European Journal of Clinical Nutrition (2005) 59, 1049–1058

& 2005 Nature Publishing Group All rights reserved 0954-3007/05 $30.00
www.nature.com/ejcn

ORIGINAL COMMUNICATION
Obesity and household food insecurity: evidence from
a sample of rural households in Malaysia
Z Mohd Shariff1* and GL Khor2

1
Department of Nutrition and Health Sciences, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Selangor,
Malaysia; and 2Department of Nutrition and Health Sciences, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia,
Serdang, Selangor, Malaysia

Objective: The study examined nutritional outcomes related to body fat accumulation of food insecurity among women from
selected rural communities in Malaysia.
Design: Cross-sectional study.
Setting: Rural communities (seven villages and two palm plantations) in a district with high percentage of welfare recipients.
Subjects and Measurements: Malay (n ¼ 140) and Indian (n ¼ 60) women were interviewed and measured for demographic,
socioeconomic, anthropometric, dietary and physical activity information. The women were measured for their body mass index
and waist circumference (WC). Energy and nutrient intakes, food group intake and food variety score were analyzed from 24 h
dietary recalls and food-frequency questionnaire. Daily physical activity of the women was examined as the number of hours
spent in economic, domestic, leisure and sport activities.
Results: Using the Radimer/Cornell Hunger and Food Insecurity Instrument, 58% of the women reported some degree of food
insecurity (household insecure 14%, adult insecure 9.5% and child hunger 34.5%). In general, food-insecure women had lower
years of education, household income and income per capita, more children and mothers as housewives. More than 50% of
food-insecure women were overweight and obese than women from food-secure households (38%). Similarly, more food-
insecure women (32–47%) had at-risk WC (Z88 cm) than food-secure women (29%). Food-insecure women spent significantly
more time in domestic and leisure activities than food-secure women. Overweight and abdominal adiposity among the women
were associated with a number of independent variables, such as women as housewives, women with more children, larger
household size, food insecurity, shorter time spent in economic activities, longer time spent in leisure activities and lower food
variety score. After adjusting for factors that are related to both adiposity and food insecurity, women from food-insecure
households were significantly more likely to have at-risk WC, but not obese.
Conclusion: Among this sample of rural women, the relationship between food insecurity and obesity is a complex one, which
involves the interaction with other factors. Nevertheless, given that obesity and food insecurity are of public health concerns in
the developing nations, the association between the two should be further investigated.
European Journal of Clinical Nutrition (2005) 59, 1049–1058. doi:10.1038/sj.ejcn.1602210; published online 29 June 2005

Keywords: food insecurity; obesity; at-risk waist circumference; food variety; women

Introduction improvements in mortality and morbidity, but also have

Since its independence in 1957, Malaysia has experienced
rapid socioeconomic and demographic changes, which are
often attributed to its stable economic growth and political
development. These changes have not only resulted in
*Correspondence: ZM Shariff, Department of Nutrition and Health
Sciences, Faculty of Medicine and Health Sciences, Universiti Putra
Malaysia, Serdang 43400 Selangor, Malaysia.
E-mail: zalilah@medic.upm.edu.my
Received 1 July 2004; revised 8 March 2005; accepted 23 May 2005;
published online 29 June 2005
led to profound shifts in dietary and physical activity
behaviors among its population. Similar to other developing
countries, sedentary lifestyle and dietary changes towards
higher consumption of refined grains, animal source foods,
added sugar and fats and lower intakes of fruits and
vegetables have been implicated in the increasing rate of
obesity and nutrition-related noncommunicable diseases in
Malaysia (Tee, 1999).
Obesity is associated with many health problems such as
diabetes mellitus type II, hypertension, heart disease, stroke
 Obesity and household food insecurity
ZM Shariff and GL Khor
1050
and certain types of cancer. The Malaysian National Health
and Morbidity Survey 1996 (NHMS 2) has reported the
prevalences of overweight and obesity as 20.7 and 5.8%,
respectively, with women being more overweight and obese
than men (Lim et al, 2000). There was also an increase in
the percentage of adults 25 y and above with elevated blood
pressure, from 14.4% in 1986 (NHMS 1) to 24.0% in 1996
(NHMS 2). The prevalence of diabetes was 8.3%, which
included both the diagnosed and undiagnosed diabetes
(Ministry of Health, 1998, 1999). In addition, mortality rate
from diseases of circulatory system has been on the rise since
1965 and, for the past three decades, it accounted for about
one-fourth or more of the total medically certified deaths in
Malaysia (Department of Statistics Malaysia, 1965–1999).
In Malaysia and other countries in the Asian region, the
prevalence of overweight and obesity is on the rise among
both the urban and rural populations (Ge & Fu, 2000; Tee,
2002). Obesity may also occur in households that are food
insecure, as evidenced by the co-existence of underweight
and overweight in the same poor households (Doak et al,
2000; Gillespie & Haddad, 2001). In these countries, while
positive associations between obesity and food insecurity
with poverty have been well documented and elucidated,
the association between food insecurity and obesity and its
plausible explanation remained elusive. However, in both
food-rich and food-insufficient countries, food insecurity has
been shown to contribute to poor health status, inadequate
food intake, lower cognitive and academic attainments,
and psychosocial problems among children (Immink &
Payongayong, 1999; Alaimo et al, 2001; Wasito et al, 2001;
Kaiser et al, 2002; Oh & Hong, 2003). While food insecurity
contributes to nutrient deficiencies (eg iron and zinc) and
poor reproductive health in women of food-insufficient
nations, women experiencing food insecurity in food-rich
countries are more likely to be overweight, obese, and have
more disordered eating behaviors (Kendall et al, 1996; Olson,
1999; Townsend et al, 2001; Adams et al, 2003). Limited food
availability and access to healthy food choices and, conse-
quently, consumption of cheaper, less nutritious, high-fat
and more energy-dense food items may be among the
possible explanations for the observed association between
obesity and food insecurity.
Improvements in household food security and promotion
of healthy lifestyle to prevent obesity and its related health
problems are among the agenda in the Malaysian National
Plan of Action on Nutrition (NPAN) by the Ministry of Health
Malaysia. Thus, the investigation of the possible association
between food insecurity and obesity among the Malaysian
population is relevant towards identification of its underlying
factors that can be addressed in future national policies,
programs and interventions. At present, there is no published
study on the relationship between obesity and food insecu-
rity among the Malaysian population. Thus, this paper aimed
to determine the factors related to household food insecurity
and the association between household food insecurity and
overweight among a sample of rural women, independent of
European Journal of Clinical Nutrition
other demographic, socioeconomic and dietary factors, and
physical activity. This study was part of a research on dual
forms of malnutrition in the same households conducted
in six countries, namely, Japan, Korea, the Philippines,
Thailand, Indonesia and Malaysia, with funding from the
Japan Ministry of Education and Welfare.
Methodolgy
Study location and subjects
The study was conducted in the Sabak Bernam district of
Selangor state, which is about 150 km north of the capital
city of Kuala Lumpur. The district was purposively chosen for
its high percentage of poor rural households. According to
the Ministry of Rural Development of Malaysia, the district
had 1150 households or 62% of the total welfare recipients
in Selangor in 2000. Monthly welfare allowance is given to
households earning below the government poverty income
line (PLI) (based on PLI of RM510 or USD 134 for a
household size of 4.6 or RM110.87 or USD 29 per capita).
There are 12 villages and two palm oil plantations in the
Sabak Bernam district, with the Malay ethnic group as the
majority of the population. Based on the information from
the District Office, seven of the largest villages in terms of
population size and both plantations were purposively
selected for the study. There were a total of 840 households
in the seven villages and two plantations. Enumerators
visited each of these 840 households to identify those
households with at least one child aged between 1 and 6 y,
and the mother aged above 20 y and nonpregnant. A total of
290 Malay and Indian households were identified; however,
as the study was part of a larger research on dual malnutri-
tion in the same households, a set of criteria was further used
to categorize the households into two groups—mother was
overweight (body mass index (BMI) Z25 kg/m2)/child was
underweight (WAZo1 s.d.) (n ¼ 133) and mother was
normal (BMI 18.5–24.9 kg/m2)/child was normal (WAZ –
1rxr2 s.d.) (n ¼ 67). A total of 200 households were
identified, having either of these with Malay and Indian
households comprising of 140 and 60, respectively. In
households with more than one child aged 1–6 y, the
youngest of the children was selected for the mother–child
pair. The design effect of mother–child pair (mother OW/
child UW (n ¼ 133); mother N/child N (n ¼ 67)) and food
security status was tested using a two-by-two table; however,
we found no significat association between the two variables
(w2 ¼ 0.754, P ¼ 0.385).
Instrument
Trained research assistants made house visits to collect all the
pertinent research information. Demographic and socio-
economic data were collected through in-depth interviews
with the women using a pre-tested questionnaire. However,
on items that the women were not able to report (eg
incomes, expenditures and assets), their husbands’ responses

