Original Article

Food and Nutrition Bulletin

1-17
ª The Author(s) 2016
Urban Household Reprints and permission:
sagepub.com/journalsPermissions.nav
Characteristics and Dietary DOI: 10.1177/0379572116631882
fnb.sagepub.com

Diversity: An Analysis of Food

Security in Accra, Ghana

Samuel Nii Ardey Codjoe, PhD1, David Okutu, MPhil1,

and Mumuni Abu, PhD1

Abstract
Background: The world’s population is increasingly becoming urbanized. If the current urban growth
rate is to continue, new and unprecedented challenges for food security will be inevitable. Dietary
diversity has been used to ascertain food security status albeit at the multicountry and country levels.
Thus, household-level studies in urban settings, particularly in sub-Sahara African, are few. Yet, it is
imperative that assessments of food security are undertaken particularly in urban settings, due to the
projected fast rate of urbanization and the challenges of attaining food security.
Objective: To examine household characteristics and dietary diversity.
Methods: The study uses data from 452 households from the second round of the Regional Institute
for Population Studies (RIPS) EDULINK urban poverty and health study. Bivariate and multivariate
analyses are undertaken.
Results: Mean dietary diversity for all households is 6.8. Vegetables have the highest diversity, fol-
lowed by cereal-based and grain products. Household characteristics that have statistically significant
associations with dietary diversity include sex and level of education of household head, household
wealth quintile, and source of food.
Conclusions: There is high dietary diversity in the study communities of Accra but low consumption
of foods rich in micronutrient, such as fruits and milk/dairy products. The study brings to fore issues
related to resource-disadvantaged entities of the urban system, namely, females, poor households, and
the non-educated who have food insecurity problems.

Keywords
household characteristics, dietary diversity, urban, food utilization, food security, Accra, Ghana

Introduction
1
The world’s population is increasingly becoming Regional Institute for Population Studies, University of
urbanized as a result of both natural increase and Ghana, Legon, Ghana
rural–urban migration. It is projected that by the
Corresponding Author:
year 2025, 61% of the world’s population will be Samuel Nii Ardey Codjoe, Regional Institute for Population
living in urban areas, and sub-Saharan Africa will Studies, University of Ghana, PO Box LG 96, Accra, Ghana.
particularly experience a very fast urban growth.1 Email: [email protected]
2 Food and Nutrition Bulletin

As a result, it is projected that many people will prepared or convenience foods (mainly from
live close to or even below the poverty line, mak- street vendors, restaurants, and kiosks) with high
ing it a challenge to feed the ever-expanding energy density but with potentially low micronu-
urban population. Urban growth rates, projected trients.25-28 In addition, a filthy environment,
to double in less than 20 years, may compound chocked gutters, and refuse dumping near places
the problems of limited space for constructions where foods are prepared and sold render urban
and farming.2,3 If this urban growth rate is to dwellers susceptible to diarrheal diseases. Thus,
continue, new and unprecedented challenges for gastro-intestinal infections are known to be more
food security will be inevitable, especially in associated with the high levels of bacterial con-
urban poor places. tamination of street foods than of homemade
Over the past decades, numerous studies have foods.29
used a dietary diversity measure to ascertain food Reasons attributed to the high patronage of
security status because the consensus is that data out-of-home foods in urban areas are that a
are easy to collect and these have been a reliable greater number of women in the workforce have
proxy indicator of quality and quantity of diet in less time to prepare meals for the family, due to
different geographic settings.4-14 Some of the the substantial amount of time spent commuting
studies have even gone a step further to link diet- to and from work. In addition, living spaces are
ary diversity to outcome variables including birth smaller and often not equipped with kitchens or
weight, child anthropometric status, hemoglobin outdoor cooking spaces, and there is decreased
concentration, hypertension, cardiovascular dis- access to natural fuel sources. Furthermore,
eases, cancers, and obesity.15-23 However, these access to urban foods depends largely on cash
studies have been mainly at multicountry levels, exchange, with a few exceptions where some
although other country-level studies also exist, urban households engage in a little backyard
and they have provided platforms for monitoring farming for household consumption, which is not
changes in dietary energy availability. the case in shanty towns.25,30 For instance, urban
Thus, specific studies in urban settings, and at residents in Mozambique purchase 83% of the
the household level, are few at the global level, food they consume,31 and in Accra, families spend
and even fewer in sub-Sahara African cities. Yet, an average of 54% of their income on food.25 A
it is imperative that assessments of food security number of studies have shown that urban diets are
are undertaken particularly in urban settings, due strongly influenced by prices and incomes, but
to the projected fast rate of urbanization and the they are also influenced by lifestyles, social rela-
challenges of attaining food security in urban set- tionships, marriage patterns, family structures, the
tings outlined in the next section of this article. availability of packaged and processed foods, and
Our main objective is to examine household char- advertising and the media.32
acteristics and dietary diversity as a measure of Processed and packaged foods are more
food security in poor communities of Accra. widely available in urban than in rural areas, in
Accra is selected not only because of very few part because food manufacturing sectors are
studies but also because the study communities based nearby. Urban markets are more likely to
used are located close to major food markets and contain imported food items. The urban poor are
populations in these areas may have a wide range particularly vulnerable to price changes due to
of food items to choose from. limited cash and cash reserves. Wholesale food
markets and discount supermarkets are mostly
located in the city outskirts. These locations may
Justification for Studying Food Security in an
not be favorable to urban poor due to lack of
Urban Context transportation and inadequate municipal trans-
Pothukuchi and Kaufman 24 have stated the portation systems.9 The urban poor are often
importance of placing the food system on the obliged to purchase food in small neighborhood
urban agenda. Due to the fast-paced urban life- shops, which are more expensive and lack fresh
style, many urban consumers rely on already produce.
Codjoe et al 3

Finally, most urban poor households neither hemoglobin concentrations,19 reduced incidence
have large food stores nor do they have access of hypertension,20 the risk of mortality from car-
to areas for their own production due to over- diovascular diseases and cancer,21 and obesity. In
crowding. By and large, even though there may addition, dietary diversity scores can be used as
be a wide variety of foods on the urban markets, it part of food security and nutritional information
is not necessarily of diverse superior nutritional systems, at the regional and national levels, and
values and may not be socially accepted foods.7 can provide an early warning system for program
targeting and interventions. At the community
level, dietary diversity scores can be used in eval-
Conceptualizing Food Security and Dietary
uating programs that aim at improving food secu-
Diversity rity and nutrition of local communities.
Food security exists when ‘‘all people, at all
times, have physical or economic access to suffi-
Conceptualizing Household Characteristics
cient, safe and nutritious food to meet their diet-
ary needs and food preferences for an active and
and Dietary Diversity
healthy life.’’33 There are 3 fundamental compo- First, older household heads are more likely to
nents of food security: food access (affordability, prefer food prepared at home than younger ones
allocation, and preference), food availability who may buy from the roadside to feed their
(production, distribution, and exchange), and households. On the other hand, younger house-
food utilization (nutritional value, social value, hold heads may have the strength to work and
and food safety), and a fourth, food stability, was provide variety of food for their households com-
added by Clover.34 This can therefore be viewed pared to older household heads. Second, members
as food systems that have a much broader consid- of households headed by females are more likely
eration than just productivity and production to eat home-cooked food and also have food
alone.35-41 diversity since per the culture of the study area,
Millions of people worldwide suffer from hun- it is the responsibility of women to cook for their
ger and undernutrition, the majority being women families. Third, regarding household size, a
and children.6 This is particularly attributed to study has shown that increasing household size
food insecurity, which continues to be a challenge means more mouths to feed, increased expendi-
not only in developing countries but also in devel- ture on food, and thus reduced levels of con-
oped countries.5 The concept of food security has sumptions on quantity, quality, and variety.45
long evolved since the World Food Summit of Thus, smaller households are more likely to have
197442 and has shifted from global and national more diverse food consumption compared to
levels to household and individual levels and larger households.
from the concept of food availability to food Fourth, wealthier households are expected to
accessibility.6,43 The right to food remains one have the resources to purchase more food and
of the most consistently mentioned international diverse ones compared to poor households. Fifth,
human rights to date. However, unfortunately, it a higher educational level is associated with bet-
is also one of the most violated in recent times.34 ter employment opportunities and higher incomes
Dietary diversity at the individual or house- and may translate into higher purchasing power
hold level is an important element of food utiliza- and better nutrition knowledge for all household
tion. It entails the consumption of foods with members through improved dietary diversity.7
basic macronutrients and micronutrients and thus Sixth, wealthier study localities are more likely
relates to diet variety or balance.4,7-9,44 It is to have better dietary diversity.
important to study dietary diversity because pre- Ethnicity is important in a dietary diversity
vious research have shown that it is a valid out- analysis, since pregnant Ewe women in Ghana,
come in itself and either directly or indirectly may for example, abhor eating okro, snails, and ripe
lead to improved health outcomes such as birth plantain. Reasons provided are that it may cause
weight, 15 child anthropometric status, 16-18 miscarriage and children born may become
4 Food and Nutrition Bulletin

