Six sub-sessions
Chair: Amos Laar, African Nutrition Society
Speakers and presentations
Chanyalew Seyoum Aweke, Haramaya University
Assessing the Contribution of Agriculture to Nutrition. Results from a Household Survey of Two Agro-Ecological Zones in Ethiopia
Slides
Anjali V. Ganpule-Rao
Association of Dietary Diversity with Nutrient Adequacy and Socio-Economic Parameters among Rural Adolescents from Pune, India
Slides
Thalia Sparling, IMMANA
Determinants of Participation in a Homestead Food Production Program in Rural Bangladesh
Slides
Demeke Senbetu, OARI
Magnitude of Fish Post-Harvest Loss and its Intervention in Small Scale Fisheries of Ethiopia: A Missing Link but Viable Approach for Food and Nutrition Security
Josh Van Vianen
The Importance of Forested Landscapes for Food Security and Nutrition Across Agricultural Transitions: A Multi-Country Comparative Analysis
Abdullah Al Mamun, FANRPAN
Impact of Different Agroecosystems on Household Food Consumption, Health Outcomes and Autonomy of Adolescent Girls in Seafood Farming Communities in Bangladesh
Slides
Q&A
Recording
Assessing the Contribution of Agriculture to Nutrition. Results from a Household Survey of Two Agro-Ecological Zones in Ethiopia
Chanyalew Seyoum Aweke, Haramaya University
Introduction
We assess the relative and absolute importance of agricultural influences on food security and nutritional outcomes in two agro-ecological zones in Ethiopia. Although there have been an increasing number of studies aimed at measuring the impact of agriculture on nutrition in recent years, many focus on only one or two agricultural influences and few have examined the influence on measures of micronutrient status. This study analyses associations between agricultural and other variables and food security and nutritional indicators using regression analysis.
Methods
The analysis uses data gathered in an 800-household survey conducted in two food-insecure Ethiopian regions. The survey was aimed at providing a detailed understanding of agricultural practices in both regions in relation to food and nutrition security outcomes. Particular emphasis was placed on maternal and child nutrition, with a more detailed survey of nutrition indicators for a subset of 200 households with mothers and children under two years of age.
Our conceptual framework focuses on the pathways by which different groups of predictor variables are hypothesised to influence food security and nutrition outcomes. We selected two less-commonly used nutritional indicators, maternal zinc and maternal ferritin, together with two standard food security measures, the household food consumption score (HFCS) and household dietary diversity score (HDDS). The selection of predictor variables was informed by the literature and those showing the strongest correlations with the dependent variables. The predictors were then categorised into four groups: household characteristics, agricultural factors (including crop and livestock production and diversity), income issues, and food consumption decisions. We then performed ordinary least squares hierarchical multiple regressions on each outcome variable using the same four groups of independent variables in the same order, as informed by the conceptual framework.
Findings and Interpretations
The results highlight a significant difference in average maternal zinc and ferritin between the two agro-ecological zones, despite similar mean food security scores.
We found that agricultural factors have a significant association with our selected food and nutrition security indicators. Our models for maternal ferritin and zinc outcomes explain some 58 and 53 percent, respectively, of their variance (adjusted R-square). The food security models explain 47 and 34 percent, respectively, of HFCS and HDDS variance. The group of agricultural variables explain about 10 percent of the variance in maternal zinc, household dietary diversity, and food consumption scores, and some 20 percent for maternal ferritin.
Statistically significant positive associations were found between total livestock units and maternal zinc outcomes, women’s empowerment in farming decisions and maternal ferritin, and area farmed with both HFCS and HDDS. Negative associations were found for larger cash crop areas in the lowland region with maternal zinc and iron outcomes, as well as for crop diversity with HFCS and HDDS. Strong positive associations were recorded for fruit and vegetable and meat per capita consumption with maternal micronutrient outcomes, as well as asset wealth and quality of water source.
Conclusions
Despite the relatively small sample size, our regression models significantly explain a large proportion of food security and micronutrient variance for the dependent variables selected. Our conceptual framework also provides a theoretical grounding for the hierarchical regression in order to evaluate the relative importance of different sets of predictor variables, including a group of agricultural factors. This approach enables us to estimate the importance of key agricultural influences on food security and nutrition in relation to other important determinants.
