51 / 74
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 28, No 5, September/October 2017
AFRICA
325
In Pikine, a suburb of Dakar, these positive perceptions of
stoutness have been observed among women.
25
However, no
study has been conducted from this perspective among urban
men, or among the rural population.
Therefore, the objectives of this study were (1) to assess and
compare the prevalence of obesity, general and central, in Dakar
and in Tessekere, a rural municipality in northern Senegal,
and to analyse trends in obesity in Dakar; (2) to determine
sociodemographic risk factors for obesity in both environments;
and (3) to compare ideal body size between urban and rural
areas.
Methods
The study was approved by the National Ethics Committee for
Health Research of Senegal (protocol SEN13/67, no 0272). The
research was conducted in accordance with the Declaration
of Helsinki, and written informed consent was obtained from
participants.
This study was conducted from February to August 2015 on a
sample of 1 000 individuals, aged 20 years and older in Dakar, and
on a sample of 500 adults of the same age bracket in the Tessekere
municipality. The samples were constructed using the combined
quota method (cross-section by age, gender and town of residence
in Dakar; only by age and gender in Tessekere municipality) in
order to strive for representativeness of the population aged 20
years and older living in the department of Dakar and in Tessekere
municipality. Data from the Agence Nationale de la Statistique et
de la Démographie dating from the last census (2013) were used.
The quota variables used were gender (male/female), age
(20–29, 30–39, 40–49, 50–59, and 60 years and over, with an
upper age limit of 100 years) and, for Dakar, town of residence.
In Dakar, the towns were grouped by the four arrondissements
making up the department: Plateau-Gorée (five towns), Grand
Dakar (six towns), Parcelles Assainies (four towns) and
Almadies (four towns). In practical terms, this method requires
constructing a sample that reflects the proportions observed
in each target population. For example, according to the last
census, men aged 20–29 years living in the town of Medina
(arrondissement of Plateau-Gorée) represented 1.9% of the
population aged 20 years and over living in the department of
Dakar. The sample was constructed to reflect this proportion
and it included 19 men aged 20–29 living in this town.
Inclusion criteria were individuals 20 years old or older, living
in the department of Dakar. Pregnant women were excluded
from the study.
Eight trained investigators (PhD students in Medicine,
Pharmacy and Sociology) started out each day from different
points in each town (Dakar) or camp (Tessekere) to interview
individuals in Wolof, Haalpulaar or French in every third home.
Investigators had a certain number of individuals to interview to
meet the quotas. Only one person was selected as a respondent
in each home. Investigators went to the house, inquired about
the inhabitants and then chose the first person they saw who met
the characteristics needed for the quotas. In-person interviews
were conducted. They ranged from 45 minutes to more than one
hour and 30 minutes, depending on respondent availability and
desire to talk.
Weight was measured using a digital scale (measurement
accuracy of 100 g), with subjects dressed in minimal clothing
and barefoot. To measure height, the subject was to stand ‘at
attention’, arms at the sides, heels together, without shoes.
FollowingWorldHealthOrganisation (WHO) recommendations,
BMI was calculated by dividing the weight (kg) by the square of
the height (m
2
). Underweight was defined as BMI
<
18.5 kg/m
2
;
overweight was defined as 25
≤
BMI
<
30 kg/m
2
; and obesity
corresponded to a BMI of
≥
30 kg/m
2
.
26
Waist circumference (WC) was measured at the narrowest
point of the abdomen at the end of a normal expiration. WC was
measured using a measuring tape with 1-mm accuracy. WC of
≥
102 cm in men and
≥
88 cm in women was considered central
obesity.
27
Waist-to-hip ratio (WHR) was also used as a criterion
of central obesity: a WHR of
≥
0.9 in men and
≥
0.8 in women
was considered central obesity.
28
Among the sociodemographic data collected during the
interviews, three variables were taken into account for this study:
age, gender and educational level. Four age groups were defined:
20–29, 30–39, 40–49 and 50 years and over. Gender was coded
as follows: 1 for women, 0 for men. In Dakar, five levels of
education were defined based on the Senegalese school system:
none, primary (one to five years of schooling), intermediate (six
to eight years), secondary (nine to 12 years), and university (13
years and over). In the Tessekere municipality, given the large
proportion of persons who have never attended school (76%),
the educational level was dichotomised: no schooling/one or
more years of schooling.
Satisfaction with body weight was assessed in one question,
with five possible responses: ‘Do you think you are: too thin,
a little too thin, average, a little too fat, too fat?’ To determine
ideal body size, we took the BMI at which the same percentage
of individuals believed they were too heavy as those who felt they
were too thin.
29
We also used the body size scale (BSS), developed and
validated by Cohen
et al
. in Senegal,
30
to assess ideal body
size (IBS) of women and men, to obtain a complementary
representation of body image assessed from the questionnaire.
This tool has two advantages: (1) it consists of a gender-specific
scale of nine models; and (2) it represents real black models
with their anthropometric characteristics to assess specific body
weight perceptions in African populations. One model represents
the underweight category, three models the normal-weight
category, two models the overweight category, and one model
each class of obesity level as defined by the WHO (30.0
<
BMI
≤
34.9 kg/m², 35.0
<
BMI
≤
39.9 kg/m², and
≥
40 kg/m²). BSS was
considered a numerical variable, as each human picture ranged
from 1 to 9 according to increasing BMI categories to measure
ideal body size.
Statistical analysis
To answer our research questions, we used the Student’s
t
-test,
ANOVA, chi-squared test and logistic regressions. Results
are expressed as mean ± standard deviation for continuous
variables or as percentages for categorical variables. Bivariate
comparisons were performed using the Student’s
t
-test, ANOVA
for continuous variables, and chi-squared tests for categorical
variables. Multivariate analyses were performed using binary
logistic regression and results are expressed as odds ratios
with 95% confidence intervals (CIs). The software used for the
statistical analysis was SPSS Statistics 22 for Windows.