Understanding access to nutritious food by poor urban pregnant women and lactating mothers and their children in Kisumu, Kenya

By Albertien van der Veen, Rik Delnoye and Femke van der Lee

Albertien van der Veen is an experienced public health nutritionist/epidemiologist and team-leader of the urban nutrition theme group at the Royal Tropical Institute (KIT) in Amsterdam. She has worked for more than 25 years in Africa, Asia and the Middle East in food and nutrition security. Her current research focuses on food-based strategies to reduce malnutrition.

Rik Delnoye is an agricultural economist specialised in agricultural development. A former senior advisor at KIT, he is now coordinating a multi-donor funded food security and economic opportunities programme in Laos.

Femke van der Lee is a rural sociologist specialised in agrarian development and food security, currently working as guest researcher at the Autonomic University of Chiapas in Mexico. She has worked and done research in various countries in Africa and Latin America.

The authors begin with thanks to staff and students of the School of Food Security, Agriculture and Biodiversity at Bondo University College, Kenya. Special thanks to Professor Monica Ayieko, whose involvement was instrumental in making this research happen. The authors also thank all key-informants from CBOs, (I)NGO’s, governmental and international organisations, as well as the traders, brokers and wholesalers, for great engagement. Particular thanks is extended to Dan Onyonga from the Pandipieri project; Monica Oyanga, assistant District Nutrition Officer; Susan Jobando, UNICEF Kisumu; Linda Beyer, UNICEF Nairobi; Rosemay Tienoras, District Nutrition Officer; Elizabeth Kimani Murage, Catherine Kyobutungi and Steven van de Vijver from APHRC; and Yacob Yishak and Koki Kyalo from Concern. A final thanks to the many more people who participated in the data collection in Kisumu.

Location: Urban slums, Kisumu, Kenya

What we know: In Kenya, 35% of the population is urban of whom nearly half lives in slums/ informal settlements. Undernutrition is a major problem, particularly amongst the urban poor.

What this adds: In this instance, there were adequate amounts of varied foods in urban slum markets but affordability limited access, especially in low season. Average results mask huge disparities in household expenditures and demographics within slums. A low Dietary Diversity Score amongst pregnant and lactating women correlated with a low per capita food expenditure and total expenditure, no bank account/savings, poor living conditions and lower education levels. Maternal dietary diversity was found to be a reliable predictor of child dietary diversity. Interventions to empower poor urban women in food supply chains are identified, including women’s collectives and food processing.

Sale of cabbage in the main market
of Kisumu during the peak season

Sub-Saharan Africa is rapidly urbanising and has the highest proportion of slum dwellers in the world. In Kenya, an estimated 35% of the population is urban of which nearly half lives in slums¹/informal settlements. The population of Kisumu, Kenya’s third largest city and the capital of Nyanza Province, is approximately 440,000 of whom 60% lives in slums.

Undernutrition is a major problem in Kenya, 35% of all children under five are stunted (short for their age due to growth faltering), and more than half of all children suffer from mineral and/ or vitamin deficiencies. Stunting among children living in urban poverty is around three times as high as among urban upper income children. Undernutrition among women of reproductive age, in particular pregnant and lactating women, is also common. At the same time, overweight and obesity in urban poor women and pre-school children is rapidly increasing, due to an unfavourable shift in consumption patterns (more fat and sugar, less polysaccharides and fibre).

The National Food Security and Nutrition Policy (FSNP) and the National Nutrition Action Plan (2012-2017) define Kenya’s policy and programmes on malnutrition. Kenya’s Nutrition Technical Forum plays an important role in the coordination of the implementation of nutrition programmes. One strategy that fits well with the strategic objectives of the Nutrition Technical Forum for urban nutrition is improving dietary intake through increasing dietary diversity. Promoting market-based solutions and their pro-poor positioning to secure good nutrition is one of the recommended strategies, but initiatives are few and limited in scale.

The Royal Tropical Institute (KIT), Netherlands, in close collaboration with Bondo University College (BUC), undertook research in Kisumu to design a model for gender and rights aware, cost-effective and sustainable strategies to better match demand (address needs) for and supply of nutritious food in urban areas. The ultimate aim was to improve food and nutrition security of pregnant women, and lactating women with children under two years.

