Prevalence of obesity and associated risk factors in chinese pre school children aged 6 to 72 months old in singapore 2

CHAPTER 2

2. METHODOLOGY
2.1

Objectives

2.1.1 General objective
To examine the prevalence of overweight and obesity of Singapore Chinese
children aged 6 to 72 months using three different criteria; “CDC BMI for age”,
“IOTF BMI for age” and “Singapore BMI for age” references.

2.1.2 Specific objectives
(1) To compare the prevalence of overweight and obesity between boys and
girls
(2) To compare the variation of overweight and obesity by age groups
(3) To evaluate the associations between overweight/obesity with potential
risk factors such as birth weight, duration of preschool hours per day,
physical activities (playing outdoors, leisure activities), sedentary
activities (watching television, playing computer/television/hand held
video games, reading, drawing and coloring activities), socioeconomic
status (father’s education, mother’s education, total combined monthly
income), presence of park or garden near to home, breastfeeding, types of
breastfeeding, maternal smoking and alcohol consumption during
pregnancy.
34


2.1.3 Hypothesis for Risk Factors
From the literature review, we hypothesized that overweight, obesity or
combined overweight and obesity is potentially associated with the following risk

factors:
(i) birth weight,
(ii) physical activities (playing outdoors, leisure activities), sedentary
activities (watching television, playing computer/television/hand held
video games, reading, drawing and coloring activities),
(iii) socioeconomic status (father’s education, mother’s education, total
combined monthly income),
(iv) breastfeeding, types of breastfeeding, maternal smoking and alcohol
consumption during pregnancy.
(v) other factors: duration of preschool hours per day, presence of park or
garden near to home.

2.2 Study Design and Study Population
2.2.1 Study Design
It is a cross-sectional study which is part of “A Study on Strabimus,
Amblyopia and Refractive Error in Singapore Chinese Preschoolers (STARS)”.
35


STARS is a population based cross-sectional study to determine the prevalence of
eye diseases (mainly myopia, strabismus and refractive error) in 3,000 Chinese
children aged 6–72 months in Singapore. The STARS study was conducted from
February 2006 to November 2008.

2.2.2 Study area and Study population
The study areas were South-Western part of Singapore (Bukit Batok,
Clementi and Queenstown) and Western parts of Singapore (Jurong East and
Jurong West) of Housing Development Board (HDB) apartments. These parts of
Singapore were chosen because they are close to the Singapore Eye Research
Institute and Jurong Medical Centre where the subject assessment and

measurements took place. Chinese children aged between 6–72 months living in
this study area were our study subjects.
In Singapore, the total resident population is 3.7 million in which there are
410,107 children (10.8%) aged below 9 years.73 The total resident population in
our study areas (South-Western and Western parts of Singapore) is 690,216 and
children below 9 years old constitutes 10.9% (75,528).73 Children aged below 1
year constitute 0.9% [n=32,788; boys=16,693 (50.9%), girls=16,095 (49.1%)] in
3.7 million Singapore resident population, compared to our study in which
children aged 6 months to below 1 year is 0.03% [n=181; boys=85 (47.0%),
girls=96 (53.0%)]. Moreover, of which 3.7 million resident population, the
proportion of children aged 1 year is 1.0% [n=38,350; boys=19,549 (51.0%),
36


girls=18,801 (49.0%), aged 2 years is 1.1% [n=41,562; boys=21,049 (50.6%),
girls=20,513 (49.4%)], aged 3 years is 1.1% [n=40,869; boys=20,853 (51.0%),
girls=20,016 (49.0%)], aged 4 years is 1.1% [n=40,863; boys=20,746 (50.8%),
girls=20,117 (49.2%)] and aged 5 years is 1.1% [n=40,915; boys=20,755 (50.7%),
girls=20,160 (49.3%)].73 In our study, aged 1 year constitutes 0.08% [n=527;
boys=304 (57.7%), girls=223 (42.3%)], aged 2 years constitutes 0.07% [n=509;
boys=257 (50.5%), girls=252 (49.5%)], aged 3 years constitutes 0.08% [n=571;
boys=291 (51.0%), girls=280 (49.0%)], aged 4 years constitutes 0.09% [n=601;
boys=321 (53.4%), girls=280 (46.6%)] and aged 5 years constitutes 0.08%
[n=575; boys=290 (50.4%), girls=285 (49.6%)].
A total of 5,648 preschool children aged 6–72 months were recruited from
households in South-Western and Western areas of Singapore and 3,009 (72.2%)
children responded.

