APPLYING THE POSITIVE DEVIANCE MODEL IN ETHIOPIA: THE MENTOR MOTHER PROJECT AND
ITS IMPACT ON CHILDHOOD MALNUTRITION IN HOLETA
Tsedey A. Tamir Master’s thesis
Public Health School of Medicine Faculty of Health Sciences University of Eastern Finland May 2017
UNIVERSITY OF EASTERN FINLAND, Faculty of Health Sciences Public Health
TAMIR, Tsedey A.: Applying the positive deviance model in Ethiopia: The Mentor Mother project and its impact on childhood malnutrition in Holeta
Master’s thesis, 66 pages
Supervisors: Adjunct Professor, Arja Erkkilä, PhD, Sohaib Khan, MBBS, MPH, PHD May 2017
Key words: childhood malnutrition, positive deviance, peer education
APPLYING THE POSITIVE DEVIANCE MODEL IN ETHIOPIA: THE MENTOR MOTHER PROJECT AND ITS IMPACT ON CHILDHOOD MALNUTRITION IN HOLETA
Malnutrition is responsible for more than 1/3 of deaths among young children in the world and Ethiopia is one of the most severely affected countries. The main objective of this study was to examine the determinants of weight-for-age Z score as well as the success of a behavioural change program in decreasing the number of underweight children using monthly records of weight measurements. The study was conducted using data from the Mentor Mother Program in Holeta, Ethiopia, running since 2014. It aims to enhance access to preventive and rehabilitative health and antenatal care among women and children under 5 through a wide scope of interventions in maternal and child health, family planning, environmental sanitation and personal hygiene.
Altogether 262 participants were included in the study of which 107 were boys and 155 were girls. The Mentor Mothers assigned to the family weighed the children and collected information on socio-economic status, health history, environmental and home sanitation using a questionnaire. The monthly weight-for-age Z scores were calculated using WHO Growth Standards. Informed by the childhood malnutrition conceptual model proposed by United Nations Children’s Fund, associations between possible causative factors and malnutrition were first established and integrated into a linear mixed model. Linear mixed method analysis was further used to examine variations in weight-for-age Z scores over a period of 7 months after entry into the program. Sub-group analyses of children who were at a normal weight and those who were underweight at the time of recruitment were carried out.
The study revealed that the weight-for-age Z score is associated with a number of underlying and immediate causes such as adequate dietary intake in the form of breast-feeding and complementary feeding, food security, sanitation and hygiene. Bivariate analyses demonstrated relationships between higher weight-for-age Z score scores and delivery at a health center, marital status, less illness frequency in the past month, better breast-feeding and complementary feeding practices, less food shortages, better environmental sanitation, better waste disposal and better personal hygiene. Multivariate analyses showed statistically significant associations between weight-for-age Z score and the children’s personal hygiene, complementary feeding and food shortages. The longitudinal multivariate analyses suggest that, in the frame of 7 months, the more time children spent in the intervention, the higher weight-for-age Z score was. However, the high proportion of children lost to follow-up signifies that these results have to be interpreted with caution.
ABBREVIATIONS
ANC Antenatal Care
CBN Community-Based Nutrition
CSA Central Statistical Agency
DHS Demographic Health Survey
EECMY–DASSC Ethiopian Evangelical Church Mekane Yesus - Development and Social Services Commission
EGST Ethiopian Graduate School of Theology
EHDHS Ethiopian Mini Demographic Health Survey
ENN Emergency Nutrition Network
FAO Food and Agriculture Organization
HEP Health Extension Program
HEW Health Extension Worker
HIV/AIDS Human Immunodeficiency Virus/Acquired Immune Deficiency Syndrome
IQ Intelligence Quotient
MMs Mentor Mothers
MUAC Mid Upper Arm Circumference
MUDH Ministry of Urban Development and Housing
NNS National Nutrition Strategy
OVC Orphans and Vulnerable Children
PANP Poverty Alleviation and Nutrition Program
PD Positive Deviance
PISP Projet Intégrée de Santé et de Population
PNC Post-natal Care
SC Save the Children
SD Standard Deviation
STD Sexually Transmitted Disease
SUN Scaling Up Nutrition
TB Tuberculosis
UNICEF United Nations Children's Fund
WASH Water and Sanitation Hygiene
WAZ Weight-for-age Z scores
WHO World Health Organization
ACKNOWLEDGEMENT
I am most grateful to my supervisors: Adjunct Professor Dr Arja Erkkilä and Dr. Sohaib Khan, at the Institute of Public Health and Clinical Nutrition (University of Eastern Finland, Kuopio, Finland) for their guidance and supervision through the process of this thesis.
I am also thankful to Dr Adamu Addissie at the Faculty of Medicine and School of Public Health,University of Addis Ababa who helped me in identifying this research project and provided me with a great number of invaluable resources, to Wosen who patiently answered all my questions and took time out of her busy days to go through the data with me, to the team at Holeta - both Siyoni and Chaltu - who regularly provided their help and input.
To my friends in the MPH program, Ana, Diana and Lucy, I owe many uplifted days during the winter.
Last but not least, I want to thank my family (Mom, Gashe, Mickey, Ghion and Priya) whose steady, firm and loving support gave me the strength needed to complete my thesis.
