Deneke et al.

BMC Public Health (2022) 22:222

https://doi.org/10.1186/s12889-022-12665-4

RESEARCH Open Access

Milk and meat consumption patterns

and the potential risk of zoonotic disease
transmission among urban and peri-urban dairy
farmers in Ethiopia
Tilaye Teklewold Deneke1*, Adam Bekele2, Henrietta L. Moore3, Tadele Mamo2, Gizat Almaw4,
Getnet Abie Mekonnen4, Adane Mihret5, Rea Tschopp5,6,7, Likawent Yeheyis1, Catherine Hodge3,
James L. N. Wood8, Stefan Berg9 and the ETHICOBOTS consortium

Abstract
Background: In the Ethiopian dairy farming system, prevalence of zoonotic diseases such as bovine tuberculosis
(bTB) is high in the cattle population. This, combined with some risky milk and meat consumption habits, such as raw
milk and uninspected raw meat consumption, poses a considerable risk of zoonotic disease transmission. A survey
was conducted to investigate milk and meat consumption patterns, and the level of exposure to urban and peri-
urban dairy-keeping households for risks of zoonotic disease transmission.
Methods: Data on milk and meat consumption behaviours and other socioeconomic and demographic variables
were collected from 480 urban and peri-urban dairy farms randomly surveyed in major towns in Ethiopia (Mekele,
Hawassa, and Gondar towns, Addis Ababa city, as well as five Oromia towns around Addis Ababa). Determinants of
raw milk consumption associated with a number of demographic and socio-economic factors were analysed using a
generalised ordered logistic model.
Results: The results indicated that about 20% the population consumed raw milk and their awareness about pasteur-
isation and its benefits were low. Location, gender of the household head, previous bTB testing of cattle on the farm,
knowledge of zoonotic risks associated with raw milk consumption, household size, and per-capita milk consumption
were found to be important determinants of the frequency of raw milk consumption. About 60% of the respondents
were exposed to the risk of zoonotic diseases through their habit of frequently consuming raw meat. This was despite
that over 90% of the respondents were aware of possible zoonotic risks of raw meat consumption. The determinants
of raw meat consumption behaviours were associated with location, gender and age of the household head, house-
hold size, meat type preference, per-capita meat consumption, knowledge about disease transmission risks, and
training on zoonoses.
Conclusion: Creating awareness about the risk factors for zoonotic transmission of diseases through training and
media campaigns, improving meat hygiene through better abattoir services, and inducing behavioural change

*Correspondence: [email protected]
1
Amhara Regional Agricultural Research Institute, Bahar Dar, Ethiopia
Full list of author information is available at the end of the article

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativeco
mmons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Deneke et al. BMC Public Health (2022) 22:222 Page 2 of 17

around meat sourcing, raw meat and raw milk consumption, are all crucial to the successful prevention and control of
the spread of zoonotic diseases, including bTB.
Keywords: Zoonosis, Raw meat, Raw milk, Consumption, Bovine tuberculosis, Mycobacterium bovis

Background for zoonotic transmission of bTB are by ingestion, aero-

Ethiopia is an agrarian country with Africa’s largest sol inhalation, or direct contact with mucous membranes
national livestock herd and over 70% of the human popu- and skin abrasions [6]. Airborne transmission is often
lation are directly engaged in the agricultural sector. In associated with respiratory disease manifestation while
such settings, peoples have close interaction with ani- ingestion of infected food (i.e. milk- or meat-borne) is
mals and animal products. Considering the high burden linked to increase risk of extra-pulmonary tuberculosis
of zoonotic diseases in the Ethiopian livestock [1], the [12].
community is at risk of zoonotic transmission through Raw milk consumption. Milk offers a huge potential to
inhalation and ingestion of pathogens. Consumption of improve nutrition and livelihoods for hundreds of mil-
uncooked or unprocessed food increases the risk for dis- lions of poor people and it is estimated that some 750
ease transmission of bovine tuberculosis (bTB), bovine million people are engaged in milk production around
leukosis, brucellosis, anthrax, campylobacteriosis, men- the world, the majority of whom are in low and middle
ingitis, typhoid fever, and gastroenteritis and the like income countries [13]. In countries that are transforming
[2–5]. from an agrarian and rural society towards an increased
The aim of this paper was to investigate milk and meat urban and ‘modern’ society, milk production often shifts
consumption patterns among urban and peri-urban from small scale cattle farming towards larger intensive
dairy-keeping households in Ethiopia, in order to under- dairies. The latter are associated with an increased risk
stand the level of exposure of these segments of society for transmission of infectious diseases, including bTB
to the transmission risk of bTB and other zoonotic dis- [8, 9, 11, 14]. In addition, pasteurisation plants in low
eases through their milk and meat consumption habits. income countries are limited and the habit of raw milk
Studying these issues and the socioeconomic factors that consumption is widespread. Thereby, these factors cre-
determine the milk and meat consumption behaviour of ate the perfect conditions for increased risks of zoonotic
dairy farmers, and consumers more generally, are impor- transmission of bTB through raw milk consump-
tant for devising strategies for controlling the spread tion, especially among dairy farmers with bTB positive
of zoonotic diseases such as bTB, and therefore it was animals.
taken on by the Ethiopia Control of Bovine Tuberculo- In many societies around the world, raw (unpasteurised
sis Strategies (ETHICOBOTS) project - a collaboration or unboiled) milk consumption is a deeply rooted cul-
between researchers in the United Kingdom, Switzerland tural habit. In the Ethiopian context, milk is often con-
and Ethiopia - with the purpose to improve the scientific sumed in its natural state or as a fermented form [15].
knowledge base on bTB and to explore control strategies Some even believe that boiling or pasteurising processes
for the disease in Ethiopia. destroy the quality of the milk [16, 17].
Although many high income countries are free of bTB Studies have shown that M. bovis has frequently been
due to comprehensive test-and-slaughter programmes isolated from unpasteurised and un-boiled milk samples
[6], the disease is endemic in cattle in many other parts [18–20]. In countries where pasteurisation is not widely
of the world [7]. The latter is also the case for Ethiopia in practiced and where there is poor milk hygiene and a
which vast epidemiological studies have shown its pres- common habit of raw milk consumption, it has for long
ence in most parts of the country [8] and the burden is been estimated that about 10–15% of all human TB cases
particularly high among intensively reared dairy cattle are caused by M. bovis [21]. Although such high rates
kept in urban and peri-urban settings [9–11]. National of zoonotic transmission may exist in certain contexts
test-and-slaughter programmes with the purpose to erad- and under certain conditions, more recent reports sug-
icate animal and zoonotic diseases are costly and often gest that the global impact on human TB by M. bovis is
not prioritised in low income countries such as Ethiopia. around 1.5% [7], much lower than previously estimated.
The main causative agent of bTB is the bacterium Raw meat consumption. Meat consumption in Ethiopia
Mycobacterium bovis (M. bovis), causing TB disease in a is also a deep-rooted cultural behaviour. Meat is often
wide range of animals as well as in humans. However, as consumed as part of the staple diet of the people and also
the name ‘bovine TB’ infers, cattle are considered as the during special occasions of festivity. Its cultural symbolic
main reservoir of this disease. The most common routes weight is greater than any other food [22]. Eating raw
Deneke et al. BMC Public Health (2022) 22:222 Page 3 of 17

