 *
* -
0 .0
1 4
5 -
0 .0
0 9
3 -
0 .0
1 1
8 0
.0 0
0 2
- 0
.0 1
1 9
- 0
.0 0
7 5
- 0
.0 1
8 8
- 0
.0 2
5 0
0 .0
0 0
2
F em
al e
5 0
– 6
5 0
.0 1
9 6
* *
0 .0
1 3
9 -
0 .0
4 8
7 -
0 .0
6 1
9 -
0 .0
0 0
9 *
* 0
.0 3
0 1
* *
* 0
.0 2
6 6
- 0
.0 0
6 6
- 0
.0 0
8 7
- 0
.0 0
0 2
* *
M ar
it al
st at
u s
M ar
ri ed
0 .0
5 4
6 *
* *
0 .1
5 3
2 0
.0 6
0 5
0 .0
7 6
9 0
.0 1
1 8
* *
* 0
.0 3
5 6
* *
* 0
.0 8
6 3
0 .0
7 3
0 0
.0 9
6 8
0 .0
0 8
4 *
D iv
o rc
ed o
r w
id o
w ed
0 .0
1 3
8 *
* *
0 .0
0 9
8 -
0 .2
0 6
8 -
0 .2
6 3
0 -
0 .0
0 2
6 0
.0 2
2 8
0 .0
0 4
1 0
.0 0
2 4
- 0
.2 1
1 4
- 0
.2 8
0 3
- 0
.0 0
0 7
0 .0
2 5
7
S in
g le
(R ef
.)
S o
ci o
ec o
n o
m ic
v ar
ia b
le s
E q
u iv
al iz
ed h
o u
se h
o ld
an n
u al
in co
m e
0 .0
0 0
1 *
* -
0 .0
3 2
3 0
.3 5
1 2
0 .4
4 6
7 -
0 .0
1 4
4 -
0 .0
1 4
4 0
.0 0
0 1
0 .0
0 2
2 0
.3 5
0 4
0 .4
6 4
8 0
.0 0
1 0
0 .0
0 1
0
H o
u se
h o
ld ar
ra n
g em
en ts
S in
g le
(R ef
.)
C o
u p
le w
it h
o u
t a
ch il
d -
0 .0
1 5
8 *
- 0
.0 1
7 9
0 .1
5 8
3 0
.2 0
1 3
- 0
.0 0
3 6
- 0
.0 0
6 8
- 0
.0 0
9 1
0 .1
6 3
2 0
.2 1
6 4
- 0
.0 0
2 0
W it
h ch
il d
(r en
)a g
ed le
ss th
an 6
- 0
.0 2
2 7
* *
- 0
.0 1
0 5
- 0
.0 9
7 4
- 0
.1 2
3 9
0 .0
0 1
3 0
.0 0
0 8
* *
0 .0
0 0
3 -
0 .0
9 6
7 -
0 .1
2 8
2 0
.0 0
0 0
W it
h ch
il d
(r en
) ag
ed b
et w
ee n
6 an
d 1
2
- 0
.0 2
2 3
* *
- 0
.0 1
3 0
- 0
.0 9
1 1
- 0
.1 1
5 9
0 .0
0 1
5 -
0 .0
0 4
9 -
0 .0
0 1
6 -
0 .1
1 0
0 -
0 .1
4 5
9 0
.0 0
0 2
W it
h ch
il d
re n
ag ed
u n
d er
6 an
d 1
2
- 0
.0 2
7 2
* *
- 0
.0 0
8 1
- 0
.1 8
1 5
- 0
.2 3
0 9
0 .0
0 1
9 0
.0 0
7 0
0 .0
0 1
6 -
0 .1
6 1
1 -
0 .2
1 3
6 -
0 .0
0 0
4
O th
er 0
.0 1
3 1
0 .0
1 3
8 0
.0 4
6 5
0 .0
5 9
1 0
.0 0
0 8
0 .0
0 1
9 0
.0 1
0 3
0 .0
0 6
6 0
.0 4
1 5
0 .0
5 5
0 0
.0 0
0 4
- 0
.0 0
1 8
H o
m e
o w
n er
sh ip
- 0
.0 2
6 1
* *
* -
0 .0
8 4
9 0
.1 1
9 5
0 .1
5 2
0 -
0 .0
1 2
9 -
0 .0
1 2
9 -
0 .0
1 2
1 -
0 .0
3 7
0 0
.1 2
4 7
0 .1
6 5
4 -
0 .0
0 6
1 *
* -
0 .0
0 6
1
E d
u ca
ti o
n le
v el
L es
s th
an se
co n
d ar
y sc
h o
o l
(R ef
.)
