Journal of Agriculture and Food Research 14 (2023) 100728

Contents lists available at ScienceDirect

Journal of Agriculture and Food Research

journal homepage: www.sciencedirect.com/journal/journal-of-agriculture-and-food-research

Urban households’ food safety knowledge and behaviour: Choice of food

markets and cooking practices
Makafui I. Dzudzor a, *, Nicolas Gerber b
a
Centre for Development Research (ZEF), University of Bonn, Genscherallee 3, D-53113, Bonn, Germany
b
Institute for Food and Resource Economics (ILR), University of Bonn, Nussallee 19, D-53115, Bonn, Germany

A R T I C L E I N F O A B S T R A C T

Keywords: Food safety is a public health issue and a shared responsibility of everyone. Traditional food markets can be high-
Food safety risk locations for the spread of foodborne diseases, especially in developing countries. The focus has been to
Food market improve institutional food handlers’ food safety knowledge and behaviour. However, the household is the last
Urban households
barrier to preventing the transmission of foodborne diseases. Households’ knowledge and proper behaviour
Knowledge
Cooking behavior
towards food safety in the home can improve their protection against foodborne diseases. Using household data
from the NOURICITY project on urban households in Ghana, the study sought to answer the primary research
questions, including; the main factors that influence urban households’ choice of food markets and the effect of
household food safety knowledge and wealth status on food safety cooking practices/behaviour. Structural
Equation Modelling (SEM) was applied to address these questions. The study results show that convenience
(proximity and availability of all products at one location) is the primary consideration for choosing a food
market, not food safety. Only 18% of respondents considered food safety one of their topmost priorities in
choosing a market. In addition, although households are food safety knowledgeable and have a positive attitude
towards food safety, neither food safety knowledge nor attitude has a statistically significant effect on food safety
cooking practices/behaviour. However, household wealth status positively affects food safety cooking practices/
behaviour. We conclude that households’ food safety cooking behaviour may improve when in addition to
appropriate food safety knowledge, households are economically better off.

1. Introduction diseases, primarily found in low- and middle-income countries [10].

The importance of food safety as a public health and socioeconomic
Food safety is fundamental to food and nutrition security and health issue must be considered, especially in developing countries like Ghana.
[1]. Unsafe food has the potential to create a vicious cycle of food and For example, institutional catering (restaurants, food vendors, “chop
nutrition insecurity, malnutrition and poor health [1]. Food safety issues bars”, and schools) are a significant source of food safety concerns in
can affect everybody and are of concern at every stage of the food system Ghana [11]. For example, the Ghanaian Food and Drugs Authority
[2,3]. Food safety is the assurance that food prepared or consumed by an (FDA) shut down restaurants and local food outlets because consumers
individual for an intended purpose will not cause harm or adverse health experienced symptoms of foodborne diseases, and investigations
effects [4]. Unsafe foods can result from various microbial and chemical revealed poor sanitation, poor food handling practices and heavy mi­
contaminants [2,5–7]. Food safety is a collective and shared re­ crobial load (pathogens) in the food sold by these eateries [12,13].
sponsibility of all stakeholders [1,8]. The consequences of unsafe food Furthermore, urban households are exposed to high risks of foodborne
are enormous, especially in developing countries. Foodborne diseases disease transmission because most urban households rely on traditional
caused by food contaminants cause productivity loss of about US$ 95.2 open-air markets to meet their food consumption needs: traditional
billion annually in low- and middle-income countries. Sub-Saharan Af­ open-air markets are prone to unsanitary and unhygienic conditions,
rica accounts for about US$16.7 billion of this total productivity loss [9]. which are fertile grounds for the spread of foodborne pathogens
Moreover, children bear the brunt of foodborne diseases. Children under [14–16].
five years account for about 40% of the global burden of foodborne The home is the final barrier to preventing and transmitting

* Corresponding author.
E-mail address: [email protected] (M.I. Dzudzor).

https://doi.org/10.1016/j.jafr.2023.100728
Received 26 February 2023; Received in revised form 6 July 2023; Accepted 3 August 2023
Available online 6 August 2023
2666-1543/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

foodborne diseases [17] but can also be an avenue for the spread of food [42]. Therefore, for example, households will have a firm intention to
pathogens and foodborne diseases [18]. Therefore, proper food handling purchase food from a hygienic food environment and practice food
in the home is critical to preventing foodborne diseases [19], and food safety cooking practices at home when they have a positive attitude
handlers are essential in implementing safe and hygienic cooking towards that behaviour, how much social pressures they feel to perform
practices in the home [20]. However, home food handlers still grapple that behaviour (subjective norms) and the belief that they can practice
with their role in ensuring food safety at home [20]. Moreover, food these behaviours comfortably. Therefore, according to Ajzen [42],
safety concerns like chemical contamination of food are generally knowledge (correct factual information) does not directly influence
beyond the capacity and ability of the home food handler to address. actual behaviour, but rather knowledge influences beliefs which intend
Notwithstanding some of these hidden food safety hazards confronting influences attitude, subjective norm and perceived behavioural control.
households, proper personal hygiene and household water, sanitation However, from the social cognitive theory by Bandura [40], knowledge
and hygiene (WASH) behaviour can prevent many foodborne diseases creates a precondition for change. A person with appropriate knowledge
[21]. So, households are urged to eat healthier and safer home-cooked and essential skills is positioned to successfully perform a behaviour
meals instead of food away from home [22,23]. Inherent in this state­ because of a high self-efficacy (confidence) in his or her ability [40].
ment is that the food handler in the home is knowledgeable in food Therefore, knowledge is just one factor that influences behaviour [35,
safety and healthy food preparation and has the tools and the environ­ 43]. However, it is a critical factor in the formation of behaviour. Other
ment to act according to their knowledge. factors (moderators) affect the strength of the relationship, whiles others
However, the literature on food-related knowledge translating into (mediators) explain the mechanisms through which knowledge and
appropriate food behaviour change is mixed [17,24–26]. Campbell et al. behaviour are related [44,45].
[27] showed that maternal knowledge of food safety and healthy diets in We present the conceptual framework in Fig. 1. Knowledge can
the home environment influences children’s food behaviour. Tabbakh either directly affect food behaviour or indirectly affect food behaviour
and Freeland-Graves [28] showed a positive relationship between a through the food attitude of the household [46]. Internal and external
mother’s nutritional knowledge and the diet of her adolescent child. factors in our study, like the food handler’s personal and household
Men’s nutritional knowledge can also improve the nutritional status of characteristics, source of information (government and private sources),
households [29,30]. However, knowledge does not necessarily translate and educational level, influence their knowledge. Subsequently, the
into appropriate behaviour [24,31,32]. For example, a food handler relevant acquired food knowledge may indirectly influence food
with proper food safety knowledge only sometimes translates this behaviour through the attitude of household members towards food
knowledge into appropriate food safety practices [33]. Nevertheless, the safety and healthy diets. The household with appropriate knowledge
lack of food safety knowledge is a significant barrier to food safety and skills will then have to overcome barriers like the cost of foodstuffs,
practices [34]. Therefore, the effect of food safety and nutrition kitchen space, cooking utensils and fuels to perform food safety cooking
knowledge on behaviour is a necessary but insufficient factor in positive practices and prepare healthy meals.
food safety and nutrition behavioural change [35]. We analyse the relationship between food safety knowledge, food
Reviewing the existing literature on household food safety behaviour safety cooking behaviour, and source of food purchases in urban areas.
shows a paucity of empirical evidence on food safety knowledge, atti­ We assume that households with the requisite food safety knowledge
tude, and practice (KAP) in Ghanaian urban homes. The focus has been will purchase food from markets or places that meet their food safety
on institutional food handlers like restaurants, food outlets, and food standards. Also, the household has the requisite cooking tools and
sellers and vendors [36,37]. Thus, this paper aims to explore the cooking area to translate the food safety cooking knowledge into
knowledge level of household food handlers on safe foods; and the effect appropriate cooking behaviour. However, the lack of cooking tools and
of food safety and nutrition knowledge in determining households’ food the cooking area may prevent households from observing appropriate
purchases and food safety cooking behaviour. We hypothesise that cooking behaviour.
households with the appropriate food safety knowledge will always
practice food safety cooking behaviour. The study answers the ques­ 2.2. Study area
tions: What are the main factors that influence urban households’ choice
of a food market; does food safety knowledge affect cooking behaviour, The study area is Ghana, located in West Africa: the study sites are
and what is the moderating effect of household wealth status on food located in three cities in Ghana-Accra, Kumasi and Tamale, in the
safety behaviour? Our study is unique because, to the best of our southern, middle and northern parts of Ghana, respectively (Fig. 2).
knowledge, there is currently a need to study Ghanaian urban household According to Ghana’s 2010 Population and Housing Census (PHC), these
food safety knowledge and cooking practices using the methods applied cities are the biggest in the southern, middle and northern parts of
in this study. Ghana based on the population size of the cities. They have large food
markets integral to the county’s food system. The three study sites
2. Material and methods provide a national picture of the urban food system investigated from
different geographic and socioeconomic perspectives. We provide
2.1. Conceptual framework further details on these unique cities surveyed in this study.

