Background
People have the right to adequate food, and the right to be free of hunger under the international law. Food security exists when all people, at all times have physical, social and economic access to sufficient, safe and nutritious food to meet their dietary needs and food preferences(Reference Kent1). Food and nutrition security is among the greatest challenges facing humanity in the present world(2). Under-five children are the main victims of food insecurity with devastating consequences in the future generation. Despite the progress in preventing and treating child undernutrition, 5 million annual deaths occurred in under-five children worldwide, nutrition-related problems account for almost half of the deaths(3,Reference Black, Victora and Walker4) . Sub-Saharan Africa carry the highest burden of child malnutrition, besides the East and West African subregions are highly affected by stunting and wasting(Reference Akombi, Agho and Merom5). In Ethiopia, child malnutrition remained one of the major public health problems which account 51 % of all childhood deaths(6). According to the Ethiopian Mini Demographic and Health Survey (EMDHS) 2019 report, the prevalence of stunting and wasting remained to be one of the highest in the world at 49 and 8 %, respectively. This prevalence is predominant in rural areas, where access to improved food sources, health facilities and other infrastructure is very limited(Reference Fantay Gebru, Mekonnen Haileselassie and Haftom Temesgen7,Reference Hirvonen8) .
Current strategies to address these problems focus on nutrition-specific interventions targeting mothers and children. The programmes aimed at intervening in major underlying factors like women's empowerment, food systems, socio-economic factors and disease prevention(Reference Bhutta, Das and Rizvi9). As farming is the primary food and income source for the ever-increasing African population(10), rapid agricultural transformation is the only way out to reach to the targets and achieve food security by 2030(11). One of the mechanisms was nutrition-sensitive agriculture. Nutrition-sensitive agriculture will narrow the gap between available and accessible food and the food needed for a healthy and balanced diet for all people, it also promotes food and nutrition sovereignty by promoting access to an affordable, diversified and more adequate diet(Reference Jaenicke and Virchow12,Reference Maluf, Burlandy and Santarelli13) .
On the other hand, one of the reasons why agriculture in sub-Saharan Africa is underperforming is mainly because women, who represent a crucial resource in agriculture and the rural economy through their multiple roles as farmers, labourer's and entrepreneurs, face more severe constraints than men in access to productive resources(Reference Hazell14). Likewise, it is important to unlock this and other critical bottlenecks to empower women to ensure rapid agricultural transformation in Africa(15).
There is now greater urgency to close gender gaps to empower women and young girls especially in vulnerable settings. Women are active agents of economic and social change, they play crucial roles ensuring food and nutrition security, eradicating rural poverty and improving the well-being of their families yet continue to face serious challenges as a result of gender-based stereotypes and discrimination that deny them equitable access to opportunities, resources, assets and services(16). Women empowerment is important not only for women but also reducing poverty and anger, improving health and nutrition(Reference Shenggen, Sivan and Rajul17). Studies show that women make essential contributions to the agricultural and rural economies in all developing countries by increasing agricultural performance(Reference Diiro, Seymour and Kassie18). Recent intervention strategies focusing on developmental indicators (such as child and household nutrition, poverty alleviation, education and improving investments in human capital) are considering women's empowerment as a determinant factor(Reference Tsiboe, Zereyesus and Popp19).
To end hunger by 2030, and to meet the sustainable development goal, the Ethiopian government implemented the National Nutrition Program. The overall goal of the programme was to facilitate and ignite the accelerated reduction of malnutrition. The main strategy was an integrated approach linking between nutritional needs and improvement of agriculture/animal breeding/fisheries and aquaculture sectors. The interventions under this programme include: optimal breastfeeding; optimal complementary feeding; mitigation and prevention of micro-nutrient deficiencies; water sanitation and hygiene; deworming; food fortification and management of acute malnutrition(20).
Despite the important roles they play in agricultural economies, rural women in Africa suffer from the highest illiteracy rates and are the most visible face of poverty(Reference Akter, Rutsaert and Luis21). Women in Africa have less access and control over all land and productive resources(Reference Hazell14). This is mainly attributable to lack of tailored and evidence-based strategies, programmes, policies or limitations of existing policies which are either misdirected or insufficient to high levels of poverty and illiteracy. Moreover, non-involvement of rural women in the design and planning of agricultural programmes and policies in order to reflect the real and specific needs is also another challenge.
