Impact of livestock support on animal milksupply and child nutrition in Ethiopia

Summary by Kate Sadler and Emily Mitchard

Kate Sadler is an Assistant Professor of Nutrition at the Feinstein International Centre, Tufts University and the Research & Development Manager at Valid International. A public nutritionist with over 15 years of experience in the design, management and evaluation of nutrition interventions in sub Saharan Africa, she completed an MSc in Public Nutrition at the London School of Hygiene & Tropical Medicine in 1997. She went on to complete her PhD, which focused on the early development of CTC/CMAM, in 2008.

Emily Mitchard is a food security and nutrition consultant currently based in Bangkok, Thailand. Previously, she worked with Feinstein International Centre on the Milk Matters projects in Somali Region, Ethiopia and Karamoja, Uganda. She holds an MSc in Agriculture, Food Security and Nutrition and an MPH in Biostatistics and Epidemiology from Tufts University.

This work was made possible with funding from the Office of Foreign Disaster Assistance (OFDA) United States Agency for International Development (USAID). The authors would like to thank Axel Weiser, Girma Tadesse, Mohammed Musse, and Nasteha Ahmed of Save the Children USA and Hassan Ibrahim and Mohammed Ali of Save the Children UK for providing considerable support and hard work throughout the implementation of the study, as well as for providing valuable feedback for this report. Thanks also go to Adrian Cullis of the Food and Agriculture Organisation (FAO) for important input on initial concepts and intervention design. Finally, many thanks go to the dedicated data collectors and community members who made the study possible.

The full report of the research summarised here, including additional methodology and the results from the other three intervention sites, can be found on the Feinstein International Centre website http://sites.Tufts.edu/Feinstein/2012/Milk-Matters

Background

Animal milk has long been recognised as an important component of pastoralist diets across the world¹. As a nutrient-dense food, milk is known to contribute a high proportion of the nutrients, such as high quality protein and micronutrients, to the pastoralist diet^2,3,4. Previous research in the Somali region of Ethiopia has found that Somali pastoralists consume, on average, between 20 and 50 percent of their energy requirement as milk and animal products in a normal rain year^5,6,7. For Somali children in particular, when milk is plentiful it was found to constitute a central pillar of the diet, providing two-thirds of the daily energy requirement and 100 percent of the daily protein requirement⁸. Yet, in recent decades, levels of global acute malnutrition among young children in the Somali region are regularly reported to rise above 15%, the level defined as a nutritional emergency by the World Health Organisation (WHO)⁹. These surveys identified a seasonal aspect to child malnutrition, with particularly high rates of acute malnutrition occurring during the dry season and periods of drought. Seasonal variation in livestock milk production has also been well described in the literature on pastoralism in Africa, with milk supply falling as the dry season advances^10,11,12. Whilst this work indicates that the main risk period for child malnutrition occurs routinely in the late dry season in many pastoralist areas, nutrition and humanitarian programming continues to take a reactive emergency approach emphasizing the delivery of food aid and therapeutic treatment of severe acute malnutrition (SAM). Moreover, food aid is typically characterized by the delivery of a food aid basket that rarely includes a protein or fatty acid source suitable for infants and young children. Finally, little has been done to understand the potential role of milk in preventing malnutrition in pastoral regions.

As part of Save the Children’s African Region Pastoral Initiative¹³, the ‘Milk Matters’ project is a joint venture between the Feinstein International Centre at Tufts University, Save the Children USA (SC US), and Save the Children UK (SC UK) in Ethiopia. The Milk Matters project aims to improve the nutritional status of children in pastoralist/semi pastoralist areas of Ethiopia through demonstrating an explicit link between livestock health, milk availability and access, and child nutrition¹⁴. It also aims to contribute to improvements in policy and programming for child nutrition by providing scientific evidence in support of more holistic and preventative approaches to malnutrition in pastoral regions. This article briefly describes the findings from two cohort studies conducted in two Zones in the Somali region of Ethiopia from July 2010 to July 2011. The studies were designed to assess the impact of small-scale livestock interventions to sustain access to and availability of animal milk at the household level over the dry season on the nutritional status of children less than five years of age. The study objective, design, and findings from one intervention site, as well as discussion and recommendations, are summarised below.