Different theories and models explain behaviour and behaviour 2.2.1. Accra metropolis
change [38–41]. These theories have shaped our understanding of the The Accra Metropolitan Assembly (AMA) is in southern Ghana. Ac­
factors influencing behaviour change [40]. For example, according to cording to the 2010 PHC, the metropolis makes up about 42% of the
Ajzen’s [41] theory of planned behaviour, behavioural intention is the total population of the Greater Accra Region. The entire metropolis is
immediate predictor of actual behaviour change. A person’s intention is urban. However, there are variations in the socioeconomic status of the
the individual’s effort to undertake a behaviour. Also, behavioural people. There are about 450,748 households in the metropolis. About
intention is influenced by an individual’s attitude, subjective norm, and 47% of the population are migrants. The informal private sector is the
perceived behavioural control. These factors are further shaped by the largest employer, with about 48% of the inhabitants self-employed. The
normative beliefs, motivation and evaluation of outcomes by the indi­ city is the economic hub of the country; and has some of the largest food
vidual [41]. markets, namely Makola and Agbogbloshie markets. The AMA has three
Additionally, internal (knowledge, skill and individual abilities and sub-metros: Ablekuma South, Ashiedu Keteke and Okaikoi South [47].
characteristics) and external (resources, money, time, equipment and
legal barriers) factors can interfere with the actualisation of behaviour

2
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

Fig. 1. Conceptual framework of the effect of food safety and nutrition knowledge on behaviour.

Fig. 2. A map of Ghana showing the study sites.

3
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

2.2.2. Kumasi metropolis “sometimes”, “not often”, and “never” statements that assessed food
The Kumasi Metropolitan Assembly (KMA) is in the middle of Ghana. safety self-reported behaviour [19].
According to the 2010 PHC, the metropolis makes up about 36% of the
total population of the Ashanti Region and has about 440,283 house­ 2.3.2. Household survey sampling design
holds. The city is a vibrant commercial centre. Strategically positioned The NOURICITY project used a multistage sampling technique in the
to link the north and the south of the country. The Kejetia central market sampling of households. A three-stage sampling procedure was applied.
is the largest open-space food market in West Africa, and the food sec­ The first stage was purposive, and the subsequent two stages were
tion is one of the largest in Ghana. The city’s food system and the rural randomisations. In the first stage, we selected the three largest cities in
food system of neighbouring districts are closely linked. Food prices in the south, middle and north of Ghana based on the 2010 Population and
the city are lower compared to other cities in the country [47]. Housing Census (PHC). The choice of these study sites was because; of
the presence of major food markets, level of development and urbani­
2.2.3. Tamale metropolis sation, food socialisation behaviour, socioeconomic characteristics and
The Tamale Metropolitan Assembly (TaMA) is in northern Ghana. agroecological characteristics. The three study sites provide a national
According to the 2010 PHC, it accommodates about 9.4% of the picture of the urban food system in large and main cities in Ghana. The
Northern Region’s population. About 80% of the metropolis is urban. consideration is to have a geographically evenly distributed sample. The
The total number of households in the city is 219,971. The metropolis is second stage of sampling was randomisation at the level of the
the centre of economic activity in the Northern Region and other regions Enumeration Area (EA). The EAs are the lowest geographical units
in northern Ghana. This city is unique because of its geographical demarcated by the Ghana Statistical Service (GSS) for national popula­
location and the socio-cultural and economic status of the people. Food tion census purposes. The Ghana Statistical Service (GSS) performed the
systems in the metropolis are linked to other national and international randomisation at the EA level. Based on our budget and geographical
food systems and the rural food system. The nature and type of food representation, the GSS randomly selected the total number of EAs we
consumed vary from those eaten in the middle and southern parts of the requested in the various study sites based on the 2010 PHC.
country [47]. The third stage of randomisation was at the household level within
each EA. Within each EA, data collectors did random walks to the
2.3. Survey data and sampling design households. They started from the EA base, the major landmark within
the EA, and moved in four opposite directions to sample the households.
2.3.1. Survey data and questionnaire Where the houses are densely populated, we sampled after every 10th
The data used in this study is part of the more extensive data house. In Accra and Tamale, we sampled 18 households from each EA,
collected under the NOURICITY project in Ghana. The NOURICITY whiles in Kumasi, we sampled 12 households each. The total sample
project is a European Union Horizon 2020 research and innovation collected was 672 households from 44 EAs. However, after data cleaning
program-funded project. The NOURICITY project studied the urban food and management, 609 responses had complete data for analysis. Table 1
systems in Ghana, South Africa and Uganda. The project ran from 2018 presents the distribution of households sampled.
to 2022. We conducted different research activities in Ghana, including
stakeholder workshops, market and household surveys, and microbial 3. Data analysis and empirical strategy
food analysis. We used in this paper the first of four rounds of household
survey data collected under the NOURICITY project. We collected the 3.1. Measurement of key variables
first round of household data in November–December 2019. The study
relied on the first round of the household survey because; we did not Food safety cooking behaviour is the primary outcome variable of
introduce any interventions between survey rounds targeted at changing interest. Food safety cooking behaviour is computed using respondents’
the households’ food safety knowledge, attitude and practices (KAP), responses to 10 statements on their food safety behaviour contained in
which are the key variables of interest in this study. the WHO’s “five keys to safer foods” [19]. Respondents indicate whether
We administered a structured questionnaire to sampled households. they “always”, “most times”, “sometimes”, “not often”, and “never”
We trained data collectors in administering the questionnaires. As part practice the stated behaviours. So, household food safety behaviour was
of the training, we trained the data collectors in administering the computed as the sum of all the “always” responses per household. The
questionnaire in the local language. Technical and key terms were higher the aggregated number, the better the implementation of
agreed upon during the training to ensure consistent communication appropriate food safety cooking behaviour of the household according
with the respondents. After the training, we pre-tested the questionnaire to WHO standards.
to ensure all questions were phrased concisely and appropriately to The explanatory variables used in this study include; household
capture the needed information. The comments and feedback from the knowledge and attitude towards food safety, household nutrition
pre-testing were discussed and appropriately incorporated into the final knowledge, source of food safety information, household wealth index,
version of the questionnaire. and household characteristics. Food safety and nutrition knowledge are
The questionnaire administered to households had sections on the computed based on the summation of correct answers to standard WHO
household roaster and demographics, food purchasing behaviour, food
safety and nutrition KAP, food security indicators, the health status of
household members, household income and expenditure module, access Table 1
Number of households sampled.
to public amenities and housing characteristics. We based the section of
the questionnaire on food safety and nutrition KAP on the “World Health Region City Sub- Number Number of Completed
metro/ of EAs households number of
Organisation’s (WHO) five keys to safer food” [19]. The WHO’s five keys
district sampled sampled household
to safer food is a manual used to evaluate food handlers’ knowledge, interviews
attitude and behaviour towards their cooking practices. The core themes
Greater Accra Ashiedu- 12 216 175
of the five keys to safer food are: keep clean; separate raw and cooked; Accra keteke
cook thoroughly; keep food at safe temperatures; and use safe water and Ashanti Kumasi Manhyia 20 240 218
raw materials. These five themes comprise eleven “true”, “false”, and & Subin
“don’t know” questions that assess food safety knowledge. In addition, Northern Tamale Tamale 12 216 216
44 672 609
there are ten “agree”, “not sure”, and “disagree” statements that tested
food safety attitudes. Finally, there are ten “always”, “most times”, Source: NOURICITY project survey, 2020