Though new evidences are emerging on the association between women empowerment and child nutrition, there is a scarcity of data relating the index to child health and nutrition outcomes. Few published studies explored Abbreviated Women's Empowerment in Agriculture Index (A-WEAI) associated with child height-for-age (HAZ) but not with weight-for-height (WHZ) using cross-sectional designs(Reference Malapit and Quisumbing22–Reference Malapit, Kadiyala and Quisumbing25). Therefore, the present research aims to fill these gaps through the generation of empirical evidence between women empowerment in general and components of the five domains of A-WEAI empowerment index (5DE) and the association with nutritional status of children(26).
Methods
Study setting and design
A community-based cross-sectional study using a multi-sectoral and multi-disciplinary study tool (A-WEAI) was conducted in selected rural districts of the Gedeo zone, Southern Nations, Nationalities, and Peoples’ Region, Ethiopia from 1 February 2019 to 15 August 2019. The zone has six districts(27) and is well known for its indigenous agroforestry. It is unique for its exceptional mixed cultural and natural landscape(Reference Talemos, Abreham and Fisseha28) that allowed the land to carry the highest population density in Africa(29). On the other hand, it is also known for unacceptably high (54 % stunting and 15–27 % wasting) rate of child undernutrition and gender-based violence (84 % lifetime prevalence of intimate partner violence)(Reference Bogale, Wondafrash and Tilahun30) (Fig. 1).
Map of the study area taken from Molla et al. (Reference Ba, Behaksra and Tadesse31)
Sample size and sampling
We used a single population proportion formula considering the following parameters: 95 % confidence level, 9⋅2 % Proportion of empowered women (P) for overall A-WEAI 5DE based on a similar study in rural Nepal(Reference Cunningham, Ploubidis and Menon32) and margin of error of 2 % for overall A-WEAI 5DE (as it is a rare event). Adding, 10 % for possible non-response rate and multiplying by 1⋅5 design effect, the total sample size required for the study is 428.
Since the present study aimed to study rural areas, six districts of the zone were eligible for this study. We selected two districts from the rural areas by simple random sampling. The same sampling technique was applied to select six kebeles from the two districts (three kebeles per each district).
Following this, the list of all households having under-five children was obtained from the Kebele administrator in consultation with local health extension workers and health development army representatives. Study subjects fulfilling the criteria were allocated proportional to each kebele's household size.
All mothers/caretakers (15–49 years) and all children (6–59 months) in the household of selected districts were eligible for the study. Out of the six rural districts in the Gedeo zone, two was drawn randomly using a lottery method and in 428 randomly selected households, all mothers/caretaker (15–49 years) and children (6–59 months) of selected households were studied.
Data collection procedure
We used a pre-tested structured interviewer-administered A-WEAI questionnaire for collecting multi-sectorial women's level of empowerment. The questionnaire used for the survey includes modules on basic demographic information; household consumption expenditure; women empowerment indicators; dietary diversity; child anthropometry and infant and young child feeding; employment, agricultural productivity and input use; and other relevant information. The questionnaire was prepared in English and then translated to Amharic and Gedeufa languages and back translated to English to maintain consistency.
Anthropometric assessment
Duplicate measurements of child weight and height (supine length of children) were taken using standardised calibrated digital weighing scales (Seca GmbH & Co. KG, model 881 1021659; precision ± 100 g) and height/length boards. Child age was derived from maternal recall or, when available, from a birth certificate. HAZ and WHZ were computed using the WHO growth reference. Children with length-for-age or WHZ below −2 sd from the median of the reference population (HAZ < −2 or WHZ <−2) was classified as stunted or wasted, respectively(Reference Nabiha, Zunaria and Farzana33). HAZ and weight-for-age (WAZ) Z-scores were computed using the World Health Organization (WHO) growth reference computed using the AnthroPlus software.
Measurement for scoring the A-WEAI
The 5DE (Table 1) is measured by a number ranging from zero to one, with higher levels indicating greater empowerment. The score has two components. First, it reflects the percentage of women who are empowered (He). Second, it reflects the percentage of domains in which those women who are not yet empowered (Hn) already have adequate achievements. In the 5DE formula, Aa is the percentage of dimensions in which disempowered women have adequate achievements:
Domains used to compute the Women's Empowerment in Agriculture Index(Reference Alkire, Foster and Seth35)
This can also be written, following the Alkire Foster methodology(Reference Alkire, Roche and Ballon34,Reference Alkire, Foster and Seth35) , as {1−(H × An)}, where A = (1a) and reflects the percentage of domains in which disempowered women on average do not have adequate achievements.