were obtained. In addition to the screening weight and
height measurements, the women were again measured for
their weights and heights using a digital SECA weighing scale
and SECA body meter. Measurements on waist and hip
circumferences were carried out using a nonstretchable fiber-
measuring tape. Each of the weight, height, waist and hip
measurements was carried out twice and the average was
taken as the final reading. In some of the households where
the husbands refused to let the research assistants measure
the waist and hip of the women, the husbands were assisted
by the research assistants to conduct the measurements.
Food intake information was obtained through a 24-h
dietary recall (2 days) and nonquantitative food-frequency
questionnaire (FFQ). The research assistants guided the
women on the estimation of food quantities, using a set of
calibrated household measurements (eg cups, glasses, bowls,
plates, spoons, ladles). In some cases, other household
measurements that were used by the women were also
measured and weighed against our calibrated household
measurements. The food intakes were then translated into
energy and nutrient intakes based on the Malaysian Food
Composition using Diet 4 software. The Malaysian Food
Guide Pyramid was utilized as the basis to calculate the
number of servings from each food group. The mean values
for energy, nutrient and food group intakes for the 2 days
were then calculated.
The FFQ consisted of 58 food items that were common
among the Malay and Indian ethnic groups. The food items
were compiled based on the various dietary studies con-
ducted among these ethnic groups in Malaysia (Zaitun &
Terry, 1990; Chee et al, 1996). The women were asked on
the frequency (daily, 2–3 times a week, once a week, once
a month and rarely/never) of each food item consumed
during the past month. Food variety score (FVS) for each
subject is calculated as the total number of food items in
the FFQ consumed regularly (daily or 2–3 times a week)
(maximum ¼ 58). The method was modified from Drew-
nowski et al (1997), Lee et al (1999) and Ogle et al (2001).
The women were also asked to recall their daily typical
activities. The activities were recorded in hours or minutes
and divided into four groups—economic (both income and
non-income-generated activities), domestic (child care, food
preparation, including cooking and cleaning, house-clean-
ing and cloth washing), leisure (sleeping, watching television
or listening to the radio, talking to or visiting the neighbors
and doing personal activities such as bathing and praying)
and sports/exercise/active work (brisk walking, playing
games and gardening). The types of activities in each group
were identified through observations prior to the study as
common activities among the Malay and Indian women in
these rural villages and plantations.
To categorize the households as food secure or food
insecure, the Radimer/Cornell Hunger and Food Insecurity
instrument was utilized, whereby the 10 items reflect four
levels of food insecurity with increasing severity—food
secure, household food insecure, individual or adult food
Obesity and household food insecurity
ZM Shariff and GL Khor
1051
insecure and child hunger. The levels of food insecurity have
been understood as a managed process (Radimer et al, 1990).
The items on the instrument were developed from the
perspective of rural women in upstate New York, who had
experienced hunger through in-depth interview with these
women. Further studies were conducted to establish the
validity and reliability of the instrument (Radimer et al,
1992; Kendall et al, 1995).
These 10 items have been translated and the translated
version has been used in previous local studies (Zalilah 1998;
Zalilah & Ang, 2001; Zalilah & Tham, 2002). The internal
consistency of the 10 items (Cronbach’s alpha) in these
previous studies were reported as in the range of 0.8–0.9. The
research assistants went through the items with the women
in the absence of the household heads, as, from our previous
research experience, the women were quite hesitant to
answer the questions in the presence of the household
heads (husbands, parents, parents-in-law).
Analysis
The statistical program used to analyze the data was SPSS
11.5. (SPSS Inc., Chicago, IL, USA). The first step in the
statistical analysis was to compare demographic and socio-
economic variables, anthropometric measurements, food
intakes and patterns, and time spent in various daily
activities according to food insecurity status. As the data
were normally distributed, w2 analysis and one-way ANOVA
were utilized for categorical and continuous variables,
respectively. Post hoc test (Bonferroni test) was utilized to
examine the significant differences of the outcome variables
among food insecurity groups.
The odds ratio (OR) for each independent variable
(demographic, socioeconomic, dietary, physical activity)
with outcome variables was determined through univariate
logistic regression to identify the individual risk factor for
overweight and at-risk waist circumference (WC) (data not
shown). Multivariate logistic regression was then run to
identify the association between food insecurity status and
the outcome variables, adjusting for the effects of the
significant individual risk factors for measurements of
adiposity and factors which were significant according to
food insecurity levels. However, factors such as ethnicity and
energy intake were also included in the multivariate
regression model, despite their insignificant associations
with measures of adiposity and food insecurity. Income per
capita was used to represent the combination of household
income and household size. The covariates included in this
final step were ethnicity, women’s age, years of schooling
and employment status, income per capita, number of
children, time spent in economic, domestic and leisure
activities, energy intake and food variety score. All of these
covariates were continuous, except for women’s employ-
ment status (working or housewife) and ethnicity (Malay or
Indian). Results from the logistic regressions were expressed
as adjusted ORs with 95% confidence intervals (CIs). An OR
European Journal of Clinical Nutrition
 Obesity and household food insecurity
ZM Shariff and GL Khor
1052
greater than 1 indicates that food insecurity is a risk factor
for obesity and at-risk WC, while lesser that 1 shows a
protective effect of the adverse health outcomes.
Results
Food insecurity and sample characteristics
Majority of the women were in the age group of 25–44 y
(80.5%), with an average of 7 y of formal education
(completed primary school) (Table 1). Women who reported
child hunger had the lowest mean years of education (6.3 y)
than the other groups. In general, more women in food-
insecure households (household insecure 68%; individual
insecure 84%; child hunger 77%) than women in food-
secure households were housewives. There were significant
mean differences in household size and the number of
children between the child hunger and food-secure house-
holds. Most of the households (91.5%) had incomes below