CONTENTS
1 INTRODUCTION ... 9
2 LITERATURE REVIEW ... 11
2.1 Nutritional status and anthropometry ... 11
2.2 Epidemiology of childhood malnutrition in Ethiopia ... 13
2.3 Aetiology of childhood malnutrition ... 15
2.3.1 The UNICEF conceptual model ... 15
2.3.2 Immediate causes: the nutrition – infection cycle ... 16
2.3.3 Underlying causes... 18
2.3.4 Macro factors ... 22
2.4 Consequences of childhood malnutrition ... 23
2.4.1 Effects of child malnutrition on child development ... 23
2.4.2 Effects of child malnutrition on adult health ... 24
2.5 The Positive Deviance model ... 24
3 THE MENTOR MOTHER PROJECT ... 28
4 OBJECTIVES OF THE STUDY ... 30
4.1 General objective ... 30
4.2 Specific objectives ... 30
5 MATERIALS AND METHODS ... 31
5.1 Study population ... 31
5.2 Data collection ... 32
5.3 Statistical Analysis ... 33
5.4 Ethical considerations ... 33
6 RESULTS ... 35
6.1 Baseline characteristics ... 35
6.2 Malnutrition and immediate/underlying factors ... 39
6.3 Malnutrition status during the program ... 43
6.3.1 WAZ in whole study population and sub-groups during the program ... 43
6.3.2 Determinants of WAZ changes in normal weight children ... 45
6.3.3 Determinants of WAZ trends in the whole sample during the program .... 46
6.3.4 Determinants of WAZ changes in underweight children ... 47
6.3.5 Characteristics of children with missing weight measurements ... 48
7 DISCUSSION ... 52
7.1 Main Findings ... 52
7.2 Discussion of findings in relation to other studies ... 53
7.3 Strengths and weaknesses ... 54
7.4 Recommendations ... 56
8 CONCLUSION ... 58
9 REFERENCES ... 59
LIST OF TABLES
Table 1: General baseline characteristics of mothers ... 35
Table 2: General baseline characteristics of children at recruitment ... 37
Table 3: Living environment characteristics... 38
Table 4: Bivariate associations between immediate/underlying factors and baseline WAZ .. 39
Table 5: Effect size of predictor variables of baseline WAZ (cross-sectional model) ... 42
Table 6: Descriptive statistics for WAZ during the first 7 months ... 44
Table 7:Effect size of fixed factors in the longitudinal model... 46
Table 8: Effect size of fixed factors in longitudinal model of normal weight children ... 47
Table 9: Effect size of fixed factors in longitudinal model of underweight children ... 48
Table 10: Bivariate analysis of children lost to follow-up and retained by month 7 ... 49
1 INTRODUCTION
Malnutrition is responsible for more than 1/3 of deaths among young children in the world (WHO 2015) and, according to the United Nations Children’s Fund (UNICEF), Ethiopia is one of the most undernourished countries with 28% of all the child mortality in the country associated with undernutrition (2015). Although the numbers have improved over recent years as shown by the Demographic Health Study (Central Statistical Agency 2012), the more recent 2014 Ethiopian Mini Demographic Health Survey (EHDHS) found that more than 2 out of every 5 children under the age of 5 were stunted and 19 % were severely stunted (Central Statistical Agency 2014). Additionally, the EHDHS survey reported that child wasting or acute malnutrition was at a national average of 9% with the phenomenon being highest (15%) in children less than 6 months of age. The survey also highlighted that 25% of children under the age of 5 were underweight while 7% were severely underweight.
By definition, malnutrition refers to both undernutrition and overnutrition (Blössner & Onis 2005) but for the sake of simplicity, this study will use the term to refer to undernutrition. In 1990, UNICEF developed a nutrition conceptual framework that highlighted three causal levels of malnutrition. First, it outlined the immediate determinants of malnutrition related to dietary quality and disease. Then it identified more distal underlying causes at the household or family level such as lack of access to basic services and infrastructure as well as inadequate maternal and child feeding practices. Lastly, it underlined the possible broader causes that occur at the societal level related to the political, economic and cultural context.
(Black et al. 2008)
In keeping with this framework, Megabiaw and Rahman (2013) found that poor maternal and demographic situations such as rural residence, place of delivery, parental educations and regional differences constituted the most important factors of stunting in Ethiopia.
Additionally, they reported that socioeconomic conditions, poor feeding practices, long-term breastfeeding and poverty were all significantly associated with severe and moderate stunting.
The consequences of malnutrition are far-reaching; some important outcomes include growth retardation (Grantham-McGregor et al. 1991) and poor cognitive development that leads to significant impact on education (UNICEF 2015) as well as on varying aspects of adult life (Black et al. 2013). Moreover, there are other more immediate health impacts on a child’s
well-being. For example, undernourished children under 5 are more at risk of anemia, acute diarrheal syndrome, acute respiratory infection, and sometimes fever (UNICEF 2015).
In light of these challenges and the aforementioned underlying risk factors that contribute to them, the Mentor Mother project was approved for implementation in Ethiopia in 2012. The program is based on the Philani child health and nutrition program piloted in South Africa, which integrates elements of a number of international behavioral change models that have proven successful in improving healthy development of children in settings with limited resources. This paraprofessional home visiting program was developed in townships near Cape Town and seeks to address the needs of mothers of underweight and at-risk children through the positive deviance (PD) model. (le Roux et al. 2010)
Although the impact of the program has been studied in the context of South Africa (Le et al.
2010; Rotheram-Borus et al. 2011; le Roux et al. 2011), only a few studies have examined its application in Ethiopia (Sjöling 2015; Andréasson 2015). This thesis will endeavor to specifically understand the impact of the Ethiopian program on the nutritional status of the children enrolled in the program.
2 LITERATURE REVIEW
2.1 Nutritional status and anthropometry
Among the many possible anthropometric measurements, the three most commonly used and internationally recommended anthropometric indices of malnutrition are height-for-age, weight-for-age (WAZ) and weight-for-height used to assess stunting, underweight and wasting, respectively (De Onis et al. 2003; Fenn et al. 2008). Stunting is used as an indicator of chronic malnutrition, wasting as an indicator of acute malnutrition and underweight as an indicator of chronic and/or acute malnutrition (WHO 2016). The health and nutritional status of a child is best defined by growth assessment since imbalances in health and nutrition inevitably impact child growth irrespective of etiology. (Blössner and Onis 2005).
These indices can be expressed in terms of Z-scores, percentiles, or percentage of median, all of which allow for comparisons between a chosen child or a group of children and a reference population. In an effort to create uniformity in monitoring worldwide trends in child malnutrition and make comparisons, the WHO, in 1986, initiated the WHO Global Database on Child Growth and Malnutrition. The database compiles, standardizes and thus makes possible the dissemination of results of nutritional surveys from many countries. (WHO 2016)
The WHO recommends the use of Z-score or standard deviations from the median of an international reference population to characterize malnutrition (Seetharaman et al. 2007). The child growth standards the organization uses were derived from children raised in optimal environments that mitigated constraints to growth such as inadequate diets and infection (Turck et al. 2013).
Measurements of standard deviations (SDs) from the mean (how far a child’s anthropometric measure is from the mean) is used to indicate the severity of malnutrition: -1 SD to -2 SD is said to be mild, -2 SD to -3 SD is said to be moderate and greater than -3 SD is said to be severe. (WHO 2016)
2.1.1 Stunting
The presence of stunting reveals a failure to reach linear growth potential reflecting poor socioeconomic conditions and increased odds of frequent and early exposure to hostile conditions such an inappropriate care and disease. It is the accumulated consequence of retarded growth (WHO 1986; WHO 2016). It is most often an indicator of long-term dietary inadequacy (de Onis et al. 2003). The prevalence of stunting ranges from 5% to 65% in developing countries. Stunting is most prevalent between three months and three years of age. After this period, mean heights run parallel to the reference indicating a state of having
“failed to grow” as opposed to a continual process of “failing to grow” reflected in the previous period. (WHO 2005)
2.1.2 Wasting
Wasting is commonly associated with acute starvation or severe infection resulting in severe weight loss. Wasting usually exists in population usually below 5% and prevalence peaks in the second year of life (WHO 2016). Wasting refers to a deficit in tissue and fat mass compared with the amount that is expected for a child of a specific weight and length (WHO 1986).