Operational definitions
• Stunting: Height-for-age less than −2 Z-score based on the WHO reference group.
• Wasting: Weight-for-height less than −2 Z-score based on the WHO reference group(Reference Nabiha, Zunaria and Farzana33).
• Body mass index (BMI): calculated by dividing weight in kilograms by the square of height in metres. A woman or man is classified as overweight if he/she has a BMI of 25⋅0 or greater, and obese if a BMI of 30⋅0 or greater and underweight if he/she has a BMI less than 18⋅5(36).
• Adequate dietary diversity: represents children meeting the minimum level of dietary diversity. This means they have consumed ≥4 food groups during the previous day. The food groups incorporated in computing dietary diversity was: (1) Grains, roots and tubers, (2) Legumes and nuts, (3) Dairy products (infant formula, milk, yogurt, cheese), (4) Flesh foods (meat, fish, poultry and liver/organ meats), (5) Eggs, (6) Vitamin A-rich fruits and vegetables and (7) Other fruits and vegetables.
• Inadequate dietary diversity: represents children who did not meet the minimum dietary diversity level. This means they have consumed <4 food groups during the previous day.
• Food secure HH: If the respondent answers ‘No’ to all, or ‘Yes’ for only one of the eight standard questions of the Household Food Insecurity Experience Scale (HHFIES), has acceptable levels of internal validity for use in SSA(Reference Wambogo, Ghattas and Leonard37).
• Food insecure HH: If the respondent answers ‘Yes’ to at least two of the questions of the HHFIES.
• Access to safe water: If the household has improved water sources as defined in(38).
• Access to latrine: Household has any type of latrine.
Statistical analysis
A statistical analysis was done using Stata version 14 statistical software. Children not measured or with anthropometric Z-score values outside the biologically plausible range (HAZ 〈−6/〉6 and WHZ 〈−5/〉5); were excluded from analysis, as recommended by the WHO. The associations between each of the six indicators of women's empowerment in agriculture and child HAZ and WHZ were tested using ordinary least-squares bivariate and multivariate regression analyses after controlling for district-level clustering. For each of these linear regressions, the dependent variable was the Z-score and the independent variables were the dichotomous empowerment variables. In multivariate logistic regressions, data were also adjusted for various child, maternal and household factors that could confound the relationship between women's empowerment in agriculture and child nutritional status based on prior literature review of similar studies and our knowledge of the local context.
Ethical considerations
Ethical approval was sought from Dilla University College of Medicine and Health Science Institutional Review Board (IRB). Investigators were certified on ethical principles from the National Institutes of Health (NIH). Investigators and all research assistants were ensured all ethical principles of the latest Helsinki declaration ethical principles for medical research involving human subjects. Confidentiality, beneficence and privacy were cornerstones of the ethics and ensured. For data collected from children, assent was made and consent from guardian/parents was obtained before any study activities.
Results
Participant description
A total of 428 households were sampled for participation, 422 households gave responses, yielding a response rate 98⋅6 %. Accordingly, it was found that the mean (±sd) age of children enrolled to the study was 25⋅1 (±13⋅5) months; whereby nearly two-thirds (64⋅1 %) of them were older children aged between 18 and 24⋅9 months. Less than half (47⋅2 %) were boys and only a quarter (25⋅8 %) had adequate dietary diversity (≥4 food groups). The mean (±sd) HAZ and WHZ Z-scores were 0⋅42 (±0⋅49) and 0⋅76 (±0⋅43), respectively. Approximately three in five (57⋅6 %) children aged 6–59 months appeared to suffer from chronic undernutrition (stunting) and one in four (23⋅8 %) children in the study area was acutely malnourished (wasted) (Table 2).
Selected child and household characteristics of the sample, by 5DE inclusion status; children under 59 months of age and their mothers (n 422) in rural Gedeo, August 2019
sd, standard deviation; DDS, dietary diversity score.
Women's empowerment, maternal and household characteristics and child nutritional status
Among maternal and household characteristics, younger (12–18 months) and older (19–59 months) children (P < 0⋅05); children born to mothers in their twenties (P < 0⋅05) or to a small household size (P < 0⋅05); children of families owning one or none livestock in the house (P < 0⋅05); children who have inadequate dietary diversity (P < 0⋅001) and those children whose families having no access to latrine (P < 0⋅001) had higher stunting compared to their counterparts. Nonetheless, women empowerment in agriculture had no significant association with childhood stunting (P = 0⋅58) (Table 3).
