Prevalence of Salmonella in raw milk, pasteurized milk, and cottage cheese
The prevalence of Salmonella in raw milk was significantly higher (19.7%) compared to pasteurized milk (11.5%; p = 0.002). When comparing Salmonella prevalence at different locations in the value chain, milk samples taken from collectors had the highest prevalence of Salmonella (21.3%; p = 0.01), whereas samples collected from retailers had the lowest prevalence (9.3%). Prevalence of Salmonella in samples collected from producers was lower (18.1%) compared to those collected from collectors, but higher compared to those collected from processors (13.7%; p = 0.01).
The overall prevalence of Salmonella was highest in Oromia (17.2%), followed by SNNP (15.6%), and Amhara (11.5%); however, the differences in the prevalence among regions were not statistically significant (p = 0.26).
Salmonella prevalence in milk samples collected in Oromia was highest at the production level (21.9%), followed by 20.8% at the collection, 15.6% ate the processing, and 9.5% at the retail level. Meanwhile, in SNNP, the highest prevalence was observed at the collection (22.9%), followed by production (16.7%), processing (14.6%), and retail (8.3%) levels. In Amhara region, the prevalence was highest at the collection (20%), followed by retail (10%), and production and processing levels (7.5%).
Table 2 presents the prevalence of Salmonella between pasteurized and raw milk. Salmonella prevalence differed significantly between raw and pasteurized milk only in the Oromia region (p = 0.015). Furthermore, the prevalence of Salmonella in raw milk was highest in Oromia (21.4%), followed by SNNPs (19.8%), and Amhara (14.1%); however, these differences were not statistically significant (Table 3). Similarly, the prevalence of Salmonella in samples from pasteurized milk was highest in Oromia (12.5%), followed by SNNP (11.5%), and Amhara (9%), but the differences were not statistically significant (Table 4).
Table 2
Prevalence of Salmonella in raw and pasteurized milk samples collected in each study region in Ethiopia.
Region | Milk type | Sample size | Salmonella positive (n) | Salmonella positive (%) | ORa | χ2 b | P-value |
Oromia | Raw milk | 192 | 42 | 21.8 | 1.6 | 5.9 | 0.015 |
Pasteurized milk | 192 | 24 | 12.5 | 1 |
Total | 384 | 66 | 17.2 | |
SNNPs | Raw milk | 96 | 19 | 19.79 | 1.5 | 2.5 | 0.112 |
Pasteurized milk | 96 | 11 | 11.45 | 1 |
Total | 192 | 30 | 15.6 | |
Amhara | Raw milk | 80 | 11 | 14.1 | 1.4 | 4.0 | 0.267 |
Pasteurized milk | 76 | 7 | 9.2 | 1 |
Total | 156 | 18 | 11.5 | |
a OR, odds ratio. |
b χ2, chi-square. |
Table 3
Overall prevalence of Salmonella in raw milk by region and within value chain.
Variable | Observation | | Sample size | Salmonella positive (n) | Salmonella positive (%) | χ2 a | P value |
Region | Amhara | | 78 | 11 | 14.1 | 1.9 | 0.391 |
Oromia | | 192 | 41 | 21.4 | | |
SNNPs | | 96 | 19 | 19.8 | | |
Total | | 366 | 71 | 19.4 | | |
Value chain | Producer | | 183 | 33 | 18.0 | 0.4 | 0.509 |
Collector | | 183 | 38 | 20.8 | | |
Total | | 366 | 71 | 19.4 | | |
a χ2, chi-square. |
Table 4
Overall prevalence of Salmonella in pasteurized milk by region and within value chain.
Variable | Observation | | Sample size | Salmonella positive (n) | Salmonella positive (%) | χ2 a | P value |
Region | Amhara | | 78 | 7 | 9.0 | 0.7 | 0.712 |
Oromia | | 192 | 24 | 12.5 | | |
SNNPs | | 96 | 11 | 11.5 | | |
Total | | 366 | 42 | 11.5 | | |
Value chain | Producer | | 183 | 25 | 13.7 | 1 | 0.190 |
Retailer | | 183 | 17 | 9.3 | | |
Total | | 366 | 42 | 11.5 | | |
a χ2, chi-square. |
The prevalence of Salmonella in cottage cheese was 6.3% in samples collected from producers, as well as from retailers in Oromia. Salmonella prevalence was similar in samples collected from producers in Amhara (6.2%) and SNNP (4.2%). No Salmonella was detected in cottage cheese samples collected from retailers in Amhara nor in SNNP (Table 5). The differences in prevalence between producers and retailers within each region were not statistically significant, as shown in Table 1.
Table 5
Prevalence of Salmonella in cottage cheese samples collected from producers and retailers in three study regions in Ethiopia.