Our results suggest that agriculture plays an important role in food security and nutrition outcomes, but that other factors, such as food consumption choices, asset wealth, and water quality are often more influential. Of the agricultural factors that were significantly associated with maternal micronutrient outcomes, the negative association between per capita cash crop areas and both maternal zinc and ferritin outcomes was somewhat surprising. This may be because cash crops such as groundnuts account for larger areas grown in the lowland region, which could be at the expense of food crops for households in that region. The models provide direction for further research.
Association of Dietary Diversity with Nutrient Adequacy and Socio-Economic Parameters among Rural Adolescents from Pune, India
Anjali V. Ganpule-Rao
Introduction
Nutrient deficiencies, excesses, and lack of dietary diversity are public health concerns in developing countries including India. Early identification and improvement of such diets will be a cost-effective preventive measure. However, this is hampered by the current nutritional assessment methods that are time consuming and require technical expertise. Dietary diversity scores (DDS) provide a rapid measure of diet quality but are under-investigated in the Indian scenario. Only few studies have validated DDS with bio-markers of circulating micronutrients. These research questions are addressed by this study among rural adolescents in India to assess validity of DDS against vitamin B12, hemoglobin concentration.
Methods
Pre-collected data of 18-year follow-up (year 2012-2014) on 650 apparently healthy adolescents of the Pune Maternal Nutrition Study (PMNS, a prospective birth cohort) was analyzed. Dietary diversity scores were calculated based on frequency of consumption of a food per day (FAO 2013 guidelines). Only foods that were consumed at least once a day and in daily quantities of more than 10g were considered. The nutrient adequacy ratio was calculated using reference recommended values by Indian Council of Medical Research (2010). Socio-economic status was assessed against the Indian National Family Health Survey standards (2012). Standard of living (SLI) scores were classified as Low= 0-14, Medium= 15-24, High = 25 to 67. Those with circulating vitamin B12 <150 pmol/L and with Hb <12 g/dL were considered deficient. Data were analyzed using Spearman correlation coefficient, ROC and analysis of variance tests.
Findings and Interpretations
At 18 years, 42 percent of boys and 55 percent of girls were undernourished (BMI<18.5 kg/m2), and 8 percent of boys and 4 percent of girls were overweight (BMI>25kg/m2) (WHO 2009). Average caloric intakes were 2964 and 2068 kcal/day in boys and girls. One third of boys consumed inadequate quantities of nutrients particularly energy and vitamin B12 while 50 percent of the girls consumed inadequate energy, protein and iron and 80 percent had low vitamin B12 intakes. Of the most frequently consumed twenty foods, ~60 percent were carbohydrate rich (cereals, legumes, and pulses).
Average dietary diversity score was 4/9 for boys and girls. Mean adequacy ratio (MAR) was 0.64 for boys, 0.45 for girls and was directly associated with DDS (r=0.58, P<0.01). Almost 60 percent of boys and girls had deficient circulating vitamin B12 levels and >50 percent of girls had hemoglobin levels <12g/dL Maximizing sensitivity and specificity, DDS of 5 was the best option among boys and girls for achieving 50 percent of MAR as well as for identifying those with deficient circulating levels of vitamin B12 and hemoglobin.
The majority of adolescents (90 percent) belonged to high socio-economic status. DDS was associated with economic status (income, type of house, occupation), place of residence (hostel vs family), and source of foods (own farm vs purchased) (p<0.05 all).
Conclusions
The current study demonstrates a high prevalence of macro as well as micronutrient deficiencies among Indian rural adolescents with low dietary diversity scores, low body mass index, and deficiencies in vitamin B12 and nutrients required for red blood cell synthesis. These deficiencies are remarkably high among adolescent girls. Dietary diversity scores were useful proxy indicators of nutrient adequacy as well as micronutrient status of rural adolescent boys and girls. Socioeconomic parameters were found to play a significant role in determining dietary diversity of rural adolescents.
Determinants of Participation in a Homestead Food Production Program in Rural Bangladesh
Thalia Sparling, IMMANA
Introduction
Many nutrition-sensitive agriculture programs rely on behavior change activated through convergent thematic pathways at several levels of engagement (participation, initial uptake, long-term uptake, and translation of knowledge, for example). For interventions where there is observed impact, and especially those without, challenging the theory of change through monitoring the implementation process will lead to more effective programs, especially through systems that can address delivery and uptake challenges early on in the program cycle. As part of an ongoing process evaluation during a four-year complex nutrition intervention, we aim to assess the predictive factors of initial program participation in rural Bangladesh.