Bivariate associations between child, maternal and household characteristics and stunting among children 6–59 months of age (n) in rural Gedeo, southern Ethiopia, August 2019
HH, household.
* Significant at P < 0⋅05.
Similarly, a bivariate analysis of the Z-scores for HAZ and WHZ have also showed a statistically significance association between input into production decisions, control over resources, control over use of income and group membership with childhood stunting (P < 0⋅05). However, only empowerment dimensions of input into production decisions and resources were statistically significant with childhood wasting (P < 0⋅05) (Table 4).
Bivariate associations between maternal empowerment with HAZ and WHZ among children 6–59 months of age (n 422) in rural Gedeo, August 2019
* Indicates significant association. The table depicts bivariate association between child nutritional status and indicators of A-WEAI. In bivariate association using ordinary least-square method; input into production decisions, resources, control over use of income and group membership were predictors of stunting. On the other hand, input into production decisions and resources were indicators of A-WEAI associated with wasting.
Multivariate analysis of WEAI and nutritional status of children
From variables fitted to the multivariable regression model, three variables were found to be independently associated with childhood stunting and only one variable associated with childhood wasting. As such, children whose mother is disempowered on input into production decisions were nine times more likely to be stunted than empowered women [AOR = 8⋅85; 95 % CI (3⋅66, 21⋅39)]. On the other hand, those children whose mother is disempowered for control over income were 65 % less likely to have stunted child than their counterparts [AOR = 0⋅35, 95 % CI (0⋅16, 0⋅75)]. Whereas, families who own between 2 and 3 livestock were 62 % less likely to have stunted child than households who have less than equal to one livestock [AOR = 0⋅38; 95 % CI (0⋅18, 0⋅80)]. Interestingly, women disempowerment in production decisions seems beneficiary for wasting, disempowered women have 84 % less likely to have wasted child than empowered women [AOR = 0⋅16; 95 % CI (0⋅05, 0⋅48)] as indicated in Table 5.
Multivariate associations between women's empowerment in agriculture and nutritional status among children 6–59 months of age in rural Gedeo, August 2019
* Indicates P-value less than 0⋅05. Multivariate logistic regression indicates the association of nutritional status in children and A-WEAI indicators as well as other child and maternal characteristics. Empowerment in input into production decisions, control over use of income and livestock density were associated with stunting, while input into production decisions was the only variable associated with child wasting.
Discussion
The present study was conducted to examine the role of multi-dimensional women's empowerment in agriculture in improving child nutrition. It turns out that the prevalence of undernutrition was high. The study revealed that children whose mother is disempowered on input into production decisions and for control over income and families who own between 2 and 3 livestock were less likely to have stunted child than their counterparts. Women disempowerment in production decisions seems beneficiary for wasting than empowered women. Different domains may have different impact on nutrition, because gender norms are culture and context-specific. WEAI domains in different countries were important in different settings(Reference Malapit and Quisumbing39).
The prevalence of stunting and wasting in the area was 57⋅6 and 23⋅8 %, respectively. The stunted percentage of children was almost comparable with neighbouring study in Halaba, and a bit higher of Zeway, 54 and 42 %, respectively(Reference Ersino, Zello and Henry40). But, the figure in both stunting and wasting was higher than the national figure; results from 2019 EMDHS show 49 % of under-five children were stunted and 8 % were wasted(41). Nutritional status in children was tied to maternal nutritional status during adolescence, pregnancy and breastfeeding periods, communities cultural background, women's empowerment, socio-economic factors and infectious diseases(Reference Tebeje and Abebe42).
Among the domains of women empowerment, input into production decisions was a decisive factor. We found that children whose mother is disempowered on input into production decisions were nine times more likely to be stunted than empowered women. Those children from disempowered mother over control of income were 65 % less likely to have stunted child than their counterparts. Whereas, families who have livestock density between 2 and 3 were 62 % less likely to have stunted child than households who have less than equal to one livestock. Women empowerment as measured by the five domains of empowerment (5DE) is significantly associated with better HAZ Z-scores for children aged between 6 and 59 months(Reference Holland and Rammohan43). Empowerment in production decisions, control over income and livestock density were factors affecting stunting. When women empowered to make production decisions about food production, crop production and livestock raising, they may make decisions in favour of the nutritional needs of young children compared to these decisions are out of their hand by their spouse or pressured by the community(Reference Cunningham, Ploubidis and Menon32,Reference Tebeje and Abebe42) .