2.1.3 Underweight
The two ideal anthropometric indices for nutritional status are wasting and stunting since these reflect distinct physiological and biological processes. Low WAZ or underweight is a composite indicator that reflects body mass relative to age and influenced by both height and weight (WHO 2016). As an index, WAZ does not differentiate between short children of appropriate body weight and tall, thin children. WAZ reflects both the long-term nutritional experience provided by stunting and the short-term nutritional changes demonstrated by wasting (de Onis et al. 2003, WHO 1986). Thus, using WAZ as the sole indicator may underestimate the true load of undernutrition (Seetharaman et al. 2007). WAZ seems to be more purposeful in giving an overview of the distribution of nutritional problems in a country or the general direction of change (WHO 1986).However, WAZ is primarily used due to the
ease of collecting a sole measurement with the caveat of having accurate age information available. WAZ is useful for tracking an anthropometric status but not comparison with a cut- off value (Gorsetein et al. 1994). Furthermore, the appeal of growth assessment is in its universality. It is generally culturally acceptable, equipment for its measuring is easily transportable, little training is required, it is an inexpensive and non-invasive tool and it is straightforward and robust as a measure. (WHO 1995).
2.2 Epidemiology of childhood malnutrition in Ethiopia
The majority of undernourished people currently live in the global developing regions where 780 million people were undernourished in the period between 2014 and 2016. The prevalence of undernourishment has decreased from 18.6% in 1990-92 to 10.9% in 2014-16 largely due to changes in heavily populated countries like China and India. In Southern Asia and Sub-Saharan Africa, progress is much slower. Approximately 23.2% of the population in Sub-Saharan Africa, was estimated to be undernourished in the 2014–16 period which is the highest prevalence of undernourishment for a region. (FAO 2015)
Although Ethiopia has achieved the Millennium Development Goal 1c target to halve the proportion of people who suffer from hunger between 1990 and 2015 (UN 2016), the poor nutritional status of children and women continues to be a serious problem in Ethiopia (EHDHS 2014). The EHDHS found that 9% of children are wasted while 3% are severely wasted. Wasting occurred most often in children under 6 months (15%) and children between 12 and 17 months (14%). The same study revealed that 25% of children under age 5 are underweight, that the proportion of underweight children is highest in the age group 24-35 months at 31% and that there are regional variations in the proportion of underweight children with rural children being more likely underweight than urban children and the northern region being more affected than the southern region (Figure 1).
As far as trends are concerned, there has been a downward trend in the proportion of children stunted and underweight over the last four DHS surveys. The prevalence of stunting was reduced by 31% between 2000 and 2014, highlighting an improvement in chronic malnutrition while the proportion of underweight children decreased by 39% during that same period. Wasting seems to be one of the biggest nutritional challenges as only a small decline was visible over the same period. (EHDHS 2014)
Figure 1. Spatial epidemiology of child underweight in Ethiopia (Alemu et al. 2016)
2.3 Aetiology of childhood malnutrition
2.3.1 The UNICEF conceptual model
In 1990, UNICEF developed a nutrition conceptual framework (Figure 2) that highlighted three causal levels of malnutrition in children; first, it outlined the immediate determinants of malnutrition related to dietary quality and diseases that act on the individual (Scrimshaw et al. 1968).
Although all causes act on the individual, other causes act indirectly through more complex causal pathways (ENN 2011). More distally impacting nutrition status are the underlying causes at the household/family and community level such as lack of access to basic services and infrastructure or inadequate maternal and child feeding practices. Lastly, the model underlined the possible broader causes at the societal level related to the political, economic, and cultural context of the population (Black et al. 2008; ENN 2011). The framework is notable for its effort in interweaving both biological and socioeconomic causes (Haddad &
Smith 2001).
Though seemingly comprehensive, the model does not take into account individual factors that affect nutritional status such as genetic and phenotypic factors, disease susceptibility, and heterogeneity in individual nutritional pathways (differences in metabolic rates for example).
Neither does the model acknowledge that seasonality may affect the different causal pathways proposed. Seasonal cycles determine agricultural cycles that in turn affect food availability and access (ENN 2011).
Figure 2. A model of the UNICEF conceptual framework on malnutrition (UNICEF 2016)
2.3.2 Immediate causes: the nutrition – infection cycle
As noted earlier, inadequate dietary intake of energy, protein, fat and micronutrients and disease are the two immediate causes of malnutrition (Scrimshaw et al 1968). This entails either consuming too few nutrients or acquiring an infection that can increase requirements and prevent the body from absorbing nutrients. However, malnutrition and infection usually occur at the same time in what is known as the infection – malnutrition cycle illustrated in Figure 3 (Cook 2002; ENN 2011).
Although, the level of interaction depends on the type of infection and the extent of malnutrition, generally poor nutrition can lead to reduced immunity to infection which in turn
increases the chances of acquiring an infection or increasing its duration and intensity (UNICEF 2016).
Figure 3. The Infection-Malnutrition Cycle (UNICEF 2016)
Indeed, malnutrition is the primary cause of immunodeficiency in the world; poor nutrition results in underweight and weakened children who are vulnerable to infections (Katona &
Katona-Apte 2008). As research from Central America, Chile, Mexico and South Africa emerged on the subject, Keusch (2003) and Scrimshaw et al. (1968) showed the synergistic, cyclical and antagonistic nature of the interplay between infections and malnutrition.
A good nutritional status is vital for a robust and effective immune system. To serve as effective protection mechanisms, the skin surface, as the naso-esophageal, gastro-intestinal and genito-urinary tracts and the mucous secretions they produce must remain intact. To remain so, they constantly require adequate nutrients for cell growth and replication. By
allowing easier access to pathogens and mitigating the ability of the host to eliminate pathogens once inside the body, undernutrition diminishes the barrier function and predisposes to infections (Calder & Jackson 2000).
Recurrent infections such as acute respiratory infections, diarrhea, and helminthes and chronic diseases such as HIV/AIDS, may heighten the risk of malnutrition through loss of appetite, by hindering the uptake of nutrients, by diverting nutrients for immune response, or increasing metabolic requirements and nutrient loss such as urinary nitrogen loss (Katona &
Katona Apte 2008). In turn, these changes could lead to further weight loss and further damage of the defense mechanism (Katona & Katona Apte 2008; Lutter et al. 2011; UNICEF 2015).
In addition to adequate nutrient intake, dietary diversity has also been shown to be associated with nutrient adequacy and nutritional status. Using data from 11 DHS to look at the association between dietary diversity and stunting for children 6-23 months old, the authors demonstrated that dietary diversity was also positively associated with stunting Z scores (Arimond 2004).
2.3.3 Underlying causes
Nutrient inadequacy and the risk of infection are due to a number of factors at the household and community level commonly known as underlying causes. They can be grouped into three broad categories: household food insecurity (“food factor”), inadequate care (“care factor”) and unhealthy household environments including lack of access to health services (“health factors”). (Haddad & Smith 2000; UNICEF 2016)
2.3.3.1 Household food insecurity
Household food security is a concept that was developed during the 1980s in international development circles (Cook 2002). It is defined as the sustainable access to safe food of sufficient quality and quantity so as to ensure adequate intake and the health for all members of the household. (UNICEF 2016)
Three main components of household food security are key: availability, access and utilization. Availability refers to a sufficient amount of appropriate food being physically available. Access refers to income or other resources such as land which allow for the acquisition of sufficient as well as appropriate food. Utilization is defined by the proper use of food through appropriate food processing and means of storage (Swindale 2006; ENN 2011).
