Stunting, or low height-for-age, is the most prevalent form of child undernutrition, affecting approximately 149 million children. It impedes them from reaching both their physical and cognitive potential, with known negative impacts to future educational and professional opportunities(1). Pathologies related to stunting and loss of physical growth potential include increased morbidity and mortality, reduced neurodevelopmental and cognitive function and elevated risk of chronic diseases in adulthood(Reference De Onis and Branca2). Over the past decade, the persistence and increasing severity of global child stunting have driven research to focus on the global objective of reducing the number of stunted children (under 5 years) by 40 % by 2025.
Research has so far focused on four main areas. The first focuses on examining combinations of pre-defined determinants, including poor socio-economic conditions, household food insecurity(Reference Wondimagegn3,Reference Titaley, Ariawan and Hapsari4) , poor maternal health and nutrition(Reference Khatun, Rasheed and Alam5), constraints on women’s decision-making, educational background, antenatal care services(Reference Khatun, Rasheed and Alam5,Reference Baye, Laillou and Chitweke6) , poor water, sanitation, hygiene, insufficient or late breast-feeding, inappropriate complementary feeding practices(Reference Argaw, Hanley-Cook and De Cock7) and cultural beliefs and socio-cultural realities(Reference Titaley, Ariawan and Hapsari4,Reference Tela, Gebremariam and Beyene8) .
A second research area focuses on establishing causal links with specific negative outcomes of child stunting and linear growth retardation across five domains(Reference Leroy and Frongillo9): delayed child cognitive development(Reference Sudfeld, Charles McCoy and Danaei10); reduced physical strength and work capacity(Reference Bossavie, Alderman and Giles11); physiological changes leading to increased risks of adult chronic, non-communicable diseases and mortality(Reference Guerrant, DeBoer and Moore12); increased risk of cephalopelvic disproportions leading to dystocia, mortality and morbidity(Reference Black, Victora and Walker13) and undesirable birth outcomes (i.e. low birth weight, small-for-gestational-age infants)(Reference Addo, Stein and Fall14). Of these, however, causal links have only been established for the latter two domains(Reference Sudfeld, Charles McCoy and Danaei10). This focus on causal links is strongly rooted in nutrition research but is reported as inadequate to produce programmes reducing stunting and ensuring children’s full developmental potential(Reference Leroy and Frongillo9).
A more recent third line of research focuses on integration of ‘preventive’, ‘support-led’ and ‘growth-mediated’ strategies based on multi-sectoral and multi-factorial approaches(Reference De Onis and Branca2,Reference Leroy and Frongillo9,15) . This research explores how different community resources, capacities and strategies are fundamental to tackle child stunting, including: social care and health workers’ training and skills(Reference Permatasari, Nurika and Damayanti16), mid-level system actors leadership and networking capacities(Reference Pelletier, Gervais and Rafeez-Ur-Rehman17), individual and community knowledge, participation and commitment, institutional and organisational capacity to create and maintain partnerships with different stakeholders(Reference Lawaceng and Rahayu18), integration of resource and capital investments on childcare investments embedded in food, education, water, sanitation and hygiene, local governance and business systems(Reference Renzaho, Chitekwe and Chen19) and strategic advocacy communication(Reference Pelletier, Haider and Hajeebhoy20).
Across this considerable number of research efforts is the premise that the most effective way to make impact on the consequences of stunting is to intervene to directly improve stunting and linear growth retardation(Reference Sudfeld, Charles McCoy and Danaei10). The problem with this direct causal link approach is that it neglects pathways via non-linear, socio-cultural, socio-economic and structural processes and circumstances that contribute to stunting, thus hindering the design of predictably effective interventions.
To improve knowledge on these, recent research has turned to qualitative and mixed-method studies to explore several indirect influences of nutrition-related determinants such as gender(Reference Muraya, Jones and Berkley21,Reference Isler, Sawadogo and Harling22) , socio-cultural practices and religion(Reference Leroy and Frongillo9,Reference Burtscher and Burza23,Reference Chan, Tay and Kembhavi24) , socio-economic(Reference Abubakar, Holding and Mwangome25), structural and environmental factors such as school systems, health communication campaigns, food chains(Reference Chan, Tay and Kembhavi24), as well as local lived experiences shaping the design, uptake, implementation and efficacy of interventions(Reference Jones26). The social context and cultural meanings within a community shape the perceptions of what is considered a child’s healthy growth(Reference Mchome, Bailey and Darak27,Reference Mchome, Bailey and Kessy28) and influence the engagement in nutrition interventions, such as family relations and gendered household roles(Reference Chan, Tay and Kembhavi24,Reference Jones26) , for example, where women have little control over family finances(Reference Huang, Wu and Xue29).
