Livestock Research for Rural Development 28 (1) 2016 Guide for preparation of papers LRRD Newsletter

Citation of this paper

Improving efficiency of the dairy value chain in Uganda; effect of action research-based interventions on milk quality and safety

J B Kaneene, P Ssajjakambwe1, S Kisaka1, P Vudriko1, R Miller and J D Kabasa1

Center for Comparative Epidemiology, Michigan State University,
736 Wilson Road Room A109, East Lansing, MI, 48824-1314,
kaneenej@cvm.msu.edu
1 College of Veterinary Medicine, Animal Resources and Biosecurity, Makerere University,
PO Box 7062, Kampala, Uganda

Abstract

The aim of this research was to measure the effectiveness of the AgShare university–led action research–based approach to educating farmers in the dairy value chain, through measures of milk safety and quality before and after participation in the program.  A longitudinal intervention study was conducted in southwestern Uganda from January to November, 2011. The intervention was to provide farmers with education and training to improve dairy production, milk quality, and milk safety, emphasizing hygiene and the importance of sub–clinical mastitis. Farm hygiene scores, milk, and sera were collected to measure milk quality and safety at the beginning of AgShare and 11 months after participation. Milk quality was determined by somatic cell count scores, and milk safety was measured by the prevalence of mastitis and brucellosis. Qualitative assessments of AgShare impact were conducted by an external reviewer at the end of the program, and through interviews with study participants.

Mastitis was prevalent at both baseline and after intervention, and prevalence of mastitis decreased after participation in AgShare. Improved milk quality and safety were significantly associated with increasing hygiene scores. Over 90% of participants indicated that AgShare was beneficial, and farmers’ communities benefitted by employing similar interventions and methods in their own dairy production systems.  AgShare had beneficial impacts on participating farms, through improving milk quality and safety and farm hygiene practices. Improvements in farmer knowledge and practices were spread to community members by the participants themselves. The external reviewer recommended that AgShare should be continued in small–scale livestock production systems.

Key Words: brucellosis, epidemiology, mastitis, milk hygiene, heard health


Introduction

The economy of Uganda is agricultural-based (Grimaud et al 2009), and the dairy value chain has been identified by the Government of Uganda as an area to improve household incomes and livelihoods. This prioritization has created employment and income for over 700,000 dairy farmers and other stakeholders in the dairy value chain (Mbowa et al 2012). However, challenges to the dairy sector in Uganda remain, including livestock diseases that lead to production losses and directly impact the safety of milk for human consumption. These diseases include brucellosis (Mwebe et al 2010; Mugabi et al 2012; Miller et al 2015), clinical mastitis (Byarugaba et al 2008; Abrahmsén et al 2014; Kasozi et al 2014; Mugenyi 2014), and subclinical mastitis (SCM), which is more common than clinical mastitis and is difficult to diagnose by farmers (Björk 2013). While the quality and safety of dairy products are extensively monitored in the developed world (Oliver et al 2009), they have not gained similar levels of attention in sub–Saharan Africa (Bonfoh et al 2003). Milk quality and safety interferes with the processing of milk products (Le Maréchal et al 2011), and reduces the shelf life of yoghurt and pasteurized milk (Hachana and Paape 2012).

The AgShare project was implemented to help address existing challenges to the improvement of the dairy value chain in Uganda (Kaneene et al 2013). AgShare used academic, public and private sector partnerships in a university–led action research–based model to improve livestock–based agriculture and communities by disseminating knowledge, skills, and community service through an information loop system for research, education, and knowledge. AgShare project was implemented in two phases. In Phase I, AgShare described the condition of the existing dairy value chain from farm to markets, and conducted needs assessments from different dairy value chain stakeholder groups, and institutions that will produce graduates to work in the dairy industry. Baseline data was collected for use in measuring the effectiveness of interventions to improve dairy production, milk quality, and the safety of milk produced by participating farmers. The goals of Phase II were to develop interventions based on information from Phase I, and implement them through the action research information loop system: students interact with participating farmers on a monthly basis, and both students and farmers provide feedback to university researchers, who use this information to create new or improve existing interventions. Results from the action research information loop were also used by AgShare to develop open education research learning modules on milk production and hygiene for use by universities and other stakeholders (Kaneene et al 2013; Ssajjakambwe et al 2013).

The purpose of this study was to measure the effectiveness of the AgShare model in improving milk quality and safety. The study hypotheses were that, for dairy farmers in Uganda, participation in student–led action research would improve milk quality and milk safety, by reducing the prevalence of mastitis and brucellosis in dairy cows. The objectives of the study were to collect metrics for milk quality and safety (total somatic cell count (SCC), prevalence of SCM and brucellosis in the milking herd), and farm milking hygiene practices from dairy farmers at the beginning and 11 months later, and assess the impact of AgShare using quantitative and qualitative assessments. For quantitative analyses, measures of milk quality and safety, and farm milking hygiene practices, were compared between baseline and after participation in AgShare. For the qualitative analyses, an external review was conducted to describe farmers’ opinions and attitudes about milking hygiene, milk quality, cattle health and milk safety, and the effectiveness of the AgShare project.