Methods

The objectives of this research were to:

• Profile the primary target group in terms of food access and consumption (intake) including underlying factors related to food habits, gender and other intra-household distribution factors.
• Identify if, what types and how strategies to better match demand and supply can contribute to improving food and nutrition security of the primary target group in a sustainable way.
• Identify opportunities for the development and/or improvement of local supply chains connecting the demand of the food and nutrition insecure to affordable and accessible supplies of nutritious food, in a need-responsive and sustainable way.
• Explore relevant public-private partnerships in the form of business propositions. This included defining the impact pathways, roles and investments and expected return on investments at target group and partner level.

A conceptual framework that focused on immediate, underlying and root causes of (in-) sufficient dietary intake was used to guide the research. The framework elaborated the three key food and nutrition security pillars (access, availability and utilisation), and for each considered demand, needs and supply.

The research consisted of two main parts. The first part focused on understanding underlying socio-cultural and economic factors influencing the demand for and access to nutritious food and ultimately food intake. The second part concentrated on understanding the functioning of the agri-food system, taking the results of the first part as an entry point. The overall set-up of the study was of quasi-experimental design involving both quantitative and qualitative research methods. Household characteristics and food consumption data were collected through a cross-sectional survey.

Field research was carried out in Kisumu in the last week of October and the first three weeks of November 2012 (in the high season of food availability). Within Kisumu, the urban slums of Manyatta and neighbouring Obunga were selected purposely for the food access and consumption survey. This choice was motivated by the fact that these areas are representative for urban slum dwellers in Kisumu according to key informants.

Background² data were collected from key informants and seven focus group discussions. The focus group discussions were held with one mixed group (men and women) and six groups of women. Participants were living in three different areas: Obunga, Manyatta and Nyalenda. The groups were divided over three age categories, each of the age categories focused either on food habits of pregnant and lactating women or child feeding practices.

The food access and consumption survey collected data from a randomly selected representative sample consisting of 295 women of reproductive age. The number of clusters per location was determined using probability proportional to size. The household was the primary sampling unit. Two stage cluster sampling was applied to determine the location of each household. Data were collected by nine local enumerators and ten students from BUC. Data collectors were supervised by staff from BUC and KIT. SPPS was used for the analysis of the data.

Prices of food commodities were collected at six retail outlets varying from the major Kibuye market to smaller outlets close to the targeted slum areas. Collection of data on the value chain was done through in-depth interviews with chain actors. A total of 50 retailers, eight wholesalers and six brokers were interviewed in a semi-structured manner. In-depth interviews were also conducted with some (poor) households to explore seasonal influence on food consumption and coping strategies.

A nutritious food selection matrix and a chain selection matrix were used to select food products that could potentially be looked into for value chain development options. Both matrixes were developed during the research. The value chains of each of the selected food items were mapped and analysed³.

Potential value chain interventions of selected food items were explored using a gender lens, in particular the concepts of ‘agency’ and ‘structure’⁴. Potential strategic interventions for each selected food item were mapped in terms of economic feasibility, potential nutritious impact, agency and structure. An opportunity matrix, also designed during the research, was used to strategise these interventions

Main findings

Demographics

The total number of households interviewed for the household survey was 295, of which 60 were in Obunga and 235 in Manyatta. The average household size was 4.5 persons per household in Manyatta and 4.4 in Obunga. Nearly one in eight (16%) of all households was a female headed household. In two out of three households (67%), the highest level of education was secondary school or higher. In 27% of the households, the highest level of education attained was college or university. Additional characteristics are summarised in Box 1.

Box 1: Household profiles in Obunga and Manyatta

Sale of Sukuma Wiki in Kisumu

A little over half (54%) of the surveyed households lived in a permanent house (Manyatta, six out of ten households, in Obunga only one in four households). Nearly half (47%) of all households had a separate place for cooking - this was significantly higher in Manyatta than in Obunga. Three quarters (75%) of the sample households had their own pit-latrine. Over half (59%) were connected to the town electricity supply, 40% of the sample households had no electricity.

One out of five households (21%) in the survey owned productive animals, 90% of whom had chickens (on average 5 chickens per household). One out of three animal owners possessed other poultry; mainly ducks. Two households owned goats and six households owned other productive animals.

Only 7% of the sample households had a vegetable garden. Thirteen households (4%) had one or more fruit trees and 15 households (5%) had a shamba⁵. Shamba’s were mainly used for growing maize. Despite its location at the edge of town bordering the country-side, Obunga households were not more likely to have a shamba or vegetable garden than households in Manyatta. The percentage of households having animals was also nearly the same.