We posit that instead of identifying and trying to model an increasing number of candidate factors and then linking them, a more ethnographic approach could be studying local shared values, which permeate local life. We define shared values as those things that people thing are ‘valuable, worthwhile and meaningful’ to groups. Upon these are constructed local perceptions and experiences which carry rich socio-cultural information. For example, in certain settings, moderate undernutrition is perceived not as a health problem but rather a ‘seasonal weight loss’(Reference Chan, Tay and Kembhavi24,Reference Hossain, Ickes and Rice30) ; the impact of time- and socio-economic costs of participating in interventions reduces their effectiveness(Reference Gillespie, Menon and Kennedy31) and lack of understanding of locally situated perceptions(Reference Harder, Velasco and Burford32) can cause ‘the creation of solutions that are neither meaningful nor beneficial to those in need’(Reference Burford, Hoover and Velasco33) (pp. 509).
We propose that much richer and interlinked contextual detail could be obtained if we started from an understanding of local shared values, and in particular those of groups involved in the socio-economic contexts influencing stunting such as food growing, sourcing, purchasing, preparing and eating, and child education. We therefore adopt the WeValue_InSitu (WVIS) approach for crystallising shared values of groups, immediately followed by a specialised form of focus group discussions (FGD), called ‘Perspectives EXploration’ (PEX:FGD)(Reference Huang, Wu and Xue29,Reference Sethamo, Masika and Harder34) , where the information which emerges is grounded in their own mini-culture, and the linking local logic is more apparent. Together, this WVIS_plus_PEX:FGD approach can provide rich information on cultural contextual factors which might influence stunting. Such information could contribute to more culturally acceptable stunting programme design and implementation practices, and localised, meaningful communication strategies(Reference Mchome, Bailey and Darak27,Reference Mchome, Bailey and Kessy28,Reference Gillespie, Menon and Kennedy31,Reference Pérez-Escamilla, Cavallera and Tomlinson35) .
Because this is the first time the WVIS approach is used in health research, we provide some background here on why it was chosen. WVIS provides ‘scaffolding’ to assist local people to better articulate what is ‘important’ to them, in their own way and to the extent of producing statements linked into a framework that can be comprehended by outsiders. WVIS has thus contributed to research and practice in sustainability via local indicators(Reference Burford, Hoover and Velasco33), participatory design and strategic planning(Reference Harder, Velasco and Burford32), sustainable development goals(Reference Podger, Piggot and Zahradnik36), action research for building evaluation capacity(Reference Harder and Burford37), identifying and assessing legacies of community projects or collaborations(Reference Brigstocke, Hoover, Harder, Facer and Pahl38) and collective meaning-making in community-based development contexts(Reference Moreno, Sanyal and Firoozmand39). WVIS is a scaffolding process which ensures high face validity(Reference Harder and Burford37), which means that the facilitator shows participants how to challenge and clarify their own statements and meaning rather than inserting content for uptake. For those interested in micro-processes, WVIS has been described as complex facilitations to take the group through a tacit-to-explicit translation in cycles of Polanyi-like meaning-making(Reference Odii, Huang and Des Bouvrie40).
In brief, this work aims to identify links between locally situated shared values and potential contextual factors of stunting, using an original two-step approach which combines methods of crystallisation of in-situ shared values with specialised FGD which then reveal links to those shared values.
Method
In this paper, we carry out a case study(Reference Yin41), which utilises the qualitative method WVIS(Reference Huang, Wu and Xue29,Reference Harder, Velasco and Burford32–Reference Sethamo, Masika and Harder34,Reference Podger, Piggot and Zahradnik36–Reference Odii, Huang and Des Bouvrie40) to understand local shared values.