meat or half cooked meat is very common and although techniques used at Ethiopian abattoirs failed to detect
Ethiopians from various cultures enjoy eating meat, they 70% of the carcasses with grossly-visible lesions of TB.
are generally very selective, in that only poultry, beef, Unlike in official abattoirs in Ethiopia, backyard
mutton, goat and fish (not including shell-fish) are cul- slaughter does not undergo any formal meat inspection.
turally and religiously acceptable. Eating other kinds of Therefore, sourcing meat through home and communal
meat, such as pork, is a cultural taboo among most Ethio- slaughter could be potential risk factors for zoonotic dis-
pians [22] while camel meat is allowed in Muslim com- ease transmission. In fact, focus group discussions with
munities [23]. Ethiopian farmers have revealed that slaughter is a com-
In high income countries, where bTB control and/ mon mechanism for dealing with chronically sick animals
or eradication programmes and effective meat inspec- [30]. Home slaughter is more readily practiced on small
tion mechanisms usually are in place, the risk of getting ruminants. Slaughter of large livestock species is often
infected with bTB due to meat consumption is very low due communally as a means of providing financial assis-
[6]. This is because viable M. bovis have rarely been iso- tance to the owner, as each participant will pay money to
lated from skeletal muscles and the risk of getting TB the animal owner in return for a share of the meat. This
from meat is apparently not from infectivity of muscle sharing praxis makes communal slaughter potentially the
meat but rather from possible contamination of meat most risky meat source in terms of zoonotic transfer of
surfaces during unhygienic slaughtering of animals with diseases between households and provide more opportu-
severe TB lesions in other organs [6, 24]. An effective nities for disease transmission through consumption of
meat inspection would condemn any such carcass and infected meat. In addition to sharing the meat between
protect the consumers from risk of infection. In addition, households, feasting together on raw meat and organs
the M. bovis bacterium is sensitive to heat and any con- such as liver and kidneys are widespread practices in
tamination could be deactivated by cooking [25]. In low Ethiopia [22], and conducted as part of the communal
income countries though, where meat inspection mech- slaughtering tradition.
anisms are likely to be less effective and where the vast With this background we set out to collected data on
majority of the meat consumed is sourced from informal different aspects of milk and meat consumption e.g.
sources without disease inspection and proper hygiene, sourcing preferences, type preferences, and consump-
the habit of raw meat consumption could be a considera- tion behaviours as well as the determinants thereof.
ble risk factor for zoonotic transfer of bTB and other dis- We focused on dairy farmers in urban and peri-urban
eases [26]. It also needs to be said that since diseases such dairy production areas where the prevalence of certain
as bTB and brucellosis are often not controlled at farm zoonotic diseases such as bTB have been reported to be
level in low income countries, it is more likely that dis- particular high [10, 11, 31–34], to better understand their
eased cattle will reach the point of slaughter without such consumption behaviour in relation to risks for zoonotic
diseases having been detected or communicated. disease transmission.
In Ethiopia, there are three sources of meat for home
consumption: purchase from local butchery (sourced Methods
from abattoir), home slaughter, or communal slaughter. The study areas
There are no official estimates of the proportions of these The study was carried out in five study sites (Fig. 1) that
meat sources, but as the abattoir service is very limited included Mekele, Hawassa, and Gondar towns, Addis
in Ethiopia (even the capital, Addis Ababa, has only one Ababa city, as well as five Oromia towns (Holeta, Sebeta,
municipal abattoir service facility), people rely mostly on Debre Zeit, Sululta and Sendafa) surrounding the capital
home or communal slaughter [27]. Communal slaughter and defined as one single study site (Addis Ababa sur-
is even a cultural tradition where approximately 4–10 roundings). The study areas contained both urban and
neighbours or close friends get together to buy an animal peri-urban settings where intensive dairy-farming activi-
and slaughter it for holidays or other festive events. ties took place. In these intensive dairy systems, Holstein-
Butcheries as a source of meat, which usually obtain Frisian (H-F) cattle and their crosses with local zebu
its meat from the official abattoirs, is likely the safest breeds were mainly kept with indoor feeding and zero
way of sourcing meat. However, it carries also the risk grazing. The number of animals in the recruited farms
of zoonotic disease transmission as meat inspection in ranged from five to over 100 dairy cattle and all were kept
abattoirs in Ethiopia is limited. A study on routine abat- for commercial milk production purpose. All the study
toir inspection in Ethiopia detected only 55% of all cat- sites are mid-altitude areas ranging from 1500 to 2500 m
tle confirmed with TB lesions [28], suggesting that meat above sea level with mild temperatures and annual rain
infected with M. bovis have a chance to enter into the fall between 800 and 1500 mm. Addis Ababa is a big city
food chain. Biffa et al. [29] also indicated that inspection with a population around 3.5 million and its surrounding
Deneke et al. BMC Public Health (2022) 22:222 Page 4 of 17

Fig. 1 Geographical locations of the study sites in Ethiopia included in this survey. Study sites of Addis Ababa city and surrounding Oromia towns
are shown as two overlapping dots on the map. Dot-size do not represent actual sample-size as shown in Table 1

suburban areas, where a sprawling intensive dairy farm- agriculture office. Farm size and estimated total number
ing is practiced, have also a large population estimated of farms in each study site were taken in to consideration
to around 1.5 million people [35]. Mekele, Hawassa, to make the sample proportional to size. Table 1 shows
and Gondar are regional towns each having a popula- the distribution of sample farms at each study site.
tion ranging between 300,000–700,000 people. The total
dairy animal population in Ethiopia with exotic blood Data collection
(mainly Pure exotic blood and exotic blood Zebu crosses) Survey data was collected through a well designed and
was estimated to be around 1.44 million [36], of which tested questionnaire by trained, both male and female,
the majority is found in these study sites. According to enumerators with the supervision of ETHICOBOTS
the last few years’ trend, this figure is likely to continue researchers. Along with survey questions, the cattle in
increasing in the coming years due to an artificial insemi- the 480 sampled dairy herds were tested for bTB using
nation programme by the Government and the increase
in demand for milk and meat in the country [37].
Table 1 Distribution of sample farms by site and herd size
The sample farms
Site Herd size of dairy cattle Total
We interviewed 480 urban and peri-urban dairy farmers
about their milk and meat consumption habits as part of small- medium- large-size
holder farm size farm farm
a wider cross-sectional study of the epidemiology of bTB 5–19 20–49 > 49
in Ethiopia, described in Almaw et al. and Mekonnen
et al. [9, 11]. In short, households/farms were selected Addis Ababa city 123 34 7 164
using a two-stage stratified sampling technique, where Oromia towns 82 36 20 138
around Addis Ababa
in the first stage five urban/peri-urban study sites were
Gondar 53 9 4 66
selected (see ‘The Study Areas’ above and Fig. 1). In the
Mekele 50 8 2 60
second stage, using random numbers, a random sample
Hawassa 29 19 4 52
of farmers was selected from a sampling frame com-
Total 337 106 37 480
posed of list of farms in the areas obtained from the local
Deneke et al. BMC Public Health (2022) 22:222 Page 5 of 17

the Single Intradermal Comparative Cervical Tuberculin & 2 versus all others, then 1, 2 & 3 versus all others, etc.
(SICCT) test [38] with PPD-A and PPD-B sourced from [42]. This is based on the assumption that the effect of the
Lelystad (The Netherlands); the results of these tests have predictor variable may vary across the range of the pre-
been published elsewhere [9, 11] but utilised for analy- dicted variable.
ses in the current study. The questionnaire was prepared As discussed further above, raw milk consumption
and tested by a multidisciplinary team composed of agri- is often considered to be a risky behaviour in terms of
cultural economists, anthropologists, veterinarians and zoonotic transmission of bTB and other diseases [2, 21,
other biomedical scientists. Computer Assisted Personal 33]. We hypothesized that raw milk consumption habit
Interview (CAPI) equipment was used for data collection. and its frequency are determined by demographic and
The questionnaire was designed to collect socioeconomic socioeconomic variables such as study site, gender, age,
characteristics of farmers and farm workers, farm man- literacy, income or wealth status, per capita consumption
agement practices, knowledge and attitude towards bTB levels, as well as knowledge about the risk of zoonosis
and other zoonotic diseases as well as milk and meat con- transfer of diseases. The dependent variable, ‘raw milk
sumption behaviour of farmers, farm workers and their consumption frequency’, was measured as an ordinal var-
families. Farm owners or managers were the respondents iable with three ordinal scales, namely:
during the interview which typical took about one and
half hours. In only one case a farmer refused to answer (A) No raw milk consumption;
all the questions and interrupted and turned back on the (B) Moderate level of raw milk consumption (only
enumerator; otherwise the farmers were cooperative and occasional consumption of raw milk); and
willing to answer the questions as well as allowing bTB (C) High frequency of raw milk consumption (at least
testing of their animals. once a fortnight)

Data analysis: generalised ordered logistic model We used Generalised Ordered Logit as that model was
Descriptive and inferential statistical analysis tools found to violate the proportional odds ratio assumption
were used to analyse the data collected. In order to ana- discussed in the methods section. As indicated in Table 4,
lyse the determinants of milk and raw meat consump- Eq. 1 compares the category of those farmers who never
tion frequency we used generalised ordered logit model consumed raw milk (A) to the category of those who did
described below. (B + C). Eq. 2 models the category of those who never
The ordered logit is a widely used model when deal- consumed raw milk or who had an occasional consump-
ing with outcome variable measured at ordinal scale. tion frequency (A + B) to those who had a high frequency
The assumption that ordered logit makes is that the gap consumption (C).
among the various scales of the ordering are equidis- Similarly in order to analyse the determinants of raw
tant to each other. It is assumed that the effect of each meat consumption habits, we hypothesized that the hab-
predictor across the categories of the ordinal dependent its and frequency of raw meat consumption among dairy
variable is the same and is known as proportional odds farmers in Ethiopia could be determined by demographic
assumption [39–41]. However, this does not often hold and socioeconomic variables such as study site, gender,
up [39, 42]. In our data, Brants’ Wald test [43] reveals age, literacy, income or wealth status, as well as having
that the effect of each explanatory variable varies across knowledge about the risk of zoonotic disease transfer
different cutting points of the ordinal outcome variable; through raw meat consumption. To explore this hypoth-
as a result, we adopted generalised ordered logit model esis within our data set, the dependent variable ‘raw meat
that relaxes proportional odds assumption [42, 44]. The consumption frequency’ was measured as an ordinal var-
generalised ordered logit can estimate models (i.e. par- iable with three ordinal scales, namely:
tial proportional odds) that are more parsimonious than (IV) No raw meat consumption;
non-ordinal alternatives, such as multinomial logit (how- (V) Moderate level of raw meat consumption (less than
ever, multinomial logit ignores the ordering of categories once a fortnight); and
and hence would not be more appropriate to deal with (VI) High frequency of raw meat consumption (at least
ordered outcomes). The model is estimated by using a once a fortnight)
constraint partial proportional odds model that enter-
tain the violation of parallel assumption to all predictors We used generalised ordered logit as that model was
and/or a certain predictors of the model [45, 46]. In such found to violate the proportional odds ratio assump-
a way, there would be as much number of binary logit tion. As shown in Table 4, Eq. 3 compares the category
equations as the number of ordinal levels but one, where of farmers who never consumed raw meat (D) to the cat-
first it is category 1 versus all others, then categories 1 egory of those who did (E + F), while Eq. 4 models the
Deneke et al. BMC Public Health (2022) 22:222 Page 6 of 17