S ec
o n
d ar
y sc
h o
o l
- 0
.0 2
8 4
* *
* -
0 .0
2 7
1 -
0 .0
5 1
1 -
0 .0
6 4
9 0
.0 0
1 8
* *
* -
0 .0
2 2
7 *
- 0
.0 1
5 8
- 0
.1 0
2 8
- 0
.1 3
6 4
0 .0
0 2
2 *
S o
m e
p o
st -s
ec o
n d
ar y
- 0
.0 3
9 4
* *
* -
0 .0
1 6
5 -
0 .0
5 4
2 -
0 .0
6 8
9 0
.0 0
1 1
* *
* -
0 .0
1 2
8 -
0 .0
0 4
3 -
0 .1
2 7
0 -
0 .1
6 8
4 0
.0 0
0 7
P o
st -s
ec o
n d
ar y
d eg
re e/
d ip
lo m
a -
0 .0
3 9
8 *
* *
- 0
.0 9
4 0
0 .1
2 9
4 0
.1 6
4 6
- 0
.0 1
5 5
* *
* -
0 .0
1 2
6 -
0 .0
4 1
7 *
* *
- 0
.1 0
6 8
0 .0
9 8
1 0
.1 3
0 0
- 0
.0 1
3 9
* *
* -
0 .0
1 1
0
Socioeconomic inequalities in adult obesity risk in Canada 215
123
T a
b le
4 co
n ti
n u
ed
2 0
0 0
/2 0
0 1
2 0
0 9
/2 0
1 0
C o
ef fi
ci en
ts E
la st
ic it
ie s
C k
C k = 1 �
l C
o n
tr ib
u ti
o n
C o
ef fi
ci en
ts E
la st
ic it
ie s
C k
C k =
1 �
l C
o n
tr ib
u ti
o n
A b
so lu
te
v al
u e
S u
m m
ed A
b so
lu te
v al
u e
S u
m m
ed
O cc
u p
at io
n st
at u
s
W h
it e-
co ll
ar (R
ef .)
B lu
e- co
ll ar
0 .0
2 0
8 *
* *
0 .0
1 9
6 -
0 .0
5 9
4 -
0 .0
7 5
5 -
0 .0
0 1
5 *
* 0
.0 2
2 6
* *
0 .0
1 5
5 -
0 .0
1 4
2 -
0 .0
1 8
8 -
0 .0
0 0
3
S al
es an
d se
rv ic
es 0
.0 1
0 5
* 0
.0 0
7 7
- 0
.1 1
4 1
- 0
.1 4
5 1
- 0
.0 0
1 1
0 .0
2 2
9 *
* 0
.0 1
3 0
- 0
.1 1
9 6
- 0
.1 5
8 6
- 0
.0 0
2 1
O th
er o
cc u
p at
io n
s 0
.0 3
1 *
* *
* 0
.0 0
6 6
- 0
.1 2
1 7
- 0
.1 5
4 8
- 0
.0 0
1 0
- 0
.0 9
6 5
* *
- 0
.0 0
1 1
0 .1
0 2
8 0
.1 3
6 3
- 0
.0 0
0 2
U n
ab le
to w
o rk
0 .0
8 8
8 *
* *
0 .0
1 1
1 -
0 .5
6 6
0 -
0 .7
1 9
9 -
0 .0
0 8
0 *
* *
0 .0
6 1
6 *
* *
0 .0
0 8
2 -
0 .5
6 3
4 -
0 .7
4 7
1 -
0 .0
0 6
2 *
* *
S tu
d en
t -
0 .0
4 9
6 *
* *
- 0
.0 1
3 1
- 0
.1 8
2 7
- 0
.2 3
2 3
0 .0
0 3
0 *
* -
0 .0
6 4
7 *
* *
- 0
.0 1
4 8
- 0
.2 1
2 5
- 0
.2 8
1 7
0 .0
0 4
2 *
* *
U n
em p
lo y
ed 0
.0 3
5 9
* *
* 0
.0 2
4 1
- 0
.3 3
5 7
- 0
.4 2
7 1
- 0
.0 1
0 3
* *
* -
0 .0
1 8
8 0
.0 2
7 4
* *
* 0
.0 2
3 1
- 0
.2 8
1 1
- 0
.3 7
2 8
- 0
.0 0
8 6
* *
- 0
.0 1
3 1
Im m
ig ra
ti o
n st
at u
s
C an
ad ia
n b
ir th
p la
ce (R
ef .)
[ 1
0 y
ea rs
- 0
.0 9
9 4
* *
* -
0 .0
4 2
0 0
.0 0
7 0
0 .0
0 8
9 -
0 .0
0 0
4 *
* *
- 0
.1 3
1 5
* *
* -
0 .0
5 1
4 -
0 .0
4 7
6 -
0 .0
6 3
2 0
.0 0
3 2
* *
*
B 1
0 y
ea rs
- 0
.3 0
9 4
* *
* -
0 .0
4 5
0 -
0 .3
3 2
2 -
0 .4
2 2
5 0
.0 1
9 0
* *
* 0
.0 1
8 7
- 0
.2 7
6 1
* *
* -
0 .0
4 3
4 -
0 .3
2 6
2 -
0 .4
3 2
6 0
.0 1
8 8
* *
* 0
.0 2
2 0
B eh
av io
ra l
v ar
ia b
le s
F ru
it an
d v
eg et
ab le
s
co n
su m
p ti
o n
- 0
.0 0
0 2
- 0
.0 0
4 3
0 .0
1 2
7 0
.0 1
6 2
- 0
.0 0
0 1
* -
0 .0
0 0
1 -
0 .0
0 2
4 *
* -
0 .0
4 7
0 0
.0 2
4 2
0 .0
3 2
0 -
0 .0
0 1
5 *
* *
- 0
.0 0
1 5
P h
y si
ca l
ac ti
v it
y
A ct
iv e
(R ef
.)