Materials and Methods

Study design

A longitudinal intervention study was conducted in the sub-counties of Kashongi and Keshunga, in Kiruhura district, in south-western Uganda from January to November 2011. Six dairy farms from Keshunga were purposively selected based on previous participation in AgShare Phase 1, and six were randomly selected by the Kashongi area veterinarian, based on a history of brucellosis in the sub-county.

The AgShare intervention

The AgShare dairy value chain project used an action research information loop (Figure 1) to deploy interventions and use stakeholder feedback to modify and improve interventions (Kaneene et al 2013). Rather than using a standardized approach to improve dairy farming, the AgShare intervention was to engage farmers through education and training (Ssajjakambwe et al 2013), develop management plans specific to their needs and concerns, and to monitor how farmers utilized these plans and collect feedback from farmers through monthly visits.

Strategies for improving the productivity, safety, and quality of dairy farmers were developed from a series of needs assessments conducted with different stakeholder groups during Phase I. Farmers and interested parties attended the meetings, and one–on–one sessions were arranged to discuss the project with potential participants. Based on findings from the needs assessments and baseline data describing dairy farms in Phase I, three issues were identified for interventions: diseases affecting milk production; milk quality and safety; and mastitis and brucellosis prevention and treatment. Educational materials were developed for student training to aid farmers in improving dairy production, quality, and safety, by emphasizing dairy production hygiene and the importance of subclinical mastitis.

Baseline information about participants’ dairy management knowledge, attitudes and practices was collected through discussion sessions, and questionnaires based on issues identified in Phase I. Educational materials on farm hygiene and SCM were provided by students and AgShare researchers to farmers. Specific issues were identified, based on the questionnaires and discussions with each farmer, and strategies for improving farm hygiene and SCM detection were developed and implemented by the farmer.

Each farm was visited by students every month during their participation in the project. During farm visits, AgShare students assessed farm hygiene and progress on the farm’s management strategy. This information, along with the farmers’ assessments of the intervention, were relayed back to university researchers to develop new interventions and improve upon the existing interventions and curriculum, thereby closing the information loop (Figure 1). During the course of the project, farmers were able to seek advice from local veterinarians, students, and faculty. Any laboratory reports on the disease status of the dairy herd were discussed with farmers on a one–to–one basis.

Figure 1. Action Research Information Loop for the AgShare Dairy Value Chain Intervention
Data collection, sample collection and processing

Hygiene scores, milk, and blood samples were collected at the beginning of farmer participation (baseline) and 11 months after the baseline sample. To determine the prevalence of SCM, 25% of lactating cows were randomly selected from each herd, excluding cows with a recent history of calving because the presence of colostrum in the milk affects the reliability of the mastitis screening test. All dairy cows and serving bulls on the farm were eligible for sampling to determine the prevalence of brucellosis.

The farm hygiene score

Farm hygiene index scores (Table 1) were generated for overall farm hygiene, and for farm environment, milking equipment, and milker personal hygiene. Hygiene scores were subjectively scored by trained study personnel on a scale that ranged from very good (4) to poor (1).

Table 1. Components of the farm hygiene score

Index

Score

Description

Index

Score

Description

Manure Disposal

4

Use of manure pit, well drained sleeping area

Other solid waste disposal (empty drug containers, polyethene bags)

4

None seen on the farm

3

Drainage not good, patches of accumulated manure and urine

3

No drug containers, polyethene bags are visible on the farm

2

No manure pit, poor drainage, 50% of sleeping area soiled but not muddy

2

Drug containers, polyethene bags visible in small amounts

1

No manure pits, very poor drainage, 50% of sleeping area heavily soiled and muddy