flabby, which could delay walking.35,46 Among partially built environments.52 The communities
the Dagaare of northern Ghana, eating of baobab used in this study are Jamestown, Usshertown,
leaves by pregnant women is forbidden,47 and and Agbogbloshie. Jamestown and Usshertown
there is fear of giving protein products such as are large neighborhoods of the Ga Mashie area
eggs and meat to children since it may instigate of Accra, whereas Agbogbloshie, on the other
them to steal.48 All these may have consequences hand, lies in the suburbs of Accra and has one
for limiting the food variety of household of the largest markets in the Accra metropolitan
members. area. It is a densely populated, permanent com-
Religious doctrines are significant in examin- munity with some residents tracing their roots
ing what is considered consumable and noncon- back to the early 1960s.53 The people in Agbog-
sumable foods and may therefore have impacts on bloshie mainly live in kiosks and makeshift woo-
dietary diversity in households. For instance, den shacks, and others live in uncompleted
Moslems do not eat pork, and among some Chris- building after carrying out minor structural work
tians, olive oil is only used for religious purposes. (eg, patching up uncompleted windows, doors,
Furthermore, the kind of work household and ceilings with scrap wood, aluminum, poly-
heads are engaged in determines the incomes, thene, and netting). The structures lack basic
access to resources and other social services, and amenities such as pipe-borne water, sanitation
access to food in terms of affordability.25,49 The (bathrooms and toilet), and electricity.54
occupation household heads are engaged in may
prevent them from having enough time to cook
food or even monitor the food given to younger Data and Methods
household members. As a result, the households
may resort to buying food from outside as a mat-
Data
ter of convenience. The study uses data from the second round of the
The study further hypothesizes that house- Regional Institute for Population Studies (RIPS)
holds that receive support and transfers are more EDULINK urban poverty and health study, con-
likely to use them to complement household ducted in November and December of 2011. Two
resources and therefore have a higher dietary sets of questionnaires were administered, one at
diversity compared to households that do not the household level (actual household or de facto
receive support and transfers. In addition, it is heads) and the other at the individual level
conceptualized that households with heads that (household members who were eligible, ie,
perceive that they have sufficient quantity and 15-49 years old for females and 15-59 years old
quality of food should have higher dietary diver- for males). All study participants signed or
sity compared to households’ heads perceive oth- thumb-printed an informed consent form before
erwise. Furthermore, households where members participating in the study. The information on the
avoid the eating of any food due to taboos are consent form was read to each volunteer and
expected to have lower dietary diversity, and those who agreed to participate in the study either
households where food is mainly prepared at signed or thumb-printed depending on their level
home are more likely to have higher dietary of education. In the case of participants who
diversity compared to households where food is could not read, a witness was made to sign in
mainly purchased. Finally, households with addition to the thumbprint of the participant to
higher total monthly expenditure on food are indicate that the project was explained to the par-
expected to have higher dietary diversity.50 ticipant and he/she agreed to participate in the
research.
This study considers the nutritional value of
Study Area food measured by a proxy variable (household
The city of Accra with a population of 1 848 614 dietary diversity [HDD]) to ascertain household
in 201051 is a sprawling metropolis with rapid nutritional adequacy after the studies by Food
regeneration of new suburbs and pockets of and Agriculture Organization (FAO) and
Codjoe et al 5

Kennedy et al. 6,9 The analysis is based on doing the actual provision of food and other sup-
responses from the actual heads of households who port). The de facto household head is the person
responded to the food list section in the individual who acts the role of household head in the (tem-
questionnaire. The food frequency questionnaire porary) absence of the actual household head.
used to measure dietary diversity was administered
only to eligible individual respondents.
Measuring Dietary Diversity and Frequency
The RIPS EDULINK study has a total sample
of 806 households distributed across 29 enumera- of Food Consumption
tion areas. A new sample size of 452 households Following Ruel10 and Kennedy et al,9 dietary
is obtained after eliminating households with diversity in this study is a measure of the number
heads younger than 15 years for males and of individual unique food groups consumed by
females and older than than 49 years for females households over a 7-day period preceding the
and 59 years for males. These household heads survey, for both foods prepared at home and out-
were eliminated because the selection criterion side the home, irrespective of the time of meals
for individual participants to which the food fre- (breakfast, lunch, or supper). Usually, the refer-
quency questions were administered did not ence period varies depending on the interest of
include individuals in these ages. the study, but generally the previous day or week
The smaller units in compound households are is used.6,44
the basis of the definition of a household for the Nine different standardized food groups are
survey, since they were defined largely in terms scored to generate the HDD score ranging
of consumption and were comparable in terms of between 0 and 9 using Food and Nutrition Tech-
composition to households in other areas of the nical Assistance Project7,8 and the Radimir/Cor-
city, as stated by Maxwell et al.25 The definition nell scale.56 There is no international consensus
of a household is a group of people who eat from on which food groups to include in the scores, and
the same pot when food is prepared at home and the results of new research could justify changing
who get money for street food from the same the groups proposed in these guidelines. Each
source. Although these households depend heav- food group has a range of food items for which
ily on cash exchange and reciprocity, overcrowd- the household head was to mention the frequency
ing means that some people, while not accurately of eating and where it was eaten (home, local
categorized as ‘‘homeless’’ in the usual sense of restaurant, street vendor, fast-food joint, and res-
the word, do not have a permanent place where taurant) in the last 7 days preceding the survey.
they sleep. They may sleep in shifts, and people In all, households reported 76 food types that
who are part of a household may not sleep with are categorized into 9 different standardized food
that household at all.55 groups. The food groups and types as shown in
For this reason, especially for the kind of Table 1 include the following: (1) cereal-based
households that predominate Accra’s migrant and grain products; (2) starch, roots/tubers, and
communities, coresidence does not define these plantain-based staples; (3) fats and oil; (4) meat/
households, and because of the heavy reliance on fish/eggs/poultry; (5) processed foods; (6) soft
consumption of street foods, consumption from a drinks/sweets/sugar; (7) milk and dairy products;
common pot alone is not an adequate definition of (8) fruits; and (9) vegetables. Alcoholic bev-
a household.25 In addition, the fact that children erages are excluded from the analysis since not
often eat somewhere else outside the home of all household members derived utility from it.
their mothers needs to be taken into account when The food groups are adapted from the FAO’s49
defining a household in this context. food composition tables for West Africa.
Household members are identified with refer- Household dietary diversity is calculated by
ence to the head of the household who is defined the summation of the number of times different
as the person with primary responsibility to see food items under each food group was eaten. This
that members are provided for in terms of food is based on the assumption that every household
and care (even though the person may not be should be able to consume a particular food group
6 Food and Nutrition Bulletin