Study objective

The objective of this study was to evaluate the impact of communitydefined livestock interventions on child nutritional status during the dry season. In doing so, the study asked the following primary research question:

What is the impact of livestock interventions on children’s consumption of animal milk and nutritional status over one calendar year, particularly during the dry season?

In addition, the study sought to compare the costs of an early intervention focusing on livestock health and milk production with an emergency feeding programme for children.

Study design

Two cohort studies, each involving three sites, were implemented over one calendar year, from July 2010 to July 2011. In each study, two sites were designated to receive livestock interventions and one site served as a control. All children aged 6 to 59 months in both intervention and control sites were enrolled in the monthly surveillance system that collected data on child health and nutrition. Overall, the surveillance system aimed to follow 940 children, 610 living in intervention sites and 330 living in control sites.

Study site, household, and milking animal selection

Six sites were selected to participate in the study, three from two separate Zones in the Somali region. Site selection criteria included:

1. Pastoral livelihoods
2. Targeted for assistance under SCUS Protective Safety Net Programme (PSNP) and SCUK Revitalising Agriculture/ Pastoral Incomes and New Markets (RAIN) programme in Liben and Shinile respectively
3. Population size greater than 200 house holds¹⁵, and
4. Vulnerability to elevated rates of child malnutrition during the dry season, as indicated through regional nutrition assessments¹⁶.

Site selection was completed in December 2009 in consultation with local government officials, SCUS (Liben) and SCUK (Shinile), and local communities. As far as possible, sites with a similar level of access to basic resources such as pasture, water, health care, and education services were selected. Selection of households was also done in consultation with local officials and community members, with a focus on inclusion of all households with children under the age of five years. Cows and small ruminants were designated as the targeted livestock species for support under the intervention¹⁷. In the four designated intervention sites, households were asked to choose either one milking cow or three to four milking goats, a ratio based on estimated milk off-take per species. The particular milking animals at each household were selected for support according to the following set of criteria: (1) recently lactating, (2) with a normal milking yield and (3) in overall good health with no problems that could compromise milk production.

The livestock interventions

Two livestock interventions were designed. In two of the four designated intervention sites (Waruf in Shinile and Biyoley in Liben), the milking animals were given a daily ration of supplementary feed over the dry season. In the other two intervention sites (Ayiliso in Shinile and Washaqabar in Liben), the milking animals were given a daily ration of supplementary feed plus a package of vaccinations and de-worming medications at the outset of the dry season¹⁸. The two remaining sites were designated as control sites and received no intervention. While the dry season, Jilaal, typically falls between February and April in the northern Shinile Zone and between January and March in the southern Liben Zone, the interventions were extended in both Zones due to the onset of a drought in 2010/11¹⁹. In total, the interventions spanned 146 days in Shinile and 135 days in Liben, but due to certain sitespecific occurrences, the total days of actual livestock feeding varied by site and ranged from 47 to 73 days.

The supplementary feed ration size for each animal species was based on the livestock feeding guidelines for drought set by the Ethiopian government²⁰ and was meant to supplement natural browse. However, the ration size was subsequently raised as a result of the drought (Table 1). Local agricultural cooperatives with irrigation capacity were contracted to grow and transport Sudan grass in Liben, while a private contractor was ultimately contracted to provide Rhodes grass in Shinile after an unexpected frost destroyed the Sudan grass crop. The care and shelter of the grass hay upon delivery to the sites was tasked to community members in each intervention site. Both SCUS and SCUK provided assistance in building the shelters for the hay. Participating households were responsible for providing the milking animals with sufficient water.

Table 1: Final feed rations (kg/day) for milking species
	Liben Sudan grass	Shinile Rhodes grass	Extended Feeding (Shinile only) Wheat bran
Adult cow	9	6	3
Adult goat	3	2	1
Calf	3	2
Kid	2	1

 

For the two sites receiving the livestock health package, vaccinations were procured through the Ethiopian Government Bureau of Livestock, Crop and Rural Development, while private vendors of veterinary drugs were contracted to maintain a consistent supply of the designated prophylactic drugs throughout the dry season (see Table 2). The drugs were then collected as needed by trained community animal health workers using the voucher system established by SCUS and SCUK in the respective regions.

