10 minute abstract-driven presentations.  

Speakers and Presentations:

Chair: Stuart Gillespie, International Food Policy Research Institute (IFPRI)

• Swetha Manohar, JHSPH, USA 
  Seasonality of consumption of non-staple micronutrient-rich foods among young children in three geographically diverse Nepali communities  
  Recording / Slides

• Gerald Shively, Purdue University, USA 
  Assessing the rainfall-agriculture and rainfall-health nutrition pathways in Uganda
  Recording / Slides

• Edward Joy, LSHTM, UK 
  Realistic and healthy dietary changes to address freshwater constraints in India: a modelling study  
  Recording / Slides

• Upasona Ghosh, Institute of Health Management Research, India 
  Pathways of child food insecurity amidst climate change A Case Study of Indian Sundarbans 
  Recording / Slides

• Nick Chisolm, University College Cork, Ireland 
  Seasonal variations in household food security and dietary diversity and associations with maternal and child nutritional status in rural Ethiopia
  Recording / Slides

• Rojee Suwal, Helen Keller International, Nepal 
  Household Food Production and Maternal and Child Dietary Diversity in Nepal: Variations in association by season and agro-ecological zone
  Slides

Abstracts:

Seasonality of consumption of non-staple micronutrient-rich foods among young children in three geographically diverse Nepali communities

Swetha Manohar, JHSPH, USA

Introduction: A diverse diet that includes micronutrient-rich foods is important for child nutrition and health. Due to the crosssectional design of most dietary diversity studies, the potential seasonality of child diets is not well understood across the different ecologies of Nepal. We examined seasonal patterns in child consumption of key micronutrient rich foods across Nepal’s three agroecological regions, and assessed whether seasonal consumption patterns vary by wealth and caste/ethnicity.

Methods: We analyzed longitudinal data from three PoSHAN Community Studies surveillance sites. Data was collected three times per year during the monsoon (May-July), harvest (September), and winter (January-February) seasons, between May 2013 and February 2015. A sevenday food frequency questionnaire was used to collect dietary data from children 6-72 months in the three sites: mountains (Jumla, N=226), hills (Arghakhanchi, N=168), and plains (Banke, N=225). For each food group (vitamin A-rich fruits and vegetables, dairy, eggs, and flesh foods), we calculated summary statistics and the relative contribution of individual foods to total food group consumption by season and region. We then fitted multivariate negative binomial models to estimate the relationships of season, wealth, and caste/ethnicity with consumption frequency for each region. These models provided coefficient estimates in the form of incident rate ratios (IRR), interpreted as the ratio of times a child in a given category consumed a food in a seven-day period compared to children in the baseline category. Finally, we calculated and plotted the average adjusted predicted consumption frequencies for each food group across seasons, wealth levels, and caste/ethnicity groups.

Findings and Interpretations: Children’s diets were very low in micronutrient-dense foods year-round, with seasonal vulnerability to particularly low consumption varying by agro-ecological region and socioeconomic group. In the mountain site (Jumla), consumption of vitamin A-rich fruits and vegetables was lower in the harvest season compared to the monsoon season (IRR: 0.5, p < 0.001), and in the winter those with lower income were particularly vulnerable to low consumption (IRR: 0.5, p < 0.001). In the plains site (Banke), weekly consumption of vitamin A-rich fruits and vegetables was lower during the harvest season compared to the other two seasons (IRR: 0.2, p < 0.001), reaching close to zero. Dairy consumption decreased during the monsoon season in the hills (Arghakhanchi), compared to the other two seasons (IRR: 1.6, IRR: 1.5, p < 0.001) and in Jumla dairy consumption decreased during the winter, but the change was only significant among Dalit (very low caste) children (IRR: 0.3, p < 0.004). Consumption of meat and fish was slightly higher during September in Banke compared to the other two seasons (IRR: 1.6, p < 0.001). Interestingly, this increase was primarily attributable to greater consumption of fish among Muslim children at that time of year (1.7, p < 0.004).