Table 1 Household demographic and socioeconomic variables according to food security status

Food secure Household food Individual food Child hunger

Variables All (n(%)) (n ¼ 84) insecure (n ¼ 28) insecure (n ¼ 19) (n ¼ 69) P-valuea

Ethnicity n(%)
Malay 140 (70) 58 (69.0) 19 (67.9) 15 (78.9) 48 (69.6) NS
Indian 60 (30) 26 (31.0) 9 (32.1) 4 (21.1) 21 (30.4)

Age of women (y)

Mean (s.d.) 33.9 (7.6) 33.1 (7.8) 33.8 (9.3) 36.0 (7.3) 34.4 (6.7) NS
20–24 19 (9.5)
25–34 101 (50.5)
35–44 60 (30.0)
Z45 20 (10.0)

Education of women (y)

Mean (s.d.) 7.2 (2.7) 7.6 (2.6)* 7.8 (3.0) 7.6 (3.6) 6.3 (2.3) 0.01
No schooling 4 (2.0)
1–6 132 (66.0)
7–9 30 (15.0)
Z10 34 (17.0)

Employment status of women n (%)

Housewife 128 (64.0) 40 (47.6) 19 (67.9) 16 (84.2) 53 (76.8) 0.00
Working 72 (36.0) 44 (52.4) 9 (32.1) 3 (15.8) 16 (23.2)

Number of children
Mean (s.d.) 3.3 (1.7) 2.8 (1.4)* 3.5 (2.0) 3.7 (1.70) 3.6 (1.7) 0.01
1–3 121 (60.5)
4–6 69 (34.5)
Z7 10 (5.0)

Household size
Mean (s.d.) 5.4 (1.8) 5.0 (1.60)* 5.8 (2.2) 5.6 (1.7) 5.8 (1.8) 0.05
3–5 112 (56.0)
6–7 74 (37.0)
Z9 14 (7.0)

Household income (USD)b

Mean (s.d.) 204 (149) 251 (175)* 250 (169)** 157 (101) 140 (71) 0.00
r452 183 (91.5)
4452 17 (8.5)

Household income per capita (USD)

Mean (s.d.) 32 (25) 53 (37)***,* 47 (40)** 30 (21) 26 (12) 0.00
o14.6c 10 (5.0)
14.6–29.2d 78 (39.0)
429.2 112 (56.0)
a
One-way ANOVA or w2 analysis.
b
Average household income in rural areas of Malaysia (Eight Malaysian Plan, 2001).
c
Hard core poor.
d
Poor.
USD 1 ¼ RM 3.8. NS ¼ nonsignificant.
*Significantly different between food secure and child hunger.
**Significantly different between household food insecure and child hunger.
***Significantly different between food secure and individual food insecure.