Table 1. Food Groups and Food Types. with the assumption that what a household con-
sumes over a week is more likely to be the food
Food Group Food Type
consumption habit of the household. A check is
Cereal-based and grain Millet, maize, rice, oats also conducted to ensure the summation of the
(porridge and staple scores fall between 0 and 9, and a score closer
based) to 0 indicates insufficient diets, whereas a score
Starch roots/tubers and Yam, plantain, cassava, closer to 9 implies high dietary diversity. There
plantain cocoyam
are no established cutoff points in terms of num-
Fats and oils Red palm oil, vegetable oil,
butter, margarine, lard, ber of food groups to indicate adequate or inade-
animal fat quate dietary diversity, although some studies
Meat/fish/eggs/poultry Mutton, beef, pork, suggest HDD score less than 4 represents poor
chicken, duck, guinea dietary diversity, 4 as medium, and 5 or more is
fowl, tuna, herring, an indication of high dietary diversity.8,9 At the
salmon, crab, lobster, bivariate level of the analysis, the number of
shrimp, egg, sausage
times food groups is consumed is divided by 7
Processed foods (baked, Bran/wheat bread, sugar/
roasted, boiled, and tea/butter bread, meat (number of days) to give an average number of
fried) pie, cake, doughnut times food groups are consumed on a daily basis.
Soft drinks/sweets/sugars Fanta, Coca-Cola, Pepsi This is done to ascertain the food groups that are
Cola, etc most or least consumed.
Milk and dairy Milk, yoghurt, ice cream,
cheese
Fruits Mango, apple, grape, Measuring Household Characteristics
avocado pear, orange,
pineapple, banana, Age of household head is categorized as less than
pawpaw, watermelon, 20 years to 50 years and above in 10-year inter-
coconut vals and later entered in the regression model as a
Vegetables (include both Leafy vegetables, cabbage, continuous variable. Sex of household head is
cooked [stews, soups, carrot, eggplant, okra, either male or female. Household size is categor-
and sauces] and raw tomato, onion ized as single member, 2 to 3 members, 4 to 5
[salads])
members, and 6 or more members and entered as
a continuous variable in the regression model.
The construction of the wealth quintile
at least once in the last 7 days. This means that if a included all household assets and utility services
household consumes different food items from rather than a section of items. This broad criter-
different food groups for the 7 days preceding the ion, with its greater number of indicator variables,
survey, then that household would have con- improves the distribution of households, with
sumed from 7 food groups for the week. It is also fewer households being concentrated on certain
possible for a household to consume from 2 or index scores.57 All variables included in the index
more different food groups in a day, and this was are dichotomized. The next step in the index con-
taken into consideration in the computation of the struction uses the principal component analysis to
HDD score. A new food group is created for those calculate an index score.
food items that needed to be aggregated—an Using this method, the indicator variables are
example being starchy roots and tuber-based sta- standardized (calculating z-scores), then the fac-
ples as one of the food groups. tor coefficient scores (factor loadings) are calcu-
The HDD score is then computed by summar- lated, and finally, for each household, the
izing all the scores from the 9 different food indicator values are multiplied by the loadings
groups created. The focus here was to have the and summed to produce the household’s index
total score for the different food groups over the value. In this process, only the first of the factors
week without considering the specific days that produced is used to represent the wealth index.
households consumed those foods. This was done The resulting sum is itself a standardized score
Codjoe et al 7

with a mean of 0 and a standard deviation of 1. survey. The amount is calculated based on the
The index score is then divided into 5 cut points currency conversion rate at the time of survey
(20 percentiles) in order to generate the house- (1.00 USD ¼ 1.57 New Ghana Cedi [GH¢]). In
hold wealth index ranging from the poorest to the the regression model, it is treated as a continuous
richest quintiles. variable.
The educational level of household heads is
categorized as heads with no education, primary, Statistical Analysis
middle/junior high school, secondary/senior high
school (SHS), and higher education. Statistical analyses are undertaken with SPSS for
Localities used in the study are Usshertown, Windows (version 20.0). Descriptive characteris-
Jamestown, and Agbogbloshie. Ethnic groups of tics of variables are assessed by means or percent-
household heads include Akan, Ga-Dangme, age distribution in order to highlight the
Ewe, Guan, Gruma, Mole-Dagbani, Grussi, important differences across all household char-
Mande, and other minor ethnic groups. However, acteristics. Bivariate analyses are undertaken to
Grussi, Gruma, Guan, and Mande are combined assess frequency of food consumption and dietary
at the bivariate and multivariate levels of analy- diversity by household characteristics. Finally, a
sis. Religious affiliation is categorized as no reli- linear regression model is used to investigate the
gion, Catholic, Protestant, Pentecostal/ relationship between household characteristics
Charismatic, other Christian, Islam, traditional/ and dietary diversity.
spiritualist, and other. Occupation of household
heads includes none, professional/technical/man-
agerial/clerical, agricultural self-employed, sales, Results and Discussion
household and domestic, service, skilled manual,
unskilled manual, and other occupations. Descriptive Statistics
Furthermore, support and transfers in the last Table 2 shows that the mean age of household
30 days preceding the survey are categorized as heads is 36.2 years, and 60.2% and 39.8% of
no support, money only, in-kind, and both money household heads are males and females, respec-
and in-kind. In the regression model, no support is tively. The mean household size is 2.9 persons,
categorized as households that did not receive any with 38.1% being single-member households.
support and the others are classified as house- Furthermore, slightly more than one-third
holds that received support. Perceptions of house- (39.8%) of the households is in the poor and the
hold heads are sought on sufficiency or otherwise poorest quintiles, and 21.2% and 17.3% are in the
of food quantity and quality. The responses are in rich and richest wealth quintiles. Although 6.5%
the affirmative (Yes) or negative (No). House- of heads of households have no formal education,
hold heads are also asked whether their household 5.2% have higher education.
members avoid the eating of any food due to More than half (51.0%) of the sampled house-
taboos. The responses are also in the affirmative holds is in Usshertown, followed by Jamestown
or negative. In addition, source of food eaten in (29.0%) and Agbogbloshie (20.0%). As expected,
the household is categorized as (those) prepared most (55.0%) of the household heads are Ga-
at home and (those) prepared outside the home Dangme, followed by Akan (28.9%). Further-
(street vendors, chop bars, fast-food joints, and more, more than one-third (37.4%) of household
restaurants). For each source, the mean number heads are Pentecostals/Charismatic, and 7.5% has
of times food was consumed is computed for the no religion. With regard to occupational status, a
various food groups and entered in the regression relatively larger proportion (35.8%) of household
model as a continuous variable. heads is involved in sales, and 14.8% engage in
Finally, household food expenditure is com- professional/technical/managerial or clerical
puted as a percentage of total household monthly work.
expenditure on food and the overall household Overall, 69.9% of all households did not
expenditure in the last 30 days preceding the receive any form of financial or in-kind support
8 Food and Nutrition Bulletin