Conclusions: Overall, these findings emphasize Nepali children’s low consumption frequency of micronutrient-rich foods and the need for program planners to be watchful for and responsive to seasonality. Seasonal variations in young children’s consumption of non-staple nutritious foods may result in poorer diet quality at certain times of year and among certain population groups. There are also, however, opportunities to identify by examining the “positive deviant” practices adopted by certain groups who adequately cope with seasonal availability of foods, or are able to substantially increase their consumption at certain times—practices that could be promoted more widely in the same population. Our findings also emphasize the need for more widespread and long-term collection of seasonal dietary intake data to identify groups most vulnerable to poor diets, and for researchers to consider the implications of when data collection occurs and its potential to mask or accentuate deficiencies in diet quality and disparities between groups.

Assessing the rainfall-agriculture and rainfall-health nutrition pathways in Uganda

Gerald Shively, Purdue University, USA 

Introduction: This paper studies nutrition outcomes of 4,463 children below age 5 in Uganda. In addition to studying traditional indicators at the child, mother, household and district levels, data on rainfall are used. Rainfall is linked to shortterm child growth through two pathways, one for agriculture and another for health. When rainfall is sufficient, crop yields are likely to be high, leading directly to more food consumption and indirectly to lower food prices. The link between health and rainfall can be traced through its influence on the household health environment and disease transmission. The analysis aims to measure these connections.

Methods: Child growth is measured using weight-for-height (WHZ) and height-for-age (HAZ) z-scores, obtained from 2006 and 2011 Uganda DHS surveys. Rainfall data come from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS), gridded satellite-based precipitation estimates from NASA and NOAA with daily time scales at 0.05 x 0.05 degree spatial resolution. Analysis of links between rainfall and WHZ follow the approach of Rabassa, Skoufias and Jacoby (2012) and Tiwari and Jacoby (2013). For HAZ the approach of Cornwell and Inder (2015) is followed. The analysis assesses the importance of cumulative rainfall received at different periods -- before and after birth, and during the main agricultural season -- and its associations with child nutrition outcomes. Two hypotheses are tested: (1: agriculture pathway) that children observed (conceived, born, or contemporaneously measured) in periods following a high rainfall season have higher WHZ and HAZ; and (2: health pathway) that children observed during or immediately following a high rainfall season have subsequently lower WHZ. Four-level hierarchical regressions are used to account for the structure of the data: nutritional outcomes (first level), nested within households (second level), nested within clusters (third level), nested within districts (fourth level).

Findings and Interpretations: Evidence supports the hypothesis of a negative health pathway. Diarrheal disease has a significant and negative association with WHZ when rainfall is included only as a moderator or as a moderator as well as a covariate. This association remains negative and significant when the production pathway (the main crop yield) is included as a control. Evidence also supports the hypothesis of a positive agricultural pathway. The main crop yield has a significant and positive association with WHZ when rainfall is not accounted for and when rainfall in the previous year is accounted for only as a moderator or as a covariate. The effect is positive and significant below 665 mm of rainfall received in the main rainfall season of the previous year -- when rainfall is accounted for only as a moderator, and is positive and significant above 403 mm of rainfall -- when rainfall is accounted for as a moderator as well as a covariate. The health and agricultural productivity effects of rainfall are somewhat offsetting: at low ends of the rainfall spectrum, crop yield effects dominate, leading to low child nutrition outcomes; at the higher end of the rainfall spectrum, negative health effects offset positive crop yield effects.

Conclusions: This study highlights the importance of rainfall in influencing the nutrition outcomes of children under age 5 in Uganda. The findings for the health pathway show that diarrheal disease has a negative and significant association with child nutrition outcomes, which becomes more negative with increasing rainfall received in the main rainfall season in a given year. This reveals the importance of rainfall in influencing the household’s disease environment and its impacts on the child health in the short-term. The findings for the production pathway suggests that at the lower end of the rainfall spectrum crop yields are low leading to lower child nutrition outcomes, but at the higher end of rainfall spectrum the benefits of rainfall on crop yields are offset by the effects of diarrheal disease. The negative disease effect dominates. This further reveals the importance of rainfall in influencing the household’s disease environment and its impacts on child health in the short-term.

Realistic and healthy dietary changes to address freshwater constraints in India: a modelling study  

Edward Joy, LSHTM, UK 

Introduction: Diets in India have a large blue water footprint (WF) relative to other countries. The blue WF represents the embedded irrigation water in crop and livestock products. Irrigation is widely used in India, and this dependence is likely to increase with population growth and dietary change, yet water resources are already stretched. This challenge may affect the availability of and access to healthy diets, particularly for poorer populations. This study aimed to identify realistic and healthy dietary changes that, if adopted at a wide scale, could enhance the resilience of the Indian food system to future decreases in water availability.