Methods
We used data from the baseline survey and program implementation monitoring system of the Food and Agricultural Approaches to Reducing Malnutrition (FAARM) cluster-randomized trial in northeastern Bangladesh. In an analysis of 1290 participants (the intervention arm who also completed the baseline survey), we investigated the factors related to attendance at training sessions on horticulture, poultry production, nutrition, care, and hygiene practices. We investigated socio-demographic characteristics and household wealth; knowledge on nutrition topics; food insecurity and diets; gardening characteristics including land size, seasonal land changes, number of different crops grown, improved horticulture practices undertaken at baseline, earnings from sale of agricultural goods; screening positive for depression; and woman's agency measured in social support, mobility, communication, and ability to make decisions. In a multilevel linear regression model adjusting for the clustered design and with fixed effects for field facilitators in charge of the area, we examined the influence of these factors on the number of training sessions attended in the previous year and a half. Sixteen group events were delivered in this time: nine nutrition and hygiene courtyard sessions, three horticulture trainings, two poultry trainings, and two marketing sessions. We then conducted a dominance analysis including all variables from the multivariable regression with p<0.2 to assess the relative contribution of these factors in predicting session attendance.
Findings and Interpretations
The mean number of training sessions attended was 10.4 (SD 5.36), ranging from 0 to 19, but 16 being attendance at all sessions. Each additional year of a woman's age was associated with a 0.18 increase in session attendance and each younger year of age at first marriage resulted in 0.27 sessions less attendance. Higher household wealth was related to less attendance (-0.88 per wealth score unit). Women whose husbands had higher education were less likely to attend (-.25 per educational level). Women who came from families where the near land was under water for longer periods of time during the year were less likely to attend sessions, and women who came from families who owned their agricultural land and made more money selling dairy and eggs were more likely to attend (0.69 more sessions for owning agricultural land and 0.66 for each level of income from dairy sales). Women with better communication with their husbands, natal family, and peers (0.36 per communication score unit), less remittances from abroad, and less money earned from sale of fruits and vegetables were more likely to attend sessions, although these factors were only significant at p<0.1. Many of the groups of field facilitators in charge of delivering the program were also a significant factor in predicting attendance at sessions.
The dominance analysis ranked the village, the field facilitators, household wealth, age of index woman, husband’s educational category, age of the woman at first marriage, total amount of agricultural land owned, household food insecurity, money earned from sale of dairy and eggs, and length of time near land underwater during the year as the most important factors contributing to attendance.
Conclusions
Attendance in program training sessions was strongly influenced by the village in which they took place (possibly encompassing socio-cultural aspects, topography, and the intervention delivery characteristics) and the field facilitators delivering the program, highlighting that staff characteristics are an important factor in participation. Other factors most associated with increased attendance in program activities were being poorer, higher age of the woman, marrying later, and husband’s lower education. Some factors of land size, land characteristics, and agricultural productivity were associated with increased attendance while other similar factors were not. These results can contribute to tailoring programs to those most likely to participate and thus benefit, as well as to understanding the importance of field staff and the community dynamics underlying any intervention.
Magnitude of Fish Post-Harvest Loss and its Intervention in Small Scale Fisheries of Ethiopia: A Missing Link but Viable Approach for Food and Nutrition Security
Demeke Senbetu, OARI
Introduction
As the world population is expected to become 9 billion in the near future, increasing food production by 70 percent becomes an ever more urgent priority. However, a vital yet neglected step toward this end is to reduce the food losses that occur through the entire food system. Ethiopia loses more than US$8 million due to post-harvest loss of 10,000 metric tons of fish, which is one third of the total production. Reducing losses is therefore an important development goal in this sector. Vital to this end is to characterize the important factors and their magnitudes of fish loss in this sector.
Methods
The data was collected monthly for seven years between 2011 and 2017 from nine water bodies, which provide about 70 percent of the total national fish production. The study was carried out by adopting three methods: Informal Fish Loss Assessment Method (IFLAM), Load Tracking (LT), and the Questionnaire Loss Assessment Method (QLAM). On the first step, IFLAM, an informal method based on participatory rural appraisal principles, was used to generate qualitative and indicative quantitative post-harvest fish losses data. Then, LT was used to quantify losses at stages along the distribution chain or losses related to specific activities such as fishing, transport, processing, and marketing. Finally, QLAM was adopted by interviewing a population sample in a community (geographical area) using a questionnaire to validate data generated by IFLAM and LT.
