Session 8B: Food environments and drivers of food choice Parallel Session
byANH Academy
Academy Week Research Conference
| Food Environments, Social Sciences
Date and Time
From: 28 June 2019, 14:30
To: 28 June 2019, 15:50
BST British Summer Time GMT+1:00
Country: India
Open Full Event


Five,10-minute abstract-driven presentations.  

Speakers and Presentations:


  • Chair: Matilda Laar, University of Ghana
  • Celine Termote, Bioversity International
    Food-based recommendations to improve dietary adequacy of women living in pastoral and agro-pastoral zones of Turkana County, Kenya
    Slides/ Recording

  • Archana Konapur, National Institute of Nutrition
    Identification of food environmental factors hindering dietary diversification: A mixed methods study in rural south India
    Slides/ Recording

  • Samira Choudhury, SOAS University of London
    Drivers of fruit and vegetable intake in India: An unconditional quantile regression approach
    Slides/ Recording

  • Solveig Cunningham
    Understanding food choices in the context of globalizing food options: Evidence from a novel picture method in India

  • Ursula Trübswasser, Wageningen University
    How do adolescents understand their food environment? A pilot study using Photovoice in urban Ethiopia
    Slides/ Recording



Food-based recommendations to improve dietary adequacy of women living in pastoral and agro-pastoral zones of Turkana County, Kenya

Celine Termote, Bioversity International 

Introduction:  The efficient use of local foods in addressing food insecurity is becoming more important as the global population increases. Dietary analysis is not only a crucial part of nutrition status assessment but also forms the basis for planning food-based recommendations (FBRs). FBRs are easily adaptable and important in enhancing sustainability of food systems and promoting healthy eating. Optifood is a linear programming tool used to test and formulate FBRs for specific target groups within the context of local constraints such as food availability and cost. This study combines dietary analysis and diet optimization using local foods to attain nutritional adequacy.

Methods:  The study adopted a stratified cross-sectional design on a sample of 240 women aged 15 -49 years from the pastoral and agro-pastoral livelihood zones of Loima sub-County, Turkana County, Kenya. Food intake data was obtained using enumerator based multiple-pass quantitative 24-hour recalls on two non-consecutive days. Analysis was done in SPSS 24 while making comparisons between the pastoral and agro-pastoral livelihood zones in terms of quantities of food items, food groups and nutrients consumed. Food items were categorized into ten food groups to allow for correlations with dietary diversity. The food items identified from the 24-hour recalls formed the basis for formulating FBRs in Optifood. Target group details, list of food items and serving sizes (median per day), food and food (sub) group constraints were entered in Optifood to test and formulate FBRs. Food and food (sub) group constraints were entered in terms of lowest, average and highest servings per week based on the 5th, 50th, and 95th percentiles.

Findings:  The dietary patterns of women in pastoral and agro-pastoral zones were comparable where nutrient intakes were below recommendations. Forty food items were identified from the 24-hour recalls which indicated limited choice in terms of food variety. Mainly four food groups were consumed by more than 50% of the women, namely fats and oils, added sugars, grain and grain products, legumes and pulses, which was linked to low dietary diversity, hence micronutrient inadequacy. Plant foods were the main source of nutrients where grain and grain products contributed 60% of energy. This indicated a monotonous staple-based diet, characteristic of most households in low-income countries. In the Optifood model, grain and grain products had the lowest constraints while animal products had the highest constraints which was associated with increased cost. The generated FBRs for both zones in terms of food groups were comparable namely: vegetables (3-4 servings), dairy (2-3 servings), fats (2 servings), meat and eggs (1-2 servings), grain and grain products (2-3 servings) and legumes (2 servings). Daily FBRs for fruits were not feasible within the current constraints, which were associated with increased cost and seasonality.

