Burden of undernutrition and its associated factors among children aged 6–59 months: findings from 2016 Ethiopian demographic health survey (EDHS) data

Abstract

Background

Despite numerous government nutrition-specific and sensitive interventions, undernutrition (e.g., underweight) remains the major public health concern among under-five-year-old children in Ethiopia. Therefore, this study aimed to assess underweight and associated factors among children in Ethiopia using 2016 EDHS data.

Method

The current study used 9,013 children under five years old. An ordinal logistic regression (e.g., proportional odds model) was applied to determine the associated risk factors of being underweight. The current study used SAS software version 9.4 at the 5% significance level.

Results

The prevalence of underweight was 25.3%. Variables such as children’s sex, place of residence, whether the child is twin at birth, breastfeeding status, size of children at birth, childbirth order, employment status of mothers, parents’ educational level, children’s age groups, the incidence of diarrhea in the past two weeks, and baby fortified food were statistically associated with underweight among under-five age in years.

Conclusions

Underweight among under-five children is predicted by place of residence. In addition, there is a regional disparity of underweight among children. Therefore, further effort is needed to improve health education in children’s welfare and health facilities to enhance patients’ understanding of proper information and feeding.

Introduction

Malnutrition refers to deficiencies, excesses, or imbalances in a person’s energy and/or nutrients [1]. This term covers both overnutrition and undernutrition, including stunting, wasting, underweight, micronutrient deficiencies, overweight, obesity, and diet-related noncommunicable diseases [2]. Undernutrition is when a child’s intake of energy and nutrients is insufficient to maintain health. In other words, these children experience micronutrient deficiencies [3]. Undernutrition remains a significant public health concern in low-income countries [4]. It is also one of the most priority areas in Sub-Saharan Africa, specifically in Ethiopia [5,6,7]. In Ethiopia, undernutrition accounts for 45% of deaths of under-five age in years [8]. The 2016 Central Statistical Agency (CSA) reported that children from rural areas are more likely to be undernourished than those from urban areas (25% vs. 13%) [9]. These statistics show that child underweight is an alarming issue in Ethiopia. The significant problem of underweight varies across different regions of the country [9,10,11,12,13]. Given the pronounced regional disparities, the 2016 Ethiopia Demographic and Health Survey (EDHS) underscores the need for a detailed analysis of the factors influencing underweight [14]. Addressing the underlying factors contributing to underweight is also essential for planning children’s growth and development interventions [10,11,12].

Although previous research has explored various associated risk factors for undernutrition (e.g., wasting and stunting) in Ethiopian children under five years of age [5, 6, 11, 15]These studies need to understand the burden of underweight among children under five. In addition, several studies have examined individual- and community-level factors of underweight among under-five age in years [5, 6, 15,16,17,18,19]. However, these studies have methodological gaps, such as small sample sizes that did not represent populations and inconsistency in measuring outcome variables [12]. Moreover, previous studies included a limited number of associated risk factors for underweight with limited independent variables [5, 6, 11, 12, 15,16,17,18,19,20,21,22]. Including a comprehensive set of factors will provide a more nuanced view of the underlying factors of underweight. Moreover, previous studies on children underweight have primarily classified children into two broad categories: underweight and normal weight [5, 16, 18,19,20, 22, 23]. However, these classifications fail to capture the severity of underweight, such as mild, moderate, and severe underweight. This limitation restricts a comprehensive understanding of the issue and hinders the development of targeted interventions that address children’s specific needs at different underweight levels.

Besides, previous studies have examined the nutritional status among children under five in specific study settings [5, 11, 22]. This study seeks to fill these gaps using a comprehensive and representative sample from 2016 EDHS data, representative data collected across nine regions and two additional administrative cities in the country. The 2016 EDHS datasets include individual and community risk factors among under-five children. Despite the resolutions tried, the issues are still evident, and studies have recommended frequent research to understand the prevalence of underweight and associated factors. Therefore, this study aimed to determine the prevalence of underweight among children aged 6–59 months in Ethiopia and their associated risk factors.