Review of secondary data, such as existing socioeconomic and production information, and Initial contacts with a community through key individuals, such as local fishery officials and fishers was done before conducting fieldwork in order to have a thorough background knowledge of key issues.
Findings and Interpretations
From about 180 fish species that are found in Ethiopia, Tilapia, African cat fish, Barbus, Bagrus, Clarias, Labeo, Lates, and Nile perch species are the major contributors and commercially important fish, yielding about 90 percent of total production. Based on IFLAM, the factors that contribute to post-harvest fish losses appraised were spoilage, size discrimination, species preferences, operational losses, less market access, controlled fish price regulation, and others. Then the procedure of LT and QLAM were followed. Accordingly, the results indicate that from the total annual 21,084 tons of fish catch, the total post-harvest fish loss constitutes 9,066.2 tons (43 percent).
Conclusions
The determinant factor for post-harvest fish losses include less market access, size and species preference, inadequate infrastructure for fish handling, processing, storage, transportation, controlled fish price regulation, and distance from the central market. Fish will become unfit for human consumption within twelve hours of capture unless subjected to some form of cold preservation or further processing. Cold preservation through refrigeration and freezing is not applicable as it is very expensive and difficult to access electric power around the production (fishing) area. Thus, from this paper, it is possible to conclude that fish post-harvest technology such as fish drying, fish marinating, fish retaining cages, fish smoking, fish meal, and fertilizer processing as alternative means for discarded and lost fish, fish processing kits, and solar ice making are very important in reducing mass and nutrient loss which plays a great role in securing food and nutrition. Therefore, it is very important to further adopt, verify, popularize, and disseminate these technologies for end users to secure food and nutrition.
The Importance of Forested Landscapes for Food Security and Nutrition Across Agricultural Transitions: A Multi-Country Comparative Analysis
Josh Van Vianen
Introduction
Smallholder family farms are increasingly becoming commercialized and transitioning away from diverse subsistence systems towards specialized market-orientated operations, leading to dramatic shifts in the scale and nature of agricultural landscapes and the associated effects on the environment, ecosystem service provisioning, livelihoods, well-being, and health of local populations. We examine how changes in agricultural landscapes manifest themselves as dietary transitions which represent an often-overlooked social dimension of tropical conservation. We show there are some complex interactions between forests, people, and landscape configurations with consequences for health and nutrition that hold true across a broad range of diverse landscapes.
Methods
We have applied a novel methodological approach as part of the Agrarian Change Project, which aims to explore the nature of forest loss and landscape-scale agricultural transitions in tropical forested areas across seven landscapes in different countries: Indonesia, Burkina Faso, Ethiopia, Nicaragua, Bangladesh, Cameroon, and Zambia. Each landscape has historically undergone and is currently undergoing agriculture-driven landscape transformation, and three experimental zones we established within each landscape represent increasing levels of deforestation and transition away from forest-based livelihoods. Zones represented different stages of transition and were paired with household surveys (n=1904) exploring the diets and livelihoods of local people. Zone 1 represents landscapes where people are carrying out subsistence farming and are heavily dependent on forests. Zone 2 represents areas with an intermediate or mixed-farming system with less access to forests. Zone 3 represents locations that have been converted to monoculture agricultural systems with very little access to forests. We conducted a comparative analysis by using a mixed model approach that allowed us to draw some broad conclusions about the impacts of agricultural intensification, forests, and diets at the landscape scale.
Findings and Interpretations
After controlling for a wide range of variables known to contribute to the status of a household’s food security and nutrition, including relative wealth, education levels, household size, land availability, etc, we found that both a households’ position along the forest transition/agricultural intensification curve and its reliance on forests can have large impacts on food security, dietary diversity, and consumption frequency of nutritionally-important individual food groups. Reduced reliance on forests can negatively impact food security to varying degrees across zones, whereas dietary diversity is significantly better for those who have an intermediate-level reliance on forests. Furthermore, in both cases, the size of the effects of forest use are mediated by an interaction with landscape configuration. The impact of reduced reliance on forests for food security is amplified for those who remain forest-dependent in the highly modified agricultural zones. Contrastingly, the positive effects of maintaining a moderate level of forest dependence are diluted when maintained in the heavily modified zone. Other factors that consistently impact food security and nutrition are wealth, the amount of land owned, and self-producing staple foods.