Methods: The blue WFs of typical Indian dietary patterns were quantified based on published dietary and WF data. A target reduction in blue WFs was based on projected declines in per capita freshwater availability to 2025 (18%) and 2050 (30%). The consumption of 36 food groups was optimised to meet blue WF reduction targets and WHO nutritional guidelines while minimising deviation from existing patterns. To ensure that results were equitable, the WF target was applied across all dietary patterns, such that greater reductions in blue WF were required for dietary patterns with higher baseline blue WFs. Resulting changes in life years lost due to coronary heart disease, stroke, diabetes and cancers, were modelled using life tables. Changes to greenhouse gas (GHG) emissions associated with the diets were also quantified.

Findings and Interpretations: The target reductions in blue WFs of diets was achievable with relatively small changes to diets. Optimised diets typically contained less wheat, dairy and poultry, and more legumes. The lower WF diets were also generally healthier. Adopting diets for the 2050 scenario would result in 6800 (95% CI 1600–13100) life-years gained per 100000 population over a 40-year period. Dietary changes were also accompanied by reductions in GHG emissions. The magnitude of the health and environmental impacts varied between dietary patterns. For example, wheatbased diets had larger blue WFs than rice-based diets (wheat is typically grown under irrigation) and wheatbased diets changed more following optimisation.

Conclusions: The projected decline in per capita freshwater supplies poses a challenge to ensuring availability and access to healthy diets in India. Policies to guide dietary choices could reduce blue WFs of diets and deliver environment and health co-benefits; however, current dietary trends run counter to some of the optimisation model outputs. For example, consumption of dairy and poultry products is rapidly rising in India, yet these products have large blue WFs and their consumption decreased in the optimisation process. However, relatively small changes to consumption patterns across food groups achieved significant reductions in dietary blue WFs demonstrating the strong potential for diet-led strategies (alongside production-side innovations) to shape a healthy and sustainable food system in India.

Pathways of child food insecurity amidst climate change A Case Study of Indian Sundarbans

Upasona Ghosh, Institute of Health Management Research, India 

Introduction: Climatic vulnerability affects food security by altering all four dimensions of food security: food availability, accessibility, utilisation and food system stability. It is hence seen as a crucial “hunger risk multiplier”. The Indian Sundarbans is an example of a badly affected region due to climate change that is at a high risk of food insecurity. The impacts of climate change manifested as extreme weather conditions, gradual land erosion and increase in land salinity have greatly affected the four dimensions of food security in the region.

Methods: We used a comparative case study approach using a cross-sectional, mixed method design. For the purpose of comparison we used a criteria based sampling to select three villages based on their proximity to the ocean and a systematic random sampling for selection of households with children below 6 years of age in each village. Villages were segmented based on catchment area of Integrated Child Development Centres. 1041 Households were selected sequentially from the list of mapped households from all segments. Anthropometric measurements like height, weight and mid-upper arm circumference were conducted with all the selected children (0-6 years). The household survey assessed the following aspects- a) Malnutrition and childhood illness in the last 30 days. b) Access to health and nutrition services c) Assessment of food security in households d) Migration and its effects on households e) Experience of climate and climatic extremities and its effect on assets, livelihood and food security. The study team also administered qualitative tools like 1) Comprehensive vulnerability analysis with a mixed group of community members 2) In-depth interviews with the mothers of the selected children 3) Key Informant Interviews with all the public and private health providers in the villages.

Findings and Interpretations: Villages close to the sea are facing the most impact of climate change on the four dimensions of food security. Across the three region 25.4% of the households faced extremely high losses to food resources during climatic emergencies. 66.64% of the households face low to very low food security and 44.53% of the households show low to very low levels of child food security. People are using coping strategies ranging from borrowing money from others (67.62%) to buy food to cut down the number of meals per day (31.67%). As per USAID food security instrument for the last 12 months the region closest to the sea has a higher percentage [32.00%] of food insecure households compared to other regions. Qualitative findings suggest not only the amount of food is decreasing but the choices of food groups are also facing a limitation to rice, pulses and potatoes and consuming less the groups like meat/fish, fruits and dairy products. Further analysis shows prevalence of malnutrition is higher (Underweight 31% and stunting-36.5%) in households facing very low or low food security and average number illnesses is 8 times in a year experienced by children in a year.