4
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

questions on household food safety and nutrition. Using principal test of sphericity and Cronbach alpha test to test the appropriateness of
component analysis (PCA), households’ wealth index (a proxy for in­ the items used to reflect the latent variables [51–53].
come) is computed based on households’ assets. After conducting the EFA to select the appropriate items, we per­
formed a CFA to confirm the relationship among the variables of interest
3.2. Estimation strategy based on the study’s conceptual framework. After this, we ran the SEM
model to find the model that best fits the theory and data of the study.
We applied Structural Equation Models (SEM) to address the ques­ After running a SEM model, we performed a goodness of fit test based on
tions on the effect of household food safety knowledge and attitude on some indices to determine the appropriateness of the model for its
food cooking practices/behaviour. Food safety knowledge, attitude and intended purpose. The recommended cut-off levels for the goodness of
behaviour (KAP) are treated as latent variables [48]. Therefore, to fit indices [54–56] include; the Root Mean Squared Error of Approxi­
measure the latent variables, a set of indicators that best explain various mation (RMSEA) and Standardised Root Mean Squared Residual (SRMR)
components of the latent variables are measured. In addition, the values less than (<) 0.08 and Comparative Fit Index (CFI) and
complex interactions between knowledge, attitude and behaviour make Tucker-Lewis Index (TLI) greater than (>) 0.9. Models that meet these
them interdependent and bidirectional. SEM is appropriate to address thresholds are close fit models and suitable for their intended purpose.
these peculiarities. A system of equations is required to establish the As presented in Fig. 3, the study extends this basic model to include
relationship between food safety knowledge and food cooking behav­ all the outcome variables of interest and the moderating indicators. The
iour [49]. A measurement model of the relationship between each in­ study specifies three models to explain the relationship between food
dicator and knowledge, attitude and behaviour was built. We then safety knowledge, attitude, healthy diet knowledge, and food safety
combined the measurement models of these latent variables to establish cooking behaviour.
their relationship while controlling for measurement errors in the
observable indicators [49]. 4. Results and discussion
The indicators of each latent variable (knowledge, attitude and
behaviour) are the observable attributes that constitute knowledge and 4.1. Household demographic and socioeconomic characteristics of
the respondent’s responses to a set of questions showing their attitude respondents
towards food safety (positive or not). The respondent’s behaviour is
based on self-reported confirmation of their activities before, during and Table 2 presents summary statistics of all households that completed
after food preparation and where the household purchases food for the household survey (N=609). About 52% of households are male-
cooking. The indicator variables (Xs) of each latent variable used in the headed. Accra (38.9%) and Kumasi (38.1%) have a relatively lower
study are in Table 9. The complex interaction of the various variables number of male-headed households. The average age of a household
and their bidirectional nature leads to endogeneity and measurement head is 47 years, with Accra (44 years) having the average youngest
error challenges. In our conceptual framework, we assume that multiple household head compared to 51 years for household heads in Tamale.
factors measure multiple variables, and the factors can be correlated and Unmarried (single) household heads constitute a relatively significant
have feedback loops. This results in non-recursive models [50]. Mod­ component of the sampled households in Accra and Kumasi. About 19
erators are also incorporated into the knowledge-behaviour models to and 21% of household heads are unmarried in Accra and Kumasi. Out of
analyse the pathways through which knowledge-behaviour models this number, a disproportionate number are female. In Accra and
interact. Kumasi, 88 and 82% of unmarried household heads are female.
We performed three activities to build the SEM for our study: The average household size is 3.9. Tamale has the highest number of
exploratory factor analysis (EFA), confirmatory factor analysis (CFA), household members, 5.0, compared to 3.4 and 3.3 for Accra and Kumasi,
and run SEM. We used STATA 15.1 to perform all the analyses. The EFA respectively. The average Household Dietary Diversity Score (HDDS) is
extracted the items/questions used to construct the latent variables of 7.05. Dietary diversity across cities is similar; the observed differences
food safety knowledge, food safety attitude, food safety cooking prac­ are not statistically significant. However, we observe statistically sig­
tices/behaviour and healthy diet knowledge. The extracted factors are nificant differences among households’ food expenditure per capita.
based on eigenvalues greater than (>) 1 using the principal factor Households in Accra (GHS 254.69) and Tamale (GHS78.07) spend the
method (pf), the communality values greater than (>) 3 and factor highest and lowest on food per capita, respectively. The average
loadings of scale items greater than (>) 0.4. In addition, we conducted household is in the middle wealth index (3.06). On average, households
the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy, Bartlett’s in Kumasi (3.4) have a higher wealth index than households in Tamale

Fig. 3. Basic model of food safety knowledge, attitude and practice.

5
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

Table 2 4.2. Source of food purchases and food safety concerns by urban
Household demographic and socioeconomic characteristics. households
Variable Accra Kumasi Tamale Total P-value
The source of food purchases is an essential component of a house­
Household head characteristics
Male headed 38.86 38.07 84.26 52.08 0.0000*** hold’s food decision-making. From Table 3, convenience is the principal
households (%) reason for the choice of market for food purchases among sampled
Age (mean) 44.191 45.873 51.174 47.270 0.0000*** households. Convenience in terms of proximity to the market and the
Education (%) 0.000*** availability of all products at one location. About 66 and 58% of
None 7.43 13.76 41.67 21.84
Primary 15.43 9.63 1.85 8.54
households considered the distance to the market and availability of all
Secondary 72.00 68.35 36.11 57.96 products at one location among their top three considerations when
Tertiary 5.14 8.26 20.37 11.66 choosing the market to purchase food items. Table 4 shows that out of
Read &write in 69.14 63.30 50.00 60.26 0.0003*** the options provided, 50 and 19% of households selected distance to the
English (%)
market and availability of all products at one location, respectively as
Marital status of 0.000***
household head their main reason for choice of market. The results show that transaction
(%) cost considerations are of high importance to households. Households in
Single 18.86 20.64 3.24 13.96 urban areas adopt time-saving mechanisms to deal with the growing
Monogamous 45.71 50.92 76.39 58.46 opportunity cost of time. They cut back on time allocated to domestic
Polygamous 0.00 0.00 10.19 3.61
Divorced 13.14 6.88 2.31 7.06
activities, including food preparation and shopping, and channel the
Widowed 16.00 14.68 6.94 12.32 time saved into other economic activities. They optimise their in­
Separated 5.14 6.42 0.93 4.11 teractions with the food environment by choosing accessibility (short
Cohabitation 1.14 0.46 0.00 0.49 distance to market) and convenience (brevity of time) [58].
N (175) (218) (216) (609)
Tables 3 and 4 also show that food safety concerns were low among
Gender of unmarried 87.88 82.22 85.71 84.71 0.7943 the considerations of respondents. Many households need to be made
household head aware of the primary state institution in charge of championing food
(female=1)
N (33) (45) (7) (85)
safety. Less than 50% of households are aware of the Food and Drugs
Other household characteristics Authority (FDA), the primary state institution to champion food safety
Household size 3.377 3.335 4.968 3.926 0.0000*** issues in Ghana. Further, only 18% of households considered food safety
(mean) among their top three considerations for the choice of market (Table 3).
Household Dietary 7.091 7.151 6.907 7.048 0.5174
Only 2% of households had food safety concerns as their topmost
Diversity Score
(HDDS) consideration in selecting a food market (Table 4). The relatively lower
Household food 254.685 209.006 78.071 175.692 0.0000*** consideration for food safety in the choice of market is not necessarily a
expenditure per lack of care for safe food. The social construct around food and cooking
capita (GHS) in Ghana may explain this observation. Consumers who have had pos­
Mean household 2.783 3.358 2.995 3.064 0.0002***
wealth index
itive previous experiences with a retailer and have developed a trust­
(1=lowest; worthy relationship may continue to purchase food items from that
5=highest) retailer, irrespective of the current food safety status of the retailer [36].
Employment status of 89.14 82.11 82.87 84.40 0.1203 Consumers may continue to patronise a particular food retailer provided
household head (%)
there are no immediate adverse effects from consuming food from that
Mean percent of 51.314 47.448 41.224 46.351 0.0017***
household members source.
employed Open-air markets are still the main markets patronised by house­
Households living in 73.14 60.55 69.91 67.49 0.0191** holds in cities. The main market in the city/community, which are open-
compound houses air markets, remains the preferred choice for food purchases. In Table 5,
(%)
about 59 and 31% of households sourced food items from the com­
N (175) (218) (216) (609)
munity’s main and satellite markets, respectively. This finding is
+ANOVA conducted across study sites. *** p<0.01, ** p<0.05, * p<0.1. consistent with Hannah et al. [59], who found that open-air markets are
the preferred option for urban households in eighteen cities in Kenya
(3.0) and Accra (2.8). Compound houses are the most common type of and Zambia because open-air markets meet households’ expectations
dwelling for households. About 67% of the total sample live in com­
pound houses. The average percent of household members employed is
less than 50%. Table 3
In summary, household characteristics vary across cities except for Choice of food market and awareness of FDA.
HDDS, employment status of household heads and the proportion of Accra Kumasi Tamale Total P-value
unmarried female household heads. Further, households in Tamale have % of households’ who consider …. as 1 of their top 3 considerations for choice of
the most male-headed households, oldest household heads, largest market
household sizes, and lowest number of single (unmarried) household Convenience (short 70.86 67.43 60.19 65.85 0.0716*
heads. The above household characteristics are mainly in tandem with distance to market)
Convenience (all products 58.86 54.59 60.65 57.96 0.4252
the latest round of the nationally representative survey of the Ghana
at one place)
Living Standards Survey 7 (GLSS 7), conducted in 2017. According to Safety standards/good 15.43 15.14 22.22 17.73 0.0992*
the GLSS 7 report [57], the national average household size is 3.8, with quality products
urban areas having an average household size of 3.5. The national mean % of households aware of 49.71 54.59 26.39 43.32 0.0000***
age of a household head is 44.2 years, and about 45.6 years in Accra. In Food and drugs authority
(FDA)
addition, nationally, about 57.3% of households live in compound % of households that have 15.43 26.61 9.72 17.41 0.0000***
houses, and 37.2% of households in urban areas live in rented dwellings received any form of
[57]. education from FDA
N 175 218 216 609

+ANOVA conducted across study sites. *** p<0.01, ** p<0.05, * p<0.1.

