Childhood malnutrition remains a public health problem in low-income settings. In developing countries, the classic presentation of malnutrition, thinness and stunting is now coupled by overweight and obesity(1)- which is referred to as double burden of malnutrition (2). This is attributed to their rapidly growing economy, ongoing demographic changes, and continued urbanization (3) . These countries are undergoing rapid urbanization and are experiencing nutrition transition, which is a continuous process of change in diet patterns and food habits (3). It is characterized by a shift from high-carbohydrate and low-fat diet towards the consumption of a more refined, energy-dense and high-fat diet as well as low levels of physical activity (46).
The simultaneous presence of under? and over-nutrition within a society is reflective of the differential distribution of resources. It is further intensified in the context of developing economies whereby lack of resources together with the increasing population growth strains planning potential and increases marginalization within cities. In such situations, under-nutrition prevails in poorer households and overweight/obesity in better-off households (1).
This is, however, not the case everywhere as the pattern of obesity is changing. Overweight/obesity is reportedly increasing in children from poorer urban households as cheaper energy-dense, nutrient-poor diets are becoming more available and readily consumed (1,79). Thus, addressing the problem of inequalities in child malnutrition remains a challenge, and is of importance for policies and programs targeting child’s health(10).
Diet diversity is an essential component to increase nutrient adequacy and to promote physical and cognitive development in children. Dietary diversity has also been considered a reflection of the economic ability of a household to access a variety of foods(11). It is an important modifiable factor, in synergy with other environmental, economic and socio-cultural factors can play a key role in fostering a childs health by preventing malnutrition(12). Studies have shown that dietary diversity is positively associated with overall diet quality, micronutrient intake, and better nutritional status of young children(12).
The social and physical environment are critical factors in ensuring child health outcomes. The social environment includes aspects such as socioeconomic status, and media exposure, are known to influence food demands and preferences (9,13,14). The physical environment involves water and sanitation, and the built environment and therefore influences infectious disease risks, food availability, and affordability(15,16).
Additionally, high socioeconomic status (SES), as measured by maternal education and employment, may be associated with healthier overall dietary patterns, dietary quality, and adequate dietary diversity. It is has been documented that children in wealthier households grow better for a number of reasons, among which improved nutrient adequacy may be one important mechanism that household wealth and resources translate into better outcomes for children (17). Children in poorer households tend to be more affected by undernutrition than their better-off counterparts (1821). Wealthier households are expected to have the resources to purchase more food and thus have diverse diets compared to poor households. A higher parental educational level is associated with better employment opportunities and higher income which may give them a higher purchasing power and better nutrition knowledge. However, the relationship between socioeconomic inequality and the nutritional status of children is not conclusive.
The aim of this study is to describe the prevalence of under and over-nutrition among children under the age of five in Addis Ababa and to evaluate the relative importance of the different socio-economic strata of the population on child nutrition.
Study design and setting
A household survey was conducted in Addis Ababa, Ethiopia. Ethiopia is second-most populous country in Africa with 109 million people as of 2018, and the fastest-growing economy in the region (22,23). Though the urban population proportion is low, it is expected to triple in the next 20 years(24).
Addis Ababa, the capital city, has more than 3.4 million inhabitants according to official projections based on the 2007 census (26), however, it is estimated that the citys population will grow to 10.7 million by 2037(25) with rural-urban migration being an important factor. The city is characterized by highrate of unemployment (31%), the concentration of slum dwellings, poor housing, and severe inequalities(26). The city comprises 10 sub-cities (Kifle Ketema) and 117 woredas (the lowest administrative unit in Ethiopia) all of which were included in this study.
This study used cluster sampling; initially, we got a complete list of all administrative districts (woredas). One cluster from each woreda was then selected. In all the 117 clusters, we visited 60 households going in an interval of every third household. Interviews were carried out in eligible households meaning: a respondent mother is available, the selected households has at least one child under the age of five years and consents to participate. If the selected household had more than one child under the age of five, one was randomly selected to serve as a reference (index child) when responding to the diet section. However anthropometric measurements were taken from all under-five children present in the selected household during the data collection.
Data for this study were collected using a structured interviewer-administered questionnaire loaded onto tablets. The items included in the questionnaire were socio-demographic information, household assets, questions on food security, and food consumption. The questionnaire was developed after a thorough review of available literature augmented by formative qualitative work by the research team. The questionnaire was prepared in English and then translated into Amharic language (the official language of the country). A bilingual (English and Amharic languages) expert panel was convened to translate the study tool (27).