Conclusions
We show that no matter what the landscape, there are some complex yet generalizable patterns that emerge when transitioning from forests and subsistence-based livelihoods to more market-based monoculture commodity production. Interactions between landscape configuration and forest use have significant impacts on food security and nutrition for local people in developing rural areas. The unique experimental design of this project allows us to gain some insights into some of the less obvious trade-offs related to agricultural intensification at the expense of forests systems. To look from the more forested parts of a landscape towards the heavily modified is in many ways like looking into the past as the frontier of deforestation has continued to encroach into more natural areas. Forests play important roles for food security and nutrition, and we have shown that as a landscape evolves to meet the needs of rural development, the importance of forests for people is maintained, or in many cases increases, as its extent within a landscape is diminished. Landscapes are complex by their nature and understanding the tradeoffs inherent in any land use change is an important step to creating relevant and high-impact policy, especially in the areas of human wellbeing.
Impact of Different Agroecosystems on Household Food Consumption, Health Outcomes and Autonomy of Adolescent Girls in Seafood Farming Communities in Bangladesh
Abdullah Al Mamun, FANRPAN
Introduction
Fish is a core item of peoples’ diets. In Bangladesh, fish provides 60 percent of animal protein. Fish and fishery products offer a rich source of protein, micronutrients (LC n-3 PUFA), and are low in saturated fats. Aquaculture is well established for ensuring food security, whilst shrimp-prawn farming in Bangladesh is dynamic. This work was designed to develop an integrated metric for the impact of access to aquatic foods on health and nutrition – the Aquatic Food for Health and Nutrition index – with a special focus on adolescent girls as an at-risk group for poor nutritional status.
Methods
On the basis of water salinity, four different agroecological zones and one seafood processing community were selected for the study. As religion may influence food consumption patterns at the household level, communities with two adjacent para (small unit of a village), one which is Muslim-dominated and the other Hindu-dominated, were selected for each agroecology. The seafood processing community (SPH) was predominantly Muslim. Village and resource maps were drawn by local participants, subsequently, a well-being ranking was executed using methodology applied in previous studies. Twelve qualitative FGD's were conducted to design a measure of female autonomy that can predict health and nutritional wellbeing in a Bangladeshi context. A pretested survey tool was developed and implemented at the household level. At the final stage, 60 households were selected from each community, ensuring equal representation from different social wellbeing classes and religious communities, including at least one unmarried adolescent girl (12-16 years old). Food frequency questionnaires were developed and administrated. Anthropometric data (height, weight, age) and blood and urine samples were collected at a central location in each community. A separate set of questions was asked to the adolescent girls under a different domain to understand their autonomy at the household level.
Findings and Interpretations
The intra-household food allocations were found to be heterogeneous across the saline gradients. Adolescent girls and mothers consumed less food than what is required and also compared with the fathers and any brothers in the same household. Education, mobility, ambition, access to food, social safety and security, sports and outdoor activities, and health and hygiene all impacted female autonomy. Adolescent girls from Hindu communities enjoyed more autonomy in terms of mobility, ambition, and food choice compared to individuals from Muslim communities across the saline gradients. Girls living in the higher saline areas typically consumed more fish with a higher content of LC n-3 PUFA, and also had a higher omega-3 index, compared with girls from lower salinity areas, irrespective of social wellbeing and religion. We are currently continuing to refine the preliminary metric with the help of local health workers to align the outcomes of local practitioners to implement the metrics in the real world.
Conclusions
Establishing the relationship between aquaculture agroecosystems producing nutritious foods and their impact on the health and nutritional status of local communities living in such dynamic aquatic eco-zones is challenging due to the complex ecological dynamics of seasonal and annual fluctuations in freshwater supply and variable salinity gradients in aquatic environments, which is especially relevant for a country like Bangladesh. Another difficulty is the accurate assessment of fish consumption, and how this relates to individual health outcomes, as this relationship can be confounded by many factors, including the health impacts of poor living conditions, the high prevalence of infectious diseases, and more generally, the poor nutritional quality of diets. Here we demonstrate a novel application of the use of a new metric and how this could be applied in different agroecological settings. The paper also highlights the advantage of mixed methods and using qualitative methodology to inform the design process when trying to distil often complex concepts. Such work provides guidance to further the development and incorporation of other, broader agricultural concepts into nutritional security and a healthy life.