Region | Value chain | Sample size | Salmonella positive (n) | Salmonella positive (%) | χ2 a | P value |
Oromia | Producer | 48 | 3 | 6.25 | 0 | 1.000 |
Retailer | 48 | 3 | 6.25 |
Total | 96 | 6 | 6.25 |
SNNPs | Producer | 24 | 1 | 4.2 | 1.4 | 0.235 |
Retailer | 24 | 0 | 0.0 |
Total | 48 | 1 | 2.1 |
Amhara | Producer | 16 | 1 | 6.2 | 1 | 1.000 |
Retailer | 16 | 0 | 0.0 |
Total | 32 | 1 | 3.1 |
a χ2, chi-square. |
Table 1
Prevalence of Salmonella in raw and pasteurized milk, along a value chain, and in different regions of Ethiopia.
Variable | Observation | Sample size | Salmonella-positive (n) | Salmonella-positive (%) | ORa | χ2 b | P-value |
Milk Type | Raw milk | 366 | 72 | 19.7 | 1.6 | 9.4 | 0.002 |
Pasteurized milk | 366 | 42 | 11.5 | 1 |
Total | 732 | 114 | 15.6 | |
Value chain | Producer | 183 | 33 | 18.1 | 1.8 | 11.4 | 0.01 |
Collector | 183 | 39 | 21.3 | 2 |
Processor | 183 | 25 | 13.7 | 1.4 |
Retailer | 183 | 17 | 9.3 | 1 |
Total | 732 | 114 | 15.6 | |
Region | Oromia | 384 | 66 | 17.2 | 1.4 | 2.7 | 0.260 |
| SNNPs | 192 | 30 | 15.6 | 1.3 |
| Amhara | 156 | 18 | 11.5 | 1 |
| Total | 732 | 114 | 15.6 | |
a OR, odds ratio. |
b χ2, chi-square. |
It is important to note that this assessment has evaluated more variables than presented in Table 6, however, many of these tested variables were statistically insignificantly. Therefore, only variables with statistical significance are summarized in Table 6 to ease the readability and data interpretation.
Table 6
Risk factors associated with Salmonella contamination at milk production and collection level.
Value chain | Variable | Response | Na | n (%)b | 95% CIc of APd | ORe | χ2 f | p-value |
Milk producers | Have you ever attended food safety training? | Yes | 103 | 26 (25.2) | 17.2–34.7 | 2.7 | 9.8 | 0.002 |
No | 81 | 6 (7.4) | 2.7–15.4 | 1 |
Total | 184 | 32 (17.4) | 12.2–23.6 | |
What water temperature do you use for cow udder washing? | Cold | 62 | 16 (25.8) | 15.5–38.5 | 1.7 | 4.5 | 0.03 |
Warm | 115 | 15 (13.1) | 7.5–20.6 | 1 |
Total | 177 | 31 (17.5) | 12.2–23.9 | |
Do you mix milk from a diseased animal with that from healthy ones? | Yes | 1 | 1 (100.0) | 2.5 | 7.2 | 4.7 | 0.029 |
No | 102 | 17 (16.7) | 10.0–25.3 | 1 |
Total | 103 | 18 (17.5) | 10.7–26.2 | |
What type of milk handling container do you use? | Aluminum can | 13 | 6 (46.1) | 19.2–74.8 | 2 | 8.1 | 0.018 |
Mazzi can | 7 | 1 (14.3) | 0.3–57.8 | 1 |
Plastic container | 164 | 25 (15.2) | 10.1–21.7 | 1.1 |
Total | 184 | 32 (17.4) | 12.2–23.6 | |
Milk collectors | How do you refrigerate milk? | Bulk tankers | 4 | 1 (5.0) | 0.1–24.8 | 1 | 6.5 | 0.039 |
Deep freezers (-20°C) | 4 | 7 (16.7) | 6.9–31.4 | 3 |
Refrigerators (+ 4°C) | 29 | 17 (30.4) | 18.8–44.1 | 4.5 |
Total | 37 | 25 (21.2) | 14.6–26.6 | |
What type of milk handling containers do you use? | Aluminum cans | 15 | 21 (20.8) | 13.3–30.0 | 1.1 | 8.6 | 0.013 |
Plastic containers | 31 | 9 (13.2) | 6.2–23.6 | 1 |
Both | 12 | 7 (46.7) | 21.3–73.4 | 1.6 |
Total | 58 | 37 (20.1) | 14.6 | |
Do you filter milk? | Yes | 49 | 29 (25.0) | 17.4–33.9 | 1.8 | 4.6 | 0.031 |
No | 9 | 8 (11.8) | 5.2–21.8 | 1 |
Total | 58 | 37 (20.1) | 14.6–26.6 | | |
aN, sample size. |
bn (%), number of positive isolates (percentage of postive isolates). |
cCI, confidence interval. |
dAP, apparent prevalence. |
eOR, odds ratio. |
f χ2, chi-square. |
Risk factors associated with Salmonella contamination of milk and cottage cheese
Risk factors for milk contamination at the production level
Among 184 surveyed milk producers, 32.6% lived in urban and 67.4% in peri-urban areas. The gender ratio among surveyed participants was nearly equal (52.2% men and 47.8% women). Among surveyed farmers, the majority had only informal education (49.6%), followed by elementary school (52.8%), high school (40.7%), preparatory (13.7%), and a college degree (30.3%). Surveyed participants were asked about hygiene practices carried out at the farm. Survey results showed that washing udders of cows before milking was performed by 94% of farmers. Sixty-seven percent of the surveyed farmers also dried washed udders utilizing pieces of clothes. The above outlined variables were not significantly associated with the prevalence of Salmonella (p > 0.05).