Table 2: Vaccinations and medications for milking species
Care Type	Cattle	Goats
Vaccination	Anthrax, blackleg, contagious bovine pleuropneumonia (CBPP), Lumpy Skin Disease (LSD	Contagious caprine pleuro-pneumonia (CCPP), capri pox, peste des petites ruminants (PPR)
Prophylactic treatment	Ivermectin (internal and external parasites)	Ivermectin (internal and external parasites)
Other curative treatment	Treatment provided with diagnosis throughout the dry season	Treatment provided with diagnosis throughout the dry season

 

The surveillance system

The surveillance system was designed to monitor selected anthropometric measurements, milk consumption patterns, and infection status of enrolled children less than 5 years of age. Thirty-two data collectors were selected from within the communities because of their unique ability to track down and follow children as they moved with their families. The data collectors were trained in May 2010 and were responsible for following between 25 and 30 children each on a monthly basis. The data collectors administered three questionnaires per child, designed to collect a range of health and nutrition data, including age and weight, from which weight-for-age z-scores (WAZ) were derived, 24 hour recall of milk consumption (plain milk and milk in tea), and infection status over the last two weeks.

Results

The data were evaluated by intervention site compared with the control site in the same Zone and then analysed according to the impact of the interventions on three key variables: (1) animal milk off-take, (2) milk availability - measured as a function of both the proportion of children in each site consuming milk and the average amount consumed by those children receiving milk, and (3) child nutritional status measured by WAZ. In summary, the results seen across all four of the intervention sites were:

• Milk off-take in the intervention sites was significantly greater during the 2011 dry season with the intervention as compared to the 2010 dry season with no intervention.
• Given the drought conditions and lack of rainfall, the increase in milk off-take is attributed to the project interventions.
• By the end of the intervention, a greater proportion of children were consuming milk in the intervention sites as compared to the control sites.
• Those children who received any milk in the intervention sites consumed, on average, more milk than children in the control site.
• There was an overall trend towards stabilized nutritional status among young children in the intervention sites over the course of the intervention compared with a decline in nutritional status in the control sites.
• Within the intervention sites, those children who continued to consume some milk throughout the intervention time period maintained higher average WAZ scores than those who did not receive any milk, a difference that was frequently significant.

Case Study: Results from Washaqabar, Liben Zone

Key findings from Washaqabar, Liben Zone are summarised in Box 1 and detailed below.

Box 1: Key Findings from Washaqabar
Milk off-take	Livestock milk off-take in Washaqabar was significantly greater during the dry season/drought in 2011 with the intervention, compared to the previous year dry season in 2010 with no intervention.
Milk availability	Milk was more available to young children in Washaqabar compared with the control site, with 94 percent of children receiving milk compared to 56 percent in the control, and each child consuming on average 366 mL more milk per day than in the control.
Nutritional Status	The nutritional status of children in Washaqabar remained relatively stable during the intervention, compared with declining nutritional status among children in the control site.

 

Milk off-take

Milk off-take increased significantly in Washaqabar, by as much as 4775%, from the 2010 dry season to the 2011 dry season with the intervention (Table 3). This increase was attributed to project inputs. The primary non-project factor that could have resulted in increased milk off-take was rainfall leading to better browse and water for livestock, but drought conditions and lack of browse over the intervention period rule out this factor²¹. A secondary factor, household purchase of feed for livestock, could also have affected attribution, but was similarly ruled out. A high level of ration was delivered by the study to study animals relative to recommended levels and feed intakes. This suggests that very limited, if any, feeding of study animals with privatelypurchased feed is likely to have taken place.

Table 3: Milk off-take in Washaqabar
Livestock Type	Stage of Lactation	Mean Daily Milk Off-take (ml)		Percent Change
		Dry season, 2010, no intervention (95% CI)	Dry season, - b 2011, with intervention (95% CI)
Goat - a	Early	224 (190.5, 257.6)	628 (473.8, 782.9)	280 - c
	Middle	54 (24.5, 84.2)	567 (428.3, 706.6)	1050 - c
	Late	8 (0.0, 20.2)	382 (317.6, 446.6)	4775 - c
a Because a goat yields less milk than a cow, three goats were considered equivalent to one cow during the intervention. The above yield estimates are for one goat. b The 2011 dry season became a drought (see Table 11 of original publication). c Significant at the 95% confidence level.