category who never consumed raw meat or had a low Table 2A, 77.5% of the respondents (n = 371) indicated
consumption frequency (D + E) to the category of high that they never drank raw milk while about 20.4% drank
frequency consumers (F). raw milk but with varying degrees of frequency. Only
8.1% (n = 39) stated that they were regular drinkers of
Results raw milk, drinking it at least once a day. Although the
Below follows the results of our interviews on milk and majority of the sample farmers indicated that they did not
meat consumption habits held with 480 urban and peri- drink raw milk, about 82% of the respondents did actually
urban dairy farmers from the five study sites shown in drink fermented milk, a yoghurt commonly called ergo in
Fig. 1. the Amharic language, which is usually made from non-
pasteurised/unboiled milk.
Milk consumption We also investigated raw milk consumption in relation-
Per capita milk consumption ship to gender, literacy, study site, religion, and age and
The average per capita milk consumption per day for our we found that it was not related to any of these socioeco-
sample was found to be 0.25 l (SD = 0.26). Farmers were nomic variables, except for study site (data not shown).
asked about their consumption on a daily and monthly A statistically significant systematic relationship between
rather than on a yearly basis, as the latter would be dif- study site and raw milk consumption habit was estab-
ficult to recall and estimate and could also be erroneous lished (likelihood-ratio chi2 (4) = 28.70, P < 0.001) with
due to there being many fasting days in the Ethiopian only 5% of the dairy farmers from Mekele (in Northern
orthodox church calendar when believers do not con- Ethiopia) indicated that they consumed raw milk as com-
sume milk. In this sample the orthodox Christians made pared to 37% of those from Hawassa (in the South).
up 83.3%. This average per capita milk consumption fig- The majority of respondents (88%; n = 424) indicated
ure is statistically significantly higher (t = 16.09; P < 0.001) that they knew that drinking raw milk can cause diseases
than the national average of 19 l per year, which corre- while only 5.6% (n = 27) indicated that it does not cause
sponds to ~ 0.05 l per day [47]. The mean per capita milk any diseases. Some 6.0% (n = 29) indicated that they did
consumption among the sample was not found to be sta- not know about such risk (data not shown), while 78%
tistically significant between sexes, religions, literacy sta- indicated that they thought drinking raw milk to be
tuses, or study sites (Supplementary Table S1). unhealthy or very unhealthy (Table 2B).
Neither general training on zoonosis transmission
Raw milk consumption mechanisms (Fisher’s exact = 0.415), nor specific training
Farmers were asked about their habits of raw (unboiled on bTB bore any relation to raw milk consumption fre-
and unpasteurised) milk consumption. As shown in quency (Fisher’s exact = 0.680). Moreover, we observed

Table 2 Milk consumption characteristics among sampled dairy farmers in Ethiopia

A. Consumption frequency of raw and processed milk reported by dairy farmers

Raw milk Pasteurised milk Boiled milk
n (%) Cum. % n (%) Cum. % n (%) Cum. %
Everyday 39 (8.1) 8.1 3 (0.6) 0.6 122 (25.5) 25.5
3–6 times a week 30 (6.3) 14.4 9 (1.9) 2.5 222 (46.3) 71.8
Once/twice a week 14 (2.9) 17.3 6 (1.3) 3.8 83 (17.3) 89.1
Once/twice a month 15 (3.1) 20.4 9 (1.9) 5.7 18 (3.8) 92.9
On Special occasions only 10 (2.1) 22.5 26 (5.4) 11.1 11 (2.3) 95.2
Not at all 371 (77.5) 100 426 (88.9) 100 23 (4.8) 100
Total 479 (100) 479 (100) 479 (100)
B. Farmers’ perception of the healthiness of drinking raw milk
How healthy is drinking raw milk? n (%) Cum. %
Very healthy 23 (4.8) 4.8
Healthy 48 (10.1) 14.9
Do not know 35 (7.4) 22.3
Unhealthy 238 (50.1) 72.4
Very unhealthy 131 (27.6) 100
Total 475 (100)
Deneke et al. BMC Public Health (2022) 22:222 Page 7 of 17

no difference in the frequency of raw milk consumption consumer. On top of this, our data show that there is sta-
between those farms whose animals tested positive for tistically significant relation between bTB status of the
bTB before our survey and those which were not. This farmers’ herd and knowledge about pasteurisation/pas-
indicates that acquiring knowledge of the bTB status of teurised milk (Chi(2) = 7.19 and P = 0.007) i.e. among
the cattle at farm level has not generally led to change in those farmers who had bTB positive cattle, 55% had no
raw milk consumption behaviour of these farmers. knowledge of pasteurisation and among those farmers
Despite their knowledge of the possible risk of dis- who did not know about pasteurisation, 42% had bTB
ease transmission, a considerable number of our sample positive animals. These results are alarming given the
(20.4%; n = 98) consumed raw milk frequently and on a high prevalence of zoonotic diseases in the area, includ-
regular basis (Table 2A). Among participants drinking ing bTB.
raw milk, 47% had bTB positive animals in their herd but
there was in fact no statistically significant difference in Boiled milk consumption
the raw milk consumption habits between farms with Although the vast majority of respondents said they did
bTB positive and negative cattle. Interestingly however, not drink raw milk (78%) or pasteurised milk (89%), 89%
we did find that there was a statistically significant rela- (n = 427) of the respondents drank boiled milk at least
tionship between raw milk consumption habit and occur- once a week (Table 2A), while only 4.8% (n = 23) indi-
rence of TB disease in the farm household in the last 3 cated that they never drank boiled milk. The frequency
years before the survey (likelihood-ratio chi2(1) = 12.09; of boiled milk consumption was found to be dependent
P = 0.001). Among those farm households which on study site (likelihood-ratio chi2(8) = 21.62; P = 0.006),
reported a TB case in the last 3 years, 41% indicated that with those in Hawassa (87%) and in Addis Ababa (75%)
they were in the habit of consuming raw milk, compared were drinking boiled milk more frequently than those
to only 20% among those farmers who reported no TB in Gondar, in Amhara region (71%), in Mekele, in Tig-
cases. This result warrants further clinical epidemiologi- ray region (68%) and in the Oromia towns surrounding
cal investigation to establish whether the confirmed TB Addis Ababa (65%).
cases might be attributable to zoonotic or human TB,
caused by M. bovis or M. tuberculosis, respectively. Meat consumption
Per capita meat consumption
Pasteurised milk consumption The published national average of per capita meat con-
As shown in Table 2A, nearly 89% of the surveyed farm- sumption in Ethiopia is 5.3 kg per annum [48], which
ers did not drink pasteurised milk. Only 38% (n = 181) corresponds to less than 0.5 kg per month. However, the
knew the benefits of pasteurisation, while 54% (n = 259) corresponding consumption for urban areas is 11.5 kg
did not and 8.1% (n = 39) had never heard about pas- (~ 1 kg per month). As shown in Table 3A, the latter
teurisation. Only 1.5% of the respondents indicated that figure is much in line with the mean rate of per capita
their main source of milk was pasteurised milk. Pasteur- monthly meat consumption of 1.25 kg (SD =1.44) among
ised milk was ranked second by 7.6% respondents, 15% the dairy farmers in our sample, with Hawassa having
ranked it third and 19% as fifth (data not shown). the lowest (0.88 kg) and Addis Ababa having the highest
Investigation of the relationship between literacy and (1.37 kg) consumption. Interestingly, the mean per capita
pasteurised milk consumption frequency showed no sys- meat consumption per month for male-headed house-
tematic relationship (Fisher’s exact test P value = 0.690). holds was found to be 1.35 kg, which was statistically
Since the majority of the large dairy farms and pas- higher (t = − 2.43, P = 0.015) than the 0.95 kg per month
teurisation plants in Ethiopia are located in the capital for female-headed households. We found no statistically
Addis Ababa and its surrounding towns, it was logical significant difference in per capita meat consumption
to expect regional differences in the use of pasteurised between study sites, religions (Christians and Muslims),
milk consumption frequency. However, contrary to what or between households with illiterate and literate heads.
was expected, no relationship was found between study
sites and pasteurised milk consumption (Fisher’s exact Meat type preference
test P value = 0.480). Similarly, there was not a statisti- For the whole sample, when respondents were asked to
cally significant relationship between gender and fre- rank their meat type preference, it was found that 48%
quency of pasteurised milk consumption (Fisher’s exact (N = 220) ranked beef as first choice, followed by mutton
test P value = 0.156). It needs to be added though that 32% (n = 144), chicken 11% (n = 48), and goat meat 9.4%
the generally low levels of consumption of pasteurised (n = 43). However, these figures varied by region with a
milk among the surveyed farmers could be because they statistically significant association between preferred
have easier access to unpasteurised milk than the average meat-type and study site (likelihood-ratio chi2(12) value
Deneke et al. BMC Public Health (2022) 22:222 Page 8 of 17