Based on the survey and microbial data, a lower prevalence of Salmonella was observed in raw milk collected from untrained study participants of milk hygienic handling-related training, while milk collected from trained study participants had a higher prevalence of Salmonella (25.2%; p = 0.002). Milk collected from individuals that used cold water for cleaning cow udder was more frequently contaminated with Salmonella (25.8%) compared to milk collected from individuals utilizing hot water (13.1%; p = 0.03). All milk samples collected from producers that mixed milk collected in the morning and evening were contaminated with Salmonella, whereas Salmonella was detected in just 16.7% of samples from participants that did not mix morning and evening milk batches (p = 0.029). Finally, Salmonella prevalence in milk collected from participants that stored milk in aluminum cans was higher (46.1%) compared to those that stored milk in plastic cans (15.2%) and mazzi cans (14.3%; p = 0.018).
Our data indicate that training related to milk handling, water temperature used for udder washing, mixing of evening and morning batches of milk, and types of milk handling containers all had a significant effect on the prevalence of Salmonella in milk (Table 6). Based on study participants’ responses, most producers had access to training related to milk production and hygienic milk handling (e.g., housing, feeding, milking, and health issues). Surprisingly, data analysis indicated a higher prevalence of Salmonella in products collected from respondents that attended a training compared to respondents that did not (p = 0.002). It is, however, important to note that this study did not evaluate the curriculum nor the effectiveness of these trainings, making it difficult to make inferences about this counter-intuitive finding. Further research on this topic could help to reveal the factors that may be contributing to this outcome. Finally, the practice of mixing milk from evening and morning milking showed a greater association with Salmonella contamination than in those who do not (p = 0.029; Table 6).
Risk factors for milk contamination at the collection level
A total of 58 milk collectors participated in the survey, with 79.3% of them living in urban areas. Among surveyed individuals, there were 51.7% women and 48.3% men. Approximately a quarter of participants (24.2%) had primary school education, 36.2% completed high school and had a diploma/college degree, and 8.6% completed preparatory school. One third of the respondents had 1–2 or 2–5 years of experience collecting milk. According to the survey, 65.9% of study participants collected milk by foot, while 20.3% used a three-wheel drive and 13.8% used a four-wheel drive. In terms of cooling systems used during transportation of milk to collection centers, 63.8% of surveyed individuals indicated that they had a cooling system (e.g., refrigerator). At the collection center, almost all (91.4%) of surveyed participants indicated that they store milk containers on a concrete floor, and 96.5% used tap water to clean the containers. The above outlined factors were not statistically significantly associated with Salmonella prevalence (p > 0.05).
Refrigeration system, milk containers, and milk filtration used by collectors were significantly and positively associated with Salmonella contamination (Table 6). In contrast, Salmonella prevalence was significantly lower in milk samples collected from collectors that use refrigerated bulk tanks (5%) but was higher for users of + 4°C refrigerators (30.4%) and for deep freezer (-20°C) users (16.7%) (p = 0.039). The prevalence of Salmonella was significantly higher in milk samples collected from collectors that used both plastic containers and aluminum cans compared to those that used mazzi can (p = 0.013). Lastly, milk filtration was associated with increased prevalence of Salmonella. Specifically, milk collected from collectors that filtered milk was two times more likely to be contaminated with Salmonella compared to milk collected from collectors that did not filter milk (p = 0.031).
Risk factors for milk contamination at processing level
A total of 12 milk processors were surveyed in this study. Seventy-five percent of the participants lived in urban areas and 92% were men. Nearly half of them (42%) had more than ten years of experience with milk processing, 17% had less than a year of experience, 16% had 5–10 years of experience and 25% had 2–5 years of experience. Almost all participating processors (92%) have had had basic food safety training. Most participants (67%) cleaned milk containers twice a day, while the rest did so once a day. The majority of respondents (67%) used tap water for cleaning milking equipment, while the rest used groundwater. Twenty-five percent of the respondents used a piece of cloth to sieve the milk, while 67% of them used microfiltration and 8% used a metal sieve. Almost all (92%) of the respondents reported having cold storage for pasteurized milk. None of these variables had a significant effect on the prevalence of Salmonella.