1

Drug containers, polyethene bags visible everywhere

Milking facility hygiene

4

Milking in clean designated facility/area

Milking utensil hygiene

4

Clean stainless steel or aluminum utensils

3

Milking in dirty designated facility/area

3

Dirty stainless steel or aluminum utensils

2

Milking in clean area, poor environmental hygiene

2

Clean plastic utensils

1

Milking in dirty area, poor environmental hygiene

1

Dirty plastic utensils

Milking cattle cleanliness

4

Clean flanks, hind legs

Udder cleanliness

4

Clean udder

3

Manure on the parts of flanks, hind legs

3

Udder has few patches of dirt

2

Extensive manure on flanks, hind legs

2

Dirty udder

1

Solid caked manure on flanks, hind legs

1

Dirty udder with solidified cakes of manure

Water for cleaning udder and utensils

4

Water is clean, in clean containers

Udder towel hygiene

4

Clean individual paper towels used

3

Water is clean, containers are not properly cleaned

3

Clean cotton towel used, shared between cows

2

Water is dirty, containers looked clean

2

Moderately dirty cotton towel used, shared between cows

1

Water and storage containers dirty

1

Very dirty cotton towel used, shared between cows

0

No udder towel used

Milker hand washing

4

Use clean water and soap before and between cows

Milker personal hygiene

4

Clean overalls, nails short and clean

3

Use clean water and soap before, rinse hands between cows

3

Clean street clothes, nails somewhat short and clean

2

Use clean water without soap before, may not wash hands between cows

2

Dirty street clothes, finger nails long and somewhat dirty

1

Use dirty water without soap, does not wash hands between cows

1

Very dirty street clothes, nails very long and dirty

Milk sample collection and testing with the CMT

Milk samples from separate quarters of each cow were analyzed for the presence of SCM using California Mastitis Test (CMT) using standard procedures (Mellenberger and Roth 2000), and were interpreted on a scale that ranged from 0–4 (Table 2). Results were also reported as an SCC index, based on CMT scores from all four quarters that was used as an indicator of SCM. The SCC index ranged from negative (all quarter CMT=0 or all quarter CMT=1), trace/possible infection (1–3 quarters CMT=1, rest CMT=0), or positive (any quarter CMT=2–4).

Table 2. Criteria for scoring the California Mastitis Test for determination of mean somatic cell counts (SCC)

CMT Score

Mean SCC
(cells / ml)

Description of reaction

0 (negative)

100,000

No thickening, homogeneous.

0.5 (trace)

300,000

Slight thickening. Reaction disappears in 10 seconds.

1

900,000

Distinct thickening, no gel formation.

2

2,700,000

Distinct thickening, slight gel formation.

3

8,100,000

Gel is formed, surface elevates with a central peak above the mass.

Blood sample collection and testing with the Rose Bengal Plate Test

Blood samples were analyzed for the presence of Brucella antibodies using the Rose Bengal plate test (RBPT) using standard protocols (OIE 2012). Test results were based on the level of sample agglutination, and were interpreted as negative (no agglutination present) or weak, strong, or very strong positive, based on level of agglutination.

Statistical analysis
Quantitative assessments

Statistical analyses for the impact of AgShare on milk quality and safety were conducted at the herd level. Descriptive statistics were generated for all measures overall, at the baseline, and after 11 months. Changes in CMT score, SCC index, SCM status, brucellosis status, and hygiene score after participation in AgShare were assessed using the Mantel–Haenszel X2 and Fisher’s Exact test, and the strength of effect was reported as relative risk (RR) with 95% confidence intervals. For mean hygiene scores, one–way analysis of variance (ANOVA) and the Kruskall–Wallis X2 statistic for the Wilcoxon Rank–Sum test were used to identify significant differences in measures between baseline and after AgShare participation.

Associations between hygiene scores and milk quality and safety were evaluated by comparing hygiene scores between samples classified by the percent of quarters with CMT > 1,SCC index, SCM status, and brucellosis status, using ANOVA. Detailed breakdowns of the hygiene score were available for samples collected after 11 months, and the Mantel–Haenszel X2 was used to identify associations between hygiene indices and milk quality and safety.

Logistic regression models were developed to evaluate the effect of hygiene practices on SCM, using a backwards hierarchical model building strategy (Kleinbaum and Klein 2010). Model outcome was the SCM status of an individual cow after participation in AgShare, and potential risk factors included subcounty, hygiene indices, and SCM status at baseline. Univariable models were generated for each risk factor, and those with p< 0.2 were considered for inclusion in the multivariable model. The final model was determined as the best combination of statistically significant risk factors (p< 0.05) and model Akaike information criterion (AIC) statistic adjusted for small samples (Hurvich and Tsai 1993).

Qualitative assessments

Qualitative assessment of the impact of AgShare was conducted by an external reviewer (Worth Consulting, Howick, KwaZulu–Natal, South Africa) after the end of AgShare (December 2011). The goals of the assessment included evaluating the effectiveness of the project, and assessing the contribution of AgShare to improved dairy farming, community well–being, and economic development. A random group of participating farmers completed a structured questionnaire based on a set of qualitative and quantitative questions. The purpose of the questionnaire was explained to respondents via informal discussion sessions, and questionnaires were administered in person by trained interviewers from Makerere University.


Results

All AgShare participants were small–scale farmers. Several (42%) became involved through the local dairy cooperative or project officer, 25% were involved by friends or family members, and the remainder by the local village chief and local veterinarians. The motivation of the majority (92%) to participate in AgShare was to learn more about dairy production systems, while the remainder were interested in dairy product value addition. Most farmers (67%) expected to learn more about dairy farming and livestock diseases, 25% expected to learn more about improving dairy product marketing, and 8% believed AgShare would provide better access to veterinary medicines.

Most AgShare intervention plans included a mastitis management program, a brucellosis testing and vaccination program, and a farm worker and milk hygiene program. Over half of study participants (~58%) made use of their local veterinarian to test cows for brucellosis and to design a vaccination program. Approximately 17% of respondents were also advised to improve their milking facilities, and 25% were advised to implement or improve their record keeping systems.