European Journal of Clinical Nutrition


the national mean income for rural households in Malaysia
(USD 452). In all, 44% of the households were living below
the PLI, with 5% as hardcore poverty households. There were
decreasing trends in the mean household income and
income per capita as household food insecurity worsened.
Food insecurity and anthropometric measurements
The mean BMI was significantly higher for women from
individual food-insecure (27.8 kg/m2) and child hunger
(27.5 kg/m2) than food-secure households (24.6 kg/m2);
however, no significant difference was observed for WC
(Table 2). While 39% of women in food-secure households
were overweight and obese, more than 50% of women in
each food-insecure group were obese and overweight.
Similarly, more women in food-insecure groups (32–47%)
than women in the food-secure group (29%) had at-risk WC
(Z88 cm).
Food insecurity and food intakes
Women from food-insecure groups had higher intakes of
energy and most of the nutrients compared to food-secure
women; however, none of the differences was statistically
significant (Table 3). All the groups had similar mean
number of serving for each food group consumed. In
general, all groups met at least 2/3 of RDA for all nutrients
and the recommended number of servings for each food
group, except for calcium, iron, grains/cereals/tubers and
milk/dairy products. Food variety score (FVS), defined as the
number of food items consumed on habitual basis, was
highest in the food-secure group. The mean for FVS was
significantly higher between food-secure and all food-
insecure groups, and between household food-insecure and
child hunger groups.
Obesity and household food insecurity
ZM Shariff and GL Khor
1053
Food insecurity and daily physical activity
As more women in food-secure households were working,
they spent on average significantly more hours in economic
activities (4.6 h) and less in domestic activities (7.6 h) than
the food-insecure groups (economic activities 2–3 h; domes-
tic activities 8–9 h). Consequently, the time spent in both
economic and domestic activities may reduce the hours
spent by food secure group in leisure activities. While
significant mean differences in time spent in economic and
leisure activities were observed between food-secure and two
of the food-insecure groups (individual food insecure and
child hunger), women from food-secure households spent
significantly less time on average in domestic activities than
the women from the child hunger group. All of the women
rarely participate in vigorous physical activity (Table 4).
Food insecurity as a risk factor for overweight and at-risk
WC
Overweight and at-risk WC are associated with a number
of independent variables. While women as housewives
(OR ¼ 2.3, CI ¼ 1.3–4.1), higher number of children
(OR ¼ 1.2, CI ¼ 1.1–1.5), larger household size (OR ¼ 1.2,
CI ¼ 1.1–1.4), longer time spent in leisure activities
(OR ¼ 1.4, CI ¼ 1.1–1.7) and food insecurity (OR ¼ 2.0,
CI ¼ 1.2–3.5) were significant risk factors for overweight,
time spent in economic activities (OR ¼ 0.8, CI ¼ 0.7–0.9)
and food variety score (OR ¼ 0.9, CI ¼ 0.8–0.9) were signifi-
cantly protective against overweight. Longer time spent in
leisure activities (OR ¼ 1.2, CI ¼ 1.1–1.5), food insecurity
(OR ¼ 1.9, CI ¼ 1.0–3.7) and lower food variety score
(OR ¼ 0.9, CI ¼ 0.8–0.9) were significantly associated with
at-risk WC. Table 5 indicated that, after adjusting for
potentially confounding effects of ethnicity, income per
capita, number of children, time spent in leisure, domestic
and economic activities, women’s age, employment status

Table 2 BMI and WC of women according to food security status

Measurement Food secure (n ¼ 84) Household food insecure (n ¼ 28) Individual food insecure (n ¼ 19) Child hunger (n ¼ 69) P-valuea

BMI (kg/m2)
Mean (s.d.) 24.6 (4.5)*,** 26.4 (4.9) 27.8 (5.5) 27.5 (8.8) 0.01
n (%)
o18.5 3 (3.6) 1 (3.6) 1 (5.3) 2 (3.0)
18.5–24.9 48 (57.1) 11 (39.3) 7 (36.8) 29 (42.0)
25.0–29.9 25 (29.8) 10 (35.7) 3 (15.8) 17 (24.6)
Z30.0 8 (9.5) 6 (21.4) 8 (42.1) 21 (30.4)

WC (cm)
Mean (s.d.) 81.0 (9.9) 81.5 (13.2) 86.4 (11.2) 84.4 (11.1) NS
n (%)
o88 60 (71.4) 19 (67.9) 10 (52.6) 40 (58.0)
Z88 24 (28.6) 9 (32.1) 9 (47.4) 29 (42.0)

NS ¼ nonsignificant.
a
One-way ANOVA.
*Significantly different between food secure and individual food insecure.
**Significantly different between food secure and child hunger.

European Journal of Clinical Nutrition

 Obesity and household food insecurity
ZM Shariff and GL Khor
1054
Table 3 Dietary intake of women according to food security status

Food secure Household food Individual food Child hunger

Variables (n ¼ 84) insecure (n ¼ 28) insecure (n ¼ 19) (n ¼ 69) P-valuea

Energy and nutrient Mean (s.d.)

Energy (kcal) 1450 (515) 1581 (467) 1517 (538) 1491 (685) NS
(%RDA) (74.7) (81.0) (78.4) (80.3)
Protein (g) 46.6 (17.6) 51.7 (15.0) 49.4 (15.1) 47.3 (20.4) NS
(%RDA) (114.9) (129.0) (109.9) (111.8)
Total fat (g) 43.4 (21.1) 47.6 (18.6) 51.5 (20.2) 43.5 (26.3) NS
% Energy carbohydrate 60.2 (7.9) 58.3 (8.7) 55.7 (5.4) 59.9 (10.6) NS
% Energy fat 26.3 (6.4) 27.1 (7.7) 30.4 (5.0) 26.0 (7.3) NS
% Energy protein 13.2 (3.6) 13.7 (3.5) 13.5 (3.1) 13.3 (3.3) NS
Vitamin C (mg) 26.3 (36.3) 24.9 (19.8) 22.1 (14.9) 23.5 (20.1) NS
(%RDA) (84.5) (93.1) (71.7) (81.1)

Energy and nutrient Mean (s.d.)

Vitamin A (ug) 601.6 (420.8) 658.2 (421.7) 676.2 (363.5) 702.5 (569.5) NS
(%RDA) (82.1) (84.9) (104.7) (88.3)
Calcium (mg) 279.9 (150.0) 344.0 (159.4) 280.0 (121.0) 293.8 (206.3) NS
(%RDA) (63.1) (76.1) (63.9) (64.7)
Iron (mg) 12.1 (8.7) 12.8 (7.0) 15.3 (12.1) 11.9 (7.6) NS
(%RDA) (62.6) (66.5) (68.2) (65.7)

Food groupb (no. serving/day) Mean (s.d.)