6
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

Table 4 regarding dietary preference, convenience, accessibility and prices of

Households’ main reason for choice of market for purchase of food items. foodstuffs.
Main reason for choice of type of market for shopping N % On the other hand, our findings show that about 1% of households
sourced their food items from supermarkets. Supermarket shopping for
Overall sample
Convenience (short distance from my house) 303 49.75 food products, especially fresh fruits and vegetables, and some local food
Convenience (all products at one place) 115 18.88 commodities still need to be higher among respondents in urban areas.
Lower price of products 103 16.91 Despite the growth in the supermarket sector in Ghana [60], we may
The products are fresh 44 7.22 attribute households’ very low patronage of supermarkets to the
Can buy in bulk 19 3.12
Safety standards/quality of product 14 2.30
uniqueness of the study areas, which are close to major traditional
Others (social construct, buy product on credit, culture/tradition) 11 1.82 open-air markets. In places where open-air markets and supermarkets
Total 609 100.00 are nearby, consumers patronise open-air markets because of lower
Accra prices, fresher products and convenience [59,61]. In addition, unlike
Convenience (short distance from my house) 93 53.14 open-air markets, supermarkets are associated with food safety because
Lower price of products 29 16.57 of the implementation of quality and safety standards throughout the
Convenience (all products at one place) 27 15.43 supply chain [62]. However, supermarkets sell more processed foods
The products are fresh 14 8.00
Safety standards/quality of product 7 4.00
from the start of operations than fresh fruits and vegetables [63],
Social construct 3 1.72 especially local and indigenous varieties. So major traditional markets,
Others (buy in bulk, culture/tradition) 2 1.14 compared to supermarkets, are more convenient (proximity and all
Total 175 100.00 products at one location) for households to get their domestic fresh fruits
Kumasi and vegetables and processed food items from other retail shops in the
Convenience (short distance from my house) 115 52.75 market.
Lower price of products 41 18.81
Convenience (all products at one place) 34 15.60
The products are fresh 17 7.80
Can buy in bulk 7 3.21 4.3. Household food safety knowledge and information
Safety standards/quality of product 3 1.37
Others (culture/tradition) 1 0.46 The results presented in Table 6 show that households’ average score
Total 218 100.00 on food safety knowledge is 60.9%. Households in Tamale (61.5%) had
Tamale the highest score on food safety knowledge. However, differences in
Convenience (short distance from my house) 95 43.98 food safety knowledge across study sites are not statistically significant.
Convenience (all products at one place) 54 25.00
Food handlers performed better on some questions than others
Lower price of products 33 15.28
The products are fresh 13 6.02 (Table A1). Over 90% of respondents know it is essential to wash hands
Can buy in bulk 11 5.09 before handling food, cooked food should be kept very hot before
Safety standards/quality of product 4 1.85 serving and wash fruit and vegetables before use. Although 83% of food
Others (buy products on credit, culture/tradition, social construct) 6 2.78 handlers know that raw food needs to be stored separately from cooked
Total 216 100.00
food, only 17% of them know that it is a false statement that the same
cutting board can be used for raw and cooked foods provided it looks
clean. Although over 96% of food handlers know it is essential to wash
Table 5 fruit and vegetables before use, only 20% of them know that safe water
Where households mostly purchase food items.
cannot be identified by how it looks. Also, only 21% of food handlers
Where respondents mostly purchase food items N % know that cooked meat cannot be left at room temperature overnight to
Overall cool before refrigerating.
Main market in the city/community 361 59.28 Regarding food handlers’ attitudes, about 75.5% of households had a
Daily market (satellite market) 186 30.54 positive attitude towards or agreed with the food safety guidelines
Periodic markets 34 5.58
presented to them (Table 6). However, food handlers have different
Sidewalk 20 3.28
Supermarkets 6 0.99 attitudes towards safe food handling activities and general hygiene
Others 2 0.33 (Table A2). Over 90% of food handlers have a positive attitude towards
Total 609 100 keeping kitchen surfaces clean to reduce the risk of illness and inspecting
Accra food for freshness and wholesomeness. However, only 31% of food
Main market in the city/community 93 53.14 handlers have a positive attitude towards meat thermometers as useful
Daily market (satellite market) 68 38.86 kitchen gadgets for ensuring food is cooked thoroughly.
Periodic markets 8 4.57
Sidewalk 6 3.43
Regarding food handlers cooking practices, only 53% of households
Total 175 100

Kumasi Table 6
Main market in the city/community 111 50.92 Food safety knowledge, attitude and self-reported behaviour.
Daily market (satellite market) 84 38.53 Food safety knowledge, Accra Kumasi Tamale Total P-value
Periodic markets 10 4.59 attitude and behaviour
Sidewalk 9 4.13
Supermarkets 4 1.83 Average accurate 61.14 60.13 61.53 60.92 0.5020
Total 218 100 percentage score
% of positive attitude 73.0 78.2 75.0 75.5 0.0144**
Tamale towards food safety
Main market in the city/community 157 72.69 guidelines
Daily market (satellite market) 34 15.74 % of practiced food safety 47.14 55.64 55.05 53.00 0.0045***
Periodic markets 16 7.41 behaviour always
Supermarkets 5 2.31 Healthy diets knowledge 16.97 17.94 18.75 17.95 0.0000***
Sidewalk 2 0.93 (mean)
Others 2 0.93 N 175 218 216 609
Total 216 100
+ANOVA conducted across study sites. *** p<0.01, ** p<0.05, * p<0.1.

7
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

practised all the safety guidelines provided “always” (Table 6). Food Table 8
handlers practised more activities than others (Table A3). Most maintain Sampling adequacy (KMO) and internal consistency (Cronbach alpha).
general hygienic conditions in their cooking spaces: they constantly Latent Kaiser- Bartlett’s test of Cronbach’s No. of
wash their hands before and during food preparation (78.8%) and wash variables Meyer- sphericity (p- alpha questions/
fruit and vegetables with safe water before eating (79.2%). The least Olkin value) items
(KMO)
practised safe food handling activities are thawing frozen food in the
refrigerator or other cool place (31.2%), using separate utensils and Food safety 0.537 0.000 0.3534 4
cutting boards when preparing raw and cooked food (36.1%) and stor­ knowledge
Food safety 0.696 0.000 0.6202 6
ing any left-over food in a cool place within 2 hours after cooking attitude
(36.3%). Based on the WHO’s five keys to a healthy diet [64], re­ Food safety 0.789 0.000 0.7732 8
spondents also scored an average of 18 out of 20 on their knowledge of behaviour
healthy diets. Respondents in Accra had the lowest average mark of 17 Nutrition 0.735 0.000 0.7277 12
knowledge
out of 20. Regarding healthy diet knowledge (Table A4), food handlers
Totala 0.799 0.000 0.8037 30
knew about the potentially harmful effects of consuming high amounts
a
of fats, oils, sugar and salts. All items (questions/statements) used to compute all latent variables.
The results clearly show that households are knowledgeable about
food safety and healthy diets and have a positive attitude towards food all the latent variables [51–53] except the Cronbach alpha value of 0.35
safety. However, fewer households practice food safety cooking activ­ for food safety knowledge. In Table 9, the factor loadings of the items
ities always. These findings are consistent with the results of Makhunga presented are above 0.4. For each latent variable, the average factor
et al. [65]. Using the WHO’s five keys to safer food, the authors found loading is above 0.5, indicating that convergent validity is present [68].
that food handlers in the eThekwini District in South Africa had good Thus, the items extracted from the EFA to the CFA to construct the model
knowledge, positive attitude and acceptable behaviour towards safe are appropriate.
food handling. However, unlike our findings, household food handlers
in Bangladesh showed insufficient food safety knowledge and handling 4.4.2. Confirmatory factor analysis
practices [66]. Also, Langiano et al. [18] observed that respondents in After running the SEM model, we performed a goodness of fit test to
Cassino, Italy had insufficient food safety knowledge on the transmission determine the appropriateness of the model for its intended purpose.
of foodborne diseases and pathogens. Our models’ goodness of fit summary statistics shows acceptable results
The home environment is the primary source of food safety infor­ based on recommended cut-off levels [54–56]. The Root Mean Squared
mation. Many household food handlers acquired food safety information Error of Approximation (RMSEA) and Standardized Root Mean Squared
from their mothers/guardian and relatives (Table 7). Mother/guardian Residual (SRMR) values are within recommended levels of less than 0.08
and other relatives account for about 63% of responses as a source of (Table 10). Specifically, RMSEA values are 0.08 and 0.05 for models 1
food safety information. The home is still an important place for food and 2, respectively. The SRMR values are 0.06 and 0.07 for models 1 and
socialisation. The home can serve as a platform to introduce food safety 2, respectively. The Comparative Fit Index (CFI) and Tucker-Lewis Index
conversations that can improve knowledge and behaviours. Our finding (TLI) are very close to the recommended levels of greater than (>) 0.9.
on the source of food safety information is similar to that of Marklinder Our models have CFIs values of 0.8 and 0.9 and TLI values of 0.8 each for
et al. [67]. The authors found that among sampled university students in models 1 and 2, respectively. Thus, the models are satisfactory for the
Sweden, a majority (45%) of them had their food safety knowledge from data, and with RMSEA and SRMR values within acceptable limits, with
family and friends. caution, the models can be used for their intended purpose.
Table 11 shows the estimated standardised results of the models and
4.4. SEM analysis their goodness of fit statistics. In model 1, we estimated the relationship
among food safety KAP. The results indicate that food safety knowledge
4.4.1. Exploratory factor analysis (β1=0.595, p>0.05) and attitude (β1=0.220, p>0.05) positively affect
We performed exploratory factor analysis (EFA) to identify the items food safety cooking practices/behaviour. However, the effect is not
that affect the structure of the latent variables (food safety knowledge, statistically significant. In addition, food safety knowledge and attitude
attitude and behaviour). Tables 8 and 9 present the sampling adequacy are positively correlated (β1=0.902, p<0.05), and this association is
and reliability and the factor loadings of the items used in the EFA, statistically significant. These findings are similar in part to Soon et al.
respectively. The number of items (indicators) used to estimate the [17], who found that the effect of food safety knowledge on food safety
latent variables are 4 and 6 for food safety knowledge and attitude, and 8 practices was negative and statistically not significant among consumers
and 12 for food safety cooking practice/behaviour and healthy diet in Malaysia, but attitude had a positive and significant effect. Further,
knowledge, respectively. In Table 8, the KMO values are 0.54, 0.70, 0.79 Akabanda et al. [37], showed that the food safety knowledge of food
and 0.74 for food safety knowledge, attitude, and behaviour, and handlers in Ghana needed to correspond with their food safety practices.
healthy diet knowledge, respectively. The corresponding Cronbach’s Mihalache et al. [68], observed the contrary. The authors observed that
alpha values are 0.35, 0.62, 0.77 and 0.73, respectively. The overall food safety knowledge and shopping attitude had a positive and statis­
KMO and Cronbach alpha values for the 30 items are 0.80 each. The tically significant effect on kitchen practices among consumers in
KMO and Cronbach’s alpha values are all within recommended levels for Romania [68].
In model 2, we include healthy diet knowledge in the food safety KAP
Table 7 model (model 1). The results show that food safety knowledge
Sources of information on food safety. (β1=0.648, p>0.05), healthy diet knowledge (β1=− 0.311, p>0.05) and
food safety attitude (β1=0.307, p>0.05), do not have a statistically
Source of food safety information % of responses
significant effect on households’ food safety cooking practice/behav­
Mother/guardian 42.15 iour. However, a statistically significant positive correlation existed
Other relatives 21.46
Friends 13.48
between food safety knowledge, attitude and healthy diet knowledge
School 9.36 (Table 10).
Media (mainstream/social) 7.21 In model 3, we include the household wealth status in the model as a
Others (public health officer, social grouping, search internet, etc.) 6.34 moderating factor of knowledge and attitude on cooking practices/
Total 100.00
behaviour. Within the household, income is a significant moderator in