Quantitative assessments
Dairy cattle

A total of 338 cattle were tested at baseline, and 320 cattle were tested after 11 months (Table 3). Milk from 275 lactating cows was collected at baseline and after AgShare started for mastitis testing. Blood samples were collected for brucellosis testing from 275 cattle at baseline, and 224 cattle after 11 months. Retrospective power analyses (SAS 9.3, SAS Systems, Inc., Cary NC) demonstrated that the sample size attained had power > 99%, and significance = 0.0315 for mastitis at baseline and after AgShare, and power = 46% and significance = 0.0640 for brucellosis.

Table 3. Numbers of cattle tested

Subcounty

Herd

Total
# Cattle

Baseline

After AgShare

# Cattle

# Sera

# Milk

# Cattle

# Sera

# Milk

Kashongi

BD

34

34

34

22

22

0

22

BN

43

43

32

20

43

32

20

IF

12

12

12

0

10

10

0

KE

15

15

15

8

15

15

8

KL

10

10

10

6

6

0

6

MB

17

17

17

17

17

12

17

Subcounty Total

131

131

120

73

113

69

73

Keshunga

KJ

49

49

32

20

49

32

20

KY

40

40

30

14

40

30

14

MF

30

30

20

17

30

20

17

OB

21

21

17

17

21

17

17

RD

28

28

25

20

28

25

20

TX

39

39

31

12

39

31

12

Subcounty Total

207

207

155

100

207

155

100

Grand Total

338

338

275

173

320

224

173

Mastitis prevalence and management practices

Subclinical mastitis was prevalent in both sub-counties at baseline and after intervention, and there were significant (p < 0.05) decreases in the prevalence of SCM after intervention (Table 4). Herds from Keshunga had significantly lower levels of SCM than herds from Kashongi (RR = 0.6, 95% C.I. = 0.4 – 0.9): this pattern was seen in samples at baseline and after participation, but was only significant for samples taken after participation. The highest SCM herd prevalence in baseline samples were from herds in Kashongi (87.5%), while the lowest prevalence were seen in two herds in Keshunga (0%). Milk collected at baseline had higher levels of CMT–positive samples than those after AgShare, while the percentage of CMT trace samples increased after AgShare participation. The percentage of quarters tested with CMT > 1 decreased from baseline to after AgShare (15.0% to 2.3%), and was seen in Kashongi (22.1% to 7.6%) and Keshunga (11.1% to 0%).

Most farmers were unaware of the significance of SCM at baseline, although the majority recognized the economic importance of clinical mastitis and requested training on mastitis control. Most farmers had no experience with the CMT, and did not understand its relevance in mastitis control programs. Only one farmer routinely used the CMT. Dry cow therapy was rarely mentioned, and milking salves  were rarely used, except in farms with a proper milking parlor. The majority of farmers did not use teat dips after milking. Whenever clinical mastitis was suspected, treatment was done mainly by the farmer or herdsmen using intramammary infusions.

Table 4. Test prevalence of subclinical mastitis as positive/negative and by the SCC Index, before and after AgShare and by subcounty

SCM

Factor

Level

N

%

Mantel–Haenszel

Relative Risk

Trace

Positive

X2

df

p

RR

95% C.I.

Positive/ Negative

Sample

Baseline

173

*

48.6

12.4

1

0.0004

0.45

0.29 – 0.71

After

171

*

30.1

Subcounty

Kashongi

146

*

46.6

5.58

1

0.0182

0.59

0.38 – 0.92

Keshunga

200

*

34.0

Kashongi by Sample Time

Baseline

73

*

54.8

3.94

1

0.0472

0.51

0.27 – 0.99

After

73

*

38.4

Keshunga by Sample Time

Baseline

100

*

44.0

8.87

1

0.0029

0.40

0.22 – 0.74

After

100

*

24.0

SCC

Index

Sample time

Baseline

173

3.5

46.8

29.1

1

< 0.0001

*

After

171

16.4

14.0

Subcounty

Kashongi

146

6.3

38.9

5.19

1

0.0227

*

Keshunga

200

12.5

24.5

Kashongi by Sample Time

Baseline

73

0

54.8

10.5

1

0.0012

*

After

73

12.7

22.5

Keshunga by Sample Time

Baseline

100

6

41.0

19.2

1

< 0.0001

*

After

100

19.0

8.0

Brucellosis prevalence and management practices

Brucellosis was prevalent at baseline and after intervention (Table 5), and there were significantly higher levels of brucellosis (p < 0.05) in Kashongi than in Keshunga. Brucellosis was identified in only one herd in Keshunga (herd prevalence = 9.7%), while the herd prevalence of brucellosis in Kashongi ranged from 22%–76% at baseline, and from 20%–67% after AgShare. Farmers in Kashongi reported the incidence of brucellosis was high on their farms, and all farms reported recent cases of abortions. There were decreases in the prevalence of brucellosis after participation in AgShare, but this trend was not significant (p = 0.064).The majority of farmers had not vaccinated their cattle for brucellosis, and 58% of study participants implemented brucellosis testing and vaccination programs after AgShare participation. Farmers expressed a strong need for brucellosis control programs by the government or the university.