Conclusions: The study presents empirical evidence on the pathways of climate change impact on child food security in climatically vulnerable Indian Sundarbans. However, the risk is still not well reflected in the malnutrition status of the children due to range of short term coping strategies taken by the parents. The populace bounces back each time after a climatic shock but worse than the previous. Simultaneously, there is no such adaptive mechanism or mitigation plans for sustainable dealing with the changing situation. Social programmes which exist to ensure food security in the region need to be adapted to cater to this vulnerable population. Even within this special climatic region there are some pockets as identified by the study which will need more support in specific times. Additionally, long term measures to increase the resilience of the population and mitigate the continuous environmental degradation due to climate change need to be urgently identified.
 

Seasonal variations in household food security and dietary diversity and associations with maternal and child nutritional status in rural Ethiopia

Nick Chisolm, University College Cork, Ireland 

Introduction: In Ethiopia women and children are severely affected by poor health and under-nutrition. Food availability and dietary diversity influence this status, and are likely to vary seasonally. Different indices have been developed to measure food insecurity (e.g. HFIAS, HDDS, FCS), but there have been few analyses which investigate the relationships between these indicators, their seasonal variation between post-harvest and lean season, and direct measures of women and children's nutritional status. This study measures nutritional status through anthropometric and bio-marker measurement and analyses the relationship with measures of food security status in both seasons.

Methods: The study was conducted in two rural areas of Ethiopia: Eastern Oromia and South-Central Tigray. Data were collected in two seasons: post-harvest and lean season. In total 800 households were surveyed and a wide range of socio-economic and food security-related data were collected. A sub-sample of 183 mothers and their children age 6-23 months were then surveyed in both seasons for nutritional, socio-economic and food security data. The analysis in this paper is based on the sub-sample. Anthropometric measurements of mothers and children were performed using standardised WHO procedures. Blood samples were taken from mothers and children enabling measurement of haemoglobin, anaemia, serum zinc and ferritin. Food security/dietary diversity indices were calculated at household level (HFIAS, HDDS, FCS), and dietary diversity was also measured separately for mothers and children. A range of analyses of statistical association were undertaken, including comparison of mean values, correlations and regression using ordered logit and probit regressions. These analyses investigated the relationships between the different food security/dietary diversity indices by season, and the measures of nutritional status of mothers and children.

Findings and Interpretations Food security indicators and most anthropometric measurements for mothers and children showed higher levels post-harvest than pre-harvest. But child wasting appeared to improve pre-harvest: this result reflects reduced wasting (and improved infant dietary diversity) in Eastern Oromia, possibly due to the increased market orientation in Oromia, mitigating the shortages of household food crops, in contrast to Tigray. There are few statistically significant correlations between food security indicators and measures of nutritional status, although household FCS is correlated with maternal BMI and haemoglobin post-harvest, and with weight-for-age and weight-for-height for infants in the lean season. Regression analyses found (controlling for other factors) maternal haemoglobin and ferritin are significantly associated with household FCS both post-harvest and in lean season, and with HDDS post-harvest only. Child stunting, wasting and underweight are all significantly associated with HDDS post-harvest. No significant relationship was found between food security indicators and bio-markers for children: the latter are partially influenced by low consumption of animal-source foods. Both HDDS and household FCS were significantly associated with maternal dietary diversity in both seasons: household dietary diversity may be an important indicator for maternal nutrition. Variation in mothers' nutritional status by season shows the need to examine dietary intake by season.