M o
d er
at el
y ac
ti v
e 0
.0 4
5 5
* *
* 0
.0 5
2 7
0 .0
7 0
3 0
.0 8
9 4
0 .0
0 4
7 *
* *
0 .0
4 9
3 *
* *
0 .0
5 2
8 0
.0 5
6 3
0 .0
7 4
6 0
.0 0
3 9
* *
*
In ac
ti v
e 0
.0 8
4 0
* *
* 0
.2 0
5 6
- 0
.0 6
3 7
- 0
.0 8
1 1
- 0
.0 1
6 7
* *
* -
0 .0
1 2
0 0
.0 9
7 4
* *
* 0
.1 8
4 7
- 0
.0 9
0 2
- 0
.1 1
9 6
- 0
.0 2
2 1
* *
* -
0 .0
1 8
2
D ri
n k
in g
h ab
it
A b
st ai
n er
(R ef
.)
R ar
el y
0 .0
0 4
5 0
.0 0
6 2
- 0
.1 1
0 7
- 0
.1 4
0 9
- 0
.0 0
0 9
0 .0
4 1
8 *
* *
0 .0
4 3
0 -
0 .1
3 0
5 -
0 .1
7 3
1 -
0 .0
0 7
4 *
* *
O cc
as io
n al
- 0
.0 4
0 9
* *
* -
0 .0
5 7
0 0
.0 6
4 0
0 .0
8 1
5 -
0 .0
0 4
6 *
* *
0 .0
0 0
0 0
.0 0
0 0
0 .0
4 6
8 0
.0 6
2 0
0 .0
0 0
1 *
* *
R eg
u la
r -
0 .0
7 7
2 *
* *
- 0
.0 8
8 7
0 .1
9 5
2 0
.2 4
8 3
- 0
.0 2
2 0
* *
* -
0 .0
2 7
6 -
0 .0
6 8
7 *
* *
- 0
.0 8
9 4
0 .1
7 5
5 0
.2 3
2 8
- 0
.0 2
0 8
* *
* -
0 .0
2 8
2
S m
o k
in g
h ab
it
N ev
er (R
ef .)
F o
rm er
0 .0
2 6
6 *
* *
0 .0
4 9
1 0
.0 9
0 6
0 .1
1 5
2 0
.0 0
5 7
* *
* 0
.0 2
6 4
* *
* 0
.0 4
4 3
0 .0
9 1
6 0
.1 2
1 5
0 .0
0 5
4 *
* *
O cc
as io
n al
- 0
.0 0
1 3
- 0
.0 0
0 3
0 .0
0 1
6 0
.0 0
2 0
0 .0
0 0
1 -0
.0 1
8 6
- 0
.0 0
4 0
- 0
.0 0
8 3
- 0
.0 1
1 0
0 .0
0 0
0
D ai
ly -
0 .0
4 4
6 *
* *
- 0
.0 5
8 2
- 0
.1 3
5 3
- 0
.1 7
2 0
0 .0
1 0
0 *
* *
0 .0
1 5
7 -
0 .0
3 7
5 *
* *
- 0
.0 3
1 7
- 0
.1 9
0 7
- 0
.2 5
2 9
0 .0
0 8
0 *
* *
0 .0
1 3
4
G eo
g ra
p h
ic v
ar ia
b le
s
U rb
an /r
u ra
l
216 M. Hajizadeh et al.
123
T a
b le
4 co
n ti
n u
ed
2 0
0 0
/2 0
0 1
2 0
0 9
/2 0
1 0
C o
ef fi
ci en
ts E
la st
ic it
ie s
C k
C k = 1 �
l C
o n
tr ib
u ti
o n
C o
ef fi
ci en
ts E
la st
ic it
ie s
C k
C k =
1 �
l C
o n
tr ib
u ti
o n
A b
so lu
te
v al
u e
S u
m m
ed A
b so
lu te
v al
u e
S u
m m
ed
U rb
an -
0 .0
1 2
4 *
* -
0 .0
4 2
9 0
.0 1
4 3
0 .0
1 8
1 -
0 .0
0 0
8 -
0 .0
2 9
9 *
* *
- 0
.0 8
9 5
0 .0
1 0
0 0
.0 1
3 3
- 0
.0 0
1 2
* *
*
R u
ra l
(R ef
.)