In turn, food insecurity is the constrained or uncertain availability or access to nutritionally sufficient and safe foods (Cook 2002). In infants and toddlers, food insecurity is associated with adverse health outcomes such as wasting, stunting and underweight (Saha 2009). In a study carried out amongst children in rural Bangladesh, a gradient in the proportion of underweight from 1 to 24 months of age was observed. Food-secure households in the study had the lowest proportion while the extremely food-insecure households the highest proportion. Similar patterns of stunting were found for the same group of children (Saha 2009). In a study in Kenya, similar results were observed; stunting among children aged between 6 and 23 months was higher among children from food insecure households than those from food secure (Mutisya 2015).
In an US study aiming to determine whether household food insecurity is associated with adverse health outcomes in a population aged less than 36 months, the authors found that children in food insecure household had significantly higher odds of having fair or poor health, and of being hospitalized since birth (Cook 2004). In another study in Bengal, multiple anthropometric failures were more likely in children living in low and very low food secure households and that there was a dose response relationship between multiple anthropometric failures and grades of food security (Mukhopadhyay & Biswas 2010).
2.3.3.2 Care-giving
Although a number of studies have found proof of a relationship between access to food and nutritional status of children (Blaney 2009), others have noted the absence of such a link (Shinsugi 2015). This absence can be explained by the relative importance of care in nutritional well-being (Pelletier et al. 1995). The concept of care relates to ‘the provision in the household and the community, of time, attention and support to meet the physical, emotional, intellectual and social needs of the growing child and of other family members’
(FAO & WHO 1992). According to Engel et al. (1997), care practices as relating to children are comprised of breast-feeding and complementary feeding, food preparation and storage, hygiene and health practices, and psychosocial care.
Differences in caring practices and resources between and, to a greater extent, within cultures may exist (Engle et al. 1999). Differences appear in how each culture and more specifically each family responds to children’s basic universal needs for food, protection, health care, shelter, and affection (Martin-Prevel 2001). Multiple studies demonstrate the significant positive association between care practices and children’s nutritional status (Ruel et al. 1999, Nti & Lartey 2008; Amugsi 2014)
2.3.3.3 Household environment and basic health services
The last category of the underlying causes of undernutrition indicates those factors influencing poor public health. This classification describes the health environment, exposure to diseases and the ability to access to basic health infrastructure. Ease of access to clean, safe water and sanitation, the presence or absence of malarial breeding areas, the quality of homes and the related levels of overcrowding, temperature, and stress characterize the health environment while the level of access to primary health services determines the extent to which infection and disease can be prevented or treated. (UNICEF 2016)
Access to healthcare is defined as ‘the timely use of services according to need’ (Peters et al.
2008). Peters et al. (2008) described four main dimensions of access, each with a supply and demand element. They highlighted geographic accessibility as defined by the physical distance or travel time from the health care facility, availability which refers to having the appropriate type of care and service providers available to those who require it, financial accessibility and lastly, acceptability determined by the responsiveness of the health service providers to the social and cultural expectations of user communities.
Of the four dimensions, distance is a recurring prime factor in access to health care (Buor 2003; Puett & Guerrero 2015). One study in the rural Ahafo-Ano South district of Ghana found that distance was the main factor affecting utilization; the greater the distance from a health facility, the lower the rate of health services utilization. Although other studies have failed to find the same results, this might be due their failure to account for travel time when considering geographical accessibility (Buor 2003).
A study carried out in the eastern province of Rwanda revealed a significant and negative association between stunting z-scores for under 5 children and travel time to the nearest health center (Aoun et al. 2015). Distance was also a frequently noted barrier in a comparative analysis of severe acute malnutrition treatment services in Pakistan and Ethiopia.
Households in both countries are faced with a variety of barriers to access such as distance, high opportunity costs and lack of awareness of available services and of malnutrition itself (Puett & Guerrero 2015). In another study in east rural Ethiopia, it was noted that mothers who had no access to the health service facilities had children whose odds of exposure to the risks of acute undernutrition were higher than those children whose mothers did have access (Egata et al. 2014). Similarly, in a study in Nigeria, a 0.016-point decrease in child stunting z- score was associated with 1% increase in women reporting significant health services information problems (Agee 2010).
The lack of adequate safe water, effective sanitation systems and the presence of unhygienic conditions all contribute to an unhealthy household environment and subsequent infectious diseases (ENN 2011). The link between unhealthy environment and adverse nutritional outcomes is said to be mediated by a number of disease pathways such as diarrhea, tropical enteropathy and gastro-intestinal disorders (Spears et al. 2013).
Some have even posited that the primary causal pathway is not diarrhea but tropical enteropathy (Humphrey 2009; Ngure et al. 2014), a subclinical disorder of the small intestine caused by fecal bacteria ingested by young children living in poor environmental conditions and resulting in malabsorption of nutrients, increased permeability and inflammation (Humphrey 2009). As children start exploring the environmental world to which they are confined by crawling, walking, mouthing objects, their risk of ingesting both animal and fecal bacteria increases, especially in rural low-income settings. This results in intestinal infections which lead to reduced appetite, malabsorption and increased nutrient loss and which in return affect early childhood development (Rah et al. 2015). Indeed, a number of studies carried out in Ethiopia and India have effectively demonstrated that poor environmental conditions (such as the presence of open defecation) may increase the odds of child stunting (Spears et al.
2013; Ngure et al. 2014; Rah et al. 2015). In a study carried out in rural central Mexico, WASH indicators were strongly associated with growth as described by height, weight, and wasting in children aged 6 and 30 months (Calloway et al. 1988). It is important to note that in reality, there is significant overlap and interaction between the three categories of
underlying causes and it is not always easy to determine the relative effect of each underlying cause (UNICEF 2015).
2.3.4 Macro factors
Macro factors or basic causes that influence the underlying determinants of child malnutrition: the human, structural and financial resources that are available and how they are utilized (this refers to the way these resources are politically, legally and culturally organized and translated into food security, care, services and healthy environments). (UNICEF 2015;
Kavle et al. 2015).
In his now renowned paper, Sen (1981) posits that political factors are at the root of all famines (1981). He demonstrated that it was government policy and not agricultural failure via natural disasters that was the cause of the 1973 famine in Wollo, Ethiopia and the 1943 famine in West Bengal (Sen 1981). Kaluski et al. (2002) join him in his claim that famine in Ethiopia is more a result of structural challenges rather than natural disasters.