2.2.3 Sampling method
All households located in the South-Western (Bukit Batok, Clementi, and

Queenstown) and Western (Jurong East and Jurong West) part of Singapore with
Chinese children aged 6–72 months formed the sampling frame. Disproportionate
stratified random sampling by 6 month age groups was performed. To obtain a
high response rate and obtain a target sample size of 3,000, total 5,648 children of
aged between 6–72 months were recruited.

37


The approval to conduct the study was obtained from the Institution
Review Board (IRB) from Singapore National Eye Centre and Domain Specific
Review Board (DSRB) from National Health Group.

2.2.4 Inclusion and Exclusion Criteria
Children were eligible if they were aged between 6 months to 6 years and
Chinese in which at least one of the parents of the child must be Chinese.
Exclusion criteria included children aged over 6 years or non-Chinese or children
with chronic diseases, children with relevant disabilities or children with
congenital abnormalities. If the family moved out from this address or the given
address was an error, the child was ineligible.
Of the 5,648 participants, 4,164 of children were eligible for the study.
Eligible group was divided into refused, non-contactable and attended groups.
Children were categorised as refused (n=1,119) if the parents were not interested
to take part in the study, too-busy, had no time or they could not attend because
they thought that the clinics were far from their home. Non-contactables (n=36)
were defined as those who could not be contacted at least 8 times. The attended
group includes 3,009 participants.
Ineligible category includes overage or underage (612 children). 384
children were ineligible because the family moved out. Races other than Chinese
were also under the ineligible group of “Non-Chinese” subset, and it constituted

38


37 children. Data error (n=152) was also included in the ineligible group, and
other unknown reason of ineligibility included 299 children.

2.3 Recruitment

2.3.1 Recruitment chart

Sampling Frame (n=5,648)

Eligible (n=4,164)

Ineligible (n=1,484)

Refused

Non-

Attended

(n=1,119)

Contactable

(n=3,009)

26.9%

(n=36)

72.3%

0.8%

Overage/
Underage
(n=612)
41.2%

Moved
(n=384)

NonChinese

35.8%

(n=37)

Data error

Others

(n=152)

(n=299)

10.3%

20.2%

2.5%

39


2.3.2 Recruitment Process
Brochures of STARS and the invitation letters of both English and
Chinese language were sent out to all the households. The invitation letter
described common eye diseases in preschool Chinese children, the objective of
this survey and invitation to participate in this survey. Parents who were
interested in the survey called back and made an appointment at one of the two
clinics. For those who did not call back, the door-to-door recruitment process was
conducted and guided by recruitment officers.
During the home visit, the recruiters explained the objective of STARS
and the processes (eye examination after putting eye drops, interview
questionnaire about the child’s life style and quality of life) which they will be
encountered during the survey. The explanation was either in the English or
Chinese language depending on the subject’s preference. The recruiter also
explained about the possible risks of eye drops (mild local allergic reactions,
slight temporary difficulty or blurring of vision in reading and mild discomfort
while looking at bright lights for 1 to 2 days) which was used to dilate pupils for
eye examination. If the parents were interested in the study and agreed to bring
their child for the survey, the recruiters arranged appointments for them at one of
the two clinics.

40

2.3.3 Clinic Visit
Registration

Taking informed
consent

Eye examination
1.
2.
3.
4.
5.
6.
7.

Glasses
Stereopsis
Accommodative lag
Bruckner Test
Eye alignment
Ductions
Fixation Preference test

8. Color vision
9. Visual acuity
10. Anterior segment evaluation
11. Ocular Dominance

Putting Eye drops

(30 minutes waiting time)

Interview Questionnaire

Measuring Height and Weight

Eye examination
1. Biometry (>30months)
2. Autorefraction (>24months) or Retinomax (<24months)
3. Retinoscopy (if autorefraction or retinomax fails)
4. Fundus photo (>47months)

Give Report and End the survey
41


2.3.4 Clinic Visit and Ethics consideration
The children were registered and the qualified optometrist of the survey
team explained about the risks and benefits of this survey they participated. The
parents of the participants were also explained about the confidentiality and
autonomy with their child in the whole survey process.
If the parents agreed to the survey procedures, they were asked to sign the
informed consent form for behalf of their child. A copy was given to the parents,
and another was kept for the study. Even the parents refused to continue
participating in the research process, the refusal was respected and the whole
entire process of the survey was stopped.