Table 5. Test prevalence of brucellosis, before and after AgShare and by subcounty

Outcome

Level

Samples

Mantel–Haenszel

Relative Risk

N

% positive

X2

df

p

RR

95% C.I.

Brucellosis

Baseline

275

19.6

3.43

1

0.0640

1.31

0.97 – 1.78

After AgShare

224

13.4

Brucellosis

Kashongi

189

41.3

130

1

<.0001

0.02

0.01 – 0.07

Keshunga

310

1.94

Brucellosis at Baseline

Kashongi

120

42.5

70.3

1

<.0001

0.02

0. 01 – 0.09

Keshunga

155

1.94

Brucellosis after AgShare

Kashongi

69

39.1

56.7

1

<.0001

0.03

0.01 – 0.11

Keshunga

155

1.94

Milk quality

Of the 1,365 quarters tested, 80.3% were negative, 4.8% were trace, and 14.9% (n=203) were positive. Of the CMT–positive quarters, scores of 1, 2, and 3 were recorded for 50.7%, 36.5%, and 12.8%, respectively. Significant differences were seen in the percentage of negative and positive quarters per cow before and after AgShare, with increases in negative quarters and decreases in positive quarters overall and within each subcounty (Table 6). There were significantly higher percentages of CMT–positive quarters in Kashongi than Keshunga. However, the percentages of CMT–negative quarters from Keshunga were not significantly higher than Kashongi (p = 0.054).

Table 6. Results of the California Mastitis Test (CMT), before and after the AgShare interventions and by subcounty

Outcome

Risk Factor

Level

n

% quarters

ANOVA

F

p

% negative quarters per cow

Sample time

Baseline

173

73.1

22.3

< 0.0001

After AgShare

171

87.6

Subcounty

Keshunga

100

78.0

5.27

0.0229

Kashongi

73

66.4

Kashongi

Baseline

73

66.4

8.71

0.0037

After AgShare

71

82.8

Keshunga

Baseline

100

78.0

15.1

0.0001

 

After AgShare

100

91.0


% quarters with
CMT > 1 per cow


Sample time


Baseline


173


24.0


47.9


< 0.0001

After AgShare

171

5.7

Subcounty

Keshunga

100

19.5

5.24

0.0232

Kashongi

73

30.1

Kashongi

Baseline

73

30.1

16.1

< 0.0001

After AgShare

71

10.9

Keshunga

Baseline

100

19.5

37.8

< 0.0001

 

After AgShare

100

2.0

Hygiene scores and milk quality and safety

Hygienic practices were generally poor in most farms at baseline. The mean hygiene score at baseline was 2.21, with farms in Keshunga having higher scores than farms in Kashongi (Table 7). Milkers rarely washed their hands between cows, and towels used for cleaning udders before milking were shared between cows. There were significant improvements in the hygiene scores after AgShare for herds in Keshunga, but there was no significant change in Kashongi.

Table 7. Differences between baseline and post–intervention hygiene scores, overall and within subcounties

Subcounty

Time

N

Mean Hygiene
Score

ANOVA

F

p

Overall

Baseline

315

2.2

3.75

0.0532

After AgShare

326

2.3

Kashongi

Baseline

108

2.0

0.02

0.901

After AgShare

119

2.0

Keshunga

Baseline

207

2.3

6.43

0.0116

After AgShare

207

2.4

Significant associations were seen between hygiene scores and SCM and brucellosis (Table 8). Cattle with brucellosis or SCM had significantly lower mean hygiene scores than cattle without disease. The association between increasing hygiene with decreasing SCM was not statistically significant at baseline, but was significant for samples taken after 11 months.

Table 8. Baseline and post–intervention mean hygiene scores for baseline and post–intervention brucellosis and subclinical mastitis (SCM)

Outcome

Status

n

Mean
Score

ANOVA

F

p

SCM

Negative

210

2.3

11.58

0.0007

Positive

136

2.1

SCM at Baseline

Negative

90

2.2

2.93

0.0890

Positive

83

2.1

SCM after AgShare

Negative

121

2.3

11.87

0.0007

Positive

52

2.1

Brucellosis

Negative

391

2.3

17.00

< 0.0001

Positive

75

2.1

Brucellosis at Baseline

Negative

205

2.3

8.61

0.0037

Positive

47

2.1

Brucellosis after AgShare

Negative

186

2.4

6.56

0.0112

Positive

28

2.2

The strength of associations between hygiene scores with CMT scores were assessed after 11 months in AgShare (Table 9). The percentage of CMT-positive quarters significantly decreased with increasing milking facility cleanliness, cow cleanliness, and milker hygiene. When hygiene scores and CMT results were evaluated within sub-county, there were no significant associations, except for cattle cleanliness in Keshuga, where the association between increasing hygiene score with decreasing percentage of CMT-positive quarters was significant.