Conflict in communities affect the nutritional status of its members, particularly children; in Chiapas Mexico chronic violence and intracommunity conflict was strongly associated with stunting in children. Conflict can decrease access to food and also increase vulnerability to diseases. (Sánchez-Pérez 2007)
The economic well-being of a country plays no small role in childhood nutritional status.
Pongou and co-workers (2006) examined the impact of economic crises and the adjustment programs of the 1990s in Cameroon on childhood malnutrition and found that the prevalence of malnutrition among children younger than 3 years, as measured by WAZ, had significantly increased from 16% to 23% from 1991 to 1998. One way to explain this change could be that households reduce the dietary diversity and energy intake as a way of coping with increases in food prices and reductions in income (Brinkman 2010).
2.4 Consequences of childhood malnutrition
2.4.1 Effects of child malnutrition on child development
Nutrition has resonating effects on health throughout a life course as well as on cognitive and social development, especially during early childhood. The consequences of malnutrition are wide in scope, ranging from decreased learning capacity in children and elevated risk of death from infections to increased risk of non-communicable diseases during adulthood.
(Black et al. 2013)
Starting from the time of conception to pregnancy, nutrition is vital for fetal growth. Babies with poor fetal growth are at an increased risk of death throughout infancy (Wu & Chen 2009). Although poor fetal growth is rarely a direct cause of death, it can contribute indirectly to neonatal death via birth asphyxia and infections such as pneumonia, sepsis and diarrhea (Black et al. 2008).
Throughout childhood, stunted, underweight, and wasted children have an increased risk of death from diarrhea, pneumonia, measles, and other infectious diseases (Wu & Chen 2009).
Additionally, malnutrition, by bringing about direct structural damage to the brain and impairing infant motor development as well as exploratory behavior, can negatively affect cognitive development. Although the effect of environmental factors weakens the association progressively, birthweight is positively associated with childhood cognitive skills. (Victora et al. 2008).
Specific types of micronutrient deficiencies have severe consequences for children both through their direct effects (iron deficiency anaemia, xerophthalmia due to vitamin A deficiency) and because they increase the risk of serious infections. Vitamin A deficiency has been shown to have an association with a higher risk of mortality while zinc deficiency increases the risk of diarrhea, malaria and pneumonia. Iron deficiency, which results from the low consumption of meat, fish and poultry has effects on child cognition. A combined analysis of 5 trials showed 1.73 lower IQ points per 10g/L decline in hemoglobin (Black et al.
2008). Iodine deficiency during pregnancy plays a role in the fetal motor and mental development and raises the risk of miscarriage and fetal growth restriction (Black et al.
2008).
2.4.2 Effects of child malnutrition on adult health
The effect of maternal and child undernutrition on adult health and body composition was studied by carrying out meta-analyses of five birth cohorts from low and middle income countries (India, the Philippines, South Africa, Guatemala, Brazil) (Lancet Series 2008). The study concluded that small size at birth and at 2 years of age were associated with reduced human capital; less schooling, shorter adult height, reduced economic productivity and lower off spring birthweight for women. Although association with adult disease indicators were less obvious (Victora et al. 2008), intrauterine growth retardation and low birth weight have been shown to be associated with increased risks of cardiovascular diseases and related disorders, including stroke, hypertension and type 2 diabetes mellitus (Wu & Chen 2009).
Furthermore, higher levels of depression, anxiety and lower self-esteem were visible in adolescents who were stunted by age 2 compared to non-stunted group. Increased risk of suicidal ideation and higher levels of hyperactivity in late adolescence and adult attention deficit was shown to be associated with lower stunting z scores at 24 months of age. (Black et al. 2013)
Famines can also effectively demonstrate the long-term effects of poor nutrition during early childhood. The Dutch famine of 1944-1945 revealed the association between prenatal exposure to famine and schizophrenia but not with human capital and cardiovascular risk factors (Black et al 2013). On the other hand, the 1956-1961 Chinese famine demonstrated the effects of famine exposure during pregnancy and during the first 2 years of life on weight, height, income, mental health and intergenerational offspring birthweight (Black et al 2013).
2.5 The Positive Deviance model
The concept of PD was first articulated by Dr. Joe D. Wray, in his 1972 Tropical Pediatrics editorial, when he asked, “Can we learn from successful mothers?” (Wray 1972). During the same period, Sam Wishik and Susan Van der Vynckt branded families whose children were in the upper quartile of stunting and underweight as “positive deviants” thus indirectly answering Wray’s query (Wishik & Vynckt 1976; Schooley & Morales 2007).
According to Marsh et al. (2004), positive deviant behavior is defined as practicing advantageous but uncommon behaviors by people who are of the same socioeconomic background as peers who do not practice these behaviors (Schooley & Morales 2007). These
behaviors are successful, adapted to the local culture and are usually affordable as well as sustainable because at-risk people already practice them.
The PD approach is one alternative to needs-based approaches to development that depend on scientific methods to identify methods for improved health. The danger in needs-based methods is that local populations will be unable to obtain or sustain what has been identified missing. PD offers an “assets-based” approach in that it takes advantage of “resilience” in communities. (Lapping et al. 2002)
The caregiving behaviors that ensure the good nutritional status of children under 5 years include but are not limited to the early initiation of breastfeeding, exclusive breastfeeding for the first six months, appropriate introduction of complementary foods, the proper management of childhood illnesses, immunizations, good hygiene, and birth spacing.
(Dearden et al. 2002)
The PD approach has been primarily used to address childhood malnutrition and it works through multiple steps. Case definitions are first elaborated and four to six high risk people who have attained good outcomes are identified. These individuals are then interviewed and observed with the purpose of unmasking the uncommon behaviors that could be responsible for the good outcome. The findings are analyzed to affirm that these behaviors are indeed uncommon but accessible for those at risk individuals in need. Activities in behavior change are developed in order to encourage the adoption of these new behaviors along with monitoring and evaluation activities. (Marsh et al. 2004;Schooley & Morales 2007)
In nutrition programs, the assessor identifies PD mothers by weighing a general population’s children and selects those weighing in the top 10% or 25% of the underweight distribution whose resources are inadequate. The selected mothers are then interviewed to identify behaviors distinct from those of mothers of malnourished children which is compiled into what is known as the PD inquiry. (Berggren & Wray 2002)
The first implementation of the approach took place in Haiti in the 1970s with the “Projet Intégrée de Santé et de Population (PISP)” under the helm of Haiti’s Division of Family Hygiene. This model required daily contact with mothers over a 3-month period in which the mothers were counselled on the use of local foods to rehabilitate malnourished children in
“Grand Foyers”, large outdoor kitchens (Bolles et al. 2002; Schooley & Morales 2007). The activities were later modified in the 1990s to include itinerant workers in the “Ti Foyer” or
Hearth model, a neighborhood model for nutritional rehabilitation in home-kitchens located geographically closer to the targeted women (Bolles et al. 2002).