Fieldwork and data collection were conducted in Kaffrine, Senegal, in December 2020, in the context of a much larger 5-year project on child stunting: the UKRI GCRF Action Against Stunting Hub (AASH) project. AASH aims to investigate the interrelationship of direct and indirect factors of stunting within and across three communities in which stunting is highly prevalent but for different cumulative reasons: East Lombok, Indonesia; Kaffrine, Senegal and Hyderabad, India (https://actionagainststunting.org/).
The primary unit of analysis was a local group, but with the research design that the data across all groups would be used for a secondary unit of analysis: the shared values of relevant stakeholders in the area of the cohort. Thus, if the participants in one group were too fatigued to be asked the full range of topics, then our data collection ensured that other groups covered them.
WVIS resembles a workshop format with set stages (Fig. 1 and Table 1). Preparation is not minimal: it requires pre-localisation of the materials at each site, involving preparation of bespoke ‘trigger lists’ from 4 to 8 local interviews of 20 min each, and approximately 60 h of analysis involving a local researcher. The photo materials are also localised. WVIS must be conducted in the local language and thus requires close collaboration between experienced facilitators and local researchers.
Fig. 1 Schematic overview of the macro-level activities carried out during the WVIS workshop session
Table 1 The stages of the WVIS_plus_PEX approach
The participant groups were strategically recruited as naturally existing local groups of people whose members have some history of shared experiences (a requirement of the WVIS approach); who live near the main cohort site of the AASH (so that they have relevant shared values) and with some link to the care environment of young children or their food supply (so their FGD would be relevant to the research topic of stunting). Recruitment was conducted by AASH Senegal researchers, resulting in eleven stakeholder groups including mothers, fathers, pre-school teachers, community health workers (CHW), farmers, market traders, farmers and public administration officials (PAO) (see Table 2). All the participants were indigenous to the area. There is no optimal group size other than more than one person and not so many that there is not enough time for every individual to share, that is, 2–18 persons. However, more time is needed for more participants, for example, 2–4 h. To keep the time requested for the participants low, the group sizes were kept small (typically 4 but maximum 12: see Table 2). The questions in the FGD were tailored to the relevance to stunting of each group type: samples are given in Table 2. Ideally, at least two groups of each type would have been recruited so that an indication of variability of results between them could be examined. Recruitment could have been continued until saturation of thematic results was achieved for each group type. However, this work took place during COVID, and local researchers indicated we should proceed with a minimum number of groups. A previous study using the WVIS approach found 6–8 groups each with 2–5 participants sufficient for theoretical saturation of a specific topic(Reference Huang, Wu and Xue29).
Table 2 Participant group types and sample questions for the WVIS_plus_PEX focus group discussions in Senegal
The eleven WVIS workshops and PEX:FGD were facilitated by local Senegal AASH researchers, directed by University of Brighton AASH researchers (one on the ground and others online, live), in the local language of Wolof. One of the local researchers was indigenous to Kaffrine: the others were indigenous to elsewhere in Senegal. A live translation into French was audio-recorded alongside the Wolof and later translated into English for initial analysis, along with researchers’ observation notes. Informed consent was obtained from all participants.
Data analysis
When WVIS_plus_PEX:FGD is typically used to elicit cultural contextual information(Reference Huang, Wu and Xue29,Reference Sethamo, Masika and Harder34) , only the summative data outputted by the WVIS are used alongside the PEX:FGD data, that is, the framework of statements of shared values which the group produces as an output (last part of Fig. 1), because they encompass the discussion material. However, in this case, it was found that those statements were abnormally brief, and it was decided to refer to their underpinning discussions within the WVIS where rich, thick descriptions had been made about the group’s cultural context. Thus, transcriptions of both WVIS and PEX:FGD full discussions from the eleven groups were analysed.
The data analysis involved three main steps in a grounded approach(Reference Heath and Cowley42). First, the transcripts of each group’s WVIS workshop and their PEX:FGD discussions were thematically open-coded for emerging themes, and those themes (with their associated extracts alongside) from all eleven groups were then clustered. Two researchers did this independently in English and compared findings (a local researcher checked findings for integrity of meaning by back-checking with the pre-English versions).
Findings
Across all the participant groups, twelve themes of contextual factors which might influence stunting emerged, presented below. Table 3 provides illustrative quotes. The relative relevance found to different stakeholder group types is indicated schematically in Figure 2.