Method

Study area

For this study, we used the 2016 EDHS data collected by the Central Statistical Agency (CSA) in Ethiopia [14]. The data were collected from January 18, 2016, to June 27, 2016, covering all nine regions and two cities. The 2016 EDHS dataset was designed to estimate key country indicators and is divided into nine geographical regions and two main towns.

Study population

The 2016 EDHS data used a stratified sample method for sample size (n) determination. The population is classified into rural and urban strata, recognising these areas’ significant health outcome differences. Using proportional representation, about 645 enumeration areas (EAs) were randomly selected from a complete list, with 443 from rural regions and 202 from urban areas. Individually selected EA, 28 households were systematically sampled, starting from a random starting point and selecting every nth household, thus establishing evenly balanced inclusion probability for all households. This method reduces the selection bias and improves sample representativeness. The data were collected using a well-structured questionnaire for women and biomarker assessments for children aged 0 to 59 months. The sample size and distinct populations enabled consistent statistical analysis, with data weighting used for appropriate sampling method variation. These comprehensive sampling techniques ensured that the 2016 EDHS findings corresponded precisely to Ethiopia’s population health and nutritional status. Furthermore, this study examined the underweight among children aged 6 to 59 months. Finally, 9013 total children were included in the study. Diagram 1 shows that the sample size was determined conceptually in 2016 EDHS data.

Diagram 1

Sample size determination

Eligibility criteria

For this study, we used the 2016 EDHS, focusing on underweight children under five, and the eligibility criteria were more detailed and included the following aspects:

• I. Age: Children must be between 6 and 59 months. This age range is critical for assessing early childhood growth and nutritional status.

• II. Household selection: Children must belong to the households selected in the survey’s sampling process, which involved a portion of randomly chosen EAs. Informed Consent: The child’s mother provided informed consent for participation in the study, ensuring that ethical standards are met.

• III. Anthropometric measurements: Children must be available for anthropometric tests, including weight and height.

• IV. Data completeness: The analysis included only children whose measurements were complete and recorded in the survey database, ensuring the reliability of the data.

• V. Mother’s interview: Ideally, the mother or caregiver should have been interviewed to provide additional context regarding the child’s health, nutrition, and feeding practices.

Exclusion criteria

• I. Children whose mothers were not interviewed or who were not present during the survey were excluded from the analysis.

Study variables

In this study, we used underweight as the dependent variable, measured by weight-for-age. According to the World Health Organization (WHO) Child Growth standards, we classified underweight into three categories: normal (z-score ≥ -2), moderately underweight (-3 ≤ z-score ≤ -2), and severely underweight (z-score< -3) [24, 25]. Furthermore, this study used several individual and community variables, such as sociodemographic characteristics and health-related factors, to assess the underweight among under-five children.

Sociodemographic factors

• Sex refers to whether the child is male or female.

• Age groups refer to the age groups of children in months, e.g., those of children younger than 6 months, 6–11 months, 12–23 months, and more than 48 months.

• Twins are births, which refers to a mother who gives birth to babies, such as the first at birth, the first of multiples, and the second of multiples.

• Size of the child at birth refers to the weight and length of the newborn baby, e.g., small, average, or large.

• Birth order refers to the number of children a woman has given birth to, e.g., between one and three, between four and six, and more than seven.

• Place of residence refers to whether individuals live in urban or rural areas.

• Mother’s education refers to a mother’s education level, e.g., illiterate, primary, secondary, and higher education.

• Father’s education level refers to the level of the husband’s education, e.g., no education (illiterate), primary education, secondary education, and higher education.

• Household members refer to the number of individuals in a family, e.g., less than three, between four and seven, and more than eight members.

• Wealth index refers to the measure of socioeconomic status, e.g., poorest, medium, and richest.

• Employment status refers to whether women are employed or unemployed.

• Marital status refers to women’s marriage status, e.g., single, married, divorced, or widowed.

Health-related factors

• Incidence of diarrhea refers to whether children have had diarrhea in the past two weeks (yes or no).

• Fortified baby food refers to whether children are provided with food enriched with essential nutrients (yes or no).

• Delivery place refers to where women give birth, such as at home, in the government, in the private sector, or NGOs.