Conclusions: The paper analyses association between food security indicators and nutritional status of mothers and their children aged 6-23 months, across seasons and locations. Significant levels of under-nutrition and food insecurity were found, particularly pre-harvest. Household FCS is a predictor of maternal BMI and haemoglobin in both seasons, while HDDS is a predictor of maternal haemoglobin and ferritin post-harvest. Children's weight-for-length was associated with household FCS in both seasons, suggesting household FCS is a useful tool to measure acute under-nutrition in both infants and mothers. Weight-for-length and length-forage are associated with HDDS post-harvest, suggesting Page | 36 greater dietary diversity at household level contributes to improved nutritional status of infants. Overall, household FCS is positively associated with nutritional status of mothers and children in both food surplus and shortage seasons, while HDDS is best in predicting maternal and child nutritional status during the food surplus season. These differences highlight the importance of using multiple indicators and measurements. Improved food security at household level appears to lead to nutritional benefits for mothers and children, suggesting intra-household allocations are managed effectively at household level. Agricultural and related interventions need to focus on ensuring access to increased quantity and diversity of food, particularly in the lean season.

Household Food Production and Maternal and Child Dietary Diversity in Nepal: Variations in association by season and agro-ecological zone

Rojee Suwal, Helen Keller International, Nepal 

Introduction: Dietary intake, an immediate determinant of undernutrition, is poor in Nepal. Many agriculture programs are production oriented and not focused on dietary quality and diversity of food produced. Suaahara, a large scale integrated nutrition program, sought to improve the nutritional status of women and children. Enhanced homestead food production (EHFP) activities, including both gardening and poultry rearing, complemented behavior change communication to promote the essential nutrition and essential hygiene actions (ENA/EHA). This study examines associations between EHFP participation and dietary diversity among mothers and children under 2 years and variation in these associations by season and agro-ecological zone.

Methods: Data from Suaahara household seasonal monitoring surveys (n=2,929 mothers; n=1837 children 6-23 months), which included dietary data from a 7-day recall for mothers and children, as well as maternal reports on participation in five EHFP activities: 1-3) received vegetable seeds, chicks, and technical support, 4) participation in EHFP training, and 5) participation in EHFP beneficiary groups. We constructed binary variables for each activity and a scale (0-5) summing participation in these 5 EHFP activities. We used the 10 food-group Women’s Dietary Diversity Score (W-DDS) for women’s diet diversity and the 7 food-group classification for child diet diversity. Multivariable linear regression analyses were used to assess associations, by season and agroecological zone, between participation in each EHFP activity and the overall EHFP scale with both maternal and child dietary diversity. We controlled for the following potential confounders: child sex; women’s age in years; women’s education; household caste; number of family members in the household; number of children under 5 years of age in the household; labor migration of any adult household member; household access to food markets; size of land owned by the household and clustering at the district-level.

Findings and Interpretations: We found several positive significant associations between participation in EHFP activities and maternal and child dietary diversity in rural Nepal and in the context of a large integrated nutrition program. Suaahara. We also found that these associations varied by EHFP activity, season and agro ecological zone when controlling for various cofounding factors at the household and individual level as well as district level clustering. In adjusted models, we found three EHFP activities to be associated with dietary diversity – chicks, technical support and EHFP groups, but the magnitude of the associations varied by season and context. The degree of participation in five EHFP activities was positively associated with maternal dietary diversity in the terai (β=0.24, P<0.001) and mountains (β=0.12, P=0.01) during the winter but not rainy season and for child dietary diversity only in the terai in the winter (β=0.35, P<0.001). Our findings highlight that the relationships between EHFP and maternal and child dietary diversity are complex and may vary by place and time.

Conclusions: Our findings highlight the potential nutritional benefits of EHFP participation among mothers and children under 2 years. However, we found variation by sub-national context and seasonality in the magnitude and significance of these relationships. Future programs and policies may need to consider these context-specific factors if agricultural programs are to improve nutrition year-round in Nepal. It is important to remember that food alone is not sufficient: food, health care and a clean environment and care are all important for nutritional wellbeing. For this reason, EHFP programs complement homestead garden and backyard poultry interventions with ENA, EHA, and other nutritional related SBC strategies to promote gender and social inclusive health and nutritional advancements. For example, Suaahara’s distribution of seeds and chicks and provision of technical support was complemented by Suaahara’s EHFP beneficiary groups in which agriculture-health-nutrition linkages were discussed, as well as other non-EHFP Suaahara activities including mass media campaigns and counselling of pregnant and lactating women. Future assessments of Page | 38 EHFP should be designed to quantify the contribution of each of these programmatic investments to better understand the context in which EHFP is most able to influence maternal and child diets and nutrition.