Grains/cereals/tubers (8–12) 6.1 (1.2) 6.0 (1.0) 6.2 (1.3) 6.2 (1.0) NS
Fruits/vegetables (3–5) 4.7 (2.1) 4.7 (1.8) 4.6 (2.2) 4.2 (1.6) NS
Meat/fish/legumes (2–3) 3.1 (1.3) 3.3 (1.6) 2.8 (1.3) 3.1 (1.4) NS
Milk/dairy products (1–2) 0.02 (0.15) 0.1 (0.3) 0.0 (0.0) 0.01 (0.12) NS

Mean (s.d.)
Food variety score 26.2 (3.6)*,**,*** 19.5 (3.9)**** 14.4 (2.8) 12.3 (4.6) 0.00
a
One-way ANOVA.
b
The recommended number of servings for each food group is based on the Malaysian Food Guide Pyramid.
FVS: number of all food items consumed by the women in the past month (0–58). NS ¼ nonsignificant.
*Significantly different between food secure and household food insecure.
**Significantly different between food secure and individual food insecure.
***Significantly different between food secure and child hunger.
****Significantly different between household food insecure and child hunger.

Table 4 Time spent (h) by women in daily activities according to food security status

Food secure Household food Individual food Child hunger

Activity (n ¼ 84) insecure (n ¼ 28) insecure (n ¼ 19) (n ¼ 69) P-valuea

Mean (s.d.)
Economic activitiesb 4.6 (3.8)*,** 3.0 (3.9) 2.1 (3.2) 2.1 (3.3) 0.00
Domestic activitiesc 7.6 (3.3)** 8.4 (3.0) 8.9 (3.1) 9.2 (2.8) 0.01
Leisure activitiesd 11.8 (1.6)*,** 12.5 (1.9) 13.0 (1.8) 12.6 (1.4) 0.00
Sports/exercise 0.05 (0.2) 0.07 (0.3) 0.06 (0.6) 0.06 (0.3) NS

NS ¼ nonsignificant.
a
One-way ANOVA.
b
Income- and non-income-generated activities.
c
Child care (feeding, bathing, teaching, playing), food preparation, house cleaning, cloth washing, attending to garden.
d
Sleep, watch TV, read, talk to neighbors/friends, personal activities (pray, bath etc).
*Significantly different between food secure and individual food insecure.
**Significantly different between food secure and child hunger.

and years of education, energy intake and food variety score, Discussion
only the women from food-insecure households were The findings of this study that more women from food-
significantly more likely to have at-risk WC (OR ¼ 1.2, insecure households were overweight and obese compared to
CI ¼ 1.0–2.5). their food-secure counterparts are similar to the findings in