8
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

Table 9 Table 10
EFA with factor loadings of items used in SEM. Results of the paths of food safety knowledge, attitude and behaviour.
Indicators Factor Pathway Model 1 Model 2
(Xs) loading
Std. p- Std. p-
Knowledge estimate value estimate value
X_K1 Raw food needs to be stored separately from cooked 0.6092
Food safety Knowledge → Food 0.595 0.257 0.648 0.143
food.
safety behaviour
X_K2 Proper cooking includes meat cooked to 40 ◦ C. 0.4777
Food safety Attitude → Food 0.220 0.671 0.307 0.383
X_K3 Cooked meat can be left at room temperature 0.6632
safety behaviour
overnight to cool before refrigerating.
Knowledge of healthy diet → Food − 0.311 0.105
X_K4 Refrigerating food only slows bacterial growth. 0.5778
safety behaviour
Attitude
Food safety Knowledge ↔ Food 0.902 0.000 0.607 0.000
X_A1 Frequent hand-washing during food preparation is 0.4961
safety Attitude
worth the extra time.
Food safety Knowledge ↔ Healthy 0.148 0.000
X_A2 Keeping raw and cooked food separate helps to 0.7022
diet knowledge
prevent illness.
Healthy diet knowledge ↔ Food 0.310 0.000
X_A3 Using different knives and cutting boards for raw and 0.4940
safety Attitude
cooked foods is worth the extra effort.
X_A4 Meat thermometers are useful for ensuring food is 0.5682 Goodness of fit statistics
cooked thoroughly. RMSEA 0.075 0.054
X_A5 Thawing food in a cool place is safer. 0.5443 SRMR 0.060 0.066
X_A6 I think it is unsafe to leave cooked food out of the 0.5246 CFI 0.825 0.851
refrigerator for more than 2 h. TLI 0.782 0.825
Observations 595 595
Self-reported behaviour
X_B1 I wash my hands before and during food preparation. 0.6834
X_B2 I use separate utensils and cutting-boards when 0.7131
preparing raw and cooked food.
Table 11
X_B3 I separate raw and cooked food during storage. 0.7617
X_B4 I check that meats are cooked thoroughly by ensuring 0.5697
Household wealth status as a moderating factor in household food safety KAP
that the juices are clear or by using a thermometer. model.
X_B5 I thaw frozen food in the refrigerator or other cool 0.5051 Pathway Model 3
place.
X_B6 After I have cooked a meal, I store any leftovers in a 0.7574 Std. estimate p-value
cool place within 2 h. Food safety Knowledge → Food safety behaviour 0.745 0.122
X_B7 I check and throw away food beyond its expiry date. 0.6838 Food safety Attitude → Food safety behaviour 0.204 0.591
X_B8 I wash fruit and vegetables with safe water before 0.4592 Knowledge of healthy diet → Food safety behaviour − 0.368 0.086
eating them. Knowledge of healthy diet → HDDS 0.039 0.459
Knowledge of Healthy diets Household wealth status → Food safety behavior 0.131 0.004
1 It is better to use unsaturated vegetable oils (eg. 0.8264 Household wealth status → Food expenditure per capita 0.069 0.095
Olive, soy, sunflower or corn oil) rather than animal Household wealth status → HDDS − 0.286 0.868
fats or oils high in saturated fats (eg. Butter, ghee, Food expenditure per capita → HDDS − 0.126 0.001
lard, coconut and palm oil) Food safety Knowledge ↔ Food safety Attitude 0.832 0.000
2 White meat (eg. Poultry) and fish are better than red 0.7524 Food safety Knowledge ↔ Healthy diet knowledge 0.621 0.000
meat because they are lower in fat Healthy diet knowledge ↔ Food safety Attitude 0.319 0.000
3 Eat only limited amounts of processed meats because 0.4360 Goodness of fit statistics
these are high in fat and salt RMSEA 0.052
4 People who eat too much saturated fat and trans-fat 0.6309 SRMR 0.065
are at higher risk of heart disease and stroke CFI 0.840
5 Choose fresh fruits instead of sweet snacks such as 0.6210 TLI 0.813
cookies, cakes and chocolate Observations 595
6 People who eat too much salt have a greater risk of 0.6071
high blood pressure which can increase their risk of
heart disease and stroke
safety behaviour.
7 People who eat too much sugar have a greater risk of 0.7976
becoming overweight or obese, and an increased risk The goodness of fit summary statistics (Table 11) shows that model 3
of tooth decay. is fit for purpose. The RMSEA and SRMR values are 0.05 and 0.07,
8 From birth to 6 months of age, feed babies exclusively 0.8236 respectively. The CFI and TLI values are each 0.8. The results show that
with breast milk (i.e. give them no other food or food safety knowledge (β1=0.745, p>0.05) and attitude (β1=0.204,
drink), and feed them “on demand” (i.e. as often as
they want, day and night)
p>0.05) have a positive but statistically insignificant effect on food
9 At 6 months of age, introduce a variety of safe and 0.5784 safety cooking practices/behaviour. Household food handlers can pay
nutritious foods to complement breastfeeding, and more attention to food safety cooking practices than currently. They are
continue to breastfeed until babies are 2 years of age knowledgeable about food safety and have a positive attitude toward
or beyond
food safety guidelines (Table 6). The absence of a statistically significant
10 Exclusively breastfed babies have better resistance 0.6035
against common childhood illnesses such as effect of knowledge and attitude on food safety cooking behaviour
diarrhoea, respiratory infections and ear infections. (models 1–3) may be due to the perceived consequence of food handlers’
11 Eat a wide variety of vegetables and fruits 0.7963 food safety cooking practices/behaviour not resulting in any immediate
12 Avoid overcooking vegetables and fruit because this 0.4526 adverse impact on their health that will cause them to change their food
can lead to the loss of important vitamins
safety cooking practices/behaviour. The perceived consequence of a
practice/behaviour will influence the level of compliance [35]. Also,
the food environment. Compliance with appropriate food safety mea­ other mediating factors like income influences the practice of appro­
sures has cost implications for the household. The appropriate cooking priate food safety cooking behaviour.
space, cooking utensils and kitchen tools, safe water and foodstuff to Household wealth status (β1=0.131, p<0.05) has a positive and
cook; constrain the household’s choice to practice appropriate food statistically significant effect on households’ food safety cooking