Table 3 Illustrative quotes from participant groups with local contextual factors of child stunting
Fig. 2 Participant stakeholders’ perceived factors of importance in relation to Child Stunting and Child Overall Development
Religion and cultural traditions
For parents, grandmothers, teachers and farmers, religion and the maintenance of cultural traditions were core to children’s education, to develop a culture and moral life path of respect, solidarity and cooperation within the community.
For mothers and grandmothers, religion had an important role in nutrition, concerning recommended and prohibited foods, breast-feeding and seeking antenatal care.
Farmers mentioned diversifying food crops could ensure more secure livelihoods: ‘we would like to be able to cultivate tobacco but […] religion oppose it[.]’ (Table 3). They mentioned praying as a strategy to deal with bad climate and pests.
CHW saw the observance of certain religious practices and beliefs as detrimental towards nutrition in pre-natal and post-natal diets. They thought the stigma around close pregnancies and religious and cultural practices also discouraged women from seeking professional help and advice during the early pregnancy; ‘[…] simply superstitious reasons […]lead many women to not want to show they are pregnant […]’ (Table 3).
Gender (roles, responsibilities and expectations)
Gender was mentioned relating to food planning and nutritional strategies, access to and use of healthcare services during pregnancy and for educational and income-generating opportunities.
Responses from parents, grandmothers and farmers indicated a clear gender-biased division of labour that extended beyond home environments. Mothers and grandmothers mentioned that they took charge of meal planning, housework and childcare, but their management choices were constrained due to lack of control over household finances; irregular income of their husbands and the lack of opportunities to access jobs.
Fathers affirmed that women oversaw housework and food planning, and household management spending. Some fathers suggested the work of mothers needed to be further recognised; others talked about women’s role in society as ‘reproductive function’ activities v. men’s ‘production’ activities.
PAO referred to the difficulties for women and especially young girls to attain higher educational levels due to socio-cultural pressures for marriage and early pregnancies.
CHW emphasised that women had limited or no control over household finances and their difficulties in accessing jobs seriously affected their autonomy and capacity to access health care. This was compounded by women’s limited autonomy to seek professional advice during the early stages of pregnancy, depending on the consent and predisposition of their husbands and/or extended families.
Health care
Fathers were concerned with access and quality of health care and hospital births. Mothers and grandmothers mentioned hospital check-ups and following doctors’ advice as requisites for a ‘good pregnancy’: they also said they often could not afford it. Teachers emphasised the importance of the health services to carry out regular trimestral checks on children and give nutrition advice.
PAO emphasised the importance of hospital births not only for health and hygiene reasons but also for administrative purposes since hospital registration produces certificates necessary for subsequent access to government aid education.
For CHW, women’s limited independent decision-making powers and health care access during early stages of pregnancies had important consequences for their family planning decisions, nutritional knowledge and childcare capacity. Inadequate pregnancy care and early childcare practices were further hindered by lack of knowledge and awareness about child stunting, which itself sometimes led to negative attitudes and stereotypes not only towards stunted children but also community doctors and health workers, especially female, who were sometimes blamed for pregnancy complications.
Food systems and dietary practices
Food systems and dietary practices were relevant to most stakeholders, shaped by several converging factors.
For mothers, food planning and nutritional strategies were based on day-to-day decisions like food prices, household budget and storage conditions (e.g. owning a fridge to keep foods fresh) which often superseded nutritional and hygiene considerations. Other influences were religions recommendations for food types and breast-feeding.
Parents and grandmothers gave conflicting comments on whether eggs were good or bad for pregnant women and/or babies: some believed it could impede a baby’s speech development and some did not. They were open to interventions providing eggs but disagreed on whether the eggs would be eaten daily or stored instead.
Teachers mentioned concerns over food affordability and hygiene at schools, saying most parents chose to cook children’s lunches at home.
Farmers worried about agriculture becoming a less-secure livelihood. Income diversification was influenced by religious observance (not to cultivate certain crops like tobacco and using prayers for rain and pests) and geographical and climatic constrains (limited fertile soil and lack of water to cultivate watermelons or cotton). Farmers said difficulties to access loans for equipment, storage facilities and fertilisers, restricted their farming choices.
Market traders referred to the poor quality of fish they could acquire to sell in Kaffrine’s market, which could hinder children’s health and cognitive development, saying that better fish went to bigger/richer cities.
CHW mentioned that religious and cultural beliefs around food practices could influence the nutritional quality of household diets strongly, including breast-feeding and food diversification. They echoed farmers’ and market traders’ concerns on food supply chain constraints, competitive markets, food storage limitations and climatic and seasonal factors.
Water, sanitation and hygiene
For parents and grandmothers, access to water was a main concern for the choice and preparation of foods, and hygiene.
Water scarcity was rated important by farmers for cultivation of choice crops like watermelon, cotton and for sanitation for market traders to evacuate food waste to drains. CHW commented that insufficient access to water impacted on nutritional strategies and health, including stunting.
Socio-economic factors and natural ecosystems
For parents and grandmothers, livelihood insecurity, access to nutritious foods at affordable prices in local markets, limited transport and infrastructure to access water, limited or distant food markets and lack of fridges/storage to preserve food were all mentioned.
For teachers, work conditions influenced their role in early education of local children, motivational (remuneration), capability (facilities and resources) and capacity (pedagogical and professional training) terms. PAO considered poverty as the main cause of child stunting and malnutrition, saying families simply did not have enough income and resources.
Farmers stated that agriculture had become an insecure livelihood due to bad climate, market speculation, water scarcity, infertile soils, lack of and difficulty to access government loans and insufficient means to pay a workforce. This meant that their families, including children, worked in the fields to support the household. Market traders said that their low purchasing power and bureaucracy prevented them from improving and diversifying their livelihoods.
Learning opportunities and early education
As stunting is linked to reduced learning, local perceptions on learning are relevant. Most stakeholders indicated early learning opportunities as important for overall child development.
While for mothers the emphasis was access to (any) good education, fathers specifically considered education for its value for future professional opportunities and livelihood security to permit children to ‘do better than their parents’.
Farmers mentioned that children’s contribution to society required an education that was comprehensive and including Quranic/Muslim values for a culture of morality and dignity, and the technical elements of French schools, for better professional opportunities.
Teachers mentioned difficulties for some families to afford early education. They emphasised the importance of parental relationships and deep engagement with their children prior to joining formal education for their development, as was ongoing communication between teachers and parents. Teachers said a varied syllabus and wide range of educational resources and materials was key to nurturing children’s cognitive and psychosocial development and emphasised the need for better and repeated pedagogical and relevant professional training for their motivation and capabilities as teachers.
PAO linked low educational attainment of girls and women to gender and socio-cultural pressures and stated teenage extra-marital pregnancies caused school dropouts.
Family and home environment
The immediate and extended family environments were emphasised by many groups.
For mothers, the focus was making sure that parents and grandparents were actively involved in enhancing child development, getting children interested in learning and participating in daily chores.
The idea of unity within the household – ‘togetherness’ – was stressed by fathers and grandmothers. Fathers mentioned negative impact of divorce, and associated risks of children being neglected. Grandmothers described their roles including supporting their children’s families financially and in-kind; caring for grandchildren and sharing their knowledge and experience. They explained how developing ‘relations capital’ (networking) was very important to create opportunities of support for their extended families.
Both parents and grandmothers considered pre-marital and early-age pregnancies negatively, as they affected families’ dignity, carried financial constraints and increased difficulties to register children at birth – a prerequisite for accessing education and health care.
Psychosocial stimulation
For fathers, learning early how to pray, carry out ablutions and accompanying parents to the Mosque were important for moral and social development, while mothers said that having their children learn to read the Quran developed language comprehension and cognitive skills. Playing, storytelling, role-playing and sports were also all considered important for development.
Teachers emphasised the importance of religious education for psychosocial stimulation and to help develop harmonious and cooperative community relations. Teachers and mothers mentioned activities with neighbours, such as cleaning of common places or sharing meals, for children’s social development. CHW said sharing meals was sometimes a strategy for slightly better-off families to help others and/or to stimulate children to eat better.
Community interactions (collaboration, solidarity, support, conflict)
Although considered an indirect impact on child stunting, respect, solidarity and collaboration within the community were mentioned by all stakeholders.
For parents, the process of collaboration between members of the community involved a sense of belonging, ‘togetherness’ and ‘peacefulness’, especially when nurtured through collective actions (taking care of other families’ children, sharing meals, cleaning neighbourhood common areas, traditional celebrations).
Mothers and grandmothers added that it was through community meetings and mothers’ groups that spaces and opportunities for sharing practical knowledge and advice about parenting and childcare were created. For mothers, farmers and market traders, such spaces created a sense of not only belonging but also mutual support and solidarity. This included sharing meals and helping each other out financially or in-kind when needed. For farmers and market traders, it was a strategy to strengthen their livelihoods during bad seasons, or when resources were scarce.
Teachers emphasised child involvement in community activities, celebrations and collective collaborations, to help them become ‘exemplary citizens’.
For CHW and PAO, good community relations depended on the development of sustainable cross-sectoral services for better provision of health care, education and local governance, and their role as mediators was to ensure these collaborations happened.
Local governance and institutional frameworks
Local governance, civic administrative and institutional frameworks were mentioned for important, but more-indirect, impacts.
PAO highlighted the importance of registering births in hospitals, to facilitate later formal access to essential public services. Fathers and farmers said it was a prerequisite for their children to access schools. Fathers added that it facilitated mothers’ access to childcare and nutrition training.
For farmers and market traders, joining formal microcredit unions gave them access to governmental aid, and thus support to strengthen and diversify livelihoods, via storage facilities and loans. For market traders, especially women, formalising their trade increased their chances of accessing financial business support.
Teachers considered local government responsible for, but not delivering on, improvement and development of much-needed school resources, working conditions and relevant pedagogical and professional training, which they felt influenced their role in maximising children’s development potentials.
CHW and teachers emphasised the importance of good leadership and efficient local governance through cross-sectoral collaboration and better communication between mid-level actors to draw resources, raise awareness, share knowledge and implement health interventions aimed at reducing infant and maternal mortality rates.
Perceptions on child stunting
In addition to the above contextual factors which emerged from more open and broad questions, our analysis also drew out some directly articulated local perceptions of stunting, which we summarise below.
Among parents, and teachers and grandmothers, child stunting was mostly perceived as a health problem impeding a child’s physical and cognitive development and was primarily recognised through physical features (short stature, low weight, hair loss) and attitudinal and behavioural signs (fatigue, lack of concentration, bad temper, unsocial behaviour, lower school performance). Teachers mentioned ‘child health check cards’ provided by the local health departments to measure the children and twice-a-year formal visits from health workers.
For grandmothers, stunting was considered due to lack of hygiene; spiritual influences during pregnancy; possibly hereditary and/or food preparation practices such as leaving food uncovered where spirits could enter. Teachers considered the family context to be relevant, including broken marriages and parental attitudes towards such as overprotection or dismissal of the issue of the stunting, due to lack of knowledge.
Market traders talked about the poor quality of fish sold in their market and the lack of hygiene and water facilities to drain out waste, as potential causes for children’s health problems: they said poor quality fish caused intestinal worms and food poisoning and hampered children’s intelligence.
PAO mentioned poverty and lack of income opportunities for families, leading to only affording low food nutritional quality, as main causes of child stunting.
Some mothers expressed not fully understanding the concept of stunting. Any existing knowledge and strategies originated from mothers’ meetings and extended family circles. PAO mentioned that, to their knowledge, there was no specific government plan or strategy on child stunting or related nutrition. CHW described a wide range of factors of stunting and undernutrition including poor diets, disease, poor hygiene, hereditary factors, socio-cultural beliefs and practices around food, non-compliance with exclusive breast-feeding and premature pregnancies. Eating of prohibited foods and/or neglecting correct foods were perceived to lead to ‘good’ or ‘bad’ pregnancies, affecting whether the babies would be stunted. CHW also mentioned that it was difficult to diagnose stunting and malnutrition cases as people did not come regularly for check-ups.
Discussion
Our findings make contributions to two complementary strands of child stunting research: those highlighting the need for more relevant determinants and outcomes of stunting(Reference Leroy and Frongillo9), and those calling for more multidisciplinary approaches that take into consideration not only nutrition-specific determinants but also cross-sectoral factors that underlie nutritional challenges(Reference De Onis and Branca2).
First, while they remain exploratory in nature and limited geographically and culturally to the context of Kaffrine, Senegal, our findings provide direct insights of local culturally contextual factors which could influence child stunting and development, drawn from discussions grounded in the shared values of local community groups with roles relevant to child upbringing, education and care (mothers, fathers, grandmothers and pre-school teachers), as well as health, local governance and the food chain (CHW, farmers, market traders and PAO).
These contextual factors included unawareness of stunting as a problem, lack of information about nutrition (e.g. limited community nutrition training programmes, lack of school resources and specialised pedagogical training of teachers); religious and spiritual beliefs and cultural traditions concerning certain food eating and growing practices; women’s lack of autonomy to access healthcare services during pregnancy, school education dropouts and limited financial and income opportunities; male-dominated authority and decision-making within the household pertaining food budgeting and education; lack of trust and stigma towards health workers for pregnancy; socio-economic factors affecting children’s food environment (e.g. school meals, household irregular income, price of quality nutritious foods, lack of fridges to keep foods fresh, water, sanitation and hygiene conditions in markets and homes) and structural factors affecting the food chain (e.g. availability and quality of foods sold at markets, climate change, access to water and soil fertility to grow desired crops, storage facilities, access to micro-loans to support small merchant and farm businesses). Pre-knowledge of these will allow the design of more locally acceptable and effective interventions in Kaffrine and flag new themes of contextual factors for consideration at other sites.
Secondly, the approach we have used provides links between these candidate factors and the complexity of daily life: contextual links. In the introduction, we outlined a large body of child stunting research, which tends to be persistently framed in terms of short-, medium- and long-term negative outcomes(Reference Leroy and Frongillo9), investigating potentially causal links between linear growth retardation and child stunting, and academic sourced factors from discipline-specific knowledge bases. This framing is problematic not only because it is insufficient to ensure overall child development and/or nutrition, but also because it implies that stunting interventions are failures even if they can significantly improve children’s situation(Reference Leroy and Frongillo9). Furthermore, this focus on linear growth retardation and stunting has meant that efforts are directed towards the first 2 years of life window, when growth failure is most rapid, and interventions are most likely to have an impact(Reference Leroy and Frongillo9), to the extent of neglecting related aspects of nutrition, health and overall child cognitive, physical and psychosocial development which extend well beyond that first 2-year window, into adolescence(Reference Leroy and Frongillo9).
Our examination of local perceptions of stunting and its contextual factors is also of relevance to larger existing frameworks, such as WHO’s 2013 conceptual framework Context, Causes and Consequences of Childhood Stunting (‘WHO Stunting Framework’)(Reference Stewart, Iannotti and Dewey43), as it helps identify missing interlinking factors that are directly relevant to local communities and thus could be used to create more specific indicators. For example, while the WHO Stunting Framework includes indicators for ‘Home Environment’ under its ‘Household and Family Factors’ section, it does not provide specific factors that have been reported to affect a child’s birth and/or growth development such as family members relationships, gender norms or household socio-economic characteristics. While some of these missing factors have already been reported(Reference Wirth, Rohner and Petry44), our value-based approach provides detailed understanding of locally important contextual factors not previously considered.
As Leroy and Frongillo(Reference Leroy and Frongillo9) (pp. 202) point out, a major pragmatic challenge is that most interventions fail to pinpoint and then address factors of child stunting or the type of interventions needed. The integration of multiple types of interventions spanning nutrition and other underlying contextual causes is needed(Reference Leroy and Frongillo9). This paper demonstrates that the WVIS approach is effective and useful for crystallising/creating local knowledge, allowing us to identify and articulate specific practices, motivating factors and strategies which are relevant to local community stakeholders in the context of child stunting and undernutrition. It is a highly localisable and culturally adaptable approach(Reference Huang, Wu and Xue29,Reference Harder, Velasco and Burford32–Reference Sethamo, Masika and Harder34,Reference Podger, Piggot and Zahradnik36–Reference Odii, Huang and Des Bouvrie40) . WVIS produces statements of core shared values of the groups, which provide a weft onto which they can weave a communicable context-rich picture of their perceptions of stunting. This picture is inherently multi-factorial: concepts are framed in local values, transcending disciplinary boundaries. But they can be interrogated subsequently in terms of disciplinary frameworks.