European Journal of Clinical Nutrition


Table 5 Crude and adjusted OR and 95% CI for overweight and at-risk WC for women by food security status
Overweighta
Variables n (%) Crude OR (95% CI)
Food secure (reference) 84 (42) 1.00
Food insecured 116 (58) 1.97*
(1.19–3.49)
*Po0.05.
a
Overweight: Z25 kg/m2.
b
At-risk WC: Z88 cm.
c
Covariates included in the adjusted model were ethnicity, income per capita, number of children, age, employment status, years of schooling, time spent in leisure,
domestic and economic activities, energy intake and food variety score. All covariates were continuous unless specified—ethnicity (Malay ¼ 0, Indian ¼ 1);
employment status (working ¼ 0, housewife ¼ 1).
d
Food insecure includes household food insecure, individual food insecure and child hunger.
the developed countries (Olson, 1999; Townsend et al, 2001;
Adams et al, 2003). When the prevalences of overweight and
obesity were tabulated according to the four levels of food
insecurity status, more than 50% of women from food-
insecure households were overweight and obese, with the
highest proportion among women from individual food-
insecure households (58%). In addition, the percentages of
women with at-risk WC were also highest among households
that are individual food insecure (47.4%). In a study of
women from the rural county in Upstate New York, the
mean BMI (28.2 kg/m2) and the prevalence of obesity,
defined as BMI429 kg/m2, were highest (37%) among
women from food-insecure households (Olson, 1999). For
other groups (food secure, individual food insecure and child
hunger), the mean for BMI ranges from 25.5 to 25.9, while
the percentages of obese women were 22–28%. Similarly,
Townsend et al (2001) found that women experiencing mild
to moderate food insufficiency were 1.3 and 1.5 times at risk
of obesity (BMIZ27.3 kg/m2) than food-sufficient women.
However, the risk of overweight for women with severe levels
of food insecurity did not differ significantly from that of
food-secure women. Using the data from the California
Women’s Health Survey 1998 and 1999, Adams et al (2003)
reported that the prevalence of obesity (BMIZ30 kg/m2)
among the Non-Hispanic Whites was highest for women
who were food insecure without hunger, but, for those who
were food insecure with hunger, their risk of being obese is
no different from that of food-secure women. Our findings
are similar to these studies, which indicated that the risk of
obesity is significantly associated only with the intermediate
levels of food insecurity (42%). However, this study also
found that more women from households experiencing
child hunger were overweight and obese (55%) compared to
those from food-secure households (39%).
In addition to the use of BMI to identify the degree of
overweight and obesity among these women, we also
measured WC to assess the intra-abdominal fat distribution.
A high WC is a risk factor for diet-related chronic diseases in
individuals with BMI between 25 and 34.9 kg/m2; however, a
recent study has shown that, even among men and women
with normal BMI, those with high WC were more likely to
Obesity and household food insecurity
ZM Shariff and GL Khor
1055
At-risk WCb
Adjusted ORc (95% CI) Crude OR (95% CI) Adjusted ORc (95% CI)
1.00 1.00 1.00
1.36 1.94* 1.18*
(0.69–2.67) (1.03–3.67) (1.02–2.54)
have type II diabetes, hypertension, dyslipidemia and
metabolic syndrome (Janssen et al, 2002). Our findings
showed that the food-insecure women were not only more
overweight and obese, but also had a higher proportion
with at-risk WC than food-secure women. Being overweight
and obese, the food-insecure women are already at greater
risk for various health problems and the risk could be
further amplified if they also have excess abdominal fat
accumulation.
Our findings also indicate that a higher proportion of
women from food-secure households had BMIo25 kg/m2
(61%) and WC o88 cm (71%) (Table 2). The differences in
time spent in various daily activities and variety of foods
consumed between the food-secure and food-insecure
groups as discussed below may provide plausible explanation
for the observed differences in the measures of adiposity.
We showed that food-secure women spent more time in
economic activities, but less time in domestic and leisure
activities than the other food-insecure groups (Table 4). The
differences in economic and domestic activities could be due
to food-insecure women being mostly housewives and
having more children compared to women from food-secure
households. From our observations, participation in leisure
activities, such as watching television and taking afternoon
nap and talking to or visiting the neighbors, was a part of
child care (besides feeding, bathing, teaching and playing
with the child). Using the Continuing Survey of Food Intakes
by Individuals (CSFII) conducted in United States, Townsend
et al (2001) reported that food insecurity was related to
television viewing in that television viewing 41 h/day and
exercising r1 time per month were among the risk factors
besides food insecurity, for overweight among women.
However, food-insecure individuals were also more likely to
have physical limitations, which may prevent them from
being active in nonrecreational activity (Gulliford et al,
2003). In our study, whether consumption of higher-energy
dense foods in combination with more time spent in
physically less vigorous activities (domestic and leisure
activities) and physical limitations is associated with house-
hold food insecurity and thus contributes to overweight and
obesity requires further examination.
European Journal of Clinical Nutrition
 Obesity and household food insecurity
ZM Shariff and GL Khor
1056
Several studies have shown that the diets of women
reporting household food insecurity or food insufficiency
were not only inadequate in nutrient and food variety, but
also in calorie (Kendall et al, 1996; Rose & Oliveira, 1997;
Tarasuk & Beaton, 1999; Rose, 1999; Dixon et al, 2001). In
the present study, none of the energy, nutrient or food group
intakes was significantly different among the food-secure
and food-insecure groups; however, the food variety score as
an indicator of dietary diversity was highest in the food-
secure group than all of the food-insecure groups. A study by
Basiotis and Lino (2002) also found that, although women
in food-insufficient households had a worse diet quality
(eg lower vegetable, fruit and milk but higher cholesterol
intakes, and lack food variety), they had similar caloric
intake as women from food-sufficient households. In a
multi-country analysis to assess dietary diversity as an
indicator of household food insecurity, it was suggested that
dietary diversity could be a promising outcome indicator of
household food insecurity (Hoddinott & Yohannes, 2002).
Both poverty and food insecurity may reduce household
food budget, which consequently limit the access and
procurement of foods of higher quality and wider variety.
For example, in impoverished areas, meats, fish, fruits,
vegetables and whole grain products may be limited and
relatively expensive; however, when these foods are avail-
able, they are of lower variety and quality. Hence, the
consumption of limited food choices and cheaper food items
which are more likely to be high in fat and energy density
can contribute to obesity among the low-income population
(Drewnowski & Specter, 2004).
It has also been hypothesized that women may be
deprived of food during bouts of food insecurity or
insufficiency, but they may engage in bingeing behaviors
during short-term food availability. Thus, the frequent
episodes of binge eating and food restraint may put the
women at risk of being overweight and obese (Marcus et al,
1985; Blackburn et al, 1989; Wadden et al, 1992). This
continuous cycle of temporary food abundance and food
restriction could result in a gradual weight gain over time.
However, the hypotheses that food insecurity might foster
overweight and obesity through cycles of weight loss and
regain need further investigation, especially in the context of
developing nations.
We proceeded to show that only the association between
food insecurity and at-risk WC was significant after control-
ling for potentially confounding variables. Several studies
have found that food insecurity measures were significantly
associated with overweight and obesity even after adjusting
for the effects of various demographic, socioeconomic and
health variables (Sarlio-Lahteenkorva & Lahelma, 2001;
Townsend et al, 2001; Adams et al, 2003). On the other
hand, Vozoris and Tarasuk (2003) reported that men in food
insufficient households were less likely to be overweight,
while no significant association was observed between food
insufficiency and measures of underweight and overweight
among women. Based on a population survey in Trinidad,
European Journal of Clinical Nutrition
food insecurity was prevalent in all BMI groups and was
significantly associated with underweight, but not over-
weight (Gulliford et al, 2003). The inconsistent findings on
the association between food insecurity and measures of
adiposity may be due not only to methodological differences
but also to the lack of understanding on the potential
interactions among food insecurity, poverty and body
composition, especially in developing countries experien-
cing rapid socioeconomic development and nutrition transi-
tion. It is also important to address that nutritional
vulnerability as measured through nutritional assessments
(ie dietary, anthropometric or biochemical) may not be a
necessary consequence of food insecurity as the relationship
is also influenced by the nature, severity and duration of
food insecurity experienced by the individual (ONPP, 2002).
In Malaysia, the National Health and Morbidity Survey
(Lim et al, 2000) has reported that, while slightly more
women (21.4%) were overweight than men (20.7%), the
percentage of obese women (7.2%) was double than that
of men (3.8%). Compared to the other ethnic groups, the
Indians and the Malays had the highest prevalence of
overweight (Indians—24.9%; Malays—22%) and obesity
(Indians—6.8%; Malays—7%). Among selected rural popula-
tion groups in Malaysia, more women were overweight
(28%) and obese (11.1%) compared to men (overweight
19.8% and obesity 4.2%) (Khor et al, 1999). In another study
of rural communities, a higher proportion of women than
men were overweight and obese, and had high waist–hip
ratio and hypercholesterolemia (Ng et al, 1995). Elucidation
for gender and ethnic differences in these reported nutri-
tional disorders will require further examination into
genetic, behavioral and environmental factors.
There are several limitations of the study which should
be considered. First, as the study utilized a cross-sectional
design, cause and effect relationship between food insecurity
and measures of adiposity must be interpreted with caution.
Although several studies have suggested a possible cause and
effect relationship between food insecurity and obesity
(Olson, 1999; Adams et al, 2003), further investigation is
still warranted. The cross-sectional design provided informa-
tion only on the current status of food insecurity and body
composition, and not on the development of these condi-
tions and their cause and effect relationship. Although food
insecurity has been shown to be associated with many
aspects of health, food insecurity may be both a contributory
factor and a consequence of poor health, including obesity.
Second, the small sample size with the inclusion criteria of
households, with children in the age group of 1–6 y old and
mother–child pair is either overweight–underweight or both
normal may only represent a small proportion of the
Malaysian population. Hence, generalization of the study
findings is fairly limited. Third, the lack of significant
differences in energy, nutrient and food group intakes
among the food-secure and food-insecure groups could be
due to the small sample size and the availability of only two
24-h recalls, which may be insufficient to detect any