9
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

practices/behaviour. A unit change in household wealth status leads to a transportation cost, time spent on food shopping) underline households’
0.13 unit increase in practising appropriate food safety cooking behav­ choice of market. Open-air markets remain the preferred food market for
iour: this implies that as a household’s wealth status improves, house­ households. Supermarket shopping for food products, especially fresh
holds practise more appropriate food safety cooking behaviour. fruits and vegetables and some local food commodities, could be higher
Furthermore, with improved wealth, households are more likely to have among respondents in the study areas.
access to cleaner cooking areas and improved water and sanitation fa­ We also confirm that although households are knowledgeable and
cilities [69,70], which are critical to food safety. On the other hand, have a positive attitude towards food safety, neither food safety
poorer households are more likely to use solid fuels like wood, animal knowledge nor attitude has a statistically significant effect on food safety
dung and charcoal which adversely affects their health [70] and cooking practices/behaviour. However, household wealth status posi­
compromise the hygiene of the cooking area. In addition, poorer tively affects food safety cooking behaviour (model 3), indicating that
households cannot practice appropriate WASH behaviours, including households’ food safety cooking behaviour improves when in addition
hand washing with soap [71,72], and therefore, the household food to appropriate food safety knowledge, households are economically
environment is compromised. better off.
A counterintuitive result is healthy diet knowledge’s statistically The demands on urban dwellers from the labour market, especially
significant negative effect (β1=− 0.368, p<0.05) on food safety cooking those in big cities like our study areas, have altered urban life and
behaviour. Food safety and healthy nutrition are complementary con­ households’ cooking practices and eating behaviour. The home is the
cepts but practically can sometimes be incompatible because food safety last point to ensure food safety. Food safety advocacy, training and
encompasses food handling, preparation and storage, and healthy public education by state institutions are critical to ensuring food safety
nutrition addresses the nutritional quality of food [73]. So, for example, for all. We recommend accompanying public education drives and other
food cooked at high temperature and longer to kill harmful food path­ interventions in promoting food safety knowledge with programs and
ogens risk destroying the nutrient value of the food [74,75]. Also, the strategies to reduce the associated cost of practising food safety mea­
knowledge of the toxic effect of trans-fatty acids in food is optional to sures in the home, especially for the urban poor.
practice personal hygiene (e.g. washing hands before and during food
preparation) when cooking. Therefore, food safety knowledge and Ethics statement
nutrition knowledge may differ. Therefore, our finding may arise
because some nutrition knowledge may be outside the skills required to The study obtained ethical clearance from the Centre for Develop­
practice appropriate food safety cooking behaviour. The model also ment Research’s (ZEF) Ethics Committee and the University of Ghana
shows that the correlation between food safety and nutrition knowledge, Ethics Committee. All respondents consented to the in-person in­
attitude and behaviour remains positive and statistically significant terviews, and their responses remained anonymous.
(Table 11).
Other pathways (model 3) were significant in the household food Author contributions
safety consideration. Household wealth status positively affects house­
hold food expenditure per capita (β1=0.069, p<0.05), but household NG conceptualised the NOURICITY project. MD and NG contributed
wealth status has a negative and statistically insignificant effect on to the study design, designed the questionnaire and critically reviewed
HDDS (β1=− 0.286, p>0.05). Food expenditure per capita also has a the manuscript. MD collected and analysed the data and wrote the first
negative and statistically significant effect on HDDS (β1=− 0.126, draft of the manuscript. All authors contributed to the article and
p<0.05). These observations may be attributed to increasing-income approved the final submitted version.
households likely shifting to consuming other processed and ultra-
processed foods high in fats, sugars and salts, but not necessarily more Funding
diversified foods [76,77]. Consumption of unhealthy ultra-processed
foods is a public health concern. However, concurrently, improved in­ The European Union’s Horizon 2020 research and innovation pro­
comes and convenience-induced motives drive the consumption of gram under grant agreement No. 727715 provided funding for the
processed and ultra-processed foods in the long run [78]. For example, NOURICITY project. In addition, the Open Access Publication Fund of
in developing countries, households may spend on relatively costly the University of Bonn supported the open-access publication of this
processed foods when their income increase, reducing the consumption paper.
of more diversified, relatively cheaper local alternatives. In Ghana,
households may reduce the consumption of cooked beans with red palm Declaration of competing interest
oil, gari and fried plantain ("red-red") and increase their consumption of
fried rice (oily rice with ready-made spices and seasoning). The former is The authors declare that they have no known competing financial
a more balanced meal than the latter. interests or personal relationships that could have appeared to influence
the work reported in this paper.
5. Conclusion
Data availability
The study answers the primary research questions of the urban
households’ choice of food markets and the effect of household food The authors will make the data supporting this article’s conclusions
safety knowledge on food safety cooking practices/behaviour of urban available upon request and subject to the conditions of the NOURICITY
households. The study relied on household data from three Ghanaian project.
cities (Accra, Kumasi and Tamale).
We conclude that many urban households must prioritise food safety Acknowledgements
when choosing food markets. Only 18% of respondents considered food
safety one of their top three considerations for the choice of market. We would like to thank our partners at the Institute of Statistical,
Convenience (68.6% of sampled households) in terms of proximity to Social and Economic Research (ISSER), University of Ghana for their
the market and availability of all products at one location was the pri­ immense support during the implementation of the project in Ghana. We
mary consideration for urban households for their choice of food mar­ would also like to thank the participants for agreeing to be part of this
kets. Economic considerations of reducing their transaction cost (e.g. study.

10
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

Appendix
Table A.1
Households’ knowledge of safe food handling

Construct Items Correct answers (%)

Accra Kumasi Tamale Total

Keep clean
It is important to wash hands before handling food 98.86 100.00 99.07 99.34
Wiping cloths can spread microorganisms 77.14 83.03 88.43 83.25
Separate raw and cooked
The same cutting board can be used for raw and cooked foods provided it looks clean 23.43 19.72 8.80 16.91
Raw food needs to be stored separately from cooked food 75.43 83.03 89.81 83.25
Cook thoroughly
Cooked foods do not need to be thoroughly reheated 42.29 38.53 11.11 29.89
Proper cooking includes meat cooked to 40 ◦ C 51.43 42.20 58.80 50.74
Keep food at safe temperatures
Cooked meat can be left at room temperature overnight to cool before refrigerating 26.86 21.10 15.28 20.69
Cooked food should be kept very hot before serving 90.29 87.16 95.37 90.97
Refrigerating food only slows bacterial growth 73.14 77.52 83.33 78.33
Use safe water and raw materials
Safe water can be identified by the way it looks 20.00 13.76 27.31 20.36
Wash fruit and vegetables 93.71 95.41 99.54 96.39
Total respondents 175 218 216 609

Table A.2
Households’ food safety attitude

Construct Items Positive attitude (%)

Accra Kumasi Tamale Total

Keep clean
Frequent hand-washing during food preparation is worth the extra time 85.14 83.49 83.80 84.07
Keeping kitchen surfaces clean reduces the risk of illness 93.71 96.33 96.76 95.73
Separate raw and cooked
Keeping raw and cooked food separate helps to prevent illness 82.86 87.16 93.06 88.01
Using different knives and cutting boards for raw and cooked foods is worth the extra effort 52.57 61.01 68.52 61.25
Cook thoroughly
Meat thermometers are useful for ensuring food is cooked thoroughly 22.29 28.44 40.28 30.87
Soups and stews should always be boiled to ensure safety 93.71 94.95 94.44 94.42
Keep food at safe temperatures
Thawing food in a cool place is safer 62.86 69.27 55.09 62.40
I think it is unsafe to leave cooked food out of the refrigerator for more than 2 h 51.43 72.94 43.98 56.49
Use safe water and raw materials
Inspecting food for freshness and wholesomeness is valuable 94.86 94.95 94.44 94.75
I think it is important to throw away foods that have reached their expiry date 93.71 95.41 79.17 89.16
Total respondents 175 218 216 609

Table A.3
Self-reported food safety cooking behaviour

Construct Items Practice always (%)

Accra Kumasi Tamale Total

Keep clean
I wash my hands before and during food preparation 73.14 81.19 81.02 78.82
I clean surfaces and equipment used for food preparation before re-using on other food 54.29 66.97 65.74 62.89
Separate raw and cooked
I use separate utensils and cutting-boards when preparing raw and cooked food 30.86 43.12 33.33 36.12
I separate raw and cooked food during storage 39.43 50.00 53.70 48.28

Cook thoroughly
I check that meats are cooked thoroughly by ensuring that the juices are clear or by using a thermometer 35.43 39.91 44.91 40.39
I reheat cooked food until it is piping hot throughout 45.14 47.25 45.83 46.14

Keep food at safe temperatures

I thaw frozen food in the refrigerator or other cool place 25.71 36.70 30.09 31.20
After I have cooked a meal I store any left-overs in a cool place within 2 h 30.29 42.66 34.72 36.29

Use safe water and raw materials

I check and throw away food beyond its expiry date 67.43 76.15 67.59 70.61
(continued on next page)

11
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

Table A.3 (continued )

Construct Items Practice always (%)

Accra Kumasi Tamale Total

I wash fruit and vegetables with safe water before eating them 69.71 72.48 93.52 79.15

Total respondents 175 218 216 609

Table A.4
Households’ healthy diet knowledge

Items Correct answers (%)

Accra Kumasi Tamale Total

Eat moderate amounts of fats and oils

1 It’s better to use unsaturated vegetable oils (eg. Olive, soy, sunflower or corn oil) rather than animal fats or oils high in saturated fats 76.57 88.99 97.22 88.34
(eg. Butter, ghee, lard, coconut and palm oil)
2 White meat (eg. Poultry) and fish are better than red meat because they are lower in fat 79.43 81.19 93.06 84.89
3 Eat only limited amounts of processed meats because these are high in fat and salt 84.00 91.74 91.20 89.33
4 People who eat too much saturated fat and trans-fat are at higher risk of heart disease and stroke 74.29 87.61 91.67 85.22

Eat less salt and sugars

5 When cooking and preparing foods, limit the amount of salt and high-sodium condiments (eg. Soy sauce, fish sauce, cubes) 91.43 94.95 97.22 94.75
6 Avoid foods (eg snacks) that are high in salt and sugars 88.00 94.50 94.44 92.61
7 Limit intake of soft drinks or soda and other drinks that are high in sugars (eg. Fruit juice, cordials and syrups, flavoured milks and 93.71 95.41 92.13 93.76
yogurt drinks)
8 Choose fresh fruits instead of sweet snacks such as cookies, cakes and chocolate 87.43 88.99 92.13 89.66
9 People who eat too much salt have a greater risk of high blood pressure which can increase their risk of heart disease and stroke 81.71 88.07 94.91 88.67
10 People who eat too much sugar have a greater risk of becoming overweight or obese, and an increased risk of tooth decay. 77.71 84.40 90.74 84.73

Breastfeed babies and young children

11 From birth to 6 months of age, feed babies exclusively with breast milk (i.e. give them no other food or drink), and feed them “on 88.57 89.91 99.54 92.94
demand” (i.e. as often as they want, day and night)
12 At 6 months of age, introduce a variety of safe and nutritious foods to complement breastfeeding, and continue to breastfeed until 89.71 93.12 99.07 94.25
babies are 2 years of age or beyond
13 Do not add salt or sugars to foods for babies and young children 69.14 77.06 85.65 77.83
14 Exclusively breastfed babies have better resistance against common childhood illnesses such as diarrhoea, respiratory infections and 90.86 88.99 87.04 88.83
ear infections.