P ro
v in
ce s
N L
0 .0
1 5
8 *
* 0
.0 0
2 4
- 0
.2 5
2 4
- 0
.3 2
1 1
- 0
.0 0
0 8
0 .0
3 5
2 *
* *
0 .0
0 4
8 -
0 .1
2 5
6 -
0 .1
6 6
6 -
0 .0
0 0
8 *
* *
N B
0 .0
1 0
5 *
0 .0
0 1
9 -
0 .1
8 1
8 -
0 .2
3 1
3 -
0 .0
0 0
4 0
.0 4
9 5
* *
* 0
.0 0
8 4
- 0
.1 2
7 1
- 0
.1 6
8 5
- 0
.0 0
1 4
* *
N S
0 .0
1 6
2 *
* 0
.0 0
3 2
- 0
.1 4
8 2
- 0
.1 8
8 5
- 0
.0 0
0 6
0 .0
3 0
1 *
* *
0 .0
0 4
8 -
0 .0
9 5
4 -
0 .1
2 6
5 -
0 .0
0 0
6
P E
0 .0
0 1
6 0
.0 0
0 2
- 0
.1 9
4 5
- 0
.2 4
7 5
- 0
.0 0
0 1
- 0
.0 0
9 7
* -
0 .0
0 0
6 -
0 .1
2 4
9 -
0 .1
6 5
6 0
.0 0
0 1
S K
0 .0
1 6
1 *
* *
0 .0
0 4
2 -
0 .0
7 8
8 -
0 .1
0 0
2 -
0 .0
0 0
4 0
.0 0
8 3
0 .0
0 1
9 0
.0 8
4 1
0 .1
1 1
6 0
.0 0
0 2
*
M B
0 .0
0 9
7 *
0 .0
0 2
8 -
0 .0
6 0
1 -
0 .0
7 6
5 -
0 .0
0 0
2 0
.0 2
9 5
* *
* 0
.0 0
6 5
- 0
.0 0
5 9
- 0
.0 0
7 8
- 0
.0 0
0 1
* *
A B
0 .0
0 8
4 0
.0 0
4 4
0 .0
5 1
3 0
.0 6
5 2
0 .0
0 0
3 0
.0 0
9 9
0 .0
0 3
9 0
.1 7
9 3
0 .2
3 7
8 0
.0 0
0 9
Q C
- 0
.0 7
5 8
* *
* -
0 .0
6 6
7 -
0 .0
9 3
5 -
0 .1
1 9
0 0
.0 0
7 9
* *
* -
0 .0
3 8
7 *
* *
- 0
.0 3
2 2
- 0
.0 7
8 1
- 0
.1 0
3 6
0 .0
0 3
3 *
* *
B C
- 0
.0 2
4 3
* *
* -
0 .0
1 5
2 0
.0 2
0 4
0 .0
2 6
0 -
0 .0
0 0
4 *
* *
0 .0
0 4
5 -
0 .0
5 4
5 *
* *
- 0
.0 2
5 9
- 0
.0 1
7 2
- 0
.0 2
2 8
0 .0
0 0
6 *
* *
0 .0
0 1
1
O N
(R ef
.)
S u
m -
0 .0
3 4
8 -
0 .0
1 6
7
R es
id u
al (T
o ta
l C
-S u
m )
- 0
.0 0
2 0
0 .0
0 2
2
T o
ta l
C fo
r o
b es
it y
- 0
.0 3
6 8
- 0
.0 1
4 5
* *
* p \
0 .0
1 ,
* *
p \
0 .0
5 ,
* p \
0 .1
; si
g n
ifi ca
n ce
le v
el s
o f
th e
co ef
fi ci
en ts
ar e
co m
p u
te d
u si
n g
ro b
u st
st an
d ar
d er
ro rs
, si
g n
ifi ca
n ce
le v
el s
o f
th e
co n
tr ib
u ti
o n
s w
er e
o b
ta in
ed th
ro u
g h
b o
o ts
tr ap
p in
g
Socioeconomic inequalities in adult obesity risk in Canada 217
123
smoking habits make a positive contribution to the
observed income-related inequality in obesity risk in both
genders. Income makes a significant negative contribution
to the C in females in 2001/2001 and 2009/2010. In con-
trast, income contributes positively to the C in males
especially in 2009/2010.
Discussion
In this study, we analyzed socioeconomic-related inequal-
ities in health among Canadian adults by investigating
income-related inequalities in obesity risk. We found that
income exhibited a consistent inverse relationship with
obesity risk among women. For men, our results indicated
that obesity is concentrated among the better off individ-
uals, especially in recent years. We also found that obesity
is concentrated among the poor throughout the study period
under investigation. Income-related inequalities in obesity
risk are also apparent between and within Canadian prov-
inces. The decomposition analysis of income-related
inequalities in obesity risk demonstrated that income had
the opposite effects on males and females. While income
increased the probability of being obese among the rich
males, it decreased this probability among the rich women.
One important limitation of our study is that some of the
health behavioral factors included in the decomposition
analysis are likely to be endogenous, and therefore inclu-
sion of these variables may introduce unknown bias into
the study. For example, endogeneity in the case of physical
activity can be due to the impossibility of distinguishing
causality. In other words, whether people who exercise
have a lower probability of being obese or people who are
not obese are more likely to participate in exercise are
impossible to distinguish [65]. In spite of this problem, we
believe that the inclusion of behavioral factors in the
decomposition analysis is important due to the demon-
strated association of these factors with obesity in the lit-
erature (e.g., [66]).