variation in the daily intake. It may also be difficult for the
women to recall the foods or food ingredients consumed,
due to the nature of mixed dishes commonly eaten among
the various ethnic groups in Malaysia. In addition, as there
were more food-insecure women who were overweight
and obese, they may be more likely to underestimate their
energy intakes. This underestimation may be related to
under-reporting of foods rich in fat and/or carbohydrates,
such as butter, fried foods, sugars and confectioneries, cakes
and pastries (Heitmann & Lissner, 1995; Lafay et al, 2000).
Thus, on average, the energy intake of food-insecure women
remained statistically similar to that of women from food-
secure households. Finally, although the Radimer/Cornell
hunger and food insecurity instrument has been extensively
validated and used in the United States, its validity in other
settings may be questionable as the understanding and
experiences related to food insecurity are specific to those
experiencing it in the setting (Wolfe & Frongillo, 2001).
Although we did not conduct extensive validity testing of
the instrument, we did revise the items (substitution and
translation of words to make them more meaningful, and
addition of non-monetary statements such as ‘not able to
provide or get enough foods’ to the existing ‘I did not have
enough money to get or buy more foods) after several focus
group discussions and in-depth interviews with women of
various ethnic groups from poor households in both urban
and rural areas prior to this study (Zalilah, 1998; Zalilah &
Ang, 2001; Zalilah & Tham, 2002). However, in all of these
studies, internal consistency (Cronbach’s alpha) of the
instrument was higher than the acceptable level of 0.7.
In conclusion, our study demonstrated that overweight,
obesity and at-risk WC were more prevalent among rural
women from food-insecure than food-secure households,
and food insecurity was related to economic disadvantages,
less diversified diets and physical inactivity. These factors,
whether individually or in combination with household
food insecurity, may consequently influence body composi-
tion. Given that obesity is an emerging public health
concern afflicting both the rural and urban populations of
the developing countries and food insecurity is among the
many possible risk factors of obesity, the relationship
between food insecurity and obesity is worth investigated.
In addition, more studies are needed to clarify whether
restricted access to healthy food choices or poor diet quality
and sedentary lifestyle or lack of physical activity are
implicated in the association between food insecurity and
obesity.
References
Adams EJ, Grummer-Strawn L & Chavez G (2003): Food insecurity
is associated with increased risk of obesity in Californian women.
J. Nutr. 133, 1070–1074.
Alaimo K, Olson CM & Frongillo EA (2001): Food insufficiency and
American school-aged children’s cognitive academic and psycho-
social development. Pediatrics 108, 44–53.
Obesity and household food insecurity
ZM Shariff and GL Khor
1057
Basiotis L & Lino M (2002): Food Insufficiency and Prevalence of
Overweight among Adult Women. Nutrition Insights. Washington,
DC: USDA Center for Nutrition Policy and Promotion.
Blackburn GL, Wilson GT & Kanders BS (1989): Weight cycling: the
experience of human dieters. Am. J. Clin. Nutr. 49, 1105–1109.
Chee SS, Zawiah H, Ismail MN & Ng KK (1996): Anthropometry,
dietary patterns and nutrient intakes of Malaysian estate workers.
Mal. J. Nutr. 2, 112–126.
Department of Statistics Malaysia (1965–1999): Vital Statistics for
Peninsular Malaysia. Kuala Lumpur: Department of Statistics.
Dixon LB, Winkleby MA & Radimer KL (2001): Dietary intakes and
serum nutrients differ between adults from food-insufficient and
food-sufficient families: Third National Health and Nutrition
Examination Survey, 1988–1994. J. Nutr. 131, 1232–1246.
Doak CM, Adair LS, Monteiro C & Popkin BM (2000): Overweight
and underweight co-exist within households in Brazil, China and
Russia. J. Nutr. 130, 2965–2971.
Drewnowski A, Henderson SA, Driscoll A & Rolls BJ (1997): The
dietary variety score: assessing diet quality in healthy young and
older adults. J. Am. Diet. Assoc. 97, 266–271.
Drewnowski A & Specter SE (2004): Poverty and obesity: the role of
energy density and energy costs. Am. J. Clin. Nutr. 79, 6–16.
Eight Malaysian Plan (2001–2005): Kuala Lumpur: Economic Plan-
ning Unit.
Ge KY & Fu DW (2000): The magnitude and trends of under- and
over-nutrition in Asian countries. Biomed. Environ. Sci. 14, 53–60.
Gillespie S & Haddad L (2001): Attacking the Double Burden of
Malnutrition in Asia and Pacific. Philippines: Asian Development
Bank.
Gulliford MS, Mahabir D & Rocke B (2003): Food insecurity, food
choices and body mass index in adults: nutrition transition in
Trinidad and Tobago. Int. J. Epidemiol. 32, 508–516.
Heitmann BL & Lissner L (1995): Dietary underreporting by obese
individuals—is it specific or non-specific? Br. Med. J. 311, 986–989.
Hoddinott J & Yohannes Y (2002): Dietary diversity as a food security
indicator. Food Consumption and Nutrition Division Discussion
Paper, No. 136. Washington, DC: International Food Policy
Research Institute.
Immink MDC & Payongayong E (1999): Risk analysis of poor health
and growth failure of children in the central highlands of
Guatemala. Soc. Sci. Med. 48, 997–1009.
Janssen I, Katzmarzyk PT & Ross R (2002): Body mass index, waist