Conclusions:  Many developing countries are yet to develop food-based dietary guidelines that can be adopted at local or national level. More research aimed at understanding dietary patterns of different population groups is required to facilitate development of feasible FBRs. FBRs formulated in terms of food groups allow for flexibility in terms of choice, seasonality, and availability of different food items. Although Optifood can be useful in informing behavioural change interventions to improve nutrient intake, modelled diets may still be inadequate in specific nutrients (problems nutrients) or the FBRs generated may be too ambitious exceeding median consumption patterns of the target group.


Identification of food environmental factors hindering dietary diversification: A mixed methods study in rural south India

Archana Konapur, National Institute of Nutrition

Introduction:  Micronutrient malnutrition is prevalent worldwide and children are at high risk in developing countries like India. The Sustainable Development Goals (SDGs) emphasize the linkages between food systems and nutrition to achieve ‘zero hunger’ and food-based approaches including promotion of dietary diversity are important to meet the challenges of micronutrient malnutrition. Measures to promote dietary diversity should be based on context-specific measures. The study was aimed to explore five dimensions viz availability, accessibility, affordability, acceptability of a diverse diet by the consumers, accommodation of consumers’ food demands by food vendors and their influence on dietary diversity of children in rural middle-income households

Methods:  Mixed methods study comprising quantitative and qualitative components was conducted in 4 randomly selected villages (>500 households) of Ghatkesar sub-district, Telangana State, south India. Availability of foods in shops and markets were assessed by observation and interview with vendors using a checklist prepared using the Indian Food Composition Table (2017) and Nutritive Value of Indian Foods (1990) of 197 food items across 14 food groups. Accessibility to food markets by the households was measured by creating buffer zones using geographic information system (ArcGIS software). A pre-tested questionnaire was administered to 160 caregivers of 6-10 year old children to elicit information regarding perception on availability, accessibility, affordability, and acceptability of the foods in the market. A validated closed ended diet diversity score (DDS) questionnaire was administered to the same caregivers to elicit information on the dietary diversity of their children. The vendor’s accommodation of consumer’s food demands was assessed employing interview method.

Findings:  Seven wholesale and 35 retail stores in 4 villages were open for 12-hours a day and sold 45 non-perishable items and 12 perishable items daily. A variety (n=60) of green leafy vegetables (GLV), other vegetables (OV), fruits, fish were available only in a weekly-market which ran for 4-hours on one day in a week. Limited varieties of GLV (n=4), OV (n=5) and fruits (n=2) were sold by hawkers thrice a week. In total, wholesale/retail store/weekly markets together sold 107 foodstuffs. Non-perishable items, milk & flesh foods stores, weekly market areas were at an accessible distance (<2km) to the households. Mothers perceived that fruits and fresh fish were not available daily and have moderate difficulty to access them in neighboring villages; children dislike consuming GLV. The mean DDS of children was 8.7. Though availability of non-perishable foods was good, villages had fruit deserts, where no fruits were available locally on a daily basis. The same was perceived by caretakers of schoolchildren. Interviews with weekly market vendors revealed that few varieties of regional GLV and OV were not sold in the weekly markets as the villagers do not buy them. It was a demand-driven market in all villages. Wholesale/retail store vendors were accommodating consumer needs of non-perishable foods.

Conclusions:  The five dimensions of the food environment were measured contextually and non-availability of perishable food items such as vegetables, fruits, and fish was a limitation that hampered their inclusion in diets. Mapping the food deserts and assessing factors that impact availability, accessibility, affordability, and use at the grassroots level is necessary for promoting dietary diversity and improving micronutrient status.


Drivers of fruit and vegetable intake in India: An unconditional quantile regression approach

Samira Choudhury, SOAS University of London

Introduction:  Dietary risks are the top risk factor for disease burden in India. Fruits and vegetables (F&V) provide essential vitamins and minerals in a setting where micronutrient deficiencies are widespread. The Indian diet is cereal-dominated and F&V intakes are only about 165-180 g/person/day despite a significant percentage of the population following vegetarian diets. This paper examines the socio-demographic drivers of F&V intakes in India and, for the first time in the literature, also explores the potential association of food system drivers (road, market infrastructure, cold storage) and relative prices of F&V with recent F&V intakes.