 

Milk availability

More children received milk and children on average consumed greater quantities of milk per day throughout the intervention in Washaqabar than in Makinajab (Figure 1). From December to May, the percent of children receiving milk in Washaqabar increased by 31% compared to 20% change in Makinajab, and the average daily amount consumed increased by 241ml compared to gradual decline in Makinajab.

Nutritional status

Pre-intervention nutritional status was relatively stable in both the intervention and the control site (Figure 2). However, with the start of the dry season in January 2011, the average WAZ of children in the Makinajab control site started to decline, with an overall decrease of 0.31 points between January and May. In contrast, the average z-score fell by only 0.07 points from January to May in the Washaqabar intervention site.

Discussion

The results of this study demonstrate that targeted livestock support can significantly increase daily milk off-take during very dry periods with direct links to improved milk consumption among young children and positive implications for nutritional status. To our knowledge, this is the first study that presents quantitative data demonstrating this link for young children in pastoralist areas. There were several factors, however, that affected the relationship between milk off-take, child milk consumption, and child nutritional status, summarized in Box 2. These factors are important to consider as they have implications for both the impact of the intervention seen here and the design of similar interventions in the future.

Box 2. Factors affecting milk off-take, child milk consumption and nutritional status
Milk off-take and consumption
Challenges of implementation delivery	Feed procurement and delivery challenges in the two Zones included: unexpected frost late in the growing season in Shinile that killed the Sudan grass crop and necessitated a new contract for Rhodes grass that was located a greater distance from the study sites, high transportation costs, occasional poor quality of feed delivered to study sites, preference of some participants to feed animals from home instead of at designated feeding centres, long distances to feeding centres that impacted milk supply and availability for young children, movement of a small proportion of targeted cattle reduced access to feeding centres.
Independent purchasing of feed	While this happened to varying extents in all sites, the provision of vouchers used for feed purchase by an external agency to households in one control site affected the validity of the control; this event also demonstrates the difficulty of implementing field research in settings not easily controlled.
Drought conditions	Limited natural browse due to drought necessitated an increase in ration size. Water and pasture shortages forced families to move further than planned resulting in some loss to follow-up.
Sharing of milk	In sites where household coverage by the interventions was lower, participating households reported sharing milk with non-project households.
Prioritization of milk	With less overall milk available, milk was prioritised to children under three years resulting in less consumption amongst the older children enrolled in the study.
Milk consumption and nutritional status
Infection status	Several misunderstandings between mothers and data collectors led to widespread underreporting and underestimation of the prevalence of childhood infections.
Activity levels	Data to reflect this was not collected or measured by the study.

 

Where the intervention worked well and intervention coverage of households was high, such as in Washaqabar, the increase in milk consumption seen (1050 mL/day compared to 650mL/day in the control site) translates into an additional 264kcal, 12.8g of protein, and considerably higher intakes of essential fatty acids, vitamins, and minerals per child each day. For a young child of two years of age, this increase in nutrient intake would meet circa 26 percent of energy and 98 percent of protein requirements. Whilst we don’t see a dramatic improvement in weight gain among intervention children, we do see an overall pattern of stabilized WAZ among these children compared with a general decline in children in control sites over the intervention period. Together, these impacts indicate that interventions such as those tested here hold potential to maintain weights of young children in times of drought and to prevent a deterioration to acute malnutrition. It is well documented in the scientific literature that preventing acute malnutrition is crucially important for a child’s survival and overall mental and physical development²². Moreover, the results of this study also suggest that it could be more cost-effective, with the cost of the livestock interventions tested here ranging from 34 to 81 USD per child to prevent weight loss versus the cost of 145 to 200 USD per child to treat severe acute malnutrition in a CMAM programme.^23,24

In addition to the impacts seen on milk consumption and nutritional status in young children, the interventions had several positive livelihood outcomes for participating households, including:

• More free time for women as a result of reduced workload - research on maternal and child health suggests that increasing maternal well-being frequently translates to improved infant and young child feeding practices (such as perceived ability to exclusively breastfeed) with important outcomes for child nutrition^25,26,27.
• Protection of critical assets during drought conditions - many households reported a high survival rate of dams and suckling calves and perceived improved rates of reproduction in some of their animals as compared to previous dry seasons and periods of drought.
• Improvement in milk off-take for local Somali indigenous breeds - this result indicates the potential to enhance production in local breeds, which are well adapted to harsh environmental conditions and disease risks in the pastoralist areas targeted here.