circumference and health risk—evidence in support of current
National Institutes of Health guidelines. Arch. Intern. Med. 162,
2074–2079.
Kaiser LL, Melgar-Quinonez HR, Lamp CL, Johns MC, Sutherlin JM &
Harwood JO (2002): Food security and nutritional outcomes of
preschool-age Mexican-American. J. Am. Diet. Assoc. 102, 924–929.
Kendall A, Olson CM & Frongillo EA (1995): Validation of the
Radimer/Cornell measures of hunger and food insecurity. J. Nutr.
125, 2793–2801.
Kendall A, Olson CM & Frongillo EA (1996): Relationship of hunger
and food insecurity to food availability and consumption. J. Am.
Diet. Assoc. 96, 1019–1024.
Khor GL, Azmi MY, Tee ES, Kandiah M & Huang MSL (1999):
Prevalence of overweight among Malaysian adults from rural
communities. Asia Pacific J. Clin. Nutr. 8, 272–279.
Lafay L, Mennen L, Basdevant A, Charles MA, Borys JM, Eschwege E,
Romon M & FLVS study group (2000): Does energy intake
underreporting involve all kinds of food or only specific food
items. Results from the Fleurbaix Laventie Ville Sante (FLVS) Study.
Int. J. Obes. 24, 1500–1506.
Lee SK, Sobal J & Frongillo EA (1999): Accultration and dietary
practices among Korean Americans. J. Am. Diet. Assoc. 99,
1084–1089.
Lim TO, Goh BL, Zaki M, Suleiman AB, Fatimah S, Siti S, Tahir A &
Maimunah AH (2000): Distribution of body weight, height and
body mass index in a national sample of Malaysian adults. Med. J.
Mal. 55, 108–128.
European Journal of Clinical Nutrition
 Obesity and household food insecurity
ZM Shariff and GL Khor
1058
Marcus MD, Wing RR & Lamparski DM (1985): Binge eating and
dietary restraint in obese patients. Addict. Behav. 10, 163–168.
Ministry of Health (1998): Malaysia’s Health 1998: Technical Report of
the Director-general of Health Malaysia. Kuala Lumpur: Ministry of
Health.
Ministry of Health (1999): Malaysia’s Health 1999: Technical Report
of the Director-General of Health Malaysia. Kuala Lumpur: Ministry
of Health.
Ng TKW, Tee ES & Rosman A (1995): Rural communities in
nutritional transition: emergence of obesity, hypertension and
hypercholesterolemia as public health problems in three kam-
pungs in Bagan Datoh, Perak. Mal. J. Nutr. 1, 129–139.
Office of Nutrition Policy and Promotion at Health Canada (2002):
Discussion paper on household and individual food insecurity. (http: //
www.hc-sc.gc.ca/hpfb-dgpsa/onpp-bppn/food_security_entire_e.
html).
Oh SY & Hong MJ (2003): Food insecurity is associated with dietary
intake and body size of Korean children from low-income families
in urban areas. Eur. J. Clin. Nutr. 57, 1598–1604.
Ogle BM, Pham HH & Ho TT (2001): Significance of wild vegetables
in micronutrient intakes of women in Vietnam and analysis of
food variety. Asia Pacific J. Clin. Nutr. 10, 21–30.
Olson CM (1999): Nutrition and health outcomes associated with
food insecurity. J. Nutr. 129, 521S–524S.
Radimer KL, Olson CM & Campbell CC (1990): Development of
indicators to assess hunger. J. Nutr. 120, 1544–1548.
Radimer KL, Olson CM, Greene JC, Campbell CC & Habicht JP
(1992): Understanding hunger and developing indicators to assess
it in women and children. J. Nutr. Educ. 24, 36S–45S.
Rose D (1999): Economic determinants and dietary consequences of
food insecurity in the United States. J. Nutr. 129, 517S–520S.
Rose D & Oliveira V (1997): Nutrient intakes of individuals from
food-insufficient households in the United States. Am. J. Publ.
Health. 87, 1956–1961.
Sarlio-Lahteenkorva S & Lahelma E (2001): Food insecurity is
associated with past and present economic disadvantage and
body mass index. J. Nutr. 131, 2880–2884.
European Journal of Clinical Nutrition
Tarasuk VS & Beaton GH (1999): Women’s dietary intakes
in the context of household food insecurity. J. Nutr. 129,
672–679.
Tee ES (2002): Obesity in Asia: prevalence and issues in
assessment methodologies. Asia Pacific J. Clin. Nutr. 11,
S694–S701.
Tee ES (1999): Nutrition of Malaysians: where are we heading? Mal. J.
Nutr. 5, 87–109.
Townsend MS, Peerson J, Love B, Achterberg C & Murphy SP (2001):
Food insecurity is positively related to overweight in women.
J. Nutr. 131, 1738–1745.
Vozoris NT & Tarasuk VS (2003): Household food insufficiency is
associated with poorer health. J. Nutr. 133, 120–126.
Wadden TA, Bartlett S, Letizia KA, Foster GD & Stunkard AJ (1992):
Relationship of dieting history to resting metabolic rate, body
composition, eating behavior and subsequent weight loss. Am. J.
Clin. Nutr. 56 (Suppl), 203S–208S.
Wasito E, Pritasari, Susilowati D, Iswarawanti DN, Schultink W &
Gross R (2001): Temporary stability of urban food and nutrition
security: The East Jakarta study. Asia Pacific J. Clin. Nutr. 10 (Suppl),
S29–S33.
Wolfe WS & Frongillo EA (2001): Building household food
security measurement tools from the ground up. Food Nutr. Bull.
22, 5–12.
Zaitun Y & Terry RD (1990): Dietary patterns of rural elderly in
Malaysia. Ecol. Food. Nutr. 24, 213–221.
Zalilah MS (1998): Growth status determinants of school age
children from primarily low-income households in the urban area
of Kuala Lumpur, Malaysia: a focus on intrahousehold factors. PhD
Dissertation, Michigan State University.
Zalilah MS & Ang M (2001): Assessment of food insecurity among
low income households in Kuala Lumpur using the Radimer/
Cornell food insecurity instrument—a validation study. Mal. J.
Nutr. 7, 15–32.
Zalilah MS & Tham KBL (2002): Food security and child nutritional
status among Orang Asli (Temuan) households in Hulu Langat,
Selangor. Med. J. Mal. 57, 36–50.