Table 3 Meat consumption characteristics among sampled dairy farmers in Ethiopia

A. Average monthly per capita meat consumption (in kg) among dairy farmers in each study site
Mean SD n (%)
Addis Ababa city 1.37 1.87 164 (36.7)
Oromia towns around Addis Ababa 1.12 1.06 135 (30.2)
Gondar 1.35 1.46 66 (14.8)
Mekele 1.33 0.95 57 (12.7)
Hawassa 0.88 0.49 25 (5.6)
Total 1.25 1.44 447 (100)
B. Frequency of meat consumption among sampled dairy farmers
General meat consumption Raw meat consumption
n (%) Cum. % n (%) Cum. %
Everyday 5 (1.0) 1.0 2 (0.4) 0.4
2–5 days a week 56.6 (271) 57.6 96 (20.0) 20.4
Once every fortnight 109 (22.8) 80.4 51 (10.7) 31.1
Once a month 66 (13.8) 94.2 66 (13.8) 44.9
Only for holidays 25 (5.2) 99.4 91 (19.0) 63.9
Never 3 (0.6) 100.0 173 (36.1) 100.0
Total 479 (100) 479 (100)
C. Ranking order of meat source by sampled dairy farmers
Rank 1 Rank 2a Rank 3a
n (%) n (%) n (%)
Butchery 256 (53.9) 109 (24.0) 58 (17.8)
Home slaughter 174 (36.6) 191 (42.2) 91 (27.8)
Communal slaughter 45 (9.5) 153 (33.8) 178 (54.4)
Total 475 (100) 453 (100) 327 (100)
D. Farmers’ views about risks of getting disease while eating raw meat
Do you think eating raw meat can cause diseases? Have you ever experienced diseases due to eat-
ing raw meat?
n (%) n (%)
No 34 (7.1) 287 (59.9)
Yes 445 (92.9) 192 (40.1)
Total 479 (100) 479 (100)
a
Not all respondents gave a second or third ranking

of 135.5 and P < 0.001). The responding dairy farmers in 46 years (SD = 15) among those who preferred beef. Edu-
Addis Ababa (62%) and in the towns of Oromia (64%), cation, which was measured as a dummy variable with
as well as in Hawassa (44%) tended to significantly pre- the two categories being literate or illiterate, showed no
fer beef, while the majority in Gondar (75%) and Mekele relationship to meat type preference (Fisher’s exact test P
(56%) in the Northern regions preferred mutton over any value = 0.830). We also tested for a relationship between
other meat. A similar association was observed between gender of the household head and meat type preference
frequency of meat consumption and meat type prefer- but found no significant correlation.
ence, with Fisher’s Exact test of 0.029 being significant at
the 0.05 level; i.e. those households which preferred beef
tended to be more frequent meat consumers (79%). A Meat consumption frequency
one-way analysis of variance in terms of mean age of the As shown in Table 3B, only 1% of the responding farm-
respondents and their preferred meat type showed a sta- ers ate meat every day while the majority of them (57%)
tistically significant difference (F value = 5.49; P = 0.001). consumed meat 2–5 days a week. Only 0.6% (3 individ-
Those households which preferred chicken meat had the uals) indicated that they did not consume meat at all.
lowest mean age of 41 years (SD = 11) and this was signif- There was no statistical difference between study site and
icantly different from the mean age of 50 years (SD = 15) meat consumption frequency, and gender of the house-
among those who preferred mutton and the mean age of hold head and meat consumption frequency. Also, no
Deneke et al. BMC Public Health (2022) 22:222 Page 9 of 17

Table 4 Generalised Ordered Logit Estimates of raw milk (Eq. 1 and Eq2) and meat (Eq3 and Eq4) consumption frequency among
dairy farmers in Ethiopia
Raw milk consumption frequency Raw meat consumption frequency
VARIABLES Eq1 (A vs B + C) Eq2 (A + B vs C) Eq3 (D vs E + F) Eq4 (D + E vs F)

Oromia towns around Addis Ababa 0.508* 0.923** 0.584** 0.395

(0.296) (0.391) (0.297) (0.304)
Gondar −0.978** − 0.372 − 0.257 1.211***
(0.491) (0.630) (0.395) (0.410)
Mekele −17.45 21.66 −2.146*** −2.324***
(1318) (5068) (0.439) (0.574)
Hawassa 0.794 0.555 0.644 −0.221
(0.483) (0.566) (0.528) (0.565)
Sex (1 = male; 0 = Female) 0.628* −0.176 0.743*** 0.448
(0.33) (0.436) (0.281) (0.328)
Meat preference (1 = beef; 0 otherwise) – – 0.507** 0.212
– – (0.251) (0.264)
Literacy (1 = Literate; 0 otherwise) 0.0833 0.929 0.339 0.315
(0.555) (0.733) (0.469) (0.558)
Can raw meat cause TB (1 = yes; 0 otherwise) – – −2.287*** −1.120**
– – (0.694) (0.521)
Raw milk cons. Has zoonotic risk (1 = Yes) − 0.412 − 1.745*** – –
(0.413) (0.531) – –
Previous bTB Test (1 = Yes; 0 otherwise) − 0.728** −1.575*** − 0.401 − 0.373
(0.307) (0.481) (0.260) (0.284)
Know benefits of pasteurisation (1 = Yes) −0.117 0.0879 – –
(0.277) (0.340) – –
Had any zoonosis training (1 = yes; 0 otherwise) 0.0656 0.698 0.0626 −0.593**
(0.305) (0.442) (0.269) (0.290)
Herd size in number – – −0.00170 −0.00929
– – (0.00503) (0.00678)
Number of milking cows −0.00236 0.000149 – –
(0.00174) (0.00280) – –
Per capita milk consumption per day 0.768* 1.086** – –
(0.435) (0.488) – –
Meat per capita consumption per month in kg – – 0.206 0.828***
– – (0.138) (0.145)
Age of the household head −0.0152 −0.00135 0.0265 0.0908*
(0.0101) (0.0147) (0.0505) (0.0537)
Age squared – – −0.000424 −0.000939*
– – (0.000484) (0.000523)
Household size 0.106* 0.190** 0.0203 0.159**
(0.0602) (0.0751) (0.0591) (0.0631)
Constant −0.633 −2.401* 1.266 −4.167***
(0.968) (1.431) (1.471) (1.496)
Observations 440 440 417 417
Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1

significant difference in mean age of respondents was level and frequency of meat consumption were found to
observed between those households which frequently be associated (Fisher’s exact value P = 0.001); 56% of the
consumed meat and those who did so less frequently illiterate household heads indicated a high frequency
(t = − 0.2779 and P = 0.7812). On the other hand, literacy of meat consumption while about 83% of the literate
Deneke et al. BMC Public Health (2022) 22:222 Page 10 of 17

households consumed meat at high frequency. 63.9% of respondents reported that they had experienced diseases
the respondents habitually consumed raw meat (mainly and symptoms after eating raw meat, such as abdominal
beef ) and about 20% were in the habit of consuming raw discomfort, tape worm, amoeba, gout and even TB.
meat either every day or 2–5 times a week. However, Farmers were also asked if they knew that TB can be
more than a third of the respondents (36.1%) indicated transferred from animals to humans through consump-
that they had never consumed raw meat (Table 3B). tion of raw meat. Out of 477 respondents, 62.3% (297)
An investigation was conducted into the relationship indicated that they thought eating raw meat could do so,
between raw meat consumption frequency and demo- while 23.1% (n = 110) indicated that they did not know
graphic factors of which statistically significant asso- whether this was the case. Only 6.3% (n = 30) stated that
ciations were found between frequency and study site TB cannot be transferred from animal to human by eat-
(likelihood-ratio chi2(20) = 120.6; P < 0.001) and reli- ing raw meat.
gion (likelihood-ratio chi2(1) = 13.34; P < 0.001), respec- We also investigated whether a relationship exists
tively. Muslims tended to avoid raw meat, with 75% (15 between having attended training on zoonotic diseases
out of 20) of the surveyed Muslims indicating that they and bTB transmission pathways, and farmers’ meat con-
had never consumed raw meat. Among the study sites sumption behaviour. The results of this test indicated
surveyed, the proportion of dairy farmers who con- that there is a statistically significant relationship (like-
sumed raw meat more frequently (at least once in a fort- lihood-ratio chi2 (2) = 7.72; P = 0.021). Among the 348
night) was 66% in Addis Ababa, 77% in Oromia, 66% farmers who had not undertaken training on zoonoses
in Hawassa, and 67% in Gondar. However, only 25% in provided by local government extension services, 33.3%
Mekele had a habit of frequent raw meat consumption. indicated that they consumed raw meat frequently. In
No relationships between raw meat consumption fre- contrast, only 24.3% of the 123 farmers who undertook
quency and gender, literacy, or age were found in this training on zoonoses indicated that they consumed raw
survey. meet frequently. Our data also show that there was a sta-
tistically significant relationship between raw meat con-
Meat source preference sumption habit and past occurrence of TB in the family
As shown in Table 3C, the most preferred source of meat (likelihood-ratio chi2 (2) = 5.68; P = 0.017). Out of the 48
across the sample of farmers was found to be butchery farm households who reported that there has been a con-
(53.9%), followed by home slaughter (36.6%) and then firmed human TB case in the last 3 years in their farm,
communal slaughter (9.5%). 20.8% indicated that they have the habit of raw meat
The relationship between meat source ranking and consumption while the 79.2% for those farm households
variables such as gender of the household head, liter- reported no TB case in the past 3 years.
acy status, religion, study site, and age were examined.
We found a statistically significant association between Determinants of raw milk and meat consumption
study site and the main source of meat (likelihood-ratio Determinants of raw milk consumption frequency
chi2(8) = 126.4; P < 0.001). Most farmers from Addis The results of this analysis indicate that the independ-
Ababa city (76%) indicated that butchery was their pri- ent variables in our model are good predictors of the fre-
mary source of meat, compared to only 6.9% among quency of raw milk consumption (LR chi square of 109.2,
farmers from Gondar in the Amhara region. Also mean significant at 1% confidence level; P < 0.001 and Pseudo
age and the primary source of meat were found to be R-square of 0.184). Among the variables entered into
statistically different (F = 4.15; P = 0.016) between those the model, we found study site, gender of the household
households using butchery (45 years) as a primary source head, previous animal bTB testing in farm, knowledge
and those using home slaughter as a primary source of zoonotic risk of milk consumption, household size,
(49 years). The results indicated that none of the other and per-capita milk consumption levels were important
socioeconomic factors were associated to the primary determinants of frequency of raw milk consumption
meat source of a household. among the studied dairy farm households (Table 4).
The results of the model suggest that, as compared
Knowledge of zoonoses to farm households in Addis Ababa, being in the Oro-
The interviewed farmers were also asked if they think mia towns surrounding Addis Ababa increased both the
that eating raw meat can cause diseases. The result shows probability of raw milk consumption as well as its fre-
that the vast majority of farmers (92.9%) believed that quency. Being from Gondar decreased the probability of
consumption of raw meat can cause diseases and about a respondent consuming raw milk.
40% had actually experienced disease symptoms which The gender of the household head was also found to
they attributed to eating raw meat (Table 3D). Many be an important determinant of raw milk consumption
Deneke et al. BMC Public Health (2022) 22:222 Page 11 of 17