Table 9. Percent of quarters with subclinical mastitis (SCM) (CMT > 1) per cow, by hygiene index factors with p < 0.6

Hygiene factor

Score

n

% quarters/with SCM

F

p

Milking facility
cleanliness

1

71

10.9

4.48

0.0047

2

68

2.21

3

20

2.5

4

12

0


Milking utensil cleanliness


3


159


6.13


1.59


0.209

4

12

0


Milking cattle cleanliness
(flank and hock)


2


51


11.3


4.79


0.0094

3

106

3.77

4

14

0


Udder towel use


0


159


6.13


1.59


0.209

2

12

0


Hand washing


1


54


9.72


2.84


0.0613

2

105

4.29

3

12

0


Milker personal hygiene


1


28


13.4


4.77


0.0096

2

94

2.93

3

49

6.63

SCC index scores were associated with hygiene scores (Table 10). Increasing milk hand washing scores were associated with decreasing SCC trace and positive scores. Increasing cattle cleanliness scores were associated with decreasing positive SCC scores, but not for trace SCC scores. Significant associations were seen between decreasing SCM with increasing scores for milker hand washing, milker hygiene, cattle cleanliness, and udder cleanliness.

Table 10. Association between the SCC index and hygiene index factors with p < 0.6

Hygiene Index

Score

N

% cows with SCC Index

Mantel–Haenszel

Trace

Positive

X2

df

p

Milking facility

cleanliness

1

71

12.7

22.5

5.05

1

0.0246

2

68

20.6

8.82

3

20

20

10

4

12

8.33

0

Milking utensil cleanliness

3

159

17

15.1

3.17

1

0.0750

4

12

8.33

0

Milking cattle cleanliness
(flank and hock)

2

51

13.7

25.5

7.00

1

0.0081

3

106

17.9

10.4

4

14

14.3

0

Udder cleanliness

2

108

19.4

14.8

1.18

1

0.277

3

63

11.1

12.7

Udder towel use

0

159

17

15.1

3.17

1

0.0750

2

12

8.33

0

Hand washing

1

54

25.9

20.4

9.03

1

0.0027

2

105

12.4

12.4

3

12

8.33

0

Milker personal hygiene

1

28

21.4

25

1.76

1

0.185

2

94

18.1

9.57

3

49

10.2

16.3

Multivariable logistic regression was used to describe associations between different hygiene practices and SCM (Table 11). After univariable analyses, milking facility hygiene, cattle cleanliness, udder cleanliness, milker hand washing, and milker hygiene entered the multivariable model. The final model for SCM included the hygiene scores for cattle cleanliness, milker hand washing, and milker hygiene.

Table 11. Multivariable logistic regression model for subclinical mastitis (SCM) in cows after AgShare participation

SCM

n

AIC

R2

Risk Factor

Wald X2

Odds Ratio

X2

DF

P

O.R.

95% C.I.

Positive

52

192

20.7%

Cattle cleanliness

3.94

1

0.0472

0.42

0.18 – 0.99

Negative

121

Hand washing

12.7

1

0.0004

0.18

0.07 – 0.46

Milker personal hygiene

2.97

1

0.0846

2.10

0.90 – 4.89

Qualitative assessments

The external impact assessment concluded that the AgShare project was successful in improving dairy production systems, and recommended that AgShare should be continued in small–scale livestock production systems. The reviewer noted that the success of AgShare was in the action research conducted by students with farmers on farms, which enabled academics to embed theory in case studies from AgShare farmers for research–based teaching. For many university faculty, this was the first time that research findings were directed back to farmers, and farmer feedback became a research component. Another key to success was the stakeholder meetings held in Phase I, followed by analysis of production systems and feedback sessions. There was positive stakeholder feedback about the impact of the meetings, with all respondents indicating that they were beneficial.

The external reviewer found convincing data that there were very positive impacts on farmers’ practices that improved milk quality and safety. All participants reported that AgShare project members assisted them with hands–on training on their farms, and reported improvements in their dairy production systems. Almost all participants (92%) reported benefits from the information and training on disease prevention, reported decreases in SCM and other common cattle diseases, and recognized that prevention was less expensive than treating disease. The most significant improvements were in disease management and prevention (brucellosis vaccination, mastitis treatment), hygiene programs and improved record keeping systems. Farmers valued the testing and laboratory feedback from the university, appreciated the interaction with project members and veterinarians.


Discussion

AgShare impacts

Analysis of both quantitative and qualitative impact indicators showed that AgShare participants benefitted from the customized interventions implemented on their farms. Based on the Phase I needs assessments, most small dairy farmers experienced diseases affecting milk production, milk quality and safety, and disease prevention and treatment (SCM and brucellosis). AgShare interventions included a mastitis management program, brucellosis testing and vaccination program, and personal and milk hygiene program. Indicators of milk safety (SCM, brucellosis) and quality (CMT, SCC) both improved after participation, and farm hygiene improved.

Feedback from farmers was overwhelmingly positive, and it was clear that all respondents understood the goals of the project. Although only 58% of participants were positive about AgShare at its inception, everyone gave positive feedback at its conclusion. All respondents benefitted from the feedback sessions with students engaged in service learning, and learned about dairy management, and disease control and prevention programs. As an indicator of the importance of information learned through AgShare, 75% of respondents shared information from AgShare with neighbors, friends and other family members. Other studies have reported that farmers are very receptive to participatory training for improving livestock health, emphasizing the importance of face–to–face contact with veterinary health professionals, and sharing the results of training with other members of their communities (Vaarst et al 2007). Respondents were grateful that the project team took time to discuss the feedback with farmers on a one–on–one basis, and some indicated that this has been a flaw of other projects.

