Citation
Boughoula, Mariem
(2020)
Reliability and validity assessment on a diet quality index among adults in Kajang and Klang, Selangor, Malaysia.
Doctoral thesis, Universiti Putra Malaysia.
Abstract
The objective of this study was to retest a previously developed diet quality index
among the general Malaysian adult population. While the index was initially developed
to assess the diet quality of Malaysian adults, it was previously tested only among
university students. More precisely, this study will build on a previous study, done by
Fokeena et. al., (2016) which examined dietary patterns among a sample of 320
university students at Universiti Putra Malaysia (UPM) using a 12-item diet quality
index. This study consisted of two phases, where the objective of the first phase was to
assess the reliability and validity of this same index, however among a sample of the
general Malaysian free-living adult population, to assess their diet quality and to
determine its relation to physical activity, body composition, and socio-demographic
factors. Three forms of reliability were used, namely, internal consistency reliability,
test-retest reliability and inter-rater reliability. Validity will be established through
construct validity, concurrent validity, and validity with relation to body weight. The
objective of the second phase is to establish construct validity using confirmatory
factor analysis, to determine whether the model fits the data. A two phase crosssectional
study was conducted among 576 free-living adults, aged 19 – 59 years and
living in Kajang and Klang, Selangor. Data were collected through face-to face
interview and were used to determine the internal consistency reliability, test-retest
reliability, construct validity, concurrent validity, as well as validity of the index with
relation to body weight status. Test-retest reliability was measured among 30 adults
from the sample. The index had a Cronbach’s alpha value of 0.268 and mean inter-item
correlation of 0.032. The intra-class correlation for test-retest reliability was above 0.7.
Principal component analysis revealed the presence of three principal components or
dimensions with eigenvalues exceeding 1, explaining 18.3%, 14.3% and 10.6% of the
variance respectively. The minimum factor loading was greater than 0.4. Results from
confirmatory factor analysis indicate that the model resembles satisfactory fitting based
on adequacy criteria GFI, RMR and RMSEA. The overall fitting results were χ2 =
146.09; cmin/df = 2.858; GFI = 0.915; AGFI = 0.856; CFI = 0.732; RMR = 0.166;
RMSEA = 0.092 and SRMR = 0.0928. The possible range of scores was 0 to 60. The
mean total diet quality score was 34.5±5.62 for the Kajang sample and 35.5±5.03 for the Klang sample. Diet quality scores showed a small correlation between diet quality
scores and intakes of monounsaturated (-0.127) and polyunsaturated fats intake (-
0.107). Only one statistically significant positive correlation was found, namely
between diet quality scores and thiamin intake (0.109). Diet quality scores did not
significantly correlate with body weight status. Using a cut-off value of 30.5, the
proportion of respondents at risk of poor diet quality and at lower risk of poor diet
quality in the Kajang sample were 23.9% and 76.1% respectively. The cutoff score for
the Klang sample was 32.5. Based on that, the proportion of respondents at risk of poor
diet quality and at lower risk of poor diet quality were 22.1% and 77.9% respectively.
In both samples, a high proportion of respondents did not meet the dietary guidelines
for cereals, cereal products and tubers; wholegrain cereals, fruits, vegetables, fish, and
poultry, meat and egg. In addition, Klang respondents did not meet the dietary
guidelines for milk and dairy products as well as legumes and their products.
Significant differences in dietary guideline adherence were observed for wholegrain
cereals, fruits, vegetables, milk and dairy products, fish, high-fat protein foods and
sugar-rich foods in the Kajang sample, but only for the cereals, cereal products and
tubers groups in the Klang sample. In both samples, respondents at lower risk of poor
diet quality showed better adherence. Among the socio-demographic factors, diet
quality was significantly correlated to age, ethnicity, educational level, employment
status. Diet quality was not significantly correlated with BMI, waist circumference,
physical activity level and understanding nutrition labels. Diet quality was significant
to reading nutrition labels and sedentary behavior. The 12-item diet quality index is a
valid and reliable instrument to measure diet quality of the general Malaysian adult
population. However, since most respondents did not meet the dietary guidelines for
most food groups, healthcare providers and future researchers who intend to use the
index might consider using a higher cutoff score to classify ‘at risk’ and ‘at lower risk
of poor diet quality’.
Download File
Additional Metadata
Actions (login required)
|
View Item |