We performed the decomposition analysis by excluding
behavioral factors from the model as a robustness check for
the main results. The results without inclusion of behav-
ioral factors in the model did not alter the sign and mag-
nitude of the contributions of non-behavioral variables with
the exception of income. In this model, the negative con-
tribution of income to the inequality is actually more
pronounced. This is indeed what we would expect based on
the overall contribution of behavioral factors to income-
related inequalities in obesity risk. As physical inactivity
(regular drinking) is negatively (positively) associated with
income (the CK in Table 4) and this factor increases
(decreases) the probability of being obese (see the elas-
ticities reported in Table 4), exclusion of this variable from
the decomposition analysis strengthens the effect of
income on obesity and therefore the negative contribution
of income to the inequality. Although the effect of the
exclusion of smoking habits on the contribution of income
is opposite to those of physical activity and drinking habit
-0.22
Dem
Hou
Imm
Drin
Res
mogra
useho
migra
nking
idual
-0.15
aphic
old ar
ation
g hab
l
5
fact
rang
statu
bit
tors
gemen
us
nts
-0.11 -0.05
I
E
F
S
5
Incom
Educ
Fruit
Smok
me
cation
s and
king
n
d veg
habit
0
etabl
t
les coonsummptio
0.05
on
5
H
O
Ph
G
0.1
Home
Occup
hysic
Geogr
own
pation
cal ac
aphic
nershi
n stat
ctivity
c fact
ip
tus
y
tors
0.155
Tota
Tota
Mal
Mal
Fem
Fem
al 200
al 200
le 200
le 200
male 2
male 2
00/0
09/10
00/01
09/10
2000/
2009/
1
0
1
0
/01
/10
Fig. 3 Relative contribution of each factor to the inequality of obesity in Canada
218 M. Hajizadeh et al.
123
factors, the overall outcome of these effects led to an
increase in the negative contribution of income factor to
the inequality.7
Another limitation of this study is that the socio-eco-
nomic factors identified in this paper do not necessarily
have a causal interpretation. And it is beyond the scope of
the present paper to ascertain causal effect of socio-eco-
nomic factors on obesity risk. Our results can safely be
interpreted in terms of observed associations between
socio-economic factors and obesity risk.
Conclusions and policy implications
This study, for the first time, examined socioeconomic
inequalities in obesity risk among Canadian adults using
five nationally representative survey datasets from
2000/2001 to 2009/2010. These large biennial survey
datasets enabled us to measure and identify income-related
inequalities in obesity risk in Canada. A decomposition
technique was applied in order to break down the observed
income-related inequality into the main factors.
The results of this study showed that there exists
income-related inequality in obesity favoring the rich. The
observed income-related inequality, however, declined
over time. Similar results were found in the US [41, 42],
Sweden [46], and Ireland [47]. The stratified analysis of
income-related inequality across gender indicates that
obesity is typically concentrated among rich men. Among
women, obesity is concentrated mainly among the poor.
Similar results were also found in the US [39] and Ireland
[47]. Our results also showed an increasing trend in the
extent of income-related inequalities in obesity risk among
all male age groups. In contrast, the degree of inequality in
obesity among women stayed roughly constant over the
study period. Nevertheless, the results revealed differential
inequality trends across different female age-groups; whilst
the level of inequality is constant among women aged
50–64, the inequality weakened for young women aged
18–34 years and widened for adults aged 35–49.
The geographic analysis showed increasing trends in the
inequality in obesity risk favoring the rich in the Atlantic
Provinces. The results showed that obesity is concentrated
mainly among the rich in Alberta. While the calculated Cs
suggested a consistent inequality favoring the better off in
rural areas, the positive association found between income
and obesity risk in urban areas weakened over time.
The decomposition analysis of income-related inequal-
ity in obesity risk suggest that factors such as demographic,
income, immigration, education, drinking habits, and
physical activity are main factors that lie behind income-
related inequality in Canada. Moreover, our findings
showed that income was a main contributor to the observed
inequality in obesity risk among adult Canadians in
2000/2001, but it did not make a significant contribution in
2009/2010. Interestingly, we found that income has the
opposite effects on the inequality in males versus females.
While income increased the probability of being obese
among rich males, it decreased this probability among rich
women.
Based on our findings, not only are obesity rates
increasing among all age groups but there also exist
socioeconomic disparities in the distribution of obesity
across gender, age, and various geographic jurisdictions in
Canada. The results suggest that income inequality is an
important factor that increases the concentration of obesity
among poor women, while income seems to be a major
factor that explains the concentration of obesity among rich
males. Therefore, if decreasing obesity rates and inequality
in obesity risk are important goals, health policies should
target not only all populations but should specifically focus
on poor females and relatively better-off males.
In addition, greater attention should be devoted to the
Atlantic Provinces, because these provinces on the one
hand have a higher prevalence of obesity and, on the other
hand, the income-related disparity in obesity which favors
the rich (unlike the rest of Canada) is widening over time.
The Atlantic provinces should thus consider implementing
policies to reduce the incidence of adult obesity among the
poor. In particular, community interventions like the pro-
motion of physical activity and higher intake of fruit and
vegetable servings among the poor may help to reduce the
obesity epidemic (see the reported negative contribution of
these variables in Table 4).