Impact of AgShare on milk quality and safety

The prevalence of subclinical mastitis in this study (49% at baseline and 30% after AgShare participation) was comparable to other studies using the CMT in Uganda. In these studies the prevalence of clinical mastitis has been reported from 49%–86% (Byarugaba et al 2008; Abrahmsén et al 2014; Mugenyi 2014). Other studies have reported the prevalence of SCM in Uganda to be from 34%–90% (Byarugaba et al 2008; Björk 2013; Kahuta 2013; Persson et al 2013; Kasozi et al 2014; Mugenyi 2014), which was higher than the 23.9% SCM found in baseline samples in this study. In addition, another study of SCM in central Uganda reported that over 70% of cattle tested had more than two quarters with CMT scores indicative of SCM (Kasozi et al 2014), which was higher than levels in the current study (15% of all cattle at baseline). The majority of AgShare farmers were unaware of subclinical mastitis, and only one AgShare farm used the CMT to detect mastitis. This agrees with one study of dairy farmers in western Uganda, which found that over 85% of 127 farm households did no testing for mastitis, and less than 2% used the CMT to detect SCM (Kahuta 2013).

The prevalence of brucellosis in this study was 20% at baseline and 13% after participation in AgShare, which agrees with earlier studies that reported the seroprevalence of brucellosis to be 12 – 26% (Mwebe et al 2010; Mugabi et al 2012; Miller et al 2015). The differences in the prevalence of brucellosis between sub-counties were expected, as farms in Kashongi were selected for participation in AgShare due to a history of brucellosis in the area. Other studies have reported that cattle from eastern districts had lower herd– and animal–level prevalence of brucellosis than districts in the central and western parts of the country (Magona et al 2009, Kashiwazaki et al 2012), while others have also found high levels of brucellosis in northern and western regions (Mwebe et al 2010). Some researchers have attributed these geographic differences to differences in cattle management practices (Magona et al 2009), but others could not confirm these associations (Nizeyimana et al 2013; Miller et al 2015). Most farmers in both sub-counties had not vaccinated their cattle: the scarcity of brucellosis vaccine in Uganda is partially due to fact that the Ministry of Agriculture, Animal Industry and Fisheries (MAAIF) does not consider brucellosis a priority disease, compared to others diseases such as foot and mouth disease. However, the need for vaccination and curtailing brucellosis was better understood by AgShare farmers, which contributed to reductions of the disease in their herds.

The associations between hygiene scores and milk safety and quality demonstrated that AgShare interventions to improve hygiene resulted in lower levels of SCM and brucellosis and decreases in CMT scores. Improving farm hygiene and udder hygiene have been associated with decreasing clinical and SCM and SCC in Uganda dairy cattle (Byarugaba et al 2008; Abrahmsén et al 2014; Kasozi et al 2014; Mugenyi 2014). However, this and other studies have found poor hygiene to be common on dairy farms (Persson et al 2013; Mugenyi 2014), indicating the need for expanding efforts to educate farmers on the benefits of good hygiene for dairy production and animal health.


Conclusions


Acknowledgments

Special thanks go to the Bill and Melinda Gates foundation that fully funded the AgShare pilot project in Uganda through Makerere University, AmateGeitu cooperative members who voluntarily accepted to participate, in addition to the Africa Institute for Strategic Animal Resource Services and Development.


Conflict of Interest

The authors declare that they have no conflict of interest.


References

Abrahmsén M, Persson Y, Kanyima B M and Båge R 2014. Prevalence of subclinical mastitis in dairy farms in urban and peri–urban areas of Kampala, Uganda. Tropical Animal Health and Production, 46, 99–105.

Björk S 2013. Clinical and subclinical mastitis in dairy cattle in Kampala, Uganda. (unpublished degree project in veterinary medicine, Swedish University for Agricultural Sciences).

Bonfoh B, Wasem A, Traoré A N, Fané A, Spillmann H, Simbé C F, Alfaroukh I O, Nicolet J, Farah Z and Zinsstag J 2003. Microbiological quality of cows’ milk taken at different intervals from the udder to the selling point in Bamako (Mali). Food Control, 14,495–500.

Byarugaba D K, Nakavuma J L, Vaarst Mand Laker C 2008. Mastitis occurrence and constraints to mastitis control in small-holder dairy farming systems in Uganda. Livestock Research for Rural Development,  Volume 20, Article #5. http://www.lrrd.org/lrrd20/1/byar20005.htm

Grimaud P, Sserunjogi M, Wesuta M, Grillet N, Kato M and Faye B 2009. Effects of season and agro–ecological zone on the microbial quality of raw milk along the various levels of the value chain in Uganda. Tropical Animal Health and Production, 41, 883–890.