Factors associated with an increased Salmonella prevalence at the milk processing level included sanitation and inspection procedures. For example, it revealed the prevalence had doubled in the evaluators of tests that include: alcohol, organoleptic, Lacto scan, and microbiology compared with alcohol, organoleptic and microbiological test analyzers (excluding Lacto scan in the latter). It was almost the same as only alcohol and organoleptic test assessors (p = 0.006). It was important to point out that not all processors used the same factors. Consequently, these processors were classified based on their various testing methodologies, and the results were analyzed accordingly. Risk factors for milk contamination at the retail level
A total of 181 milk retailers from Oromia, Amhara, and SNNP were interviewed. Fifty-three percent of the respondents were male. The assessment of the educational background revealed that 3.3%, 24.3%, 30.4%, 8.3%, and 33.7% had informal, primary school, high school, preparatory, and college diploma/degree, respectively. The majority of the respondents (79.6%) had less than five years of experience in retail. Almost none of the respondents had access to hygienic/safe milk handling training.
Over half of surveyed participants (51.4%) acquired milk from whole sellers, 45.8% from factories, and 2.8% at the factories gate. Most surveyed retailers (78.3%) used trucks with cooling systems to deliver milk to the retailing station, whereas around 21.7% used four-wheel vehicles without a cooling system. A temperature-controlled cooling system was used by 49.7% of retailers. The majority of retailers (70.7%) had a separate refrigerator for milk storage until sale. None of the above-listed factors were statistically significantly associated with Salmonella contamination.
In this study, factors such as training on milk quality and safety, the use of hot or cold water to clean the udder of the animals, the practice of mixing the morning and night milk, the storage of milk containers, and the refrigeration of milk were significantly associated with Salmonella contamination. The conclusion from this point is that dairy farmers should not only focus on the money they earn for the production of milk but also protect the health of the community and themselves. Otherwise, salmonellosis will indeed take root and affect the health and economy of the community, reducing production and productivity.
Risk factors for contamination of cottage cheese at the production level
A total of 89 cottage cheese producers were surveyed and 56% of them were from the peri-urban area. Ninety-two percent were women. Around one-third of the respondents were illiterate and 30% had primary school education, 27% had high school education, and 5% had preparatory school education and had diplomas/degrees. The respondents' experience in years was less than a year, 1–2 years, 2–5 years, 5–10 years, and more than ten years for 7%, 20%, 32%, 15%, and 27% of respondents, respectively.
Almost all of the surveyed cottage cheese producers (92%) received training on proper handling and storage of cottage cheese. The cottage cheese was placed in whey by 60% of the respondents, while 34% used a refrigerator, and 7% used a vessel of cold water. The majority (54%) walked to the market, 21% used public transportation, 18% drove a three-wheel drive, and 7% rode in an animal-drawn cart. For 36% of them, the time it took to get to the marketplace was less than 30 minutes, for 43% it took 30 minutes to one hour and for 21% it took more than one hour. Most respondents (97%) cleaned the cottage cheese containers regularly. The most frequent container washers (77%) used water and detergents, while the rest used different herbs. None of the factors outlined above were statistically significantly associated with contamination of cottage cheese with Salmonella.
Risk factors for cottage cheese contamination at the retailer level
The production system was the only variable in the value chain of cottage cheese retailers in this study that revealed statistical significance. Milk from peri-urban residents were 10 times more exposed to Salmonella than urban dwellers (85.6%; p = 0.009). Females comprised almost four-fifths of the respondents, while males were 20%. In informal, primary school, high school, preparatory, and diploma/degree, the educational background evaluation indicated 2.2%, 24.5%, 22.2%, 4.4%, and 46.7%, respectively. The respondents' experience in years was less than a year, 1–2 years, 2–5 years, 5–10 years, and more than ten years accounted for 17.8%, 18.9%, 42.2%, 8.9%, and 12.2% of the respondents, respectively.
The majority (77.8%) of individuals interviewed in this value chain did not train on how to handle and store cottage cheeses properly. A quarter of them (26.7%) walked to the market, 20% used public transportation, 15.6% drove three-wheel drives, 5.5% used four-wheels, 28.9% used refrigerated vehicles, and 6.7% traveled in animal-drawn carts. Over half of them (52.2%) kept cottage cheese in a separate refrigerator until sold. The majority (71.1%) had backup generators, which they used when the power went out.