Table 2. Percentage Distribution of Households by Characteristics.a

Variable Percentage Variable Percentage

Age of household head Religious affiliation of household head

Less than 20 years 1.1 No religion 7.5
20-29 years 26.1 Catholic 5.8
30-39 years 34.3 Protestant 22.8
40-49 years 30.1 Pentecostal/Charismatic 37.4
50 years and above 8.4 Other Christian 10.6
Mean age (years) 36.2 Islam 12.6
Sex of household head Traditionalist/spiritualist 2.4
Male 60.2 Other 0.9
Female 39.8 Occupation of household head
Household size None 6.6
1 38.1 Professional/technical/managerial/clerical 14.8
2-3 27.0 Agricultural self-employed 3.8
4-5 24.3 Sales 35.8
6 or more 10.6 Household and domestics 1.3
Mean household size 2.9 Service 9.3
Household wealth quintile Skilled manual 14.1
Poorest 18.8 Unskilled manual 10.2
Poor 21.0 Other 4.2
Middle 21.7 Support and transfers
Rich 21.2 No support 69.9
Richest 17.3 Money only 21.0
Level of education of household head In-kind only 0.9
No education 6.5 Both money and in-kind 8.2
Primary 21.1 Perception of sufficient quantity of food
JHS/middle 45.0 Yes 64.8
SHS/secondary 22.2 No 35.2
Higher 5.2 Perception of sufficient quality of food
Locality of household Yes 69.7
Usshertown 51.0 No 30.3
Jamestown 29.0 Food avoidance due to taboo
Agbogbloshie 20.0 Yes 0.6
Ethnicity of household head No 99.4
Akan 28.9 Source of food
Ga-Dangme 55.0 Home 46.6
Ewe 6.0 Out of home 53.4
Guan 0.2 Household monthly food expenditure
Mole-Dagbani 3.5 No expenditure 13.4
Grussi 0.4 1%-19% 5.5
Mande 0.2 20%-49% 26.1
Other 5.8 50%-79% 37.8
80% and above 17.2
Mean proportion of income on food (%) 50.0
Mean household expenditure (GH¢) 572.08

Abbreviations: JHS, junior high school; SHS, senior high school.

a
Adapted from Fieldwork (2011).

within the reference period. Of the 30.1% that household heads perceive that they have suffi-
received support, 21.0% was in the form of cient quantity of food, 69.7% perceive that they
money only, whereas 8.2% received both money have sufficient quality of food. In addition, 99.4%
and in-kind. Furthermore, although 64.8% of of household heads indicated that they do not
Codjoe et al 9

Table 3. Household Dietary Diversity Score by Food head had statistically significant association with
Category.a dietary diversity, and the rest of the independent
Food Groups All
variables were not significant.
As shown in Table 4, households with heads in
Cereal-based and grain products 3.4 their 30s have the highest dietary diversity (7.2) and
Starchy roots/tuber and plantain-based staples 1.3 households with heads in the 30- to 39-year age
Fats and oils 1.3 category have the highest frequency of consumption
Meat/fish/eggs/poultry 2.6
for all food categories. Households with male heads
Processed foods (baked/roasted/boiled/fried 2.4
snacks) have a higher dietary diversity (7.5) compared to
Soft drinks/sweets/sugar 1.2 households with female heads (6.8). However,
Milk and dairy products 0.7 households with female heads have the highest fre-
Fruits 2.2 quency of consumption for all food groups.
Vegetables (cooked and raw) 3.8 Single-member households have the highest
Mean household dietary diversity score (HDD 6.8 dietary diversity (7.0), and as expected, house-
score)
holds in the richest wealth quintile have the high-
Total number of households 452 est dietary diversity (7.4). In addition, households
Abbreviation: HDD, household dietary diversity. with heads having higher education have the
a
Adapted from EDULINK/AARC data (2011). highest dietary diversity of 7.1, and consumption
frequency of all food categories is highest among
abhor any food due to cultural taboos. Regarding households with heads having primary education.
source of food, 53.4% of the households consumed Regarding locality of residence, Usshertown (6.9)
foods prepared outside the home, and 46.6% con- and Jamestown (6.8) are generally more advan-
sumed foods prepared at home. The analysis fur- taged than Agbogbloshie (6.3) in terms of dietary
ther shows that 13.4% of households have no diversity and consumption frequency of food
expenditure on food and depend on their harvests categories.
from fishing and farming. The mean proportion of In terms of ethnicity, households with heads
income on food is 50%, and the mean monthly that belong to the Mande ethnic group have the
household food expenditure is GH¢ 572.08. highest dietary diversity (8.0), followed by
Table 3 shows that the mean HDD for all Gruma (7.0) Akan (6.9), and Ga-Dangme (6.9).
households is 6.8, which suggests that on average, Furthermore, households with heads who are Ga-
every household consumed almost 7 different Dangme have the highest consumption frequency
food groups in the 7-day period preceding the for all the food categories. The results further
survey. Vegetables have the highest diversity indicate that households with heads who are of
(3.8), followed by cereal-based and grain prod- the category ‘‘other Christians’’ have the highest
ucts (3.4). Consumption of foods rich in micro- dietary diversity of 7.2, however, households
nutrient such as fruits and milk/dairy products with heads having no religion have the highest
was relatively low, and this could be due to high consumption frequency for all the food
prices attached to such commodities,58,59 lack of categories.
nutritional knowledge, and social factors such as Households with heads who have professional/
taste and preference. technical/managerial/clerical occupations have
the highest dietary diversity (7.5) and the highest
consumption frequency. As expected, households
Bivariate Analysis with heads who perceived that their households
This section presents the bivariate relationship have sufficient quality and quantity of food
between household characteristics and dietary recorded higher dietary diversities and consump-
diversity. A test of statistical association was per- tion frequencies. Table 4 also shows that house-
formed between all the independent variables and holds that mainly derive their meals cooked
the dependent variable. In all, only the age and outside have a higher dietary diversity, however,
the educational level attained by the household households that eat home-cooked meals have a
10 Food and Nutrition Bulletin

Table 4. Frequency of Food Consumption and Dietary Diversity by Household Characteristics.a

Baked/
Cereal- Meat/ Boiled/ Milk/ Mean
Based Tuber/ Fats/ Fish/ Fried Soft Diary HDD
Variable Staples Roots Oils Egg Snacks Drinks Products Fruits Vegetables Score