habits. Our results indicate that a household head being also indicates that being based in Mekele reduced both
male increased the probability of raw milk consuming, the probability of consumption of raw meat as well as its
but not the frequency of that consumption. frequency.
Awareness of bTB due to previous testing of cattle at The gender of the household head was also found to
farm decreased both the probability of raw milk con- affect the probability of raw meat consumption, but not
sumption and its frequency. Its effect was more pro- the frequency of that consumption; a household head
nounced on decreasing the frequency of consumption, being male significantly increased the probability of
indicating that although there has been change in behav- members of that household consuming raw meat. Train-
iour regarding raw milk consumption due to awareness of ing on zoonotic disease transmission risks was found to
a farm’s bTB status, this change seems to have impacted have an effect on the frequency, rather than the probabil-
more in decreasing the frequency at which raw milk was ity of consumption, meaning those households which had
consumed, rather than halting its consumption alto- access to zoonosis training tended to report a lower fre-
gether. In addition, knowledge of the possible zoonotic quency of raw meat consumption as compared to those
risk associated with raw milk consumption had its own who did not have access to zoonosis training.
effect on the raw milk consumption behaviour of farm- The result showed that age of dairy farmers had a posi-
ers, in that it was found to significantly decrease the fre- tive effect on raw meat consumption frequency up to
quency of raw milk consumption but not the probability some limits but the effect of age on the frequency of raw
that raw milk would be consumed at all. meat consumption turned to be negative as farmers got
Household size and per capita milk consumption were older. However, age did not have a significant effect on
found to be important determinants of raw milk con- the probability of raw meat consumption. Young farmers
sumption habits. With an increase in household size, the tended to have higher raw meat consumption frequency
probability of raw milk consumption and the frequency and as farmers got old they tended to decrease the fre-
of raw milk consumption were found to increase signifi- quency of raw meat consumption.
cantly. Similarly, with higher per capita milk consump- Interestingly, having ‘knowledge of the effects of raw
tion, both the probability of raw milk consumption and meat consumption on the risk of zoonotic transmis-
its frequency were found to increase significantly. The sion of diseases’ had a statistically significant effect on
reasons behind this are not clear but could be due to the both the probability of raw meat consumption and its
probable increased costs associated with boiling more frequency, and with a higher impact on the former. This
milk or purchasing more pasteurised milk consumed by means having knowledge about the risks involved in con-
more people per capita. sumption of raw meat negatively affected both the deci-
sion to consume raw meat as well as its frequency, but it
Determinants of raw meat consumption frequency affected the former much more than the later.
The results of the analysis (Table 4) indicate that some of Our data also suggests that the effect of meat type pref-
the independent variables in our model are good predic- erence, i.e. a farmer who preferred beef meat, also had
tors of the frequency of raw meat consumption (LR chi positive and significant effect on the probability of eat-
square = 156.3, p-value = 0.000 (SD = 99%)). Study site, ing raw meat; however, no effect on the probability of the
gender of household head, knowledge about zoonotic raw meat consumption frequency was seen. This might
risks associated with raw meat consumption, training be due to the suitability of beef meat for raw meat-based
on zoonoses, age squared, household size, and per capita meals such as kitfo, kurt, and gored gored, their local
consumption of meat (of all types, either raw or cooked) names in the Amharic language.
were all found to be significant variables when predicting In the model, erd size was entered as a proxy variable
the frequency of an individual’s raw meat consumption, to capture the effect of wealth on raw meat consumption
i.e. which of the three stated categories (D-F) they would habits. The result indicated that the habit of consuming
fall into. raw meat was similar across the different wealth catego-
In terms of study site as a predictor, respondents based ries in Ethiopia and there was no statistically significant
in the Oromia towns surrounding Addis Ababa were difference between the behaviour of the rich and the poor
more likely to consume raw meat than those in Addis in this regard.
Ababa city, however, the frequency of such consump- The data showed that high consumption of meat in
tion was not significant. As compared to respondents general (expressed as ‘Per capita meat consumption’) did
from the capital, an average household based in Gondar not affect the probability of raw meat consumption but it
did not consume more raw meat, but dairy households did increase the probability of doing so more frequently.
in Gondar tended to consume raw meat more fre- Increased family size was also found to be linked with
quently than those in Addis Ababa. However, the data increased frequency of raw meat consumption.
Deneke et al. BMC Public Health (2022) 22:222 Page 12 of 17

Discussion rather than demographic variables such as sex and lit-

Milk consumption eracy, differences in raw milk consumption by study site
Compared to the national average, the higher per capita might be related to differences in facilities available for
milk consumption observed in this study is not surpris- milk processing in different locations across Ethiopia.
ing as we surveyed dairy farmers who should have better Absence of any relationship between training on zoon-
access to milk and who are likely to consume more milk oses given and raw milk consumption could suggest that
than the general public. About 20% of the study popu- the training given were not adequate for precipitating
lation had the habit of drinking raw milk at least once a behavioural change among the people, and/or because of
month while the vast majority drank the milk boiled. As their positive perceptions of nutritional qualities, good
many as a 25% drank boiled milk on a daily basis while taste or health benefits of consuming raw milk as indi-
nearly 90% drank it on a weekly basis. Less than 10% cated by Oliver et al. [17]. The findings of this study on
drank pasteurised milk with any frequency. Although farmers’ perception of the healthiness of drinking raw
nearly four out of five sampled farmers said that they milk is in line with a previous study that suggested that,
never drank raw milk, we found that over 80% consumed even in cases where a considerable proportion of the
fermented milk, the ergo yoghurt. This result is similar to society have knowledge of zoonotic diseases, the prac-
other studies that also found high rates of yoghurt con- tice of consumption of boiled or pasteurised milk was
sumption in Ethiopia, especially among adults [49, 50]. found to be low [16]. This could be related to the fact
Whether ergo can still contain live pathogens such as that although people are aware of the risk of infection,
M. bovis after being fermented has yet to been proven, they may not always notice an infection after drinking
but a study from South Africa have shown that M. bovis raw milk. In the case of bTB in particular, transmission
can survive in both fresh and souring milk for periods of of the causative agent M. bovis may occur from unpro-
time that represent a risk of exposure to people consum- cessed milk, but an infected person may only develop a
ing these products. However, the conditions for survival latent infection and not display apparent clinical symp-
were dependent on both storage temperature and dose toms until later in life. In those cases, people may not
of pathogen in the milk products [51]. In Ethiopia, raw associate TB with their raw milk consumption and they
milk consumption is a common practice in various parts may also ignore their awareness of the associated risk and
of the country. According to Negash et al. [52], 50% of continue drinking raw milk, especially if they believe that
the milk produced by smallholder farmers in the Ethio- it also has positive qualities, such as a superior taste.
pian Rift Valley areas was consumed at home, in its fresh The low level of pasteurised milk consumption in our
form, without being boiled or pasteurised. According to study is likely associated with the under-developed milk
Ayele et al. [53], 35% of dairy farmers included in a sur- pasteurisation practice in Ethiopia which has resulted
vey around Sebeta in central Ethiopia, which is also one in that the vast majority (95%) of all milk produced in
of the study areas in the current study, indicated that they the country is sold through informal marketing systems
drank raw milk and only 13% of these farmers were aware without passing through pasteurisation plants [55]. In
of food borne diseases, which can be transmitted through such informal market systems, farmers often sell their
drinking raw milk. Another study showed that as many as milk directly to consumers and there is no mechanism for
67% of the interviewed farmers in North Western Ethio- regulating bacteriological quality standards of the milk.
pia drank raw milk [54]. Tolosa et al. [50] also indicated Peoples’ knowledge of the importance of pasteurisation is
that in the Jimma area of Western Ethiopia, 57% of the also limited. According to Girma [56], in a survey con-
adults drank fermented milk sporadically and 14% of ducted in North Shewa area in Ethiopia, only 3.5% of the
the interviewees did not boil the milk for their children. respondents knew about pasteurisation as a means of
Therefore, these findings suggest that some considerable preventing milk borne zoonosis. The high level of boiled
proportion of the society in Ethiopia drink raw milk in milk consumption practice in this study is in line with the
its fresh or fermented form, and if not on a regular basis, findings by Lemma et al. [57] and Duguma and Janssens
at least occasionally. In comparison to these figures, our [49]. However, the difference in frequency of boiled milk
result showed a relatively low level of raw milk consump- consumption by study site could be due to differing levels
tion frequency, possibly because our study sites were of awareness about the prevalence and health impacts of
urban and peri-urban areas where people have better bTB and other zoonotic diseases, but would warrant fur-
access to information regarding the zoonotic risks associ- ther study.
ated with drinking raw milk. According to the results of our generalised ordered
Among all socioeconomic variables studied, the statis- logit model there are differences in milk consumption
tically significant systematic relationship seen between behaviour across study areas which might be attrib-
study site and raw milk consumption habit implies that, uted to differences in access to relevant information and
Deneke et al. BMC Public Health (2022) 22:222 Page 13 of 17