Methods:  Using round 68 of the nationally representative national sample survey (NSS) household data (2011-2012) combined with food systems data (Village Dynamics in South Asia data, Indian Horticulture Database), we apply unconditional quantile regression methods (UQR) to determine the differential effects of these drivers at various parts of the intake distribution (unlike OLS estimation), with a special focus on severely inadequate intakes. The UQR methodology flexibly allows the impact of a change in an explanatory variable (e.g., income) in a population of individuals with different characteristics (unconditional effects) rather than the effect of sub-groups with specific values of covariates (conditional effects).

Findings:  The results of UQR at the 10th, 25th, 50th and 75th quantiles are compared with standard OLS (conditional mean) estimates. The UQR results (excluding relative prices) suggest that district-level market density has stronger positive effects at the lower end of the distribution (at the 10th and 25th quantiles) while road density has no significant impact on intakes. In contrast, the coefficient on state-level storage capacity relative to production area is the highest at the 90th quantile. Socio-demographic drivers, in particular, income and female head years of education boost intakes while relative prices of F&V reduce intakes only at the higher end of the distribution (at the 75th and 90th quantiles) without controlling for food system drivers.

Conclusions:  Our findings highlight the importance of allowing the effects of socio-economic characteristics, relative prices and food system drivers to vary across the intake distribution. The results indicate that any policy or program interventions that aim to improve market infrastructure for agriculture is important for improving F&V intakes, particularly those consuming the least F&V. But, evidence in support of road infrastructure on F&V intakes is limited. However, the main finding, that the impact of relative prices on F&V intakes is weak at the lower tail, i.e. those with the most inadequate intakes, is cause for concern.


Understanding food choices in the context of globalizing food options: Evidence from a novel picture method in India

Solveig Cunningham

Introduction:  Globalization of food supplies is reaching even remote parts of the Global South, introducing new foods in places that have largely relied on locally-grown foods until recently. This may increase the stability and diversity of food supplies, but also introduces calorie-rich-nutrient-poor items. Through a project in a remote district in southern India that globalization is just reaching, we document how people make decisions about food consumption when faced with new food options competing with traditional foods. We quantify drivers of food choice in adults’ and youths’ food selection, identifying the conditions under which global vs. traditional items are selected.

Methods:  The study is representative of households in an urban and a rural community in Vijayapura district in the state of Karnataka (n=400). We developed and tested choice experiments with picture-cards, based on our database of >1000 foods and beverages available locally, including foods which have just begun to enter the local market. Items were organized into six food groups: fruit+vegetables; cereals+pulses; snacks+sweets; animal products; oils+sweeteners+condiments; and beverages. Each group contains items categorized as local/traditional; national (non-local but of Indian origin) and global/modern based on our previous work. Each respondent was shown randomly selected items in sets of 3 from each: 1 local, 1 national, 1 global. Respondents were asked about their familiarity with and consumption of each. They identified which foods they consumed most frequently and why, and responded to a series of scenarios about which they would select, e.g., if they had more money, less time, were hungry, were considering their health. We identified which factors were most often selected and the frequency with which local, national, and global items were eaten. We calculated under which conditions people most frequently changed their food item selection. We estimated multivariate ordered logistic models, accounting for individual and contextual characteristics.

Findings:  Processed and packaged foods are increasingly available, but consumption is still low. We show which drivers of food choice are more salient among adolescents and adults; rural and urban households; poorer and wealthier households; men and women. We examine preferences for global vs. local and what circumstances or information would nudge people to shift towards or away from global/modern alternatives. We summarize preliminary patterns only for this abstract. Among adults, 44% had seen the global foods and 28% had consumed them; adolescents were much more aware of global foods: 75% had seen them and 35% had consumed them. For both adults and adolescents, local/traditional foods were most frequently eaten. When asked about reasons for selecting the most frequently consumed items, adults cited ease of availability (40%) while adolescents focused on taste (41%); cost was not a leading consideration. In experiments with altering purchasing conditions, when presented with a scenario of having more money, adults most often chose local/traditional foods, whereas adolescents more frequently opted for non-local Indian foods. These items were again selected when respondents were asked to focus on taste. Both adults and adolescents selected local items when asked to consider their health, ease of preparation, and hunger.