In 96% of the households, at least one person had been working over the last six months. In Manyatta, over half of the households was employed as business owner (defined as a person employing others) or doing skilled work for a regular wage during the last six months. In Obunga, only one out of three people working were similarly employed. In Obunga, 38% were working as an unskilled employee (for either a regular or irregular wage).

Expenditure

Average per capita expenditure of all households was 3,800 KES per month or 127 KES per day. Differences between households were huge, with expenditure ranging from less than 20 KES per person per day to more than 1,150 KES per person per day. Most households spent about 100 KES per day. Households spent on average 2,100 KES per person per month on food, or 70 KES per person per day. Again this average masks large disparities. Nearly 10% of all households spent only 20 KES per person per day on food, while 5% of all households spent 150 KES person per day or more. Expenditure on food was on average 55% of the total household expenditure (in smaller as well as in larger families). The share varied hugely between individual households however, ranging from 5% to 92%. One out of four families spent more than 66% of its budget on food. Rent, costs of utilities (electricity, water etc.) and transport made up the bulk of the remaining expenditure.

Food consumption and diet diversity

Nearly half of the children under two years and one-quarter of the sample women consumed a diet of poor quality, lacking vitamins and minerals due to insufficient diet diversity (individual dietary diversity scores (DDS)⁶). Based on 24 hour recall, approximately 30% of the children and women didn’t eat any fruits or vegetables and only half of the children consumed green leafy vegetables or other vitamin A rich vegetables such as tomatoes. Maternal diet diversity was found to be a reliable predictor of child diet diversity. Children aged 6 to 24 months whose mothers didn’t eat sufficiently varied in the survey were nearly four times as likely to have a low DDS as children whose mothers had an adequately varied diet.

A higher DDS was significantly correlated to a higher per capita total expenditure in general and food expenditure in particular. Interestingly, half of the women whose per capita food expenditure was less than average had an adequate DDS (5 or higher). The only significant difference between those with and without an adequate DDS in the low expenditure groups was the household size. Women in larger households were twice as likely to have an adequate DDS as those in smaller households, possibly because their total daily budget for food was higher, enabling them to purchase more different types of food items. .

A low maternal DDS (and children’s DDS) was correlated with a low per capita food expenditure and total expenditure; not having a bank-account and/or savings; living in a house without electricity and/or a separate place to cook; not owning a shamba; and an education level of unfinished secondary school or less. Findings were inconclusive on intra-household distribution and decision making, e.g. prioritisation /availability of money for nutritious food, warranting further research.

Food availability

There was enough and sufficient varied nutritious food available at the local market(s) in Kisumu, but affordability (and therefore access) was a problem. Price was an important determinant of / reason for choosing where to buy food, but the possibility to buy on credit or in small quantities were factors of influence as well. Affordability, in particular in the off-season is a problem for local nutritious green leafy vegetables (sukuma, etc), fruits (oranges, mango) and other commonly consumed vegetables (cabbage, carrots, tomatoes), pulses and seeds (beans, peanuts) and fish (in particular big fish). In the off-season, the cost of a daily food basket is approximately 70% higher than in the peak season, and out of reach of the poor.

In both the high and low seasons, the cost of a nutritiously well balanced and less balanced diet is about the same. Interestingly, if a household spends 70 KES per person per day (the average found in the household survey), it is quite possible to consume a nutritious diet in the high season, but not in the low season (see Table 1). Households said they coped with the high prices by buying smaller quantities, lower quality and/or consuming substitutes (Omena (small fish)⁷ instead of bigger fish, lemon instead of oranges). Vendors were found to accommodate these strategies. Insects, particularly termites, were consumed by over 40% of surveyed households in times of availability. However, outside the season, few traders/ vendors offered insects and usually termites only. Supply was subsistence oriented and commercialised production hardly exists.

Based on average prices. Prices in the off-season may be higher or lower, depending on supply. **3 tomatoes or 2 small mangoes as substitute for pumpkin.

The local food markets are typical spot markets, characterised by lose, informal and often one-off trade relations, causing high levels of default and high transaction costs. Buying power is in the hands of wholesalers often working in cartels allowing them to influence pricing and determine other conditions (such as packaging, measurements) in the trade. The Nairobi wholesale cartels, in particular, are well known for their market power. Kisumu is strategically positioned along major trade routes for agricultural products.

Value chains

Four leading indicators were combined to select value chains that could improve availability of affordable nutritious food at the short/medium long term. These were: (i) the extent to which the food item responded to nutritional deficits, nutritional value and availability in periods of shortage, (ii) socio-cultural acceptance, preference and pricing, (iii) potential for chain interventions to extent the periods of availability and (iv) competitiveness at the consumer-end.