Ageb
Less than 30 3.3 1.1 1.2 2.5 2.2 1.1 0.9 2.1 3.4 6.9
years
30-39 years 5.3 2.2 2.2 3.9 4.3 2.1 1.6 4.4 6.1 7.2
40-49 years 1.8 0.6 0.5 1.4 0.8 0.3 0.2 0.2 1.9 6.7
50 years and 1.9 0.6 0.5 1.4 0.7 0.3 0.3 0.1 2.1 6.4
above
Sex
Male 3.2 1.2 1.1 2.4 2.1 1.0 0.7 1.8 3.4 7.5
Female 3.8 1.5 1.4 2.8 2.8 1.3 0.9 2.7 4.3 6.8
Household size
1 3.9 1.6 1.5 2.9 2.9 1.4 1.0 2.7 4.4 7.0
2-3 4.6 1.9 1.8 3.2 3.7 1.7 1.1 3.7 5.2 6.8
4-5 1.8 0.6 0.5 1.7 0.6 0.4 0.2 0.3 1.9 6.9
6 or more 1.9 0.5 0.5 1.5 0.9 0.4 0.2 0.1 1.8 6.4
Wealth quintile
Poorest 4.6 1.9 1.8 3.3 3.6 1.7 1.1 3.7 5.2 6.4
Poor 4.8 2.1 2.0 3.5 4.0 1.9 1.3 3.9 6.3 6.8
Middle 5.7 2.5 2.5 4.3 5.3 2.4 1.7 5.6 7.0 7.2
Rich 1.9 0.6 0.5 1.6 0.7 0.4 0.2 0.3 1.9 6.6
Richest 2.1 0.6 0.5 1.7 0.8 0.4 0.3 0.3 1.7 7.4
Level of educationc
No education 1.8 0.6 0.3 1.6 0.6 0.2 0.1 0.9 1.0 6.3
Primary 7.3 3.3 3.1 5.1 6.7 3.0 2.5 6.0 4.4 6.9
JHS/middle 1.9 0.6 0.5 1.5 0.7 0.4 0.3 1.0 0.4 6.8
SHS/ 3.6 1.4 1.4 2.8 2.5 1.3 0.9 2.7 2.4 6.8
secondary
Higher 2.0 0.5 0.5 1.5 0.6 0.4 0.2 1.3 1.0 7.1
Locality
Agbogbloshie 1.9 0.6 0.4 1.5 0.7 0.4 0.3 0.3 1.7 6.1
Jamestown 3.2 1.2 1.2 2.4 2.1 1.1 0.7 1.9 3.7 6.8
Usshertown 4.1 1.6 1.6 3.1 3.2 1.5 0.9 3.0 4.7 6.9
Ethnicity
Akan 2.9 1.3 1.2 2.3 2.1 1.0 0.7 1.9 3.4 6.9
Ga-Dangme 4.1 1.6 1.5 3.0 3.0 1.5 0.9 2.9 4.5 6.9
Ewe 1.9 0.5 0.5 1.3 0.5 0.6 0.2 0.4 1.7 6.5
Guan 0.9 0.0 0.0 0.6 0.0 0.6 0.0 0.0 0.3 4.0
Gruma 2.0 0.7 0.4 1.3 0.7 0.3 0.0 0.0 3.9 7.0
Mole-Dagbani 2.3 0.5 0.2 1.3 0.7 0.3 0.2 0.0 1.4 5.8
Grussi 1.0 0.1 0.0 0.9 0.0 0.0 0.0 0.0 0.4 4.0
Mande 2.4 0.4 0.4 1.7 1.1 0.1 1.0 0.0 1.6 8.0
Other 2.0 0.6 0.5 1.7 0.9 0.5 0.4 0.3 2.0 7.3
Religious affiliation
No religion 6.8 3.1 3.0 5.0 6.1 2.8 2.2 5.5 8.2 6.6
Catholic 1.9 0.5 0.3 1.4 0.8 0.3 0.3 1.0 1.5 6.5
Protestant 3.6 1.4 1.4 2.8 2.5 1.3 0.9 2.5 3.9 6.8
Pentecostal/ 3.8 1.6 1.5 2.9 2.9 1.4 1.1 2.9 4.5 6.9
Charismatic

(continued)
Codjoe et al 11

Table 4. (continued)
Baked/
Cereal- Meat/ Boiled/ Milk/ Mean
Based Tuber/ Fats/ Fish/ Fried Soft Diary HDD
Variable Staples Roots Oils Egg Snacks Drinks Products Fruits Vegetables Score

Other 2.0 0.5 0.6 1.7 0.8 0.3 0.3 0.4 2.0 7.2
Christian
Islam 2.0 0.5 0.5 1.5 0.8 0.6 0.4 0.3 1.8 7.0
Traditionalist/ 1.8 0.7 0.4 1.7 0.7 0.1 0.0 0.2 1.8 6.2
spiritualist
Other 1.4 0.2 0.2 0.7 0.8 0.8 0.0 0.0 0.8 5.9
Occupation
None 2.2 0.7 0.6 1.5 0.8 0.4 0.3 0.2 1.2 5.9
Professional 6.8 3.0 3.0 4.7 6.1 3.0 2.3 5.7 8.6 7.5
Sales 1.9 0.6 0.5 1.6 0.8 0.4 0.3 0.4 1.6 6.4
Agricultural 2.1 0.5 0.3 1.6 0.7 0.3 0.2 0.2 1.3 6.2
self-employed
Household and 1.9 0.5 0.5 1.7 0.6 0.3 0.1 0.0 0.6 6.8
domestics
Service 1.9 0.5 0.6 1.6 0.7 0.4 0.2 0.4 2.0 6.8
Skilled manual 4.7 2.0 2.0 3.5 3.8 1.8 1.5 3.7 3.2 7.1
Unskilled 6.0 2.5 2.3 4.1 4.9 2.3 1.8 4.6 3.7 6.6
manual
Other 1.8 0.3 0.4 1.6 0.6 0.4 0.2 0.0 0.6 6.5
Support and transfer
No support 3.6 1.4 1.3 2.7 2.5 1.2 0.8 2.3 4.0 6.9
Money only 3.5 1.4 1.4 2.6 2.7 1.3 0.8 2.5 4.0 6.7
In-kind only 1.5 0.7 0.6 1.6 0.5 0.4 0.1 0.0 2.4 6.5
Both money 2.1 0.6 0.5 1.7 0.7 0.5 0.3 0.1 1.5 6.7
and in-kind
Sufficient quantity of food
Yes 4.2 1.7 1.7 3.1 3.2 1.6 1.2 3.2 4.7 6.9
No 1.9 0.6 0.5 1.6 0.8 0.4 0.3 0.7 2.0 6.7
Sufficient quality of food
Yes 4.1 1.7 1.6 3.0 3.2 1.5 1.0 3.1 4.8 6.9
No 1.9 0.6 0.5 1.6 0.7 0.3 0.2 0.2 1.7 6.6
Source of food
Home 2.2 1.1 1.1 1.7 1.8 0.9 0.6 1.6 3.2 4.3
Outside home 1.2 0.2 0.2 0.9 0.5 0.3 0.1 0.7 0.9 5.0
Expenditure on food
No 2.1 0.6 0.4 1.7 0.7 0.4 0.2 1.9 3.2 6.8
expenditure
1%-19% 1.6 0.5 0.6 1.5 0.7 0.5 0.2 0.3 2.1 8.4
20%-49% 2.0 0.6 0.6 1.7 0.9 0.4 0.3 2.1 3.8 6.9
50%-79% 3.9 1.5 1.5 2.8 2.8 1.3 1.0 2.7 4.3 6.8
80% and above 6.1 2.9 2.7 4.3 5.6 2.6 2.1 0.2 1.8 6.2
Abbreviations: HDD, household dietary diversity; JHS, junior high school; SHS, senior high school.
a
Adapted from Fieldwork (2011).
b
P < .01.
c
P < .00.
12 Food and Nutrition Bulletin

Table 5. A Linear Regression Model Showing the Table 5. (continued)