Women empowerment in control over income improved children nutritional status generally and stunting in particularly. Control over income has found to have negative association with stunting(Reference Shroff, Griffiths and Adair44,Reference Kamiya, Nomura and Ogino45) , mothers with greater control over expenditures were more likely to have children with better long-term nutritional status evidenced by stunting status of their children(Reference Malapit, Kadiyala and Quisumbing25). Mothers who can use money as their wish tend to spend it on their child's welfare. Mothers’ financial resources were important for children's growth regardless of women social status in different societies(Reference Kamiya, Nomura and Ogino45).
Interestingly, women disempowerment in production decisions seems to promote wasting, disempowered women have 0⋅16 times less likely to have wasted child than empowered women. Though this finding disagrees with previous studies, the possibility of less care to their children of empowered women should not be ruled out. Empowered women might also spend their time out of the household giving less time to their children's nutrition. Seasonal variations and consequently maternal habits (i.e. women tend to drink alcohol in groceries in the study area) could affect children's nutritional status in short period. This association of acute undernutrition with disempowerment in production decisions should be verified in different seasons of the year.
Livestock resources were determinant over nutritional status of the children. But this crucial role of livestock resource on child nutrition is overly threatened by women's control over resources and income of the household(Reference Tiberti, Zezza and Azzarri46). While food insecurity is the main driving factor for chronic undernutrition, livestock ownership provides some protection to the sustained food security of the household and children. Small ruminants accelerate child growth and were associated with initial height trajectory. Livestock ownership shown to be crucial to remote households with limited access to purchased foods, where it become the only sources of certain nutrients(Reference Tiberti, Zezza and Azzarri46,Reference Randolph, Katjiuongua and Robinson47) .
Livestock ownership through animal source food consumption could be the primary source of high-quality protein and key micronutrients, which are especially important young children's health and development(Reference Murphy and Allen48–Reference Leroy and Frongillo53). On the other hand, livestock ownership is contributing to household income, is becoming increasingly important in developing countries as the level of development improves(Reference Smith, Sones and Grace54,Reference Adams, Ohene-Yankyera and Aidoo55) . Given the incipient link between livestock production and better nutritional outcomes, livestock farming is frequently encouraged as a means of expanding children's access to and consumption of high-quality animal source food worldwide(Reference Hetherington, Wiethoelter and Negin49,Reference Smith, Sones and Grace54,Reference Dominguez-Salas, Kauffmann and Breyne56) . This could be related to the fact that livestock ownership can provide households with greater opportunities to boost animal source food consumption if it results in cheaper or more reliable food supply(Reference Hetherington, Wiethoelter and Negin49). This is particularly likely when markets are underdeveloped, especially for perishable products like milk and dairy, which require expenditures in refrigeration and other equipment that may not be economically justifiable in the face of low effective demand.
In conclusion, the multi-dimensional women's empowerment in agriculture index was not significantly associated with nutritional status of under-five children. Disempowerment on input into production decisions and control over income were among the domains negatively affect child nutritional status. Disempowered women on these two domains tend to have higher stunting rates of their children. Household livestock density shown to be crucial to remote households with limited access to purchased foods, where it becomes the only sources of certain food groups and nutrients.
Strengthen the level of empowerment of women about input into production decisions that give the opportunity for women to make a self or joint decision-making over food and cash crop farming as well as ownership of major household assets. Concerned bodies, to their levels and scopes, need to pay attention to all aspects and domains of women empowerment in agriculture, but a special attention should be paid to their time allocation to productive and domestic tasks. Cash producing farmers in rural Ethiopia could benefit from diversified their agriculture, particularly from livestock farming. Government and other stakeholders should also emphasise agricultural diversification in rural cash crop producing areas to end the unacceptably high level of undernutrition (stunting and wasting).
Acknowledgements
The authors acknowledge the study participants and Tufts University.
The data that support the findings of this study are available from the corresponding author upon reasonable request.
T. A. Z. conceived and designed the study; T. A. Z. and M. J. conducted the research and performed statistical analysis. T. A. Z., M. J. and A. A. T. wrote the paper. S. B., T. T., and A. A. T. read, edited, and cleaned the draft. All authors have read and approved the final manuscript.
This work is supported by Tufts University under small research grant.
All authors declare no conflicts of interest.
Ethical approval was sought from Dilla University College of Medicine and Health Science Institutional Review Board (IRB) [ref no 007/2018]. For data collected from children, assent was made and consent from guardian/parents was obtained before doing the study activities.
Open access