pasteurised milk, as well as differences in perceived risk to in Addis Ababa and nearby areas in central Ethiopia, as
of contracting diseases due to raw milk consumption. well as in Hawassa in the southern part.
Further studies could usefully investigate the detailed Overall, the sampled farmers favoured butcheries, who
knowledge of farmers in different areas, as well as docu- tend to source their meat from abattoirs, as their first
menting what kinds of training they have received. The choice for sourcing meat. However, location or study site
regional differences on milk consumption should also be affected their expressed preferences surrounding meat
noted by policy makers seeking to design cost-effective sources. A significant majority of farmers in Addis Ababa,
strategies for preventing and controlling zoonotic dis- over 75%, stated that they preferred to source meat from
eases through public behaviour change and/or invest- butchers, but this proportion was extremely low, at only
ment in milk pasteurisation technology. 6.9%, in Gondar. This finding is also likely to be related to
Our analysis suggests that awareness and knowledge butcheries in Addis Ababa being both greater in number
about bTB and/or other zonootic risks of consuming raw and in their capacity. It is also possible that those farm-
milk had the effect of decreasing raw milk consumption ers who live in the growing metropolis of Addis Ababa,
but without stopping the consumption completely. This where land is at a premium, may have less space in which
could be due to perceptions about the nutritional quali- to slaughter animals themselves.
ties, good taste or health benefits of raw milk, but it could The results of the analysis using the generalised ordered
also be due to the fact that raw milk was easily, quickly, logit model showed relationships between a number of
and conveniently available to the population that we sur- variables and both the probability of any raw meat con-
veyed. Boiling, fermenting, or pasteurising milk takes sumption and also the frequency of raw meat consump-
time and energy and may sometimes be abandoned in tion amongst those who reported that they were in the
favour of the most easily available option of drinking the habit of eating raw meat. Identifying such relationships
milk raw, despite the associated risks. This implies that may prove useful to policy-makers and veterinary and
repeated and multichannel education and information medical professionals seeking to understand and influ-
dissemination need to be given in order to bring about ence raw meat consumption behaviours in Ethiopia, par-
permanent behavioural change among farmers in terms ticularly in the context of preventing and controlling the
of raw milk consumption. transmission of zoonotic diseases, of which raw meat
consumption is a recognised risk factor (e.g. [59–61]).
Meat consumption In terms of study site, our findings suggest that farmers
Our overall finding that the per capita meat consump- living in the Oromia towns surrounding Addis Ababa are
tion across the sampled farmers was considerably higher more likely to consume raw meat than those in the capital
than the national average was not surprising, given that itself. We also found that those farmers who do eat raw
our study sites were all located in urban or peri-urban meat are more likely to do so at high frequency if they
areas where meat consumption rates tend to be higher live in Gondar. For policy-makers and public health offi-
[58]. Within this population, with an overall higher meat cials seeking to encourage reduced raw meat consump-
consumption, households with female heads were found tion, these findings suggest that their resources might
to consume less meat per capita than those with male be best used in the Oromia towns and in Gondar, as
heads. This may be attributed to differences in income opposed to in Addis Ababa, or indeed in Mekelle, where
between those with household heads of different genders raw meat consumption is lower. However, as it is likely
and/or different nutritional priorities, but it was not pos- that raw meat consumption rates in Mekelle are lower
sible to investigate this further using our current dataset. because of the relatively underdeveloped beef abattoir
In this regard, it should be added that whilst it might be industry in Mekelle, it would be sensible to monitor raw
expected that those households with literate heads might meat consumption in that town as the industry develops,
have higher per capita meat consumption on account of as meat type preferences are likely to shift as a result.
potentially having higher incomes, this was not the case Across the different research study sites, the results
in our sample. of the generalised ordered logistic model analysis show
In terms of meat source preference, we found that those that older people are less likely to consume raw meat
farmers living in Hawassa, Addis Ababa city, and the at high frequency. This may be due to people accumu-
Oromia towns surrounding Addis Ababa were consider- lating knowledge of the health risks surrounding raw
ably more likely to prefer beef as their first-choice meat, meat consumption over the life course and/or a general
while those in Gondar and Mekelle were more likely to shift in diet as people age. In focus group discussions
prefer mutton. This difference can be linked to the rela- carried out as part of the ETHICOBOTS project, dairy
tively slow development of beef abattoirs and associated farmers in the Oromia towns around Addis Ababa
industry in the Amhara and Tigray regions, as compared reported that some older people no longer consume
Deneke et al. BMC Public Health (2022) 22:222 Page 14 of 17

any meat because it is believed to speed up the aging Conclusions

process [30]. With the assumption that the study areas Implications for prevention and control of Zoonoses
represent urban and peri-urban centres in Ethiopia, the The present study interviewed nearly 500 urban and peri-
result indicates that being vegetarian in Ethiopia is not urban dairy farmers in major towns of Ethiopia about
a common habit (0.6%). Most people eat meat, yet they their milk and meat consumption patterns to understand
do not eat much. The reasons behind the low level of the potential risks of zoonotic disease transmission.
vegetarianism might be that the generally low level of On their milk consumption behaviours, although con-
meat consumption did not yet cause higher prevalence sumption of heat-treated milk was the most common,
of obesity an associated health risks in Ethiopia which we concluded that around 20% of the study population
in turn made people not to be serious about watching still drank raw milk at least on a monthly basis. Given
diets and adopt the habit of vegetarianism. the high prevalence of bTB in the cattle population in
Although having received training on zoonoses did the explored study areas, the habit of raw milk consump-
not seem to affect the probability that someone ever tion may expose farmers to zoonotic TB transmission.
ate raw meat, it was found that having knowledge of Despite the need for additional clinical enquiry, empirical
the specific risks of zoonotic transmission from eating evidence from this study suggest that there was a statisti-
raw meat did have a statistically significant effect on cally significant link between self-reported TB infection
both the probability of raw meat consumption and its amongst humans on the farm and the bTB status of its
frequency. In fact, our data show that it had a higher cattle. Therefore, bTB control strategies should aim to
impact on the probability of consumption of raw meat raise awareness among the dairy farming as well as in the
than the probability of high frequency of raw meat general population of the possible zoonotic risks involved
consumption. This indicates that while training on in raw milk consumption and the importance of pasteuri-
zoonotic disease control in general seemed to have an sation and milk boiling in mitigating these risks. Moreo-
impact of the amount of raw meat eaten, special care ver, variables such as location, gender, household size
should be taken to include information about risks and per capita milk consumption need to be considered
associated with raw meat consumption in these train- in any effort to induce voluntary behavioural change sur-
ings. Further training could play an important role in rounding raw milk consumption habits to tackle the risks
shaping consumption behaviours and consequently of transmission of zoonotic diseases.
represents a solid investment for public health pol- With regards to meat consumption behaviours, given
icy makers concerned about transmission of disease that consumption of raw meat presents a risk of contract-
through raw meat, due to poor inspection or handling ing diseases such as bTB, salmonella, taeniasis and others
of infected carcasses [62]. through zoonotic transmission, and that frequent con-
Among| those who reported a confirmed TB case in sumption increases this risk, the urban and peri-urban
humans in their family in the last 3 years prior to our dairy farming population are exposed to a considerable
survey, 82.8% indicated a high raw meat consump- level of zoonotic risk. However, the level of exposure to
tion frequency, suggesting that the consumption of such risks may vary from town to town and is based on
raw meat may increase the risk of TB transmission in disease prevalence, local cultural orientations, and rela-
Ethiopia. However, it should also be noted that 61.2% tive availability of infrastructure in place, such as abat-
of those respondents with no reported TB cases in toir services including their meat inspection praxises. It
the family also consumed raw meat at high frequency. should also be stressed that, as most animals in Ethio-
Those delivering training on the risks associated with pia are likely to not be slaughtered based on controlled
consuming raw meat should therefore also pay atten- hygienic practices, but rather at home or at communal
tion to the cultural value of such practices and commu- slaughtering, the risk of contaminating the meat used for
nicate to trainees that a risk of disease transmission by consumption during slaughter, and thereby the risk of
no means represents a certainty of that transmission. zoonotic transmission, increases. Increased capacity for
If this subtlety is not effectively conveyed, it is possi- slaughter at controlled abattoirs as well as improvement
ble for distrust to emerge between trainers and train- of routine meat inspections at such abattoirs are there-
ees who, going on their own experience of contact with fore recommended.
supposed risk factors without any obvious harm, may Besides designing control strategies for reducing dis-
come to the conclusion that the risk is not genuine. ease prevalence in animals at farm level in general, cre-
It should also be made clear that the consumption of ating risk awareness about zoonotic disease transmission
raw meat is not only associated with active, confirmed to consumers through training and media campaigns,
TB cases, but also with latent TB, which may not show increase availability of pasteurised or heat-treated milk,
symptoms for many years [61]. improving meat hygiene through better abattoir services,
Deneke et al. BMC Public Health (2022) 22:222 Page 15 of 17