The results from different applications of the model were promising. Monitoring results from the first cycle of the program showed that 100% and 66% of children in eight and five targeted villages respectively gained weight as fast as or faster than the international standard median six months after completing the PD program (Bolles et al. 2002). Similar results were identified in other countries where the PD approach was applied during the 1990s (Marsh et al. 2004). After implementation of the PANP program (Poverty Alleviation and Nutrition Program) in Thanh Hoa Province, Vietnam by Save the Children Federation/US (SC), severe malnutrition had been reduced from 23% to 4% (Sternin et al. 1996; Mackintosh et al. 2003).
When Mackintosh et al. (2003) carried out two follow-up surveys in the communes that had been part of the program, they found out that children from the SC study group had better nutritional statuses than those in the comparison group. The nutritional benefits even extended to their younger siblings who had not been directly implicated in the intervention.
Another study by Schroeder et al. (2002) also examining a later integrated nutrition program by SC in Vietnam found that although there was no overall effect on growth (in fact anthropometric indicators worsened as the children aged), there were improvements in dietary intakes and morbidity. Additionally, they found that the younger the child, the greater the effect of the intervention. This, the authors surmised, was due to three factors: that the highest growth rates occur during early infancy, that growth faltering occurs primarily between 6 and 15 months of age and that it is easier to prevent growth faltering in the early stages of growth than to prevent it during a later time period. Another study conducted in Delhi, India found that after four weeks of induction into a PD study, adding 1 teaspoon of ghee (milk fat) in the family pot food and providing an extra mid-day cereal snack for the child were two PD behaviors that were taken up by the study targets. (Sethi et al. 2003) What differentiates caregivers who have adopted these beneficial behaviors from those who do not? In Dearden et al. (2002), authors interviewed caregivers of children 6 to 17.9 months of age in five communes and looked at four behaviors centered on caregiving: feeding children “PD food”, feeding during diarrheal episodes, hand-washing and health-seeking behavior. By using elicitation procedures in their qualitative study, the authors discovered that, for each of these behaviors, favorable social norms determined who practiced them
while encouraging and reinforcing beliefs and attitudes were significant determinants of every studied behavior with the exception of hand-washing.
The PD approach has a number of limitations as well. For example, uncommon practices exist at a prevalence of 1-10% in communities and they are often costly to detect.
Additionally, even though training materials, workshops and technical assistance are more and more readily available, scale-up may be limited as it requires individuals skilled in community mobilization, PD theory, and participatory research. Furthermore, scale-up presents its own set of challenges. An evaluation of a PD nutrition program scale-up in Viet Nam demonstrated that the quality of interventions decreased as the program expanded (Lapping et al. 2002). There could also be a limited generalizability of findings as inquiries are local and a vital aspect of the approach is mobilization through self-discovery. Lastly, the absence of culturally-appropriate methods to ensure not only fathers’ participation but to also allow their wives to participate is not unimportant. (Lapping et al. 2002)
3 THE MENTOR MOTHER PROJECT
The primary model used in the program is the PD/Hearth model targeting childhood malnutrition implemented in Haiti in the early 1990s (Bolles et al. 2002). This model relies on “positively deviant” mothers raising relatively healthy and well-nourished children while living in resource-poor environments. These mothers are trained to lead gatherings of malnourished children and their mothers (Wollinka et al. 1997).
The program also has elements of nurse home visits as these “model mothers” visit the targeted mothers and their children in their homes. The findings from the Philani Program have been promising and include improved health behaviors, increased nutritional knowledge, and rehabilitation of undernourished children (le Roux et al. 2010).
In the Ethiopian context, the program commenced in June 2014 and will continue to December 2017. Implementation of the program in Holeta is under the helm of the Ethiopian Evangelical Church Mekane Yesus - Development and Social Services Commission (EECMY - DASSC). The general objective of the program is to enhance access to preventive and rehabilitative health and antenatal care among women and children under 5 with a wide scope of intervention, targeting multiple groups (pregnant women, mothers and children) and outlining action in many focus areas of which nutrition is just one aspect. It encompasses basic maternal health, basic child health, family planning, environmental sanitation, personal hygiene and other related focus areas. The Mentor Mothers work in collaboration with the Health Extension Workers (HEW) assigned to the area (EGST 2014). The project design and operation are closely aligned to the Ethiopian government’s health extension package and incorporates its four pillars:
1. Hygiene and environmental sanitations (safe excreta disposal, waste removal, water quality control, food hygiene, personal hygiene, proper housing)
2. Disease prevention and control (Tuberculosis (TB) prevention and control; First Aid, malarial prevention and control, Sexually Transmitted Diseases (STDs) prevention and control)
3. Family health service (maternal and child health, immunization, nutrition, adolescent reproductive health, family planning)
4. Health education and communication which integrates the previous pillars
In the realm of nutrition, the Mentor Mothers seek to provide proper counselling and support for pregnant mothers and children through ante-natal care (ANC) follow up, post-natal care (PNC), proper nutrition of pregnant women, education on breastfeeding (exclusive breastfeeding, complementary feeding and continuing breastfeeding till 2 years), preparation of nutritious food for children above 6 months, provide feeding support to households deemed most in need and regular monitoring of child growth and development. (EGST 2014) The Mentor Mothers are chosen according to a specific set of criteria of which completion of Grade 10 or Grade 12, willingness to work for 6 hours a day and experience in childcare are requirements. Additionally, the chosen Mentor Mother is an individual between the ages of 22 and 45, the owner of an ID from an area adjacent to the project site and has lived there for two years. The Mentor Mother also has to be a church member and is able to provide a letter of proof to this effect. (EGST 2014)
Based on these criteria, 28 Mentor Mothers were initially selected for the two sites of the program: 15 Mentor Mothers for the Addis Ababa site and 13 Mentor Mothers for the Holeta site. The 28 Mentor Mothers were trained for 20 days in four phases. First, they watched experienced Mentor Mothers in action. They then attended training sessions covering info about nutrition, basic child health, weighing babies and the completion of growth charts, how to recognize danger signs and crisis situations and how to handle depressed mothers. They also learnt how to build trust with mothers and, as a final step to their training, they carried out the first round of home visits on their own. (EGST 2014)
After training, the Mentor Mothers visit 6 houses a day and a single household is visited every 15 days but the frequency may increase based on the nature of individual cases. During visits, the Mentor Mothers weighed the child and discussed the developmental progress with the mother, as well as proper nutrition and hygiene, stress, breastfeeding, proper time to introduce solid food, frequent feeding and mixed diet, immunizations progress/status and deworming. (EGST 2014).
The Mentor Mothers receive a monthly salary of Birr 1500 as well as a stipend of 300 Ethiopian Birr for telephone charges and 2000 Birr for healthcare costs. (EGST 2014)
4 OBJECTIVES OF THE STUDY
4.1 General objective
This study primarily aimed to determine whether the Mentor Mother program has had an impact on 0 to 5 years old children’s WAZ scores in the 7-month period after enrolment into the program. In addition, determinants of WAZ were analyzed.