Eat a variety of foods

15 It is better to eat a combination of different foods, including staple foods, legumes, vegetables, fruits and food from animal sources 92.57 97.25 96.30 95.57
than just focusing on a particular food
16 Eating a healthy, balanced diet is especially important for young children’s growth and development 92.57 94.50 98.15 95.24
17 Eating a variety of whole (ie unprocessed) and fresh foods every day helps children and adults to obtain the right amounts of essential 93.71 95.87 94.44 94.75
nutrients.

Eat plenty of vegetables and fruits

18 Eat a wide variety of vegetables and fruits 92.00 95.87 98.15 95.57
19 For snacks, choose raw vegetables and fresh fruit, rather than foods that are high in sugars, fats or salt 83.43 85.78 91.20 87.03
20 Avoid overcooking vegetables and fruit because this can lead to the loss of important vitamins 70.86 79.82 91.20 81.28

Total respondents 175 218 216 609

References [8] FAO & WHO, A Guide to World Food Safety Day 2021. Safe Food Now for a
Healthy Tomorrow, 2021. Rome, https://www.fao.org/3/cb3404en/cb3404en.
pdf.
[1] World Health Organization, WHO Global Strategy for Food Safety 2022-2030:
[9] S. Jaffee, S. Henson, L. Unnevehr, D. Grace, E. Cassou, The Safe Food Imperative:
towards Stronger Food Safety Systems and Global Cooperation, World Health
Accelerating Progress in Low-And Middle-Income Countries, World Bank
Organization, Geneva, 2022, 2022. Licence: CC BY-NC-SA 3.0 IGO.
Publications, 2018.
[2] Z. Gizaw, Public health risks related to food safety issues in the food market: a
[10] WHO, WHO Estimates of the Global Burden of Foodborne Diseases: Foodborne
systematic literature review, Environ. Health Prev. Med. 24 (1) (2019) 1–21.
Disease Burden Epidemiology Reference Group 2007-2015, World Health
[3] HLPE, Nutrition and Food Systems. A Report by the High Level Panel of Experts on
Organization, 2015. https://apps.who.int/iris/bitstream/handle/10665/199350/
Food Security and Nutrition of the Committee on, World Food Security, Rome,
9789241565165_eng.pdf?sequence=1.
2017.
[11] P.F. Ababio, P. Lovatt, A review on food safety and food hygiene studies in Ghana,
[4] Codex Alimentarius, General Principles of Food Hygiene, CXC 1-1969, Food and
Food Control 47 (2015) 92–97.
Agriculture Organization of the United Nations, Rome, 2020 (FAO)/World Health
[12] Food and Drugs Authority (FDA), January 27, FDA Conducts Investigations into
Organization (WHO) Codex Alimentarius; 2020, https://www.fao.org/fao-who-c
Foodborne Disease Outbreak in Oyibi, Accra, 2023. http://www.fdaghana.gov.gh
odexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworksp
/img/press/FDA%20CONDUCTS%20INVESTIGATIONS%20INTO%20FOOD%20
ace.fao.org%252Fsites%252Fcodex%252FStandards%252FCXC%2B1-1969%252F
BORNE%20DISEASE%20OUTBREAK%20IN%20OYIBI-ACCRA.pdf. (Accessed 10
CXC_001e.pdf.
February 2023).
[5] A. Thakali, J.D. MacRae, A review of chemical and microbial contamination in
[13] F.D.A.( June 13, Public Notice. Update: Marwako Food Poisoning Incident, FDA,
food: what are the threats to a circular food system? Environ. Res. 194 (2021),
Accra-Ghana, 2022. http://www.fdaghana.gov.gh/img/press/MAR_11-01.pdf.
110635.
(Accessed 10 February 2023).
[6] I.A. Rather, W.Y. Koh, W.K. Paek, J. Lim, The sources of chemical contaminants in
[14] WHO, WHO Global Strategy for Food Safety. Reducing Public Health Risks
food and their health implications, Front. Pharmacol. 8 (2017) 830.
Associated with the Sale of Live Wild Animals of Mammalian Species in Traditional
[7] World Health Organization, Food Safety (No. EM/RC46/6), 1999.
Food Markets-Infection Prevention and Control, 2022. EB 150/26, January 2022,
https://apps.who.int/gb/ebwha/pdf_files/EB150/B150_26-en.pdf. (Accessed 9
April 2023).