Relationship Between Household Characteristics and Standardized Standard
Dietary Diversity.a,b Coefficients of b Error
Standardized Standard Unskilled manual 0.027 0.528
Coefficients of b Error Other 0.031 0.681
Intercept (a) 3.313c 0.627 Household support and transfers
Age of household head 0.022 0.012 Did not receive household support (RC)
Sex of household head Received household 0.033 0.251
Male 0.068d 0.800 support
Female (RC) Perception of sufficient quantity of food
Household size 0.034 0.055 Food quantity is not 0.060 0.270
Household wealth quintile sufficient
Poorest (RC) Food quantity is sufficient (RC)
Poor 0.102d 0.377 Perception of sufficient quality of food
Middle 0.022 0.425 Food quality is not 0.035 0.283
Rich 0.094 0.296 sufficient
Richest 0.149e 0.385 Food quality is sufficient (RC)
Level of education of household head Food avoidance due to taboo
No education (RC) No food taboo 0.059 0.276
Primary 0.075 0.492 Food taboo present (RC)
JHS/middle 0.094 0.474 Source of food
SHS/secondary 0.169d 0.517 Food eaten at home 0.231c 0.002
Higher 0.289e 0.688 (mean)
Locality of household Food eaten outside 0.764c 0.045
Jamestown 0.020 0.396 home (mean)
Usshertown 0.029 0.385 Household monthly 0.269 0.012
Agbogbloshie (RC) food expenditure
Ethnicity of household head Adjusted R2 ¼ .409
Ga-Dangme (RC) Abbreviations: JHS, junior high school; RC, reference category;
Akan 0.068 0.277 SHS, senior high school.
Ewe 0.047 0.488 a
Adapted from EDULINK/AARC data (2011). Sample size
Mole-Dagbani 0.000 1.670 (N) ¼ 451 excluding 1 missing case.
b
Grussi/Gruma/ 0.051 0.656 F-statistics ¼ 36.901,c and mean sum of squares ¼ 29.966.
c
Guan/Mande P < .01.
d
Other 0.025 1.386 P < .10.
e
Religious affiliation of household head P < .05.
Christian (RC)
No religion 0.099 1.349
Muslims 0.100 1.400 higher consumption frequency. Finally, house-
Traditional/spiritualist 0.000 1.522 holds that spend less than 20% of their household
Other 0.003 2.688
Occupation of household head
income on food have the highest dietary diversity,
No occupation (RC) and households that spend 80% or more of their
Professional/technical/ 0.033 0.482 income on food have the highest consumption
managerial/clerical frequency.
Agriculture 0.029 0.453
self-employed
Sales 0.053 0.713 Multivariate Analysis
Household and 0.008 1.123
domestics The model shows that 40.9% of the changes in
Service 0.052 0.546 HDD are explained by household characteristics.
Skilled manual 0.086 0.498 The F-statistics of the model is significant at 99%
confidence interval, indicating that the model is a
(continued)
good fit. Table 5 shows that the independent
Codjoe et al 13

variables that have statistically significant asso- hand, foods prepared outside the home are always
ciations with dietary diversity are sex of house- available and in varieties.
hold head, household wealth quintile, level of
education of household head, and source of food.
The remaining household characteristics were not
Conclusion
significant predictors of HDD. Male-headed As urban food demand rises, food supply and
households are likely to have 0.068 units more distribution systems have to supply the inhabi-
of dietary diversity compared to female-headed tants of cities with increasing amounts of food
households. This could be attributed to differ- coming from ever more distant production zones
ences in access to resources and income, which and/or from more intensive production systems.60
culminates in better purchasing power. Other Urban areas will provide the necessary incentives
studies have shown that female-headed house- for increased national production made possible
holds are more vulnerable to food insecurity, by more remunerative producer prices. The chal-
compared to male-headed households.25,35,49 lenges facing decision makers in the years to
Table 5 further shows that households in the come is how to meet the rapidly increasing urban
poor and richest wealth quintiles have higher diet- food demand, while reducing dependence on
ary diversity compared to households in the poor- imports, and achieving an efficient and dynamic
est wealth quintile. Households in the poorest distribution of nutritious foods at reasonable
quintile, because of their financial situation, are prices to the poorest sectors of urban populations,
unable to afford varieties of foods. However, while creating jobs in the food marketing and
households in the richest wealth quintile have distribution sectors.61
more financial resources for diverse foods. The Numerous studies in the past have used a diet-
analysis further shows that households with heads ary diversity measure to ascertain food security
having SHS and higher education have more diet- status, and others have gone on to link dietary
ary diversity compared to households with heads diversity to several outcome variables. However,
having no education. This may be due to the fact those studies have been mainly undertaken at the
that households whose heads have SHS/second- multicountry and country levels to the detriment
ary and higher education are those engaged in of the household. In addition, studies in urban
professional occupations, which are associated settings have been very few, especially in sub-
with better employment opportunities and subse- Sahara African cities. Yet, it is imperative that
quently higher incomes. Thus, those households assessments of food security are undertaken par-
have better purchasing power and can afford a ticularly in urban settings, due to the projected
variety of foods. Education may have also pro- fast rate of urbanization and the challenges of
vided these household heads with knowledge on attaining food security in urban settings. This arti-
nutritional benefits of a balanced diet. cle examined household characteristics and diet-
Finally, although households that eat home- ary diversity as a measure of food security in
prepared meals have less dietary diversity, house- Accra. The objective is to contribute to the dis-
holds that eat food prepared outside their homes cussion on dietary diversity and food security
have more dietary diversity. This could be due to within the context of rapid urbanization occurring
the fact that home-made foods may only be of a globally, but especially in sub-Saharan Africa.
specific type, and therefore, it limits a household The study shows that the mean dietary diver-
to the consumption of certain types of food. Due sity for all households is 6.8, which suggests that
to the busy nature of urban life, some household on average, every household consumed almost 7
heads stated that they preserve home-prepared different food groups in the 7-day period preced-
stews, soups, sauces, kenkey, banku, and fufu in ing the survey. Vegetables have the highest diver-
refrigerators for days and, in some instances, for sity (3.8), followed by cereal-based and grain
weeks. Until the stored foods are depleted, new products (3.4). Consumption of foods rich in
ones are not prepared. This practice limits the micronutrient, such as fruits and milk/dairy prod-
consumption of other food types.45 On the other ucts, was relatively low, and this could be due to
14 Food and Nutrition Bulletin