4.2 Specific objectives
Three specific objectives were outlined:
1. Analyze the association between environmental and living conditions, feeding practices, mother’s health status and education, access to health facilities, socio- economic status and child’s weight status
2. Study the changes in proportions of underweight (WAZ < -2) for all children under 5 during the 7-month period after the start of the intervention.
3. To compare the differences in proportions of underweight (WAZ < -2) among children deemed underweight and those deemed at a normal weight at the beginning of the intervention during the intervention.
5 MATERIALS AND METHODS
5.1 Study population
The Mentor Mother program was implemented in Holeta, a town found in the Oromia National Regional State, covering 5549 ha and situated 28 km from Addis Ababa. The town was founded in 1901 and as capital of Wolemera district and it serves as an administrative center for the surrounding area (MUDH 2016).
According to the National Population and Housing Census carried out by the CSA in 2007 (CSA 2007), the population of the town was 23,296 of which 11,512 (49.4%) were males and 11,784 (50.58%) were females. 28% of the population is within the age group of 0-15 years, 66.88% in the age group 16-60 years and the remaining 4.85% is 60 years and above. The average household size in the town was calculated to be 3.3. According to a baseline survey conducted prior to the launch of the program, almost half of the households (46.6%) income was obtained through daily labor. Holeta has one government health center with a catchment area of 15 000, six government health posts and eleven private clinics. A total of 281 children and their mothers were included in the study.
The Mentor Mothers are chosen according to a specific set of criteria of which some are:
completion of Grade 10 or Grade 12, willingness to work for 6h/day and experience in childcare. Additionally, the Mentor Mother is an individual between the ages of 22 and 45 owner of an ID the adjacent area of the project site and has lived there for two years. The Mentor Mother also has to be a church member and she must be able to provide a letter of proof to this effect (EGST 2014). Inclusion in the program was granted once a number of criteria were established.The program included:
1. Mothers/women who do not practice ANC and PNC properly.
2. Pregnant and lactating women with mild and moderate malnutrition and with a family size of 3-10 persons.
3. Pregnant women who do not have a balanced diet due to lack of awareness or a lack of means.
4. Mothers with poor hygienic practices in regards to their children and environmental sanitation.
5. The most economically vulnerable families without income and reliable livelihood.
6. Malnourished children and mothers as approved by the health centre (one requirement was that the children and mothers were unsupported by other organizations or schemes).
7. Orphans and Vulnerable Children (OVC) and mothers or pregnant women living with HIV/AIDS and other chronic illnesses such as TB or cancer.
5.2 Data collection
Data was available in the form of “blue folders” which are questionnaires that contain baseline information and are periodically updated by the Mentor Mothers with weight measurements and general well-being assessments. The questionnaire covers baseline characteristics of the children and their mothers: age, weight at recruitment, diet, neonatal health status, current health status, socio-economic status, hygiene and sanitation. Sex and the exact dates of birth were not included in the questionnaire so the study had to resort to assigning sex based on the cultural norm of gendered names and approximate birthday calculations based on a question that asked how old the child is. The questionnaire also serves as a tool for child monitoring via anthropometry in the form of monthly weight measurements from recruitment to 36 months. The study was restricted to 7 months of weight measurements as the number of weight data values decreased greatly after 7 months. The Mentor Mothers carried out the monthly weight measurements with personal weight scales.
The data from the ”blue folders” questionnaires were transcribed to an anonymized datasheet where the indicators of interest were retained and WAZ for each child’s weight measurement during the 7 months was calculated using WHO Anthro II for Windows™ using an input of sex, age and weight. Weight data was read off weight charts that were integrated into the questionnaires.
5.3 Statistical Analysis
The characteristics of the sample are first described in frequencies and percentages. Secondly, bivariate analysis using Pearson’s Chi-square test for categorical variables and independent t- tests for continuous variables were carried out to test the association between those independent variables suspected of association with malnutrition in literature and the outcome, WAZ during the first month. Sex and age were not included in the analysis as WAZ calculations already take these two variables into account. Using the independent variables that showed statistical significance (p-value >0.05) in the bivariate analyses, a linear mixed model of WAZ at month 0, was then created.
In the second part of the analysis, the sample was grouped into weight categories (children of normal weight with WAZ >= -2 and children who were underweight with WAZ < -2). Using the statistically significant independent variables for food shortage, complementary feeding, home sanitation and personal hygiene from the cross-sectional model to adjust for confounding, 3 linear mixed models were created in order to understand the effect of time spent in the intervention on WAZ over the course of the first 7 months for the whole sample (longitudinal model) and the two subgroups of underweight and normal weight. Data analysis was performed using IBM SPSS Statistics version 21 for Windows™.
5.4 Ethical considerations
Secondary analysis of existing research data is an avenue that offers new prospective for knowledge production. Secondary analysis ensures the confidence in the outcomes of pre- existing research (Law 2005).
Scientific literature does highlight a number of concerns about the use of secondary data and digital forms of data, largely related to the potential for harm to participants by way of confidentiality and privacy breach and the lack of informed consent (Law 2005; Morrow et al. 2014). These concerns can be allayed by a number of strategies that will be undertaken during the present analysis. One of these strategies is full anonymization to guarantee the safety, rights and well-being of participants (Law 2005). In this study, the anonymization process consisted of removing names and replacing them with numeric characters and removing unique outlier data values that could facilitate the identification of specific participants. Through this process of anonymization, the risk of identification was reduced
(Collins 1991), potential harm circumvented and benefits to the community optimized. Data collection was performed once the consent was obtained from the mothers and the health extension workers who are community health workers responsible for ensuring access to primary health care in the community (EGST 2014).
6 RESULTS
6.1 Baseline characteristics
Table 1, Table 2 and Table 3 present the general characteristics of the mothers sampled, their children and the living environment, respectively. Of the mothers included in the program, 62.5% were between 16 and 27 and 30 % between 27 and 38 (Table 1). The mean age of the women at the time of recruitment was 26.91 years (SD 6.94, range 16 – 60 years).
78.6% of the mothers had between 1 and 3 pregnancies at the time of questioning with a mean of 2.39 (SD: 1.70, range: 1 – 10 pregnancies). In reference to education, the largest group did not receive any form of educational instruction (40%) followed by the group consisting of those who had completed primary education. Access to healthcare during time of delivery was relatively good with 67.9% having given birth at a health centre. 78.9% of the women had a spouse.