2.4 Height and Weight Measurements
Weight was measured in kilograms (kg) to the nearest one decimal point,
and height was measured in centimeters (cm) to nearest one decimal point.

Recruiters who were assigned as examiners in height and weight measurements
were trained to perform the same standardized measurement procedures by an
experienced general practitioner.
The equipments (Seca model 220) used for measuring height and weight for
children 2 years old and above, were calibrated once a year.

42


2.4.1 Measuring Height and Weight of Child 2 years old and
above
For a child 2 years of age and above who can stand without support, the
column digital weighing scale with attached telescopic measuring rod (Seca-220)
was used. The tar weighing function was set to zero the scale before each child
was weighed. The child was wearing light clothings without shoes, heavy objects
in the pockets, and hair accessories. The child was instructed to stand still in the
middle of the scale platform, with arms relaxed and hanging down by the side of
the body, both feet slightly apart in order to distributed the weight equally.
Height was measured using the same machine. The child was asked to
look straight ahead along the Frankfort Plane. The Frankfort plane is the line
connecting the superior border of the external auditory meatus with the lower
edge of the eye.

The measuring slide of the telescopic rod was moved in

horizontal position till it firmly abut the crown of the head, slightly compressing
the hair, without bending the slide. The height is noted at the read off mark.
Accuracy of measurement better than +/- 5 mm can be achieved according to the
manufacturer.

2.4.2 Measuring Height and Weight of Child less than 2 years
If the child is less than 2 years old, or unable or unwilling to stand
unsupported on the scale, the child was measured in the parent’s arms. The parent
was weighed alone, and then weighed again with child in arms. The combined
43


weight of parent and child was recorded, and the actual weight of the child was
obtained by subtracting these two measurements.
The supine length measurement was taken with the Kiddimeter (a
recumbent length board for infant with a fixed head board and movable
footboard), placed on a sturdy table. The accuracy of the Kiddimeter is 1 mm
according to the manufacturer (Raven Equipment, Castlemead, UK). First, the
child was placed with his face upward and with the body parallel to the long axis
of the board. One examiner brought the child’s head gently to contact with the
fixed headboard. The second examiner held the child’s feet and toes perpendicular
to the measuring length board, keeping the knees straight and bringing the
moveable footboard against the heels. If the child was restless, only one leg was
positioned for the measurement.

2.5 Skinfold Measurements
For child who is 2 years old and above, subcutaneous fat thickness was
measured using the Holtain skinfold caliper. This caliper was designed to give a
constant pressure of 10gms/sq.mm according to manufacturer and marked in
divisions of 0.2mm, starting from 0 mm up to 40 mm. The dial of the caliper was
checked to ensure at zero each time before use. Only the triceps site was
measured in the study.
The right arm was used for this skinfold measurement. The child was
asked to stand, with back facing the examiner with the arms were relaxed. Mid44

way between tip of acromion and the tip of the olecranon process, a fold of skin,
deep enough to get the subcutaneous fat but without picking the triceps muscle
was lifted perpendicularly to the surrounding skin with the thumb and forefinger
of the examiner’s left hand. The calipers were applied to the fold midway between
the crest and the base of the fold, maintaining this fold throughout the measuring
process. The reading was noted down and the process was repeated another two
times to obtain three measurements at the same site. An average value was
obtained.

2.6 Questionnaire
The questionnaire in this study was administered by well-trained
interviewers of our research team. The questionnaire was adapted and modified
from both the SCORM (Singapore Cohort Study of the Risk Factors for
Myopia)74 and the quality of life questionnaire of MEPEDS (The Multi-Ethnic
Pediatric Eye Disease Study)75.
The questionnaire includes family history, birth weight, duration of
preschool hours, physical activities, sedentary activities, presence or absence of
park or garden near to home, maternal smoking and alcohol consumption history
during pregnancy, types of breastfeeding in breastfed mother.
Before proceeding with the interview questionnaire to a parent or
guardian, the interviewer assured the confidentiality of the information, which
was answered by the parent or guardian, and we respected any refusal to answer
45


any of the questions if they do not want. Informed consents were obtained from
the parents or guardians.
The questionnaire took about half an hour to complete. For those parents
or guardians who are not conversant with the English language, the interviewer

used the Chinese language version of the questionnaire.