high prices attached to such commodities, lack of products while subsidizing the cost of locally pro-
nutritional knowledge, and social factors such as duced food commodities and infrastructural
taste and preference. improvements will go a long way in regulating
Results from a multivariate analysis show that distribution and market prices, making those
household characteristics that have statistically foods not only available but also affordable.
significant associations with dietary diversity In Accra, food consumption is deeply
include sex of household head, household wealth entrenched in people’s culture, making attempts
quintile, level of education of household head, to introduce new but more nutritious foods diffi-
and source of food. The analysis indicated that cult. Promoting a shift to richer natural organic
female-headed households, the poorest house- foods should be at the core of a long-term strat-
holds, and households with noneducated heads egy. Nevertheless, there is a further need for a
had lower dietary diversity compared to male- more rigorous sensitization on nutrition to
headed households, richest households, and address changes in consumption behaviors and
households with heads educated at the SHS/ to promote awareness of the dangers involved
higher levels. in dietary inadequacy or excessive consumption
The study recommends a few pointers for pol- of specific diets, such as obesity, cardiovascular
icy. First, a closer look at the significant variables diseases, and diabetes.
in the multivariate analysis brings to the fore Finally, further research should be undertaken
issues related to resource-disadvantaged entities to address the paucity of data on food consump-
of the urban system, namely, females, poor tion at the intrahousehold level, which will allow
households, and the noneducated. It is recom- for the development and testing of a more com-
mended that the Government of Ghana should prehensive, reliable, and valid measure of food
intensify efforts being made to improve the living security. Household dietary diversity measures,
conditions of females and poor households. Edu- if appropriately used alongside other components
cation has also been demonstrated here as in of food security measures such as food access and
numerous studies as a key socioeconomic indica- food availability, may yield a substantive proxy
tor of success or otherwise. Some members of measure to food security at individual, household,
these study communities of Accra with substan- community, and national levels.
tial proportions of migrants have peculiar chal-
lenges with education. This is due to the fact
that to be able to meet the heavy financial Acknowledgments
demands of city life as well as remunerating fam- The authors are grateful to the International Develop-
ily members back at home, these migrants engage ment Research Centre of Canada for providing funding
in the brisk business activities that the city busi- for this study through the Climate Change Adaptation
ness district offers to the detriment of their edu- Research Training Capacity for Development
(CCARTCD) Project.
cation. Thus, efforts being made to improve
educational levels in the country should also spe-
cifically target members of these migrant com- Authors’ Note
munities in the cities.
S.N.A.C., D.O., and M.A. contributed equally to the
Second, in the short run, there is a need to
conceptualization, analysis, and write-up of the manu-
encourage households to increase intake of other script. Any underlying research materials related to this
food groups, for example, fruits and milk/dairy article can be accessed from the lead author at
products other than concentrating on the con- [email protected].
sumption of cereal-based foods. This can be
achieved with an increase in the production and
supply of such food commodities (horticulture) in Declaration of Conflicting Interests
the long run rather than encouraging massive pro- The author(s) declared no potential conflicts of interest
duction of low-nutrient energy-dense foods. As with respect to the research, authorship, and/or publi-
well, regulating the influx of imported food cation of this article.
Codjoe et al 15

Funding 9. Kennedy G, Razes M, Ballard T, Dop M. Food-

The author(s) disclosed receipt of the following finan- based approaches for improving diets and raising
cial support for the research, authorship, and/or publi- levels of nutrition: a simple tool for developing
cation of this article: This study received funding and evaluating food and nutrition security inter-
through the Climate Change Adaptation Research ventions. 2010.
Training Capacity for Development (CCARTCD) Proj- 10. Ruel MT. Operationalizing dietary diversity: a
ect of the International Development Research Centre review of measurement issues and research
(IDRC) with Component Number: 106548-001. priorities. J Nutr. 2003;133(11 suppl 2):
3911S-3926S.
References 11. Savy M, Martin-Prével Y, Sawadogo P, Kameli Y,
1. UN Population Division. World Urbanisation Delpeuch F. Use of variety/diversity scores for diet
Prospects: the 2001 Revision. March 20 2002. quality measurement: relation with nutritional sta-
Web site. http://www.un.org/esa/population/publi tus of women in a rural area in Burkina Faso. Eur J
cations/wup2001/WUP2001_CH1.pdf. Accessed Clin Nutr. 2005;59(5):703-716.
December 14, 2015. 12. Steyn NP, Nel JH, Nantel G, Kennedy G, Labadar-
2. Food and Agriculture Organization. Identifying ios D. Food variety and dietary diversity scores in
and addressing food security among the urban children: are they good indicators of dietary ade-
poor: summary of the FSN discussion, No 35. quacy? Public Health Nutr. 2006;9(5):644-650.
2009 13. Rashid DA, Smith L, Rahman T. Determinants of
3. UN-Habitat. The state of the World’s cities, 2001. dietary quality: evidence from Bangladesh. Paper
Nairobi: United Nations Centre for Human Settle- presented at: American Agricultural Economics
ments; 2001:125. Association Annual Meeting; July 23-26, 2006;
4. Food and Agriculture Organization. International Long Beach, CA. Web site. http://ageconsearch.
workshop. Simple tools for measuring household umn.edu/bitstream/21326/1/sp06ra11.pdf.
access to food and dietary diversity (final report). Accessed June 20, 2012.
Nairobi: FAO, EC/FAO Programme Food Insecur- 14. Wiesmann D, Hoddinott J, Aberman NL, Ruel M.
ity Information for Action, and FANTA; 2007. Review and validation of dietary diversity, food
Web site. http://www.foodsec.org/tr_nut_05.htm. frequency, and other proxy indicators of household
Updated 2011, Accessed July 30, 2012. food security. Report submitted to the World Food
5. Food and Agriculture Organization. Report on use Programme, Rome, by International Food Policy
of the household food security access scale and Research Institute. 2006.
household dietary diversity score in two survey 15. Rao S, Yajnik CS, Kanade A, et al. Intake of
rounds in Manica and Sofala provinces, Mozambi- micronutrient-rich foods in rural Indian mothers
que, 2006-2007. FAO, Food Security Project GCP/ is associated with the size of their babies at birth:
MOZ/079/BL. Version 2. 2008. Pune Maternal Nutrition Study. J Nutr. 2001;
6. Food and Agriculture Organization. Guidelines for 131(4):1217-1224.
Measuring Household and Individual Dietary 16. Allen L, Black A, Backstrand J, Pelto G, Ely R,
Diversity. Rome, Italy: Food and Agriculture Molina E, Chavez A. An analytical approach for
Organization of the United Nations; 2011. exploring the importance of dietary quality versus
7. Hoddinott J, Yohannes Y. Dietary Diversity as a quantity in the growth of Mexican children. Food
Food Security Indicator. Washington, DC: Food Nutr Bull. 1991;13(2):95-104.
and Nutrition Technical Assistance Project, Acad- 17. Hatløy A, Hallund J, Diarra MM, Oshaug A. Food
emy for Educational Development; 2002. variety, socioeconomic status and nutritional sta-
8. Swindale A, Bilinsky P. Household Dietary Diver- tus in urban and rural areas in Koutiala (Mali).
sity Score (HDDS) for Measurement of Household Public Health Nutr. 2000;3(1):57-65.
Food Access: Indicator Guide, Version 2. 18. Onyango A, Koski K, Tucker K. Food diversity
Washington, DC: Food and Nutrition Technical versus breastfeeding choice in determining anthro-
Assistance Project, (FANTA), Academy for Edu- pometric status in rural Kenyan toddlers. Int J Epi-
cational Development (AED); 2006. demiol. 1998;27(3):484-489.
16 Food and Nutrition Bulletin