The community level interventions can be implemented
along with public health policies which target physical,
environmental, and social determinants of obesity. This is
because these factors influence the effectiveness of the
interventions to promote a healthy weight. In other words,
individuals’ healthy choices are often bounded by factors
in the social, physical, and economic environments [17].
Indeed, a growing body of evidence suggests that some of
the local factors (e.g., neighborhood walkability, neigh-
borhood safety, and access to physical/recreational activi-
ties) can have a profound impact on obesity through their
effects on physical activity or healthy eating habits. For
instance, a number of studies document the link between
neighborhood safety and obesity through a physical activ-
ity mechanism. Fear of victimization (e.g., being robbed,
attacked, or injured) was found to decrease outdoor phys-
ical activity, especially walking [67]. Parkes and Kearns
[68] showed that walking for pleasure or exercise is asso-
ciated with safety, and lower neighborhood safety reduces
7 Detailed decomposition results with the exclusion of behavioral
variables are available upon request.
Socioeconomic inequalities in adult obesity risk in Canada 219
123
levels of walking. Therefore, a comprehensive and multi-
faceted strategy is required to address the higher concen-
tration of obesity among certain socioeconomic groups.
Acknowledgments We thank three anonymous reviewers of this
journal for their thoughtful comments and suggestion which have
substantially improved the manuscript. We would also like to thank
comments of participants of the 46th Annual Conference of the
Canadian Economics Association held at the University of Calgary,
June 7-10, 2012, where a preliminary version of this paper was pre-
sented. Funding for this research by the Canadian Institutes of Health
Research (CIHR) operating grant (reference number: MOP–97763) is
gratefully acknowledged. The views expressed in this paper are the
views of the authors and do not necessarily reflect the views of any
affiliated organization.
References
1. Marmot, M.: Social determinants of health inequalities. Lancet
365(9464), 1099–1104 (2005)
2. Hosseinpoor, A.R., Stewart Williams, J.A., Itani, L., Chatterji, S.:
Socioeconomic inequality in domains of health: results from the
world health surveys. BMC Public Health 12, 198 (2012)
3. Initiative, Canadian Population Health: Reducing gaps in health:
a focus on socio-economic status in urban Canada. Canadian
Population Health Initiative, Ottawa (2008)
4. Initiative, Canadian Population Health: Improving the health of
Canadians. Canadian Population Health Initiative, Ottawa (2004)
5. Raphael, D.: Poverty and Health: mechanisms and pathways. In
poverty and policy in Canada: implications for health and quality
of life, D. Raphael, Ed. Toronto: Canadian Scholars’ Press,
239–268 (2007)
6. Nettleton, S.: Surveillance, health promotion and the formation of
a risk identity. In Debates and dilemmas in promoting Health, M.
Sidell, L. Jones, J. Katz, and A. Peberdy, (eds) London, UK:
Open University Press, 314–324 (1997)
7. Marmot, M., Shipley, M., Brunner, E., Hemingway, H.: Relative
contribution of early life and adult socioeconomic factors to adult
morbidity in the Whitehall II study. J. Epidemiol. Commun.
Health 55(5), 301–307 (2001)
8. Wagstaff, A., Paci, P., Van Doorslaer, E.: On the measurement of
inequalities in health. Soc. Sci. Med. 33(5), 545–557 (1991)
9. Gwatkin, D.R.: Health inequalities and the health of the poor:
what do we know? What can we do? Bull. World Health Organ.
78(1), 3–18 (2000)
10. Kawachi, I., Subramanian, S.V., Almeida-Filho, N.: A glossary
for health inequalities. J. Epidemiol. Commun. Health 56(9),
647–652 (2002)
11. Marmot, M., Wilkinson, R.G.: The social determinants of health.
Oxford University Press, Oxford (2006)
12. O’Donnell, O., Van Doorslaer, E., Wagstaff, A., Lindelow, M.:
Analyzing health equity using household survey data—A guide to
techniques and their implementation. The World Bank, Geneva
(2008)
13. Costa-Font, J., Hernández-Quevedo, C.: Measuring inequalities
in health: what do we know? What do we need to know? Health
Policy 106(2), 195–206 (2012)
14. World Health Organization: The world health report 2002:
reducing risks, promoting healthy life. WHO, Geneva (2002)
15. WHO: Obesity and overweight. http://www.who.int/mediacentre/
factsheets/fs311/en/13. Accessed Jan 2012
16. Starky, S.: The obesity epidemic in Canada. Parliamentary
Information and Research Service, Library of Parliament Canada
(2005)
17. CIHI and PHAC: Obesity in Canada. A joint report from the
public health agency of Canada and the Canadian institute for
health, Information (2011)
18. McLaren, L.: Socioeconomic status and obesity. Epidemiol. Rev.
29(1), 29–48 (2007)
19. Rashad, I.: Assessing the underlying economic causes and con-
sequences of obesity. Gender Issues 21(3), 17–29 (2003)
20. Ross, N.A., Tremblay, S., Khan, S., Crouse, D., Tremblay, M.,
Berthelot, J.M.: Body mass index in urban Canada: neighborhood
and metropolitan area effects. Am. J. Public Health 97(3),
500–508 (2007)
21. Tjepkema, M.: Adult obesity. Health Rep. 17(3), 9–25 (2006)
22. Klarenbach, S., Padwal, R., Chuck, A., Jacobs, P.: Population-
based analysis of obesity and workforce participation. Obes.