2.6.1 Family History
2.6.1.1 Income
The total income per month of the family were asked for (“What is your
total

combined

monthly household

income?

[Singapore dollars]”),

by

interviewing either the father or mother, and classified into four categories: Less
than S$1,000, S$1,000–S$2,999, S$3,000–S$4,999, S$5,000 and above. There
were also ‘Refuse’ and ‘Don’t know’ categories for that question if either father
or mother did not want to answer, or could not estimate the actual the total
combined monthly income.

2.6.1.2 Education of parents
Educational level of father (“What’s the child father’s completed
educational level?”) was classified into seven categories: None, Primary,
Secondary, “O”/”N” levels, “A” levels/Polytechnic/Diploma/ITE/Certificate,
University education (degree and above, including bachelor, master and PhD) and
46

Others. If the educational level was chosen as ‘Others’, it was specified and
recorded. Similarly with the total combined monthly income, there were also
‘Refuse’ and ‘Don’t know’ categories for that question if the child’s father did not
want to answer his/her education, or did not know which category fit in his
educational level.
The educational level of mother (“What’s the child mother’s completed
educational level?”) was also asked as the same way as the father’s education
status.

2.6.2 Clinic Questionnaire
2.6.2.1 Birth Weight
The parents were asked to bring the child’s health booklet on the
appointment day. From the health booklet, the information of the child’s birth
weight was recorded in grams as an answer to the question “How much did your
child weigh at birth?” If the parents cannot remember their child’s birth weight,
‘Don’t know’ category for that question was selected. There was also ‘Refuse’
category for those parents did not want to answer that question.

2.6.2.2 Preschool Hours
Many children in Singapore attend pre-nursery (3 years old and younger),
nursery (4 years old), kindergarten 1st year (5 years old) and kindergarten 2nd year
47


(6 years old) before they enter primary schools. If the child was attending one of
these preschools, the duration of time in preschool was asked by “How many
hours per day does your child spend in pre-school?”

2.6.2.3 Physical Activities

2.6.2.3.1 Playing outdoors
The interviewer asked for the number of hours per day on weekdays
(outside of regular school hours) during which the child played outdoors, such as
walking and biking. Similarly, this same question was asked for weekends. If the
child was too young to play outdoors, ‘Not applicable’ category was selected. If
the parents could not estimate the actual activities of the child, ‘Don’t know’
category was chosen and if the questions were refused by the parents, ‘Refused’
category was ticked.
The weighted variable of ‘Playing outdoors’ was calculated as follows:
Playing outdoors= (time

hours of playing outdoors per day in weekdays

playing outdoors per day in weekends x

x 5/7) + (time

hours of

2/7)

2.6.2.3.2 Leisure Activities
The time spent on leisure activities (hours per day) such as going to park,
beach or picnic for weekdays and weekends (separately, as described above),
48


outside the regular school hours, was asked. There were also ‘Don’t know’ and
‘Refused’ categories if the parents could not estimate the actual activities of the
child or if the questions were refused by the parents, respectively.

The weighted variable of ‘Leisure activities’ was calculated as follows:
Leisure activities= (time hours of leisure activities per day in weekdays x 5/7) + (time hours of leisure
activities per day in weekends x

2/7)

2.6.2.4 Sedentary Activities
2.6.2.4.1 Watching Television
The hours of watching television per day in weekdays as well as in
weekends were asked separately (as described above). There were ‘Not
applicable’ category for which if the child was too young to watch television,
‘Don’t know’ category for which if the parents could not estimate about it and
‘Refused’ category for which if the parents did not want to answer that question.
The weighted variable of ‘Watching television’ was calculated as follows:
Watching television= (time hours of watching television per day in weekdays x 5/7) + (time hours of
watching television per day in weekends x

2/7)