19. Bhargava A, Bouis H, Scrimshaw N. Dietary 31. Garrett J, Ruel M. Are determinants of rural and
intakes and socioeconomic factors are associated urban food security and nutritional status differ-
with the haemoglobin concentration of Banglade- ent? Some insights from Mozambique. World Dev.
shi women. J Nutr. 2001;131(3):758-764. 1999;27(11):1955-1975.
20. Miller W, Crabtree B, Evans D. Exploratory study 32. Atkinson S. Approaches and actors in urban food
of the relationship between hypertension and diet security in developing countries. Habitat Int.
diversity among Saba islanders. Public Health 1995;19(2):151-163.
Reports. 1992;107(4):426-432. 33. Food and Agriculture Organization. Rome
21. Kant A, Schatzkin A, Ziegler R. Dietary diversity Declaration and World Food Summit Plan of
and subsequent cause-specific mortality in the Action. Rome. Web site. www.fao.org/docrep/
NHANES I epidemiologic follow-up study. J Am 003/X8346E/x8346e02.htm#P1_10. Updated
Coll Nutr. 1995;14(3):233-238. 1996, Accessed December 10, 2015.
22. Oh SY, Hong MJ. Food security is associated diet- 34. Clover J. Food security in sub-Saharan Africa. Afr
ary intake and body size of Korean children from Secur Rev. 2003;12(1):1-11.
low-income families in urban areas. Eur J Clin 35. Codjoe SNA, Owusu G. Climate change/variabil-
Nutr. 2003;57:1598-1604. ity and food systems: evidence from the Afram
23. Arimond M, Wiesmann D, Becquey E, et al. Simple Plains, Ghana. Regional Environ Change. 2011;
food group diversity indicators predict micronutrient 11(4):753-765.
adequacy of women’s diets in 5 diverse, resource- 36. Downing TE. Linking sustainable livelihoods and
poor settings. J Nutr. 2010;140(11):2059S-2069S. global climate change in vulnerable food systems.
24. Pothukuchi K, Kaufman JL. Placing the food sys- Die Erde. 2002;133:363-378.
tem on the urban agenda: the role of municipal 37. Ziervogel G, Calder R. Climate variability and
institutions in food systems planning. Agriculture rural livelihoods: assessing the impact of seasonal
and Human Values. 1999;16(2):213-224. climate forecasts. Area. 2003;35(4):403-417.
25. Maxwell D, Levin C, Amer-Klemesu M, Reul M, 38. United States Agency for International Develop-
Morris S, Ahiadeke C. Urban livelihoods and food ment. Policy Determination: Definition of Food
and nutrition security in Greater Accra, Ghana Security, PD-19. Washington, DC: USAID; 1992:4.
Research report 112. Washington DC: Interna- 39. Ingram JSI, Gregory PJ, Brklacich M, eds.
tional Food Policy Research Institute; 2000. GECAFS science plan and implementation strat-
26. van’t Riet H, den Hartog AP, Mwangi AM, Mwa- egy. Volume 2. ESSP report. Wallingford; 2005.
dime RKN, Foeken DWJ, van Staveren WA. The 40. Gregory PJ, Ingram JSI, Brklacich M. Climate
role of street foods in the dietary pattern of two low change and food security. Philos Trans R Soc Lond
income groups in Nairobi. Eur J Clin Nutr. 2001; B Biol Sci. 2005;360(1463):2139-2148.
55:562-570 41. Ericksen PJ. Conceptualising food systems for glo-
27. Mazengo M, Simell O, Lukmanji Z, Shirima R, bal environmental change research. Global
Karvetti R. Food consumption in rural and urban Environ Change. 2008;18(1):234-245.
Tanzania. Acta Tropica. 1997;68(3):313-326. 42. Maxwell S, Smith M. ‘‘Household Food Security’’.
28. Mennen L, Mbanya J, Cade J, et al. The habitual Part I. In: Maxwell S, Frankenberger TR, eds.
diet in rural and urban Cameroon. Eur J Clin Nutr. Household Food Security: Concepts, Indicators,
2000;54(2):150-154. Measurements; A Technical Review. New York:
29. World Health Organization. Obesity: preventing UNICEF and IFAD; 1992.
and managing the global epidemic. WHO Techni- 43. Sen A. Poverty and Famine: An Essay on Entitlement
cal Report Series 894. 2002. and Deprivation. Oxford, UK: Clarendon Press; 1981.
30. Kennedy G.Food security in the context of the 44. World Food Programme. Comprehensive Food
urban sub-Saharan Africa: consultant, nutrition Security & Vulnerability Analysis Guidelines.
planning, assessment and evaluation service, food Rome, Italy: United Nations World Food Pro-
and nutrition division, Food and Agriculture Orga- gramme; 2009.
nization of the United Nations. Internet paper for 45. Anarfi JK, Ahiadeke C. Improving the Health of
food security theme. 2003. Children Among the Urban Poor in the City of
Codjoe et al 17

Accra, Ghana. Accra, Ghana: African Population 54. Codjoe SNA, Owusu G, Burkett V. Perception,
and Health Research Center, ISSER; 2006. experience and indigenous knowledge of climate
46. Badasu DM. Child care among the Ewe migrants change and variability: the case of Accra, a sub-
in the city of Accra: cases of crisis. Res Rev Suppl. Saharan African city. Reg Environ Change. 2014;
2004;16:17-37. 14(1):369-383.
47. Songsore J, Denkabe A. Challenging rural poverty 55. Fayorsey C. Ga women’s autonomy: a critique of
in Northern Ghana: The case of Upper West the concepts and economy of the household and
region. Report No. 6/95, The University of Trond- family. Afr Anthropol. 1995;2(1):91-130.
heim Centre for Environment and Development, 56. Kendall A, Olson CM, Frongillo EA. Relationship
UNIT-SMU. 1995. of hunger and food insecurity to food availability
48. Nanbigne E. Helping them grow teeth: care and and consumption. J Am Diet Assoc. 1996;96(10):
conflict among the Dagaba of Northern Ghana. 1019-1024.
Res Rev Suppl. 2004;16:77-83. 57. Rutstein SO, Johnson K. The DHS Wealth Index,
49. Levin C, Ruel M, Morris S, Maxwell D, Armar- DHS Comparative Reports No. 6. Calverton, MD:
Klemesu M, Ahiadeke C. Working women in an ORC Macro; 2004.
urban setting: traders, vendors, and food security 58. Mwaniki A. The utilisation of locally grown plant
in Accra. World Dev. 1999;27(11):1977-1991; also materials in the production of an intervention for-
available as IFPRI Food Consumption and Nutri- mulation for malnourished children in marginal
tion Division Discussion Paper 66. areas. The case of location Makueni district [mas-
50. Kennedy G, Nantel G, Shetty P. Globalization of ters’ thesis]. University of Nairobi; 2003.
food systems in developing countries: a synthesis 59. Ghana Shared Growth and Development Agenda
of country case studies. Dev Policy Rev. 2004; (GSGDA), 2010-2013. Medium Term and
21(5-6):2003. National Development Policy Framework: Gov-
51. Ghana Statistical Service. 2010 Population and ernment of Ghana, National Development Plan-
Housing Census. Summary Report of Final ning Commission (NDPC). Volume I. 2012.
Results. Accra, Ghana: Sakora Press; 2012. 60. Kennedy G, Nantel G, Shetty P; Food and Agri-
52. Owusu G, Afutu-Kotey RL. Poor urban commu- culture Organization of the United Nations. Glo-
nities and municipal interface in Ghana: a case balization of food systems in developing countries:
study of Accra and Sekondi-Takoradi metropolis. impact on food security and nutrition. FAO Food
Afr Stud Q. 2010;12(1):1-16. Nutr Pap. 2004;83:1-300.
53. Codjoe SNA, Afuduo S. Geophysical, socio- 61. Nakabo-Ssewanyana S. Food Security and Child
demographic characteristics and perception of Nutrition Status Among Urban Poor Households
flood vulnerability in Accra, Ghana. Natural in Uganda: Implications for Poverty Alleviation.
Hazards. 2015;77(2):787-804. doi:10.1007/ Nairobi, Kenya: African Economic Research Con-
s11069-015-1624-y. sortium; 2003.