Towards a healthier food system for rural Ethiopia: integrating food production, diets and nutrition

Malnutrition, in all its forms, affects the health and well-being of people, and poses significant concerns to human development and economic growth worldwide. Overt hunger affects one in nine people globally, while hidden hunger disproportionately affects women and children, with over half of the children and two-thirds of the women being deficient for at least one micronutrient. Three billion people worldwide cannot afford a healthy diet. Stunting and anemia in children are important public health concern in most developing regions of the world, particularly in rural areas.
Despite a high level of commitment to reduce undernutrition globally and countrywide, Ethiopia is making some progress but off-track to achieve the global targets. This is partly because the causes of these problems are still poorly understood owing to their diverse and complex interaction with several factors nested at the community, household, and individual levels. This thesis evaluates adequacy of nutrients from local food supply, it relates food production to the household and intrahousehold consumption, it assesses the prevalence and determinants of stunting and anemia in children and of micronutrient deficiencies in women. It focuses on three agroecological zones in rural Ethiopia.
Chapter 2 considers the adequacy of essential nutrients from local production in the midland (ML), highland (HL), and upper highland (UHL) agroecological zones (AEZs). Data were collected at the village level from the agriculture office, at the farm and household levels through surveys. Annual crop and livestock production were converted into energy and nutrient supply using local crop energy and nutrient composition data. The study shows that the total energy production met significant proportions of per capita energy demand in the HL and UHL, while the supply had a 50% energy deficit in the ML. Shortfalls in per capita vitamin A supply decreased across the agroecological gradient from ML (46%) to UHL (31%). The estimated shortfall in folate supply was significantly higher in the UHL (63%) and negligible in the HL (2%). The risk of deficient iron and zinc supply was relatively low across all AEZs (<10%), but the deficiency risk of calcium was unacceptably high. The study concludes that agroecology determines the choice of crop produced and in this way affects the available supply of energy and nutrients.
Chapter 3 evaluates the nature and strength of associations between Food Production Diversity (FPD) and Dietary Diversity (DD) from agroecologically diverse rural areas in Ethiopia. It also evaluates intrahousehold DD in relation to FPD. A Poisson regression model related FPD and DD at household and individual levels. The study showed that food consumption greatly varies within households and across the different agroecological zones. FPD has a significant positive relation with DD, but the magnitude of this effect is small. Differences in agroecological zones affect DD at the household level by 23%. Additionally, non-farm household income, market access, and nutrition knowledge of women significantly affect the association. The study recommends that promoting DD in rural farming households should take these factors into account rather than solely increasing FPD. Intervention should also consider variations in agroecology and intrahousehold consumption.
Chapter 4 assesses the prevalence and evaluates determinants of stunting and anemia in children aged 6-23 months. It spans factors at the individual, household and community level including agroecology and antinutrients as a unique set of predictors. The study showed that 49.1% and 49.7% of children were stunted and anemic, respectively. Children living in ML are approximately twice more likely to be stunted than those living in UHL. The risk of stunting increases by 16.3% and 41.9% for every unit increase in phytate-to-zinc and phytate-to-iron molar ratios, respectively. A 10% increase in aggregated crop yield reduces the occurrence of stunting by 13.6%. Households with opportunities to generate non-farm income, with shorter travel time to the marketplace and utilizing health service have a decreased risk of stunting. Child diet of a low diversity, age of the child (18-23 months) and mothers at a younger age are significantly associated with increased stunting. Risk of anemia is high among households with unsafe drinking water, poor sanitation and hygiene (WASH) practices, younger age (6-11 months), and anemia mostly occurs among boys. Children living in ML have a 55% reduced risk of being anemic as compared to those living in UHL. Therefore, these factors should tailor strategies for reducing undernutrition among children in rural Ethiopia.
Chapter 5 assesses the prevalence and determinants of micronutrient deficiencies in women of reproductive age from 660 rural households representing the three AEZs. A multivariable logistic regression model was used to relate predictors of micronutrient deficiencies in women.
Deficiencies for Ca are observed among 77.8%, Zn among 76.1%, vitamin A among 67%, and low Fe store among 54.3% of the women. In addition, 22% of the women were anemic. Both dietary and non-dietary factors are associated with micronutrient deficiencies in women. Prevalence of deficiencies vary significantly across the AEZs. For most micronutrients, women living in ML have reduced risk of deficiencies as compared to those living in other AEZs. Agronomic practices, use of inorganic fertilizers and crop rotation, are associated with a reduced likelihood of Zn deficiency. Access to the market, health facility, safe water, improved WASH practices, and low household size are associated with a reduced risk of anemia and with Fe deficiency. With increased phytate supply, the likelihood of women being deficient for Fe and Zn increases. Therefore, nutrition interventions should consider variation in agroecology. Along with strategies to increase the micronutrient content of staple crops, interventions to lower the concentration of antinutrient could improve serum micronutrient status. Infrastructural intervention that facilitates easy access to the market and health facilities, and improved WASH practices could further alleviate the problem.
Chapter 6 summarizes and discusses the main findings of the thesis. It presents findings into a broader perspective and highlights implications for practice and future research directions.
In summary, the study shows that agroecology determines the choice of foods produced. In this way it affects production diversity, the available supply of gross energy, and the essential nutrients as well as nutrition outcomes in women and children. Intervention strategies should take these factors into account and go well beyond the administrative boundaries into targeting agroecological variation.