(Silver Spring) 14(5), 920–927 (2006)
23. Janssen, I., Diener, A.: Economic Burden of obesity in Canada in
2005. Public Health Agency of Canada, Ottawa (2009)
24. Black, J.L., Macinko, J.: Neighborhoods and obesity. Nutr. Rew.
66(1), 2–20 (2008)
25. Huot, I., Paradis, G., Ledoux, M.: Factors associated with over-
weight and obesity in Quebec adults. Int. J. Obes. Relat. Metab.
Disord. 28(6), 766–774 (2004)
26. Yach, D., Stuckler, D., Brownell, K.D.: Epidemiologic and eco-
nomic consequences of the global epidemics of obesity and
diabetes. Nat. Med. 12(1), 62–66 (2006)
27. Hedley-Vickers, M., Cinda-Lee, C.: Developmental program-
ming of Obesity and Type 2 Diabetes. Fetal Matern. Med. Rev.
18, 1–23 (2007)
28. Wamala, S.P., Wolk, A., Orth-Gomér, K.: Determinants of
obesity in relation to socioeconomic status among middle-aged
Swedish women. Prev. Med. 26(5), 734–744 (1997)
29. Sobal, J., Stunkard, A.J.: Socioeconomic status and obesity: a
review of the literature. Psych. Bull. 105(2), 260–275 (1989)
30. Cairney, J., Wade, T.J.: Correlates of body weight in the 1994
national population health survey. Int. J. Obes. Relat. Metab.
Disord. 22(6), 584–591 (1998)
31. Ostbye, T., Pomerleau, J., Speechley, M., Pederson, L.L., Speech-
ley, K.N.: Correlates of body mass index in the 1990 Ontario Health
Survey. Can. Med. Assoc. J. 52(11), 1811–1817 (1995)
32. Shields, M., Tjepkema, M.: Trends in adult obesity. Health Rep.
17(3), 53–59 (2006)
33. Willms, J.D., Tremblay, M.S., Katzmarzyk, P.T.: Geographic and
demographic variation in the prevalence of overweight Canadian
children. Obesity 11(5), 668–673 (2003)
34. Cranfield, J.A.L.: Factors influencing the body mass index of
adults in Canada. University of Alberta, Department of Rural
Economy (2007)
35. Moore, S., Hall, J.N., Harper, S., Lynch, J.W.: Global and
national socioeconomic disparities in obesity, overweight, and
underweight status. J. Obes. (2010). doi:10.1155/2010/514674
36. Pouliou, T., Elliott, S.J.: Individual and socio-environmental
determinants of overweight and obesity in Urban Canada. Health
Place. 16(2), 389–398 (2010)
37. Canada, Statistics: CCHS 2007–2008: Data dictionary. Public
Use Mirco Data File, Statistics Canada (2009)
38. Berghofer, A., Pischon, T., Reinhold, T., Apovian, C., Sharma,
A., Willich, S.: Obesity prevalence from a European perspective:
a systematic review. BMC Public Health 8(1), 200 (2008)
39. Zhang, Q., Wang, Y.: Socioeconomic inequality of obesity in the
United States: do gender, age, and ethnicity matter? Soc. Sci.
Med. 58(6), 1171–1180 (2004)
40. Nikolaou, A., Nikolaou, D.: Income-related inequality in the
distribution of obesity among Europeans. J. Public Health 16(6),
403–411 (2008)
41. Zhang, Q., Wang, Y.: Secular trends in socioeconomic inequality
of obesity in the United States. In studies on Economic
220 M. Hajizadeh et al.
123
Well-Being: essays in the honor of John P. Formby, Emerald
Group, 481-499 (2004)
42. Zhang, Q., Wang, Y.: Using concentration index to study changes
in socio-economic inequality of overweight among US adoles-
cents between 1971 and 2002. Int. J. Epidemiol. 36(4), 916–925
(2007)
43. Wagstaff, A.: The bounds of the concentration index when the
variable of interest is binary, with an application to immunization
inequality. Health Econ. 14(4), 429–432 (2005)
44. Erreygers, G.: Correcting the concentration index. J. Health Econ.
28(2), 504–515 (2009)
45. Costa-Font, J., Gil, J.: What lies behind socio-economic
inequalities in obesity in Spain?. A decomposition approach.