12
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

[15] WHO, A Guide to Healthy Food Markets, World Health Organization, Geneva, [47] Ghana Statistical Service [GSS] (2013), Population and Housing Census. National
2006, 2006, http://apps.who.int/iris/bitstream/handle/10665/43393/9241593 Analytical Report, 2010.
938_eng.pdf. (Accessed 9 April 2023). [48] M. Tolvanen, S. Lahti, J. Miettunen, H. Hausen, Relationship between oral health-
[16] International Food Policy Research Institute (IFPRI), Global Food Policy Report, related knowledge, attitudes and behavior among 15–16-year-old adolescents—a
International Food Policy Research Institute, Washington, DC, 2017, https://doi. structural equation modeling approach, Acta Odontol. Scand. 70 (2) (2012)
org/10.2499/9780896292529, 2017. 169–176.
[17] J.M. Soon, I.R.A. Wahab, R.H. Hamdan, M.H. Jamaludin, Structural equation [49] K.A. Bollen, M.D. Noble, Structural equation models and the quantification of
modelling of food safety knowledge, attitude and practices among consumers in behavior, Proc. Natl. Acad. Sci. USA 108 (Supplement 3) (2011) 15639–15646.
Malaysia, PLoS One 15 (7) (2020), e0235870. [50] R.B. Kline, Principles and Practice of Structural Equation Modeling, Guilford
[18] E. Langiano, M. Ferrara, L. Lanni, V. Viscardi, A.M. Abbatecola, E. De Vito, Food publications, 2011.
safety at home: knowledge and practices of consumers, J. Public Health 20 (2012) [51] J.F. Hair, J.J. Risher, M. Sarstedt, C.M. Ringle, When to Use and How to Report the
47–57. Results of PLS-SEM. European Business Review, 2019.
[19] WHO, Five Keys to Safer Food Manual, 2006. https://www.who.int/foodsafety/pu [52] J.F. Hair, R.E. Anderson, B.J. Babin, W.C. Black, Multivariate Data Analysis: A
blications/consumer/manual_keys.pdf. (Accessed 10 October 2019). Global Perspective, vol. 7, 2010.
[20] M. Jevšnik, V. Hlebec, P. Raspor, Consumers’ awareness of food safety from [53] J.A. Gliem, R.R. Gliem, Calculating, interpreting, and reporting Cronbach’s alpha
shopping to eating, Food Control 19 (8) (2008) 737–745. reliability coefficient for Likert-type scales, in: Midwest Research-To-Practice
[21] S. Brockett, M.K. Wolfe, A. Hamot, G.D. Appiah, E.D. Mintz, D. Lantagne, Conference in Adult, Continuing, and Community Education, 2003.
Associations among water, sanitation, and hygiene, and food exposures and [54] T.A. Brown, Confirmatory Factor Analysis for Applied Research, Guilford
typhoid fever in case–control studies: a systematic review and meta-analysis, Am. publications, 2015.
J. Trop. Med. Hyg. 103 (3) (2020) 1020–1031. [55] C.Y. Yu, Evaluating Cutoff Criteria of Model Fit Indices for Latent Variable Models
[22] L.P. Smith, S.W. Ng, B.M. Popkin, Trends in US home food preparation and with Binary and Continuous Outcomes, University of California, Los Angeles, 2002.
consumption: analysis of national nutrition surveys and time use studies from [56] L.T. Hu, P.M. Bentler, Cutoff criteria for fit indexes in covariance structure analysis:
1965–1966 to 2007–2008, Nutr. J. 12 (1) (2013) 1–10. conventional criteria versus new alternatives, Struct. Equ. Model.: A Multidiscip. J.
[23] K. Glanz, J.J. Metcalfe, S.C. Folta, A. Brown, B. Fiese, Diet and health benefits 6 (1) (1999) 1–55.
associated with in-home eating and sharing meals at home: a systematic review, [57] Ghana Statistical Service (GSS), Ghana living standards survey (GLSS) 7-main
Int. J. Environ. Res. Publ. Health 18 (4) (2021) 1577. report, June, 2019, in: Ghana Statistical Service, 2019 (Accra).
[24] I. Spronk, C. Kullen, C. Burdon, H. O’Connor, Relationship between nutrition [58] C. Turner, S. Kalamatianou, A. Drewnowski, B. Kulkarni, S. Kinra, S. Kadiyala,
knowledge and dietary intake, Br. J. Nutr. 111 (10) (2014) 1713–1726. Food environment research in low-and middle-income countries: a systematic
[25] K. Asakura, H. Todoriki, S. Sasaki, Relationship between nutrition knowledge and scoping review, Adv. Nutr. 11 (2) (2020) 387–397.
dietary intake among primary school children in Japan: combined effect of [59] C. Hannah, J. Davies, R. Green, A. Zimmer, P. Anderson, J. Battersby, T.P. Evans,
children’s and their guardians’ knowledge, J. Epidemiol. 27 (10) (2017) 483–491. Persistence of open-air markets in the food systems of Africa’s secondary cities,
[26] T. De Vriendt, C. Matthys, W. Verbeke, I. Pynaert, S. De Henauw, Determinants of Cities 124 (2022), 103608.
nutrition knowledge in young and middle-aged Belgian women and the association [60] K.S. Andam, D. Tschirley, S.B. Asante, R.M. Al-Hassan, X. Diao, The transformation
with their dietary behaviour, Appetite 52 (3) (2009) 788–792. of urban food systems in Ghana: findings from inventories of processed products,
[27] K.J. Campbell, G. Abbott, A.C. Spence, D.A. Crawford, S.A. McNaughton, K. Ball, Outlook Agric. 47 (3) (2018) 233–243.
Home food availability mediates associations between mothers’ nutrition [61] S. Wertheim-Heck, J.E. Raneri, P. Oosterveer, Food safety and nutrition for low-
knowledge and child diet, Appetite 71 (2013) 1–6. income urbanites: exploring a social justice dilemma in consumption policy,
[28] T. Tabbakh, J.H. Freeland-Graves, The home environment: a mediator of nutrition Environ. Urbanization 31 (2) (2019) 397–420.
knowledge and diet quality in adolescents, Appetite 105 (2016) 46–52. [62] T. Reardon, S. Henson, A. Gulati, Links between supermarkets and food prices, diet
[29] R. Ambikapathi, S. Passarelli, I. Madzorera, et al., Men’s nutrition knowledge is diversity and food safety in developing countries, Trade, food, diet and health:
important for women’s and children’s nutrition in Ethiopia, Matern. Child Nutr. 17 perspectives and policy options (2010) 111–130.
(2021), e13062, https://doi.org/10.1111/mcn.13062, 2021. [63] E.J.O. Rao, M. Qaim, Supermarket growth and rural welfare: evidence from Kenya,
[30] J. Ochieng, V. Afari-Sefa, P.J. Lukumay, T. Dubois, Determinants of dietary in: Organised Retailing and Agri-Business: Implications of New Supply Chains on
diversity and the potential role of men in improving household nutrition in the Indian Farm Economy, Springer India, New Delhi, 2015, pp. 81–97.
Tanzania, PLoS One 12 (12) (2017), e0189022. [64] WHO, Five keys to a healthy diet (n.d), https://apps.who.int/nutrition/topics/5k
[31] D.S. Kenkel, Health behavior, health knowledge, and schooling, J. Polit. Econ. 99 eys_healthydiet/en/index.html. (Accessed 10 October 2019).
(2) (1991) 287–305. [65] S.E. Makhunga, M. Macherera, K. Hlongwana, Food handlers’ knowledge, attitudes
[32] A. McCluskey, M. Lovarini, Providing education on evidence-based practice and self-reported practices regarding safe food handling in charitable food
improved knowledge but did not change behaviour: a before and after study, BMC assistance programmes in the eThekwini District, South Africa: cross-sectional
Med. Educ. 5 (1) (2005) 1–12. study, BMJ Open 13 (4) (2023), e065357.
[33] W.H. Ko, The relationship among food safety knowledge, attitudes and self- [66] M.N. Islam, N. Roy, M.B. Amin, F.K. Madilo, K. Karmakar, E. Hossain, N.J. Airin,
reported HACCP practices in restaurant employees, Food Control 29 (1) (2013) Food safety knowledge and handling practices among household food handlers in
192–197. Bangladesh: a cross-sectional study, Food Control 147 (2023), 109578.
[34] R. Meysenburg, J.A. Albrecht, R. Litchfield, P.K. Ritter-Gooder, Food safety [67] I. Marklinder, R. Ahlgren, A. Blücher, S.M.E. Börjesson, F. Hellkvist, M. Moazzami,
knowledge, practices and beliefs of primary food preparers in families with young M.L. Danielsson-Tham, Food safety knowledge, sources thereof and self-reported
children. A mixed methods study, Appetite 73 (2014) 121–131. behaviour among university students in Sweden, Food Control 113 (2020),
[35] A. Worsley, Nutrition knowledge and food consumption: can nutrition knowledge 107130.
change food behaviour? Asia Pac. J. Clin. Nutr. 11 (2002) S579–S585. [68] O.A. Mihalache, L. Dumitraşcu, A.I. Nicolau, D. Borda, Food safety knowledge,
[36] T. Rheinländer, M. Olsen, J.A. Bakang, H. Takyi, F. Konradsen, H. Samuelsen, food shopping attitude and safety kitchen practices among Romanian consumers: a
Keeping up appearances: perceptions of street food safety in urban Kumasi, Ghana, structural modelling approach, Food Control 120 (2021), 107545.
J. Urban Health 85 (2008) 952–964. [69] E.A. Adams, G.O. Boateng, J.A. Amoyaw, Socioeconomic and demographic
[37] F. Akabanda, E.H. Hlortsi, J. Owusu-Kwarteng, Food safety knowledge, attitudes predictors of potable water and sanitation access in Ghana, Soc. Indicat. Res. 126
and practices of institutional food-handlers in Ghana, BMC Publ. Health 17 (1) (2016) 673–687.
(2017) 1–9. [70] B. Behera, A. Ali, Patterns and determinants of household use of fuels for cooking:
[38] M. Gorton, D. Barjolle, Theories of food choice, in: Food Consumer Science, empirical evidence from sub-Saharan Africa, Energy 117 (2016) 93–104.
Springer, Dordrecht, 2013, pp. 15–26. [71] World Health Organization (WHO) & United Nations Children’s Fund (UNICEF),
[39] S. Michie, M. Johnston, C. Abraham, R. Lawton, D. Parker, A. Walker, Making Progress on Household Drinking Water, Sanitation and Hygiene 2000–2020: Five
psychological theory useful for implementing evidence based practice: a consensus Years into the SDGs, WHO, Geneva, 2021, 2021.
approach, BMJ Qual. Saf. 14 (1) (2005) 26–33. [72] N. Gaffan, A. Kpozèhouen, C. Dégbey, Y. Glèlè Ahanhanzo, R. Glèlè Kakaï,
[40] A. Bandura, Health promotion by social cognitive means, Health Educ. Behav. 31 R. Salamon, Household access to basic drinking water, sanitation and hygiene
(2) (2004) 143–164. facilities: secondary analysis of data from the demographic and health survey V,
[41] I. Ajzen, The theory of planned behavior, Organ. Behav. Hum. Decis. Process. 50 2017–2018, BMC Publ. Health 22 (1) (2022) 1–16.
(2) (1991) 179–211. [73] H. Walls, P. Baker, E. Chirwa, B. Hawkins, Food security, food safety & healthy
[42] I. Ajzen, The theory of planned behavior: frequently asked questions, Human nutrition: are they compatible? Global Food Secur. 21 (2019) 69–71.
Behavior and Emerging Technologies 2 (4) (2020) 314–324. [74] S. Coe, A. Spiro, Cooking at home to retain nutritional quality and minimise
[43] J. Sobal, C.A. Bisogni, Constructing food choice decisions, Ann. Behav. Med. 38 nutrient losses: a focus on vegetables, potatoes and pulses, Nutr. Bull. 47 (4) (2022)
(suppl_1) (2009) s37–s46. 538–562.
[44] S. Bamberg, G. Möser, Twenty years after Hines, Hungerford, and Tomera: a new [75] S. Prabhu, D.M. Barrett, Effects of storage condition and domestic cooking on the
meta-analysis of psycho-social determinants of pro-environmental behaviour, quality and nutrient content of African leafy vegetables (Cassia tora and Corchorus
J. Environ. Psychol. 27 (1) (2007) 14–25. tridens), J. Sci. Food Agric. 89 (10) (2009) 1709–1721.
[45] J.L. Meinhold, A.J. Malkus, Adolescent environmental behaviors: can knowledge,
attitudes, and self-efficacy make a difference? Environ. Behav. 37 (4) (2005)
511–532.
[46] C. Liu, C. Liu, D. Wang, X. Zhang, Knowledge, attitudes and intentions to prescribe
antibiotics: a structural equation modeling study of primary care institutions in
Hubei, China, Int. J. Environ. Res. Publ. Health 16 (13) (2019) 2385.

13
M.I. Dzudzor and N. Gerber Journal of Agriculture and Food Research 14 (2023) 100728

[76] J. Kearney, Food consumption trends and drivers, Phil. Trans. Biol. Sci. 365 (1554) [78] T. Reardon, D. Tschirley, L.S.O. Liverpool-Tasie, T. Awokuse, J. Fanzo, B. Minten,
(2010) 2793–2807. B.M. Popkin, The processed food revolution in African food systems and the double
[77] C.A. Monteiro, R.B. Levy, R.M. Claro, I.R.R. de Castro, G. Cannon, Increasing burden of malnutrition, Global Food Secur. 28 (2021), 100466.
consumption of ultra-processed foods and likely impact on human health: evidence
from Brazil, Publ. Health Nutr. 14 (1) (2010) 5–13.

14