Analysis of the chain selection matrix revealed that mangoes, traditional green leafy vegetables, amaranth, groundnuts and edible insects could have high potential, particularly if the focus were on increased availability at affordable prices to the target group in the off-season. The following value chain intervention strategies serving this purpose were identified: (i) increasing efficiency in chain operation, for example through the establishment of preferred supplier relations to target groups and collective purchase of larger quantities (through Community Based Organisations (CBOs)) (ii) processing perishable foods to increase shelf life and periods of availability and (iii) focusing on the commercial production of ‘wild’ foods assuring year round supply.

Possible interventions

Since women are involved in both purchasing and producing food, the focus was on interventions that favoured building women’s capacity. Some potential intervention strategies linking poor urban women to selected food value chains include:

• Increasing purchasing power through collective action by consumers (women in particular).
• Processing perishables that are crucial in diets but are hardly available outside peak seasons.
• Identifying economic participation opportunities for women throughout the value chain.

Sale of Omena fish in Kisumu market

Seasonal fruits (mango), groundnuts and vegetables (traditional green leafy vegetables) that are largely absent from diets during parts of the year can be easily processed at household or local industrial level and made available at affordable prices in low season. Opportunities for mango processing, with a focus on women’s leadership, collective purchase and women’s organisations, could be explored together with existing CBOs. Additionally, direct supply relations between producer groups and consumers (CBOs / specific women groups), which already exist for crops like amaranth, would facilitate improved chain coordination and more efficient supply to the target group.

The opportunity matrix used to strategise interventions set-out required investments (in terms of financial support, capacity building, awareness raising, and time-frames) of possible interventions against the expected impact on increased availability/affordability of nutritious food that responds to identified deficits in the current diet. Investing in commercialisation of edible insects would best increase the level of nutrient intake and address malnutrition in the target group, but would require significant investments and a longer term perspective. By contrast, interventions in, for example, the mango value chain would be relatively easy and fast to accomplish, yet have a lesser impact on nutritional intake and the reduction of malnutrition among the target group.

Next steps

Obunga slum

The results of the research and the limitations were presented to a multi-stakeholders’ meeting in November 2012. Possible interventions were discussed with the aim to identify ways to enhance availability through increased production, processing, and/or collective purchase and potential stakeholders for the various options. In March 2013, KIT signed an agreement with Cordaid to collaborate in food and nutrition security interventions in Kisumu. In May 2013 potential stakeholders, including CBOs from Manyatta and Obunga, NGOs, local authorities and the private sector were interviewed. During a consecutive multi-stakeholder workshop, potential partners assessed the feasibility of increasing the availability at affordable prices of (processed or not) mangoes, green leafy vegetables (amaranth and sukuma wiki), insects, groundnuts and fish (included because of recent successful pilots and its high popularity and nutritious value). Workshop participants concluded that green-leafy vegetables and mango had the highest potential in terms of existing opportunities to increase availability, available resources and expertise and necessary activities. Omena and insects could be interesting but not at this stage, while (processed) peanuts were not considered an interesting opportunity given the current demand. KIT and Cordaid in close collaboration with implementing partners will translate the outcome of the workshop into fundable proposals to be submitted later this year.

For more information, contact: Albertien van der Veen, email: A.v.d.Veen@kit.nl

¹A slum is a settlement close to an urban centre (i.e. town or city) where inhabitants are characterised as having inadequate housing and basic services. A slum is often not recognised and addressed by the public as an integral or equal part of the city (UN Interagency Expert Group Meeting, Nairobi 2002). The term includes informal settlements. Both the term slum and informal settlement are used in this report.

²Using the snow-ball method for identifying consecutive key informants.

³Matrixes available in full report once finalised. See author contact details.

⁴Agency’ refers to the capacity of an individual woman to act independently and to make her own free choices (Kabeer 1999). ‘Structure’ refers to the socio-cultural environment such as rules, habits and customs.

⁵Swahili word for field. These are mostly located in the (rural) outskirts of Kisumu.

⁶Kennedy, G, Ballard, T and Dop, M. (2011). Guidelines for measuring household and individual diet diversity. FAO 2011. Available from: http://www.fao.org/docrep/014/i1983e/i1983e00.pdf

⁷Omena is consumed during the entire year and can be regarded as the protein safety net for poor households in Kisumu slums (located close to the lake).