Hachana Y and Paape M J 2012. Physical and chemical characteristics of yoghurt produced from whole milk with different levels of somatic cell counts. International Journal of Food Sciences and Nutrition, 63, 303–309.

Hurvich C M and Tsai C L 1993. A corrected Akaike information criterion for vector autoregressive model selection. Journal of Time Series Analysis, 14, 271–279.

Kahuta G 2013. Milk quality and on-farm factors leading to milk spoilage in Bugaaki sub county, Kyenjojo district. (unpublished Msc. Thesis, Makerere University).

Kaneene J B, Ssajjakambwe P, Kisaka S, Miller R, and Kabasa J D 2013. Creating Open Education Resources for Teaching and Community Development through Action Research: An Overview of the Makerere AgShare Project. Journal of Asynchronous Learning Networks, 17(2).

Kashiwazaki Y, Ecewu E, Imaligat J O, Mawejje R, Kirunda M, Kato M, Musoke G M and Ademun R A O 2012. Epidemiology of Bovine Brucellosis by a Combination of Rose Bengal Test and Indirect ELISA in the Five Districts of Uganda. Journal of Veterinary Medical Science, 74, 1417–1422.

Kasozi K I, Tingiira J B and Vudriko P 2014. High prevalence of subclinical mastitis and multidrug resistant Staphylococcus aureus are a threat to dairy cattle production in Kiboga District (Uganda). Open Journal of Veterinary Medicine, 4, 35–43.

Kleinbaum D G and Klein M 2010. Logistic regression: a self–learning text. Springer Science and Business Media.

Le Maréchal C, Thiéry R, Vautor E and le Loir Y 2011. Mastitis impact on technological properties of milk and quality of milk products – a review. Dairy Science and Technology,91, 247–282.

Magona J W, Walubengo J, Galiwango T and Etoori A 2009. Seroprevalence and potential risk of bovine brucellosis in zero-grazing and pastoral dairy systems in Uganda. Tropical Animal Health and Production, 41, 1765–1771.

Mbowa S, Shineykwa I and Lwanga M M 2012. Dairy Sector Reforms and Transformation in Uganda since the 1990s. Economic Policy Research Centre (EPRC) Research Report No. 4.

Mellenberger R and Roth C J 2000. California Mastitis Test (CMT) Fact Sheet. Available at http://www.uwex.edu/milkquality/PDF/045cmt_factsheet.pdf.

Miller R, Nakavuma J L, Ssajjakambwe P, Vudriko P, Musisi N and Kaneene J B 2015. The prevalence of brucellosis in cattle, goats and humans in rural Uganda: a comparative study. Transboundary and Emerging Diseases, doi: 10.1111/tbed.12332.

Mugabi R, Khaitsa M L, Miller R, Nakavuma J L, Ssajjakambwe P, Kaneene J B and Barigye R 2012. Seroprevalence of brucellosis in selected herds of cattle and goats in Kiruhura district, Uganda. African Journal of Animal and Biomedical Sciences, 7, 28–30.

Mugenyi K C 2014. Prevalence, control and antimicrobial susceptibility patterns of bovine mastitis causing bacteria in Mityana District. (unpublished MS thesis, Makerere University).

Mwebe R, Nakavuma J and Moriyón I 2010. Brucellosis seroprevalence in livestock in Uganda from 1998 to 2008: a retrospective study. Tropical Animal Health and Production,43, 603–608.

Nizeyimana G, Mwiine F N and Ayebazibwe C 2013. Comparative Brucella abortus antibody prevalence in cattle under contrasting husbandry practices in Uganda. Journal of the South African Veterinary Association, 84, E1–5.

OIE (World Organization for Animal Health) 2009. Chapter 2.4.3, Bovine Brucellosis. In: Manual of Diagnostic Tests and Vaccines for Terrestrial Animals, 7th edn. OIE, Paris, France, 1–35.

Oliver S P, Boor K J, Murphy S C and Murinda S E 2009. Food safety hazards associated with consumption of raw milk. Foodborne Pathogens and Diseases, 6, 793–806.

Persson Y, Abrahmsén M, Björ S and Kanyima BandBåge R 2013. Clinical and subclinical mastitis in dairy cattle in Kampala, Uganda. IDF Animal Health Newsletter, 7, 5–6.

SAS 9.3 (SAS Systems, Inc., Cary NC).

Ssajjakambwe P, Kisaka S, Vudriko P, Setumba C, Bahizi G, Kabasa J D and Kaneene J B 2013. Creating Open Education Resources for Teaching and Community Development through Action Research: The Milk Production and Hygiene Module. Journal of Asynchronous Learning Networks, 17, 43.

Vaarst M, Byarugaba D K, Nakavuma J and Laker C 2007. Participatory Livestock Farmer Training for improvement of animal health in rural and peri–urban smallholder dairy herds in Jinja, Uganda. Tropical Animal Health and Production, 39, 1–11.


Received 25 November 2015; Accepted 15 December 2015; Published 2 January 2016

Go to top