2.6.2.4.2 Playing television/computer/hand held video games
In the questionnaire, time spent playing games (television/computer/hand
held video) was also asked as hours per day for both weekdays and weekends,
49


separately. If the child was too young to play these games, ‘Not applicable’
category was selected.
The weighted variable of ‘Playing television/computer/hand held video
games’ was calculated as follows:
Playing television/computer/hand held video games = (time

games per day in weekdays+

time

hours of playing television

hours of playing computer games per day in weekdays

hand held video games per day in weekdays

+ time

hours of playing

x 5/7) + (time hours of playing television games per day in weekends

+ time hours of playing computer games per day in weekends + time hours of playing hand held video games per
day in weekends

x 2/7)

2.6.2.4.3 Reading, drawing and coloring
The hours per day of reading, writing (for school work and read for
pleasure) and coloring or drawing for fun in weekdays and weekends were also
asked separately in the questionnaire.
The weighted variable of ‘Reading and writing, coloring or drawing’ was
calculated as follows:
Reading and writing, coloring or drawing = (time
weekdays+

hours of reading and writing per day in

time hours of coloring or drawing per day in weekdays x 5/7) + (time hours of reading and writing

per day in weekdays+ time hours of coloring or drawing per day in weekdays

x 2/7)

50


2.6.2.4.4 Total sedentary activities
In our study, watching television, playing television/computer/hand held
video games, reading, drawing and coloring were combined and regarded as total
sedentary activities.
The weighted variable of ‘Total sedentary activities’ was calculated in the
following expression:
Total sedentary activities = (time
playing television games per day in weekdays+

hours of watching television per day in weekdays

hours of coloring or drawing per day in weekdays)

weekends

hours of

time hours of playing computer games per day in weekdays + time

hours of playing hand held video games per day in weekdays+

time

+ time

time hours of reading and writing per day in weekdays+

x 5/7 + (time

hours of watching television per day in

+ time hours of playing television games per day in weekends + time hours of playing computer games

per day in weekends

+ time hours of playing hand held video games per day in weekends + time hours of reading

and writing per day in weekdays+ time hours of coloring or drawing per day in weekdays

) x 2/7

2.6.2.5 Presence of park or garden near home
To assess the presence of park or garden near to the child’s home, the
question of “Is there a park or garden near to your home where your child could
play outdoors?” was asked. However, the question was not defined for the
distance to park or garden. The answers provided for this question were ‘Yes’,
’No’, ‘Refused’ and ‘Don’t know’. If the family lives near park or garden, ‘Yes’
category was selected. If there was no park or garden near to their home, ‘No’
category was chosen.

51


2.6.2.6 Maternal smoking and alcohol consumption during
pregnancy
History of maternal smoking and alcohol consumption during pregnancy
were assessed separately (“At any time during the pregnancy with the child, did
you smoke?” and “At any time during the pregnancy with the child, did you drink
alcohol?”). There were options of ‘Yes’, ’No’, ‘Refused’ and ‘Don’t know’ as
answers for each question.

2.6.2.7 Breastfeeding
Regarding breastfeeding questionnaire, the mother was asked “Was your
child ever breastfed or fed breast milk?” and if she replied she did, the rest of the
questionnaire regarding breastfeeding were continued. If she replied the answer of
‘No’ or ‘Don’t know’, the questionnaire regarding breastfeeding were skipped
and proceeded to others.
If the mother breastfed the child, the interviewer asked the question of
“How long did you breastfeed this child?” There were 7 categories to answer:
‘less than 1 week’, ‘1 to 4 weeks’, ‘1 to 3 months’, ‘4 to 6 months’, ‘6 to 12
months’, ‘More than 12 months’ and ‘Still breastfeeding’. Types of breastfeeding
were also assessed by asking the mother “Which type of breastfeeding best
describes what you practiced/received at that time?” The answers provided for
this question were ‘Exclusive breastfeeding’, ‘Mostly breastfeeding’ and ‘Partly
breastfeeding’. If the child only received breast milk (may include medicines and
52


vitamins), ‘Exclusive breastfeeding’ category was selected. If the mother fed her
child with breast milk and water, sweetened water, or juices but not formula milk,

the interviewer selected ‘Mostly breastfeeding’ category. However, if the child
received formula milk or other complementary foods in addition to breast milk, it
was regarded as ‘Partly breastfeeding’.

2.7 Referral
If there was a pathologic eye condition or high/low blood pressure
detected which needed medical attention, a referral letter to the ophthalmologist
or general practitioner was given to the caregiver to seek appropriate treatment. A
gift was given to the child and incentives were given to the parents in appreciation
of their participation. The completed case files were transferred to the survey
office for data entry and management.