Of these women, 10.7% were diagnosed as malnourished by a nurse, 13.2% as anaemic and 16.4% as having experienced household food shortage. Alcohol usage and smoking rates were quite low in the study population at 9.3% and 0.4%, respectively. Income data was only available for 110 women and of these 48.2% had less than 500 Birr per month. Health- seeking rates were quite high for the group at 96.1%. (Table 1)
Table 1. General baseline characteristics of the mothers
Variables Categories Frequency (%)
Mother's age, y (n=276) 16 - 26.99 175 (62.5)
27 - 37.99 84 (30)
38 - 48.99 11 (3.9)
49 – 60 6 (2.1)
Number of pregnancies (n=279) 1 – 3 220 (78.6)
4 – 7 54 (19.3)
8 – 12 5 (1.8)
Mother's education level (n=269) None 112 (40)
Literate 12 (4.3)
Primary 74 (26.4)
Secondary 44 (15.7)
Post-secondary 27 (9.6)
Place of delivery (n=278) Health center 190 (67.9)
Home 88 (31.4)
Mother's marital status (n=275) Unmarried 54 (19.3)
Married 221 (78.9)
Health status during pregnancy (n=276) Healthy 216 (77.1)
Not healthy 60 (21.4)
Self-reported anemia (n=276) Yes 37 (13.2)
No 239 (85.4)
Alcohol Use (n=278) Yes 26 (9.3)
No 252 (90)
Smoking (n=276) Yes 1 (0.4)
No 275 (98.2)
Feeding support1 (n=273) Yes 21 (7.5)
No 252 (90)
Income (in Birr) (n=110) <500 53 (48.2)
500 – 1000 18 (16.4)
1000 – 2000 21 (19.1)
>2000 18 (16.4)
Food shortage (n=269) Yes 46 (16.4)
No 223 (79.6)
Income support (n=271) Yes 28 (10)
No 243(86.8)
Health-seeking (n=279) Health professional 269 (96.1)
Traditional healer 10 (3.6)
1Feeding support refers to any food assistance that a household might receive from any governmental or non-governmental body
The mean age of the children was 10.82 months (SD 12.38, range 0.03 – 58 months) with a mean weight of 6.25 kg (SD 2.37, range 1 – 13 kg) at recruitment (Table 2). There was an imbalance in the representation of sexes with 59.2 % being female and 40.8 % male. The mean WAZ was -1.56 (SD 2.03, range: -6.8 – 5.03). About 42.5 % of the children had a WAZ value greater than -2 and 46.8 % had a WAZ value less than -2 indicating underweight (Table 2). Roughly more than half (51.8%) of the children have experienced some sort of sickness in the month before recruitment. About 86.4 % were breastfed and 42.5 % received some form of complementary feeding (Table 2).
Table 2. General characteristics of children at recruitment
Variables Categories Frequency (%)
Age (in months) (n=278) 0 - 14.99 190 (67.9)
15 - 29.99 66 (23.6)
30 - 44.99 15 (5.4)
45 – 60 7 (2.5)
Weight (in kg) (n=262) 0 - 3.99 62 (22.1)
4 - 7.99 147 (52.5)
8 - 11.99 50 (17.9)
12 – 13 3 (1.1)
Sex (n=262) F 155 (59.2)
M 105 (40.8)
Sickness in the past month (n=270) Yes 145 (51.8)
No 125 (44.6)
Diarrhoea (n=272) Yes 36 (12.9)
No 236 (84.3)
Vomiting (n=272) Yes 38 (13.6)
No 234 (83.6)
Cough (n=272) Yes 33 (11.8)
No 239 (85.4)
Fever (n=272) Yes 34 (12.1)
No 238 (85)
Other (n=268) Yes 40 (14.3)
No 228 (81.4)
Breastfeeding (n=277) Yes 242 (86.4)
No 35 (12.5)
Complementary feeding (n=270) Yes 119 (42.5)
No 151 (53.9)
Immunization (n=274) Complete 85 (30.4)
Incomplete 186 (66.4)
Unvaccinated 3 (1.1)
WAZ (n=250) >= -2 SD
< - 2 SD
119 (42.5) 131 (46.8)
Mean -1.5637
Median -1.8550
Std. Error of Mean .12852
Table 3 presents the living environment characteristics of the households in which the mothers and children find themselves in at the time of recruitment. 87.1 % were found to be renting while water supply and toilet access are shared in majority of the households (75 % and 66%, respectively). Poor waste disposal was registered for 37.1 % of the households, poor environmental sanitation for 36.1 % of the households and poor personal hygiene for 29.3 % of the households. Mentor Mothers evaluated these characteristics.
Table 3. Living environment characteristics
Variables Categories Frequency (%)
Residence (n=276) Own 32 (11.4)
Rental 244 (87.1)
Water supply (n=275) Private 42 (15)
Shared 201 (75)
Well 23 (8.2)
Toilet (n=276) Private 11 (3.9)
Shared 185 (66.1)
Pit 80 (28.6)
Home sanitation (n=275) Good 85 (30.4)
Average 106 (37.9)
Poor 84 (30)
Environmental sanitation (n=275) Good 69(24.6)
Average 105(37.5)
Poor 101(36.1)
Waste disposal (n=266) Good 43(15.4)
Average 119(42.5)
Poor 104(37.1)
Personal hygiene (n=276) Good 86(30.7)
Average 108(38.6)
Poor 82(29.3)
Odor (n=276) Good 65(23.2)
Average 126(45)
Poor 85(30.4)
6.2 Malnutrition and immediate/underlying factors
Using Pearson’s Chi-Square test, bivariate associations (p-value<0.05) between a number of immediate/underlying categorical factors of malnutrition and WAZ at the time of recruitment (whether WAZ was greater or less than -2 SD) were shown in Table 4.
Children who were underweight had mothers who were more prone to have not received any formal education while those who were at a normal weight were more prone to have mothers who had undergone secondary and post-secondary education (p-value=0.001). Normal weight was associated with delivery at health centres (p=0.001), having mothers who were married (p-value=0.039), not being sick during the past month (p-value=0.01), being breastfed (p- value=0.003), receiving no complementary feeding (p-value<0.001) and having no food shortages (p-value<0.001). Underweight children experienced poorer home and environmental sanitation than children at normal weight (p-value<0.001). They were also more likely to come from households with poor waste disposal and have poor personal hygiene (p-value<0.001) (Table 4).
Using independent samples t-tests, the relationship between the number of pregnancies and WAZ as well as between mother’s age and WAZ were tested and found to be non-significant.
Similarly, sex, health during pregnancy, mothers’ nutritional status, alcohol intake, toilet access, water supply, residential type and health-seeking habits were not associated with WAZ.
Table 4. Bivariate associations between immediate and underlying factors and WAZat baseline
Factors
Weight Category
Significance Underweight
n (%)
Normal Weight n (%)
Sex
0.139
Female 62 (25.1) 81 (32.8)
Male 55 (22.3) 49 (19.8)
Education level (mother)
0.001
Literate 4 (1.6) 6 (2.5)
None 61 (25.1) 38 (15.6)
Primary 34 (14.0) 32 (13.2)
Secondary 13 (5.3) 28 (11.5)
Post-Secondary 6 (2.5) 21 (8.6)
Delivery location
Health centre 72 (29) 102 (41.1)