Food Policy. 33(1), 61–73 (2008)
46. Ljungvall, A., Gerdtham, U.G.: More equal but heavier: a lon-
gitudinal analysis of income-related obesity inequalities in an
adult Swedish cohort. Soc. Sci. Med. 70(2), 221–231 (2010)
47. Madden, D.: The Socioeconomic Gradient of Obesity in Ireland.
University College Dublin, Geary Institute (2010)
48. Connor Gorber, S., Shields, M., Tremblay, M.S., McDowell, I.:
The feasibility of establishing correction factors to adjust self-
reported estimates of obesity. Health Rep. 19(3), 71–82 (2008)
49. Krieger, N., Chen, J.T., Selby, J.V.: Comparing individual-based
and household-based measures of social class to assess class
inequalities in women’s health: a methodological study of 684 US
women. J. Epidemiol. Commun. Health 53(10), 612–623 (1999)
50. MacIntyre, S., Hunt, K.: Socio-economic Position, Gender and
Health: how Do They Interact? J. Health Psychol. 2(3), 315–334
(1997)
51. OECD.: Growing Unequal?: Income Distribution and Poverty in
OECD Countries. The organisation for economic cooperation and
development publishing, 1-7 (2008)
52. Chou, S.Y., Grossman, M., Saffer, H.: An economic analysis of
adult obesity: results from the behavioral risk factor surveillance
system. J. Health Econ. 23(3), 565–587 (2004)
53. Frank, L.D., Andresen, M.A., Schmid, T.L.: Obesity relationships
with community design, physical activity, and time spent in cars.
Am. J. Prev. Med. 27(2), 87–96 (2004)
54. Ewing, R., Schmid, T., Killingsworth, R., Zlot, A., Raudenbush,
S.: Relationship between urban sprawl and physical activity,
obesity, and morbidity. in urban ecology, J. M. Marzluff et al.,
Eds. University of Maryland National Center for Smart Growth
Preinkert Field House College Park MD 20742 USA: Springer
US, 567–582 (2008)
55. Newey, W.K., West, K.D.: Automatic lag selection in covariance
matrix estimation. Rev. Econ. Stud. 61(4), 631–653 (1994)
56. World Bank: Quantitative techniques for health equity analysis:
the concentration index http://siteresources.worldbank.org/
EXTEDSTATS/Resources/3232763-1171296378756/concentra
tion.pdf. Accessed July 2012
57. Bruner, M.W., Lawson, J., Pickett, W., Boyce, W., Janssen, I.:
Rural Canadian adolescents are more likely to be obese compared
with urban adolescents. Int. Pediatr. Obes. 3(4), 205–211 (2008)
58. M. Shields, M., Tjepkema, M.: Regional differences in obesity.
Health Reports, Statistics Canada, 17(3), 61 (2006)
59. Ball, K., Mishra, G., Crawford, D.: Which aspects of socioeco-
nomic status are related to obesity among men and women? Int.
J. Obes. Relat. Metab. Disord. 26(4), 559–565 (2002)
60. Public health agency of Canada: Obesity in Canada (2012)
http://www.phac-aspc.gc.ca/hp-ps/hl-mvs/oic-oac/determ-eng.php.
Accessed May 2012
61. Kuhle, S., Veugelers, P.J.: Why does the social gradient in health
not apply to overweight. Health Rep. 19(4), 7–15 (2008)
62. McLaren, L., Godley, J.: Social class and BMI among Canadian
adults: a focus on occupational prestige. Obesity 17(2), 290–299
(2009)
63. Wagstaff, A., van Doorslaer, E., Watanabe, N.: On decomposing
the causes of health sector inequalities with an application to
malnutrition inequalities in Vietnam. J. Econ. 112(1), 207–223
(2003)
64. Van Doorslaer, E., Clarke, P., Savage, E., Hall, J.: Horizontal
inequities in Australia’s mixed public/private health care system.
Health Policy 86(1), 97–108 (2008) 65. Dutton, D.J., McLaren, L.: Explained and unexplained regional
variation in Canadian obesity prevalence. Obesity 19(7),
1460–1468 (2011)
66. Finkelstein, E.A., Ruhm, C.J., Kosa, K.M.: Economic causes and
consequences of obesity. Annu Rev Publ. Health 26(1), 239–257
(2005)
67. Ross, C.E.: Fear of Victimization and Health. J. Quant. Criminol.
9(2), 159–175 (1993)
68. Parkes, A., Kearns, A.: The multi-dimensional neighbourhood
and health: a cross-sectional analysis of the Scottish Household
Survey, 2001. Health place 12(1), 1–18 (2006)
69. Kakwani, N., Wagstaff, A., Van Doorslaer, E.: Socioeconomic
inequalities in health: measurement, computation, and statistical
inference. J. Econ. 77(1), 87–103 (1997)
70. Lambert, P. J.: The distribution and redistribution of income.
Manchester University Press, (2001)
71. Yen, S.T., Chen, Z., Eastwood, D.B.: Lifestyles, demographics,
dietary behavior, and obesity: a switching regression analysis.
Health Serv. Res. 44(4), 1345–1369 (2009)
Socioeconomic inequalities in adult obesity risk in Canada 221
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
- c.10198_2013_Article_469.pdf
- Socioeconomic inequalities in adult obesity risk in Canada: trends and decomposition analyses
- Abstract
- Introduction
- Previous literature
- Data and variables
- Methods
- Calculation of the C
- Decomposition of the C
- Results
- Prevalence and trends in obesity
- Trends in socioeconomic-related inequities in obesity risk
- Decomposition of socioeconomic-related inequalities in obesity risk
- Discussion
- Conclusions and policy implications
- Acknowledgments
- References