2.8 Data Management
For the safety of data and confidentiality, the participating children were
identified by their specific identification numbers which were set up in the
database and in all the forms relating to subjects. Moreover, the password to
access the study data files was only given to the staff in the survey team. To
facilitate data entry from measurements, pre-coding was done. Microsoft access
2003 was used to setup the database.
53


Four members of the STARS team were designated to enter data into the
database. After double data entries, data were checked for inconsistencies
between the first and second data entry by using specified software. If there was a
mismatch, we checked which variable was involved and the study ID was noted.
The file was searched according to study ID and the variable of mismatch was
checked with the data inside the case file. The correct data was noted and
corrected in the database.

2.9 Data Analysis
Among 3,009 participants, data of 2,964 participating children
were involved in the analysis after dropping the missing values and excluding two
participating children with extreme values of BMI above 90 kg/m2 which are
assumed as errors during taking measurements or recording the data.

2.9.1 Definitions of overweight and obesity
BMI was calculated as weight in kilograms divided by height in meters
squared (kg/m2). To compare the prevalence of overweight and obesity, using
three different criteria, we analysed the BMI data, according to (1) the CDC
reference which defines a BMI above 95th percentile for age as “overweight” and
that of above 85th percentile for age as “at-risk-for-overweight”, (2) the IOTF
reference which defines overweight/obesity as above the extrapolated percentile
lines which intersect the adult cut-off points of BMI of 25kg/m² and 30kg/m2 at
54


age 18 year as overweight and obesity, respectively, and (3) the “Singapore BMI
for age” reference which defines BMI at and above 90th to 97th percentile for age
as “overweight”, and above 97th percentile for age as “obese”. At the time of
writing this thesis, the Health Promotion Board of Singapore has not defined
childhood overweight and obesity using this BMI percentile chart yet.
Unfortunately, this BMI for age chart does not have the 85th and 95th percentile
lines. Thus, we decided to arbitrarily us the 90th and 97th percentile curves (which
are plotted in this chart) to define overweight and obesity for this study. Although
the CDC used the terms “at-risk-for-overweight” and “overweight”, “overweight”
and “obesity” terms were only used in this study to standardize for all three
criteria.
In the analysis of associations, the weight statuses of the children
(overweight and obese) were used as dependent variables and relevant familyrelated and behaviour-related determinants were independent variables. We

compared “overweight” category with “Not overweight/obese” which includes
normal weight as well as underweight children but excludes obesity. Moreover, as
for “obesity”, we compared it with “Not overweight/obese” category, excluding
the “overweight” category. This study also combined overweight and obesity to
create the category of “Overweight/Obesity” and also compared with the category
“Not overweight/obese” for the association analysis.

55


2.9.2 Statistical analysis
All variables used in the analysis were changed to categorical data to
facilitate interpretation. The χ²-test was used to evaluate the difference between
two groups of potential variables of interest according to the outcome. However,
if the variables of interest were categorised into more than two groups (age
groups, quartiles groups and categories of parental education, total monthly
income and types of breastfeeding), the differences among the groups were
analysed by using one-way ANOVA test. The χ2-trend test was used to assess
trends in the prevalence of overweight and obesity. Odds ratios (ORs) and 95%
confidence intervals (95% CI) were calculated. Means were compared using ttests. Statistical significance was set at 0.05 and SPSS 17.0 software was used to
compute all the analyses.

2.9.3 Construction of multiple regression models
Among the variables of interest, only the potential variables with
statistically significance value less than 0.02 based on the results of univariate
analysis were placed in our multivariate models.
Age, gender and father’s educational level were introduced into the
models as confounders. The categorical variable of ‘father’s educational level’
was converted in dichotomous dummy variable.
In the multivariate linear regression analysis, BMI (kg/m2) was used as the

dependent variable. Birth weight, duration of preschool hours per day and
56


duration of watching television hours per day were used as independent variables,
and age, gender and father’s education were adjusted in this model.
In the multivariate logistic regression analysis, combined overweight and
obesity (‘Overweight/Obesity’) was used as the dependent variable and birth
weight and duration of total sedentary activities per day were independent
variables, and age, gender and father’s education were adjusted.

57