Ten teams, each consisting of five data collectors and one supervisor, were deployed to collect the data. Two weeks of training on interviewing techniques, contents of the questionnaire, use of tablet and how to take anthropometric measurements were provided for the field team. Field personnel in charge of taking anthropometric measurements had undergone standardization training. The trainees are standardized in anthropometry against the facilitator (gold standard) and assessed for inter- and intra-measurer variability. Equipment was also standardized prior to data collection. The entire procedure was pretested and necessary modifications to replace ambiguous words and edit to the electronic data capture program were made. Close supervision was provided by the field supervisors and the researchers at every stage of the fieldwork. Once the raw data was extrapolated from the server to STATA; data were cleaned by running simple frequencies to check for consistency before carrying out analysis.
The participants socio-demographic characteristics were summarized as sex (male or female), age of mother (in years), age of the child (in months), family size (2-4, 5-7, 8+) and current marital status (married, single). Other socio-demographic covariates included: education (Primary school and lower (? grade8)), secondary school and higher (?9 grade)), household head (female /male), and mother involved in an income-earning activity (yes/no).
The household wealth index was constructed using Principal Component Analysis from several key variables such as ownership of house, type of housing unit, housing material (floor, roof, wall material), access to separate toilet facility and clean drinking water as well as assets such as bicycle, motorbike, car, cell-phone, radio, TV, refrigerator, bed, Metad (electric stove used for making local bread called Injera) and a saving account. Principal components with eigenvalues greater than one were retained to construct wealth index values and then categorized into wealth tertiles (lowest, medium and highest) to serve as relative measures of household socio-economic status.
For this study mothers/caretakers were asked to provide a 24-h recall of foods consumed by the child both inside and outside the house regardless of the amount consumed. Once the mother completed the recall, we showed her pictures of food groups representing common food items in each category to get the full list of foods consumed by the child. These food items were grouped into seven food groups following the WHO IYCF guidelines. For each child, a binary variable of diet diversity was constructed from the total dietary diversity score to indicate whether the child diet over the last 24?h was adequately diverse. Children were considered to have adequately diversified dietary intake if they had food items from at least four of the seven food groups, while a score of 3 or less was considered to be inadequate(28) in accordance with the IYCF guideline.
Anthropometric measurements, weight and length/height, were done using the procedure stipulated by the World Health Organization(29). The weight of the children was measured to the nearest 0.1kg using the UNICEF electronic scale. Recumbent length and height were measured to the nearest 0.1cm using UNICEF recommended model wooden board as per WHO protocol.
Data analysis was done using STATA (version 14). Anthropometric indices were calculated using the WHO Anthro software (30). The Z-scores of indices, height-for-age Z-score (HAZ), weight-for-age (WAZ) and weight-for-height Z-score (WHZ) were categorized using the WHO Multicenter Growth Reference Standard. A child with a HAZ or WAZ less than ?2 SD from the reference population was defined as stunted or underweight, respectively. While those with a WHZ less than ?2 SD from the reference population were classified as wasted; and +2SD as overweight/obese(31).
Frequencies and percentages were calculated for all categorical variables and continuous variables; maternal and child age was summarized as means (± standard deviation).. Poisson regression model with robust variance was used to estimate the prevalence ratios (PR) and their respective 95% confidence intervals (95% CI). Similar analysis procedure was followed to test the association of select socioeconomic variables with child diet diversity.
The study protocol was approved by the institutional review board of Addis Continental Institute of Public Health Addis Ababa, Ethiopia. Necessary support letters were received from all the sub-cities and woreda level health offices. Study objectives were explained to participants and their informed consent was obtained prior to the data collection. During the MUAC measurement; those children who were severely malnourished (MUAC 10%), odds ratios do not approximate risk ratio(48).
The other strength of the study is, it has a very large representative sample; hence the results could be representative/replicated to other large cities within the country as well other metropolitan cities within Africa. Anthropometric measurements were performed under standardized measurement protocol. The use proxy measure for wealth is both the strength and limitation of the study; it is a limitation that we were not able to measure household income or expenditure directly which might have skewed our measure either to the right or left. However, the use of standardized variables adapted from the DHS to construct the wealth index is a strength.
The relatively high prevalence of overweight or obesity in the midst of persistent undernutrition among under-five children in sub-Saharan African is a major public health concern. Household wealth status has been found to be associated with both conditions; making it a critical variable that needs to be considered when designing interventions to alleviate child malnutrition. To deal with the newly emerging double burden of nutrition problems Ethiopia, understanding the inequalities could provide a good insight in designing focused interventions that simultaneously target both under and over nutrition in different groups.