Conclusions and recommendations

This study has demonstrated that through targeted livestock support to milking animals that stay close to women and children during dry season and/or drought (overall a relatively small proportion of the whole herd), milk production and consumption among children is improved, and their nutritional status benefits. There is some consensus in the programming literature at present that the humanitarian community tends to spend much more time before humanitarian disasters preparing to treat acute malnutrition rather than trying to prevent it²⁸. The interventions presented here provide us with the opportunity to change this focus and reconnect food security interventions and nutrition outcomes in these areas with the potential of creating substantial aid cost savings by preventing the need for large selective feed programmes such as CMAM. Recommendations therefore include:

• Applying a nutrition lens to the drought management cycle by, for example, supporting preservation of milk surplus during the rainy season, community level feed production/ storage, feed interventions for reproductive/ milking stock, and designing cash/ food-for-work activities that do not negatively impact on women’s time and ability to maintain their own or their children’s nutrition status.
• Scaling up similar feed interventions as that conducted in this project, but with adjustments to control costs from feed supply and transport. This could be done through the provision of vouchers or other financial mechanisms where market supply for feed is adequate²⁹.
• If private sector provision of feed is not adequate, investigating opportunities to support local livestock feed production³⁰.

Finally, food security and livelihood programmes must start monitoring more systematically their impact on nutrition outcomes. This need not involve frequent collection of anthropometric data and the measurement of nutritional status itself, which, as discussed above, can be challenging in these environs. But simple tools for measuring nutrition impact, such as participatory impact assessment³¹ and the dietary diversity index³², need to be used as standard if we are to create the momentum for investment in food security interventions to prevent increases in rates of malnutrition where these have been shown to be effective.

For more information, contact: Kate Sadler, email: kate@validinternational.org

---

¹Sadler, K., C. Kerven, et al. (2010). The fat and the lean: a review of production and use of milk by pastoralists in press. Pastoralism: research, policy and practice

²Galvin, K. A., D. L. Coppock, et al. (1994). Diet, nutrition, and the pastoral strategy. African pastoral systems: An integrated approach. E. Fratkin, K. A. Galvin and E. A. Roth. London, Lynne Rienner: 113-132.

³Fratkin, E., E. A. Roth, et al. (2004). Pastoral sedentarization and its effects on children's diet, health, and growth among Rendille of Northern Kenya. Human Ecology 32(5): 531-559.

⁴Barasa, M., A. Catley, et al. (2008). Foot-and-mouth disease vaccination in South Sudan: benefit-cost analysis and livelihoods impact. Transboundary and Emerging Diseases 55:339-351.

⁵Webb, P. and J. Braun (1994). The Pastoral Experience. Famine and food Security in Ethiopia: lessons for Africa. Chichester UK, Wiley & Sons Ltd: 85-98.

⁶SCUK and DPPA (2002). Filtu-Dolow Pastoral Livelihood Zone: a HEA baseline study. Addis Ababa, Save the Children UK and Disaster Prevention and Preparedness Bureau.

⁷SCUK (2007). The causes of malnutrition in children under 3 in the Somali Region of Ethiopia related to household caring practices: Preliminary Report. A.-M. Mayer. Ethiopia, Report on research findings from Somali Caring Practices research project in Shinile and Dambal districts of Shinile zone, Somali region, Ethiopia.

⁸Sadler, K. and A. Catley (2009). Milk Matters: the role and value of milk in the diets of Somali pastoralist children in Liben and Shinile, Ethiopia. Addis Ababa, Feinstein International Center. See also footnote 7.

⁹Ethiopian Health and Nutrition Research Institute, UNICEF, et al. (2009). Final Report from Nutrition and Mortality Surveys conducted in Seven Mega Livelihood Zones in Somali Regional State, Ethiopia. Addis Ababa, Ethiopian Health and Nutrition Research Institute.