and inducing behavioural change around meat sourcing (NRERC No. 3.10/800/07) for the period September 2014 until end of Janu-
ary 2021. This approval letter is renewed every year (The first and the latest
and raw meat consumption, are all crucial to the success- approval letters are submitted).
ful prevention and control of the spread of zoonotic dis- Farmers who took part in the interview were household heads. Hence, we
eases, including bTB. confirm that none of the subjects who participated in the study are under
sixteen.
We confirm that the data collected through interview were anonymous, con-
Supplementary Information fidential and in accordance with international and national ethical guidelines.
The online version contains supplementary material available at https://doi. Informed consent was obtained verbally from dairy farm owners who were
org/10.1186/s12889-022-12665-4. briefed in the presence of a witnesses (local experts) on the tuberculin skin
testing procedure and the study questionnaire. Farmers were briefed that
there is no known risks to the animal associated with tuberculin skin testing,
Additional file 1. their participation in study is voluntary and that confidentiality on test and
Additional file 2. survey results will be maintained. The vast majority of the farmers interviewed
were literate and could understand the nature of our study and verbally gave
their willingness to participate in the study; even those small number of
Acknowledgements farmers who were illiterate were literate enough to understand the nature of
The members of the Ethiopia Control of Bovine Tuberculosis Strategies our study and gave their informed verbal consent. The ethical approval we got
(ETHICOBOTS) consortium are: Abraham Aseffa, Adane Mihret, Bamlak Tes- from AHRI-ALERT is for the entire project that also covers collection of data
sema, Bizuneh Belachew, Eshcolewyene Fekadu, Fantanesh Melese, Gizachew from participant farmers using informed verbal consent.
Gemechu, Hawult Taye, Rea Tschopp, Shewit Haile, Sosina Ayalew, Tsegaye
Hailu, all from Armauer Hansen Research Institute, Ethiopia; Rea Tschopp from Consent for publication
Swiss Tropical and Public Health Institute, Switzerland; Adam Bekele, Chilot Not applicable.
Yirga, Mulualem Ambaw, Tadele Mamo, Tesfaye Solomon, all from Ethiopian
Institute of Agricultural Research, Ethiopia; Tilaye Teklewold from Amhara Competing interests
Regional Agricultural Research Institute, Ethiopia; Solomon Gebre, Getachew The authors declare that they have no competing interests.
Gari, Mesfin Sahle, Abde Aliy, Abebe Olani, Asegedech Sirak, Gizat Almaw,
Getnet Mekonnen, Mekdes Tamiru, Sintayehu Guta, all from National Animal Author details
1
Health Diagnostic and Investigation Center, Ethiopia; James Wood, Andrew Amhara Regional Agricultural Research Institute, Bahar Dar, Ethiopia.
2
Conlan, Alan Clarke, all from Cambridge University, United Kingdom; Henrietta Ethiopian Institute of Agricultural Research, Addis Ababa, Ethiopia. 3 Institute
L. Moore and Catherine Hodge, both from University College London, United for Global Prosperity, University College London, London, UK. 4 National Animal
Kingdom; Constance Smith at University of Manchester, United Kingdom; R. Health Diagnostic and Investigation Center, Sebeta, Ethiopia. 5 Armauer
Glyn Hewinson, Stefan Berg, Martin Vordermeier, Javier Nunez-Garcia, all from Hansen Research Institute, Addis Ababa, Ethiopia. 6 Swiss Tropical and Public
Animal and Plant Health Agency, United Kingdom; Gobena Ameni, Berecha Health Institute, Basel, Switzerland. 7 University of Basel, Basel, Switzerland.
8
Bayissa, Aboma Zewude, Adane Worku, Lemma Terfassa, Mahlet Chanyalew, Cambridge University, Cambridge, UK. 9 Animal and Plant Health Agency,
Temesgen Mohammed, Miserach Zeleke, all from Addis ababa University, Weybridge, UK.
Ethiopia.
Received: 17 August 2021 Accepted: 25 January 2022
Authors’ contributions
The authors contributed mainly to this research in the following ways:
Conception and design: TTD, AB, RT, JLNW, HLM, SB. Field data collection: TM,
GA, GAM, AM, RT. Data analysis: TTD, AB, TM, LY, CH, SB. Manuscript writing and
editing: All authors. All authors have read and approved the final manuscript. References
All authors have agreed both to be personally accountable for the author’s 1. Grace D, Mutua F, Ochungo P, Kruska R, Jones K, Brierley L, et al. Mapping
own contributions and to ensure that questions related to the accuracy or of poverty and likely zoonoses hotspots: Zoonoses project 4. Kenya: ILRI;
integrity of any part of the work, even ones in which the author was not per- 2012.
sonally involved, are appropriately investigated, resolved, and the resolution 2. Etter E, Donado P, Jori F, Caron A, Goutard F, Roger F. Risk analysis
documented in the literature. and bovine tuberculosis, a re-emerging zoonosis. Ann N Y Acad Sci.
2006;1081:61–73.
Funding 3. Ishihara A, Hashimoto E, Ishioka H, Kobayashi H, Gomi H. Campylobacter
The Biotechnology and Biological Sciences Research Council, the Department fetus meningitis associated with eating habits of raw meat and raw
for International Development, and the Economic and Social Research Council liver in a healthy patient: a case report and literature review. IDCases.
funded this work. The Medical Research Council, the Natural Environment 2018;11:97–100.
Research Council and the Defense Science & Technology Laboratory were 4. Molla B, Alemayehu D, Salah W. Sources and distribution of Salmo-
also part of the funding. The funding was under the Zoonoses and Emerging nella serotypes isolated from food animals, slaughterhouse personnel
Livestock Systems (ZELS) Programme, ref.: BB/L018977/1. SB was also partly and retail meat products in Ethiopia: 1997-2002. Ethiop J Health Dev.
funded by the Department for Environment, Food & Rural Affairs, United 2003;17(1):63–70.
Kingdom, ref.: TBSE3294. 5. Olaya-Galan NN, Corredor-Figueroa AP, Guzman-Garzon TC, Rios-
The funders had no role in study design, data collection and analysis, or prepa- Hernandez KS, Salas-Cardenas SP, Patarroyo MA, et al. Bovine leukaemia
ration of the manuscript. virus DNA in fresh milk and raw beef for human consumption. Epidemiol
Infect. 2017;145(15):3125–30.
Availability of data and materials 6. de la Rua-Domenech R. Human Mycobacterium bovis infection in
All raw data collected and used for this study are available as supplementary the United Kingdom: incidence, risks, control measures and review
material. of the zoonotic aspects of bovine tuberculosis. Tuberculosis (Edinb).
2006;86(2):77–109.
7. WHO, OIE, FAO, Union T. The Roadmap for Zoonotic Tuberculosis. A Call to
Declarations Action: WHO, OIE, FAO, and The Union; 2017. p. 24. https://www.fao.org/
documents/card/en/c/f6091ab7-cb91-4639-9a94-89b7137428a5/.
Ethics approval and consent to participate 8. Sibhat B, Asmare K, Demissie K, Ayelet G, Mamo G, Ameni G. Bovine
This study was approved by AHRI-ALERT Ethics Review Committee (Project tuberculosis in Ethiopia: a systematic review and meta-analysis. Prev Vet
Reg.No PO46/14) and Ethiopia’s National Research Ethics Review Committee Med. 2017;147:149–57.
Deneke et al. BMC Public Health (2022) 22:222 Page 16 of 17