Conclusions:  Understanding food choice and availability in the context of India’s dual burden of underweight and overweight is relevant for the health of 1/5 of the world’s population. As the population is increasingly exposed to the nutrition transition, some factors like taste, health and ease of availability guide people in selecting which traditional items to retain in their diets and which new items to add. Adolescents are more on the brim of learning about and accepting non-local and global foods. It is important to understand the salience of these factors in developing programs to promote healthy diets and preventing malnutrition


How do adolescents understand their food environment? A pilot study using Photovoice in urban Ethiopia

Ursula Trübswasser, Wageningen University

Introduction:  Adolescence is a period in which nutritional needs increase and lifelong nutrition behaviors are formed, which will also affect future generations. With changing food environments, it is expected that adolescent nutrition in Ethiopia will deteriorate as research from high-income countries have linked overweight rates and unhealthy diets with changing food environments. Perceptions of the food environment can also have an impact on dietary choices, but need to be better understood. Research in Ethiopia has mostly neglected adolescent nutrition particularly in urban areas, leaving significant knowledge gaps to be addressed in the proposed study.

Methods:  The overall objective of this pilot study was to test the feasibility of the Photovoice method to understand adolescents’ perceptions of their food environment. The pilot was conducted in a public school of Addis Ababa with sixteen students (15 -19 years old) assessing their diets with qualitative 24-hour recall, anthropometry, as well as their knowledge related to nutrition, hygiene and perceptions of their body weight and their food environment. Students were also taking photographs in their environment of challenges and opportunities to eat healthy. In parallel, availability and advertising of foods and drinks has been assessed in a 1 km radius around the school, adapting INFORMAS protocols. Photographs have been taken of front views of 138 out of 207 kiosks to assess the types of foods on display and interviews with 60 food outlets have been conducted to get a full list of available foods. Foods consumed by students as well as on display and available at stores have been categorized into ten food groups and the four groups of the NOVA classification according to their level of processing.

Findings:  All students had low dietary diversity (3.4 food groups on average) and about half (n=7) of the students consumed ultra-processed foods. The majority (n=15) were normal weight using BMI-for-age, nevertheless seven students perceived their body as underweight. Most students (n=15) knew about links between water, hygiene and diarrhea. Links between consumption of salt, animal fat, sugar and overweight and noncommunicable diseases were known to about half of the students. Findings from the qualitative interviews and discussions indicate that food safety concerns, limited resources, and availability of certain foods influence adolescents’ dietary choices. Unhealthy and unsafe foods being more available and/or cheap in their neighbourhoods was mentioned by 12 students and unsafe foods or food outlets were also depicted in one-third of the photographs taken by students . In the environment surrounding the school, ultra-processed foods are widely available (78% of assessed shops), visible (in 70% of all photographed kiosks on display), and advertised (78% of all advertising). Students considered foods available in their environments as generally unsafe, calling for more packaged food (4 students). When asked what students would buy on the way to school if they had additional money, the majority (n=14) would buy either ultra-processed or deep-fried foods.

Conclusions:  Dietary behaviour of students appears poor in diversity and negatively affected by what is available in their neighbourhoods. Limited nutrition knowledge, preference for packaged foods and high availability and advertising of ultra-processed foods, could make those foods more desirable. Solutions proposed by students included improving safety, prices, and availability of healthy foods and getting their voices heard. The Photovoice method in combination with interviews and objective assessments of the food environment can be a useful approach to get a deeper understanding of how the perceived and objective food environment are related. However bigger studies are needed to confirm these findings.

Login or Join ANH Academy to be part of the conversation