¹⁰Chell, D. and A. Chell (1979). A detailed evaluation of the primary health care programme among nomadic tribes of North Kordofan Province, Sudan. El Obeid, Nomadic Health Project, Euro-Action Accord.

¹¹Arhem, K. (1985). Pastoral Man in the Garden of Eden: The Maasai of the Ngorogoro Conservation Area, Tanzania. Uppsala Research Reports in Cultural Anthropology. Uppsala, Sweden, Uppsala University.

¹²Catley, A. (1999). The Herd Instinct: Children and Livestock in the Horn of Africa. Save the Children Working Paper 21. London, Save the Children UK.

¹³The goal of Save the Children’s Africa Region Pastoral Initiative is to "deepen and replicate innovative approaches to improve access to basic services and reduce vulnerability to drought in pastoralist populations in order to create positive change for children in this unique and harsh environment."

¹⁴Previous work under the Milk Matters project, described extensively in the report, Milk Matters: The Role and Value of Milk in the Diets of Somali Pastoralist Children in Liben and Shinile, Ethiopia, laid the foundation for the results discussed in this article and can be found at http://sites.tufts.edu/feinstein/2009/milk-matters-phase-one

¹⁵It was determined that a community with at least 200 households would ensure the desired sample size of 150 children/site.

¹⁶See footnote 9.

¹⁷The decision to support small ruminants and cows was based on the results of previous work under Milk Matters which found that households kept a cow or a few goats at the household during a typical dry season while the men took the larger stock far from the settlement site in search of pasture and water.

¹⁸Animal feed was provided in all sites based on the fact that animal health care alone would do little to improve milk off-take if the animal remained malnourished, but that extra health provisions in addition to feed may improve milk off-take compared to animals receiving only feed.

¹⁹The Somali Region is characterized by a bimodal rainfall pattern with one long and one short rainy season, and one long and one short dry season. Jilaal is the Somali term for the long dry season.

²⁰Ministry of Agriculture and Rural Development (2008). National Guidelines for Livestock Relief Interventions in Pastoralist Areas of Ethiopia. Addis Ababa, Ethiopia, Ministry of Agriculture and Rural Development

²¹Please refer to the original publication for a detailed presentation of rainfall data for the referenced time period provided by the Ethiopian National Meteorology Agency, Addis Ababa, Ethiopia.

²²Bhutta, Z. A. (2009). Addressing severe acute malnutrition where it matters. Lancet 374: 94-96.

²³Puett, C., K. Sadler, et al. (2012 ). Cost-effectiveness of the community-based management of severe acute malnutrition by community health workers in southern Bangladesh. Health Policy and Planning in press.

²⁴Please refer to original document for more indepth presentation of the cost comparison.

²⁵Shell-Duncan, B. and S. A. Yung (2004). The maternal depletion transition in northern Kenya: the effects of settlement, development and disparity. Social Science and Medicine 58(12): 2485-2498.

²⁶Nyaruhucha, C., J. Msuya, et al. (2006). Nutritional status and feeding practices of underfive children in Simanjiro District, Tanzania. Tanzania Health Research Bulletin 8(3):162-167.

²⁷See Footnote 7

²⁸Levine, S. and C. Chastre (2011).Nutrition and food security response analysis in emergency contexts. London, Humanitarian Policy Group.

²⁹In the participatory impact assessments for this project, households stated that they frequently purchased animal feed from local sources as a "normal strategy" during drought, a very positive trend that could be supported.

³⁰The role of aid projects in supporting local livestock feed production is not clear and this recommendation is dependent on good analysis of existing private sector production and apparent growth of this activity, as well as on trends in private enclosures and commercial fodder production. Please refer to full report for additional information: http://sites.tufts.edu/feinstein/2009/milk-matters-phase-one

³¹Catley, A., J. Burns, et al. (2008). Participatory Impact Assessment: A guide for practitioners. Medford, MA, Feinstein International Centre.

³²Drescher, L. S., S. Thiele, et al. (2007). A New Index to Measure Healthy Food Diversity Better Reflects a Healthy Diet Than Traditional Measures. Journal of Nutrition 137:647-651.