9. Almaw G, Conlan AJK, Ameni G, Gumi B, Alemu A, Guta S, et al. The 33. Regassa A, Medhin G, Ameni G. Bovine tuberculosis is more prevalent in
variable prevalence of bovine tuberculosis among dairy herds in Central cattle owned by farmers with active tuberculosis in Central Ethiopia. Vet J.
Ethiopia provides opportunities for targeted intervention. PLoS One. 2008;178(1):119–25.
2021;16(7):e0254091. 34. Tigre W, Alemayehu G, Abetu T, Ameni G. Preliminary study on the epi-
10. Firdessa R, Tschopp R, Wubete A, Sombo M, Hailu E, Erenso G, et al. High demiology of bovine tuberculosis in Jimma town and its surroundings,
prevalence of bovine tuberculosis in dairy cattle in Central Ethio- Southwestern Ethiopia. Afr J Microbiol Res. 2012;6(11):2591–7.
pia: implications for the dairy industry and public health. PLoS One. 35. Agency CS. Population projections for Ethiopia 2007–2037. Addis Ababa:
2012;7(12):e52851. Central Statistics Agency of Ethiopia; 2013.
11. Mekonnen GA, Conlan AJK, Berg S, Ayele BT, Alemu A, Guta S, et al. Preva- 36. CSA. Report on livestock and livestock characteristics (private peasant
lence of bovine tuberculosis and its associated risk factors in the emerg- holdings). vol. II. Addis Ababa: Central Statistical Agency of Ethiopia; 2020.
ing dairy belts of regional cities in Ethiopia. Prev Vet Med. 2019;168:81–9. 37. Shapiro BI, Gebru G, Desta S, Negassa A, Nigussie K, Aboset G, et al.
12. Kleeberg HH. Human tuberculosis of bovine origin in relation to public Ethiopia livestock master plan. Nairobi: International Livestock Research
health. Rev Sci Tech Off Int Epiz. 1984;3(1):11–32. Institute (ILRI); 2015.
13. FAO: Status of and Prospects for Smallholder Milk Production - A Global 38. Skuce RA, Brittain D, Hughes MS, Beck LA, Neill SD. Genomic fingerprint-
Perspective: FAO (Food and Agriculture Organization of the United ing of Mycobacterium bovis from cattle by restriction fragment length
Nations); 2010. polymorphism analysis. J Clin Microbiol. 1994;32(10):2387–92.
14. Kimman T, Hoek M, de Jong MCM. Assessing and controlling health risks 39. Long JS, Freese J. Regression models for categorical dependent variables
from animal husbandry. Netherlands J Agric Sci Wageningen J Life Sci. using stata. 3rd ed. College Station: Stata Press; 2014.
2013;66(1):7–14. 40. O’Connell AA. Logistic regression models for ordinal response variables.
15. Yilma Z, Emannuelle GB, Ameha S. A Review of the Ethiopian Dairy Sec- Thousand Oaks: SAGE; 2006.
tor. In: Fombad R, editor. Food and Agriculture Organization of the United 41. Agresti A. An introduction to categorical data analysis. New York: Wiley;
Nations, Sub Regional Office for Eastern Africa (FAO/SFE), Addis Ababa, 2007.
Ethiopia; 2011. p. 81. 42. Williams R. Generalized ordered logit/partial proportional odds models
16. Seyoum ET, Mekonen TK, Kebede N, Gezahegn HA, Mehirete TS, for ordinal dependent variables. Stata J. 2006;6(1):58–82.
Mengesha ZT. Knowledge, attitude and practice among small scale dairy 43. Brant R. Assessing proportionality in the proportional odds model for
farmers on MilkBorne zoonotic diseases, north Showa zone, Ethiopia. J ordinal logistic regression. Biometrics. 1990;46(4):1171–8.
Foodborne Zoonotic Dis. 2016;4(2):19–28. 44. Fu VK. Estimating generalized ordered logit models. In: Stata Technical
17. Oliver SP, Boor KJ, Murphy SC, Murinda SE. Food safety hazards associated Bulletin 44: 27–30. In Stata Technical Bulletin Reprints, vol. 8; 1998. p.
with consumption of raw milk. Foodborne Pathog Dis. 2009;6(7):793–806. 160–4.
18. Aydin FE, Ulger M, Emekdas G, Aslan G, Gunal S. Isolation and identifica- 45. Peterson B, Harrell FE. Partial proportional odds models for ordinal
tion of Mycobacterium bovis and non-tuberculous mycobacteria in raw response variables. Appl Stat. 1990;39(2):205–17.
milk samples in Mersin province. Mikrobiyol Bul. 2012;46(2):283–9. 46. Liu X, Koirala H. Ordinal regression analysis: using generalized ordinal
19. Ben IK, Boschiroli M, Souissi F, Cherif N, Benzarti M, Boukadida J, et al. logistic regression models to estimate educational data. J Mod Appl Stat
Isolation and molecular characterisation of Mycobacterium bovis from Methods. 2012;11(1):242–54.
raw milk in Tunisia. Afr Health Sci. 2011;11(Suppl 1):S2–5. 47. FAO, NZAGGRC. Supporting low emissions development in the Ethiopian
20. Kazwala RR, Daborn CJ, Kusiluka LJ, Jiwa SF, Sharp JM, Kambarage DM. dairy cattle sector – reducing enteric methane for food security and liveli-
Isolation of Mycobacterium species from raw milk of pastoral cattle of the hoods. Rome: FAO & New Zealand Agricultural Greenhouse Gas Research
southern highlands of Tanzania. Trop Anim Health Prod. 1998;30(4):233–9. Centre; 2017.
21. Ashford DA, Whitney E, Raghunathan P, Cosivi O. Epidemiology of 48. Tafere K, Worku I. Consumption Patterns of Livestock Products in Ethiopia:
selected mycobacteria that infect humans and other animals. Rev Sci Elasticity Estimates Using HICES (2004/05) Data. In: Working Paper 38:
Tech. 2001;20(1):325–37. International Food Policy Research Institute; 2012.
22. Seleshe S, Jo C, Lee M. Meat consumption culture in Ethiopia. Korean J 49. Duguma B, Janssens GPJ. Assessment of dairy farmers’ hygienic milking
Food Sci Anim Resour. 2014;34(1):7–13. practices and awareness of cattle and Milk-borne Zoonoses in Jimma,
23. Insoll T. The archaeology of Islam in sub-sharan Africa: Cambridge World Ethiopia. Food Sci Qual Manage. 2015;45:114-21.
Archeology: Cambridge University Press; 2003. 50. Tolosa T, Verbeke J, Piepers S, Tefera M, Getachew Y, Supre K, et al.
24. Francis J. Letter: very small public health risk from flesh of tuberculous Milk production, quality, and consumption in Jimma (Ethiopia): facts
cattle. Aust Vet J. 1973;49(10):496–7. and producers’, retailers’, and consumers’ perspectives. Prev Vet Med.
25. Merkal RS, Whipple DL. Inactivation of Mycobacterium bovis in meat 2016;124:9–14.
products. Appl Environ Microbiol. 1980;40(2):282–4. 51. Michel AL, Geoghegan C, Hlokwe T, Raseleka K, Getz WM, Marcotty
26. Ayele WY, Neill SD, Zinsstag J, Weiss MG, Pavlik I. Bovine tuberculo- T. Longevity of Mycobacterium bovis in raw and traditional sour-
sis: an old disease but a new threat to Africa. Int J Tuberc Lung Dis. ing Milk as a function of storage temperature and dose. PLoS One.
2004;8(8):924–37. 2015;10(6):e0129926.
27. (AACCSA) AACoCaSA. Value chain study on meat processing industry in 52. Negash F, Estefanos T, Esayas A, Chali Y, Feyisa H. Production, handling,
Ethiopia. In: Addis Ababa Chamber of Commerce and Sectoral Associa- processing, utilization and marketing of milk in the mid Rift Valley of
tions (AACCSA); 2015. Ethiopia. Livest Res Rural Dev. 2012;24:152. Retrieved February 2, 2022,
28. Asseged B, Woldesenbet Z, Yimer E, Lemma E. Evaluation of abattoir from http://www.lrrd.org/lrrd24/9/nega24152.htm.
inspection for the diagnosis of Mycobacterium bovis infection in cattle at 53. Ayele Y, Gutema FD, Edao BM, Girma R, Tufa TB, Beyene TJ, et al. Assess-
Addis Ababa abattoir. Trop Anim Health Prod. 2004;36(6):537–46. ment of Staphylococcus aureus along milk value chain and its public
29. Biffa D, Bogale A, Skjerve E. Diagnostic efficiency of abattoir meat inspec- health importance in Sebeta, Central Oromia, Ethiopia. BMC Microbiol.
tion service in Ethiopia to detect carcasses infected with Mycobacterium 2017;17(1):141.
bovis: implications for public health. BMC Public Health. 2010;10:462. 54. Nuru A, Mamo G, Zewude A, Mulat Y, Yitayew G, Admasu A, et al. Prelimi-
30. Hodge C, Deneke TT, Endalew MA, Moore HL, Consortium E. Acceptability nary investigation of the transmission of tuberculosis between farmers
and feasibility of proposed control and prevention strategies for bovine and their cattle in smallholder farms in northwestern Ethiopia: a cross-
tuberculosis among Ethiopian dairy farmers and associated professionals. sectional study. BMC Res Notes. 2017;10(1):31.
Prev Vet Med. 2020;185:105184. 55. Getabalew M, Alemneh T, Zewdie D. The milk processing: status, chal-
31. Asseged B, Lubke-Beker A, Lemma E, Kiros T, Britton S. Bovine tubercu- lenges and opportunities in Ethiopia. Int J Vet Sci Res. 2020;6(1):52–7.
losis: a cross sectional and epidemiological study in and around Addis 56. Girma Y. Assessment of community knowledge, attitude and practice on
Ababa. Bull Anim Health Prod Afr. 2001;48:71–80. Milk born Zoonoses disease in Debre-Birhan town, north Shewa, Ethiopia.
32. Ameni G, Amenu K, Tibbo M. Prevalence and risk factor assessment in J Vet Sci Technol. 2017;8:482.
cattle and cattle owners in Wuchale-Jida District, Central Ethiopia. J Appl
Res Vet Med. 2003;1(1):17–26.
Deneke et al. BMC Public Health (2022) 22:222 Page 17 of 17

57. Lemma HD, Mengistu A, Kuma T, Kuma B. The potential of milk produc-
tion and consumption for improving nutrition of smallholder dairy
households in Ethiopia. Milk Sci Int. 2017;70:10–6.
58. Betru S, Kawashima H. Pattern and determinants of meat consumption
in urban and rural Ethiopia. Livest Res Rural Dev. 2009;21:143. Retrieved
February 2, 2022, from http://www.lrrd.org/lrrd21/9/betr21143.htm.
59. Wabeto W, Abraham Y, Anjulo AA. Detection and identification of
antimicrobial-resistant Salmonella in raw beef at Wolaita Sodo municipal
abattoir, southern Ethiopia. J Health Popul Nutr. 2017;36(1):52.
60. van Bree FPJ, Bokken G, Mineur R, Franssen F, Opsteegh M, van der Gies-
sen JWB, et al. Zoonotic bacteria and parasites found in raw meat-based
diets for cats and dogs. Vet Rec. 2018;182(2):50.
61. Teklu T, Legesse M, Medhin G, Zewude A, Chanyalew M, Zewdie M, et al.
Latent tuberculosis infection and associated risk indicators in pastoral
communities in southern Ethiopia: a community based cross-sectional
study. BMC Public Health. 2018;18(1):266. https://doi.org/10.1186/
s12889-018-5149-7.
62. Mhlanga A. Assessing the impact of optimal health education programs
on the control of zoonotic diseases. Comput Math Methods Med.
2020;2020:6584323.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional affiliations.

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission

• thorough peer review by experienced researchers in your field
• rapid publication on acceptance
• support for research data, including large and complex data types
• gold Open Access which fosters wider collaboration and increased citations
• maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions