Undernutrition risk factors and their interplay with nutrition outcomes: nutrition causal analysis pilot in Kenya

By Kristy Manners, Muriel Calo, Imelda Awino and Jacob Korir

Kristy Manners specialises in Nutrition and Public Health and has worked extensively in the horn of Africa. She is currently working for ICRC Somalia as the Nutrition Manager. Previously she worked with WHO in Albania, Department of Health in Australia and with local organisations in Cambodia.

Muriel Calo was Senior Food Security and Livelihoods Advisor in ACF’s New York office at the time of writing. She has a decade of domestic and international experience working on food security and livelihood issues in vulnerable settings, supporting ACF-USA country programmes in East Africa and Asia. 

Imelda Awino is the Regional SMART Coordinator, former Head of the Nutrition Department in ACF-USA Kenya Mission.. She has previously worked as nutrition surveillance programme manager and nutritionist in ACF Kenya and South Sudan Missions, and for the Kenya Ministry of Health and GIZ in various nutrition related capacities. 

Jacob Korir is Deputy Head of the Nutrition Department in ACF-USA Kenya Mission and has over 5 years’ experience in nutrition and health programming. He previously worked for World Vision Somalia, Kenya Medical Research Institute and Christian Child Care International in various health and nutrition capacities. 

This study was undertaken with financial assistance from UNICEF. The views expressed herein should not be taken, in any way, to reflect the official opinion of the United Nations Children’s Fund. This study is made possible by the generous support, time and energy of the communities in Isiolo county, Kenya national and county level stakeholders, donor agencies, Action Against Hunger Kenya Mission staff, and the NCA team. Julien Chalimbaud (Head of the NCA Research Project) and various HQ technical advisors provided technical guidance through the course of this study.  Cecile Basquin provided peer review on an earlier draft of this article.

This article represents the views of its authors only, and not ACF as a whole.

Location: Kenya

What we know: Acute and chronic malnutrition rates remains prevalent in Isiolo County, Kenya despite international and national initiatives. Understanding contributing factors to undernutrition and their interplay in a given context can inform advocacy and help direct interventions.

What this article adds: ACF conducted a Nutrition Causal Analysis in Isiolo County, Kenya using a new participatory methodology. All stakeholders concurred on major causal factors underpinning acute malnutrition (high child morbidity, poor access to safe household water and to appropriate age-specific foods). Increased women’s workload was an important consequence of drought effects - and sometimes of interventions - to the detriment of child and maternal health. Communities suggested insightful solutions to address risks.

Isiolo County lies in the semi-arid and arid lands (ASAL) of Kenya. Action Against Hunger |ACF International (ACF) has been operating in most parts of Isiolo County (Garbatulla and Merti Sub-Counties) for a number of years. Programmatic adjustments by government and non-governmental organisations (NGOs) following on from the findings of regular quantitative surveys and surveillance data have contributed to decreased acute malnutrition prevalence below the  Horn of Africa pre-crisis level. Adjustments have included strengthening of local and national capacities (Ministry of Health, National Drought Management Authority); scaling up health and nutrition outreach to ensure that services reach remote areas not covered by facilities; scaling up Maternal, Infant and Young Child Nutrition interventions; linking beneficiaries to Food Security and Livelihoods interventions and creating synergies with WASH (waster, sanitation and hygiene) interventions; strengthening nutrition and food security information systems to prompt timely humanitarian response; on the job training; mentorship and advocacy. However, prevalence of acute and chronic malnutrition has since plateaued at around 10% and 20% respectively¹. There was therefore great interest from stakeholders to understand further the causal pathways of malnutrition so as to be able to provide recommendations to county stakeholders and develop more effective programmes to curb malnutrition.

Nutrition Causal Analysis (NCA) approach and method

The core objective of a Nutrition Causal Analysis (NCA) study is to understand the risk factors of undernutrition and their interplay with nutrition outcomes for a given study population. The generic method and history of this approach is described in Box 1 of the accompanying article on Bangladesh (see page 74).  

ACF conducted an NCA study in Isiolo County, Kenya in late 2013 using this newly developed methodology. The NCA study had the objectives of: i) identifying the main causes of wasting in Isiolo County; ii) understanding the local seasonal and historical pathways to wasting; iii) developing local causal models for acute malnutrition²; and iv) using these results to support advocacy on causes of wasting. 

The qualitative inquiry approach to primary data collection was used to complement already existing sources of information and to develop an ‘emic’ (local) definition and understanding of undernutrition. It was also used to characterize food security, health and care practice in the community and explore local perceptions of the causes of poor food security, heath and care, as well as identify seasonal and historical trends and understand how the community prioritises risk factors related to undernutrition. 

Adaptation of the ACF NCA survey methodology to the Isiolo County context involved a participatory process with community, county and national stakeholders including government ministries. Methods included secondary data review, key informant interviews, focus group discussions and an iterative analytical process based on consultations with key stakeholders. Key steps included:

1. Identifying risk factors and establishing causal ‘pathways of malnourishment’ for acute malnutrition/wasting through a literature review; formulation and preliminary rating of factors during an initial technical expert workshop. 
2. Gathering evidence of causality through a qualitative inquiry conducted during four weeks of intensive fieldwork, where communities ranked priority risk factors of acute malnutrition. 
3. Rating causal factors according to their relative contribution to undernutrition in a final stakeholder workshop, based on literature review, international scientific sources, quantitative survey results, analysis of seasonality; ranking by communities and technical experts. 
4. Validating results through a participatory and consensus-driven process where initial hypothesised risk factors were revised and validated in accordance with qualitative inquiry results and confidence notes³ delivered. 

The sample was stratified into Pastoral, Agropastoral and Labour/charcoal livelihood groups so results could be disaggregated by population groups with potentially different sets of underlying causality for undernutrition. Four villages or study sites were selected in the three Sub-Counties of Isiolo County (Isiolo, Garbatulla and Merti). 

Key findings

Findings showed that high child morbidity, linked to inadequate access to safe water for household use coupled with poor access to appropriate age-specific foods, including milk in the dry season, are the three major causal factors underpinning acute malnutrition in Isiolo County. This was agreed by communities, technical experts and other critical stakeholders and is reflected in the high confidence of stakeholders in these three major risk factors identified through the NCA process. Findings also confirmed that the impacts of recurrent drought are increasing women’s workload in significant and detrimental ways and in turn affecting maternal health and care of young children. This was reflected across the other major causal factors identified in the study process.

While the three livelihood groups under study have unique features affecting the health and nutrition of their community, many of the risk factors to acute malnutrition that have been prioritised are very similar across the groups.  Findings from the fieldwork indicated that acute malnutrition typically occurs during the dry season or during droughts, peaking at the commencement of rains, and is linked to seasonal reductions in access to and consumption of milk. However, with increasing recurrence of droughts, increasingly sporadic rainfall and unreliable climatic patterns, communities are having less time to recover during these “typical” peaks. As a result, other risk factors are becoming more important than just the availability of milk during the dry seasons. For example, as water shortages are occurring more frequently, disease and women’s workload are increasing. The community raised concerns that the health of their animals and families are not as strong as previously. They attribute it to recurrent droughts and the cycle of poverty as they cannot recover the number of animals to sustain them as before. Thus, though wasting is considered by technical experts to be the major area of concern as regards undernutrition in Isiolo County, stunting is also emerging as an area of concern to experts as well as the community, as recurrent shocks lead to chronic afflictions and long term problems. 

Table 1 summarises the main results, presenting the twelve major risk factors identified in the NCA study. It illustrates preliminary ratings given to individual risk factors in the initial technical expert workshop⁴, classification of risk factors into categories based on field investigation findings (major/important/minor/untested/rejected) and final confidence notes assigned by stakeholders in the last stage of the process⁵. 

* 1= minor contributor to undernutrition, 5= major contributor to undernutrition. 

** major, important, minor, rejected, untested  *** 1=low, 3=high

Conclusions

The agreed major factors underlying acute malnutrition reflect the impact that recurrent drought is having on the economic capacity of communities (loss of livestock, vulnerability to food price hikes, reduced access to food and markets), as well as on access to water (loss of livestock, splitting of families for migration, increased workload of women, hygiene and sanitation, disease). These factors, underpinned by women’s increasing workload, are having a detrimental effect on the care of U5 children and maternal health. 

Ongoing social transformation is resulting in families placing increased value on education compared to other generations. However, impact on households is mixed. On a positive note, literacy levels and knowledge of nutrition, health and hygiene are increasing and girls are being kept longer in school to help prevent early pregnancies. However, families split as men migrate, due to the increased tendency of communities to settle in order to keep their children in school (as a result of low numbers of mobile schools), and to counteract the effects of livestock depletion from recurrent droughts through diversification of livelihoods (agriculture, labour/charcoal). Splitting of families has, in turn, led to increased workload of mothers as they work in the shamba (homestead) or burn charcoal, and longer distances travelled for pasture and water for livestock; consequently there is more limited access to milk and meat for children and women.

Ultimately, the workload of women is recognized by both men and women in communities to have increased dramatically: working in the shamba or burning charcoal, collecting water, conducting normal domestic duties, and caring for children. With inadequate time to care and feed children properly, and with reduced access to milk for children, poor nutritional status of children is compounded.

Recommendations

The UNICEF Conceptual Framework illustrates that the causes of malnutrition are multi-factorial and that all three levels (basic, underlying and immediate causes) are vital to ensuring a healthy outcome for mother and child. Depending on the mandate of an organisation or a government agency, programming often aims to address the immediate and underlying causes and assumes that the impact of the programme towards reducing undernutrition has directly been achieved by their single-pronged approach. This approach may overlook the vital needs for, and role of, advocacy for policy change, infrastructure support and provision of basic services. 

Through the field study of this NCA, it was evident that often what are believed to be appropriate interventions, for example, increasing food security or educational knowledge, may in fact be affecting the care of a child in a detrimental way, e.g. increased maternal workload, dividing families to keep children in school, etc. Although household and community food security may be improved on one level, the workload of a woman may be increased through this strategy. Programming and strategies often neglect the importance of the workload of women who bear much of the family responsibilities, and rarely if at all, make the reduction of women’s workload an explicit objective.

Communities proposed a number of insightful solutions to address the risk factors for acute malnutrition that were shared with county level actors and included in the report⁶. ACF also developed a detailed county plan which is being used as a key advocacy tool and the county is keen for the plan to be accomplished. However, there is a clear need better to streamline this process so that there can be a joint effort towards identification of solutions rather than two parallel processes that may not always align and therefore translate into a clear and coherent action plan. A response analysis module has been added into the guidelines to support follow-up.  Many of the community solutions required external input; this is where investment of time would be valuable to come up with solutions that involve local resources, organisations, mobilisation, etc. Interestingly, the village site that had never received any outside aid came up with the most solutions and which involved minimal external assistance.  

With limited budgets, it is even more vital to involve the community from the outset in any decision-making. For this reason, to follow through from the information sharing at community level and the wealth of solutions proposed by communities, a Participatory Vulnerability and Capacity Analysis (PVCA) would be appropriate whereby communities design their own plans for longer term disaster risk reduction, resilience and nutrition security for integration into other local and county level planning processes. The NCA also provides the platform to advocate to the government and agencies to collaborate further with these communities in supporting them in this process. 

In line with key findings of the study, an Isiolo County NCA Dissemination Workshop held in January 2014 provided a forum for multi-sectorial groups at county level to enrich proposed community solutions with key recommendations around reducing the workload of women and minimising the splitting of families. 

Lessons learned

In addition to gathering evidence of causality and identifying risk factors for malnutrition, the NCA process was also used as a learning platform to improve future NCAs. Below is a synopsis of key lessons learned:

• Involve government ministries and other key stakeholders. Government ministries and other stakeholders from a range of technical areas should be fully engaged in every step of the process to ensure a thorough understanding of the process and a capacity to follow-up on findings. In this NCA, government was involved; However, ACF would seek more representation from other sectors outside nutrition in future NCAs.
• Start from the bottom up to minimise divergence between expert conclusions and community priorities. While major risk factors identified by both community and experts aligned fairly well, there is a need to engage communities early in the identification of their priorities, risks and solutions to avoid divergence with priorities set by donors, governments and agencies. This approach will allow for proactive agreement on final causal models that reflect both local priorities and expert opinion. 
• Explore the possibility of using the NCA to examine causal factors driving undernutrition as a whole. Opportunities to compare NCA outcomes in similar contexts that focus on just wasting or just stunting should be seized to understand whether there is continuity or coherence between the risk factors driving different kinds of undernutrition. A new review of evidence⁷ suggests that the current programmatic and policy divide between wasting and stunting may be losing its relevance and, ultimately, contributing to the lack of nutritional impact seen in responses that address only one part of the undernutrition problem. The NCA can be used as a tool to gain a better understanding of the relationship between the factors driving both wasting and stunting. 
• Limit the number of risk factors under study. The number of risk factors affects the time required in workshops, length of primary data collection, amount of data to analyse, report length and time required for presentation. Selection of a limited number of key risk factors would allow greater depth of understanding on the most important determinants of undernutrition.
• Use the formative research approach of the qualitative inquiry to link other sources of nutrition information available at local and national levels with local priorities. While quantitative survey information is valuable, it only paints half of the picture. Formative research can support the development of a culturally sensitive definition of undernutrition and can be the basis for elaborating more effective behaviour change strategies around key risk factors, for example. This ensures accountability of expert opinion and action, as well as improved ownership and effectiveness of change processes at every level.
• Enhance uptake of recommendations through community dialogue and stakeholder engagement. While the primary objective of the NCA is the analysis of causes, the study should also be paired with a discussion of solutions and analysis of response options so as to leave communities with some direction and motivation regarding how they can address locally identified problems. This response analysis can also lead into a community mobilisation and planning process at a later stage. Similarly, other key actors at local and national level need to be guided in next steps to enhance the uptake and impact of NCA findings.

For more information, contact: Cécile Basquin, ACF USA Nutrition Technical Advisor, email:  cbasquin@actionagainsthunger.org and tel: +1 212 967 7800

The full length report on which this article is based is available at: http://www.actionagainsthunger.org/publication/2014/02/kenya-nutrition-causal-analysis-qualitative-inquiry 

¹ACF Integrated Surveillance report for Garbatulla and Merti sub-counties, October 2013

²A local causal model is a set of interlinked risk factors and pathways posited to underlie a particular form of undernutrition. Linkages between individual pathways together constitute a local causal model.

³ The confidence note is an indication of how reliable a risk factor is considered to be, based on the perceived strength of information gathered for that risk factor and the number of pathways through which it operates. It is generated during the final stakeholder workshop where risk factor ratings are debated and stakeholders assign a confidence note for each studied risk factor; this is later averaged across the group to validate final results. The confidence note has three levels: high (participant is convinced by the rating, based on the quality of the process and the coherence of the information collected as well as her/his technical and field experience); medium (participant is fairly convinced by the rating although some points require clarification); and low (participant is not convinced by the rating and has doubts about the quality of the process and/or about the information collected. Or: key information is missing).

⁴ During the initial technical expert workshop, participants considered 39 different risk factors pertaining to undernutrition and delivered a preliminary rating on a scale of 1 to 5, five illustrating a major effect on acute malnutrition, one illustrating minimal/no effect. Factors were subsequently grouped by thematic area. The highest rated risk factors reveal that infant and young child nutrition (IYCN), child and maternal health, water, and maternal education were considered major contributors to U5 and maternal malnutrition in the study area.

⁵ See footnote 3 regarding confidence note rating.

⁶ Some of these were: creation of women’s groups (businesses e.g. beading, community children’s crèche, communal water boiling, communal domestic animals for milk, forming markets through buying/selling of goods and livestock); creation of youth groups for funding of businesses; grants to restock livestock; training on market development to develop livestock and agriculture produce markets to reduce the need to travel; ; educate children on farming; teach men about the importance of responsibility sharing; training on budgeting and time management skills e.g. cooking in morning before going to collect water or burn charcoal.  See also the main report, p58. http://www.actionagainsthunger.org/publication/2014/02/kenya-nutrition-causal-analysis-qualitative-inquiry

⁷  ENN 2014. Technical Briefing Paper: The relationship between wasting and stunting: policy, programming and research implications. Available at: http://www.ennonline.net/waststuntreview2014