This work thus demonstrates a holistic approach (WVIS_plus_PEX:FGD) to generating grounded, multi-factorial and multi-sectoral determinants of stunting, and reveals their pathways within local contexts. Application at other sites will allow comparisons and patterns of such contextual factors to be built up, contributing to an understanding of factors of stunting in those sites. The AASH project presumes that there is not one cause nor one type of stunting but a typology with different active factors in different contexts, and it is collecting a wide range of objective data – genetic, biome, gut health, diet, nutrition, food environment, food systems, child cognition and water, sanitation and hygiene – across Indonesia, Senegal and India sites, to try develop the specifics of a typology. The WVIS_plus_PEX:FGD shared values data from each site will be used to help understand possible cultural linkages between those objective data, and to inform intervention planning.
In the introduction, we drew out the trends in current research in child stunting of moving from studies of single and direct factors to consider multi-factorial issues. When we designed and wrote this study, we envisaged the shared values of local groups to be mini-cultures which, if probed appropriately, could provide not only more complete and authentic contextual factors but also an underlying understanding of why they existed in that place. We did not have anthropological training but came from a background of developing authentic indicators for local sustainability(Reference Burford, Hoover and Velasco33,Reference Podger, Piggot and Zahradnik36) , co-developed over years of action research(Reference Harder and Burford37). But reviewer feedback pointed us to related anthropological work, and we now realise that the WVIS_plus_PEX:FGD approach has much in common with the established anthropological concept of cultural schema(Reference D’Andrade, Shweder and LeVine45), the methods of focused ethnographic studies(Reference Pelto, Gove, Scrimshaw and Gleason46) and community-based participatory research(Reference Wilson, Kenny and Dickson-Swift47) and that these have been applied in nutrition studies already(Reference Mchome, Bailey and Darak27,Reference Mchome, Bailey and Kessy28,Reference Pelto, Armar-Klemesu and Siekmann48) .
Both WVIS and focused ethnographic study retain the characteristic of traditional ethnography in their intent to obtain the emic view within communities(Reference Yin41,Reference Pelto, Armar-Klemesu and Siekmann48) , while aiming to reduce the time taken to attain cultural data compared with traditional ethnography(Reference Pelto, Gove, Scrimshaw and Gleason46,Reference Pelto, Armar-Klemesu and Siekmann48) . However, they do this in diverse ways. Focused ethnographic study focuses the questions on the topic under study(Reference Pelto, Armar-Klemesu and Siekmann48) such as an intervention or policy design and planning(Reference Pelto and Armar-Klemesu49). WVIS uses the mechanisms within the highly reflective and introspective workshops sessions which involve tacit-to-explicit translations(Reference Odii, Huang and Des Bouvrie40), to facilitate deep discussions about shared experiences and articulate them into statements reflecting what is ‘important’ to that group (emic). Only after this is a FGD used to elicit information on a specific topic. WVIS can thus be used to address a broader range of research questions. Furthermore, the cycles of meaning-making and introspection involved facilitate deep, authentic participation(Reference Sethamo, Masika and Harder34), which reduces the risk of tokenistic or ‘tick box’ engagement that is common pitfall in community-based participatory research(Reference Pelto, Armar-Klemesu and Siekmann48).
Future work will include investigating in detail the overlaps between the approaches. We note that the WVIS_plus_PEX approach includes the notion that each type of stakeholder group has their own mini-culture, which is particularly effective at providing a boundary of contextualisation which can facilitate the articulation of information which is intricately linked to the shared values of that mini-culture. If true, this would be an interesting development for cultural schema, and the short time needed for WVIS_plus_PEX of 2–4 h might prove it a useful accelerated quasi-anthropological method: this can be investigated in future work.
Acknowledgements
Acknowledgements: We would like to thank Claire Heffernan for a final editing. Authorship: M.K.H. and C.H. formulated the initial research question and study design. J.M.M. developed the specific research question. Localisation of the study design involved M.K.H., N.M.S., A.H.D., A.J.C. and C.C.E. Data collection involved C.C.E., N.M.S., A.H.D., F.B.D., R.N.T., J.M., C.N. and M.F. and cultural interpretation involved A.H.D., F.B.D., N.M.S., R.N.T., J.M., C.N. and M.F. Analysis involved J.M.M. and A.J.C., and J.M.M. and M.K.H. wrote the paper. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved by the University of Brighton ethics committee. Written informed consent was obtained from all subjects.
Conflict of interest:
There are no conflicts of interest.
Financial support:
UKRI Global Challenges Research Fund, grant no.: MR/S01313X/1.
Open access