• Breastfeeding status refers to whether exclusive breastfeeding has been practised for the past six months (breastfed or not).

Statistical methods

This study used descriptive statistics for categorical variables, such as percentage or frequency. We also used the proportional odds model (POM) to determine the associations between dependent and independent variables. Ordinal logistic regression models are typically used to analyse ordered events [26]. In this study, the outcome variables reflected the severity of the underweight status of the children, categorised as normal, moderately underweight, and severely underweight. Therefore, it is essential to consider the ordered nature of these variables. Because this assumption is satisfied, we use the POM. Testing whether logit surfaces are parallel is vital. A non-significant test confirmed that it was indeed parallel, which permits interpreting the odds ratio (OR) as consistent across all potential cutoffs for children’s severity of underweight. Applying the cumulative logit model to the ordered outcome variable can represent the levels of the categorical response (normal, moderately, and severely underweight) and the explanatory variables [27]. A similar approach can be adopted when fitting a model with the relevant explanatory variables.

Model diagnostics and goodness of fit model

First, we need to compare the best model to be fitted. Model comparison aims to choose the simplest model that best fits the data. Common criteria include Akaike’s Information Criterion (AIC) and Bayesian Information Criterion (BIC). Models with lower AIC and BIC values are preferred as they indicate a better fit [28]. Moreover, several tests evaluate how well the model describes the outcome variable. The Likelihood Ratio Test (LRT) assesses the overall fit of the logistic regression model by comparing the likelihood of the entire model to a reduced model [28]. The Wald Test examines the significance of individual regression coefficients, with a significant result indicating that the variable should be included in the model. Furthermore, model diagnostics help identify and address outliers or influential observations that may affect the model’s conclusions. Leverage values measure an observation’s distance from others regarding the independent variables, with high values indicating potentially influential observations. Cook’s Distance assesses the overall influence of an observation on the regression coefficients, with values greater than one suggesting influential observations. Deviance residuals are used to identify potential outliers or misspecified cases, with standardised residuals outside the range of (-3, 3) indicating outliers [28].

Data analysis and management

The collected data were edited, cleaned, coded with values, and entered Statistical Analysis Software version 9.4. However, there is a problem of missing data. Managing missing data in a cross-sectional study is essential to maintaining the validity and reliability of our results. We used multiple imputations to handle missingness in our dataset. This includes replacing missing values with estimated ones. Multiple imputations are advanced handling missingness, which generates several plausible values based on the distribution of the observed data. This method is particularly beneficial as it accounts for the uncertainty surrounding the missing values, leading to more reliable statistical inferences. Descriptive statistics and proportional odds model (POM) were applied. The maximum likelihood estimation method was used to estimate values for the unknown parameters, which maximised the probability of obtaining the observed set of data (2016 EDHS). The edited data were exported to SAS, and the final analysis was done using it. A 5% level of significance was used for the conclusion.

Results

The study analysed data from 9013 children aged 6–59 months. Table 1 shows the prevalence of underweight among under-five children in Ethiopia. Out of the total number of under-five children, 74.70% were of normal weight (not underweight), 17.01% were moderately underweight, and 8.29% were severely underweight.

Table 1 Prevalence of underweight among under-five children (n = 9,013)

Additionally, Fig. 1 illustrates the prevalence of underweight across nine regions and two main administrative cities. We observed significant regional disparities in underweight prevalence. Somali, Afar, Oromia, Amhara, Tigray, and the Southern Nations, Nationalities, and Peoples’ Region (SNNPR) exhibit higher moderate and severe underweight rates. Conversely, Addis Ababa shows a lower prevalence of moderate and severe underweight.

Fig. 1

Prevalence of underweight among children across nine regions and two cities

Moreover, Fig. 2 shows the prevalence of underweight based on children’s sex, revealing significant disparities. Boys had a higher prevalence of underweight compared to girls. Specifically, 9.16% of boys were moderately underweight compared to 7.84% of girls, and 4.44% of boys were severely underweight compared to 3.85% of girls.

Fig. 2

Prevalence of underweight based on sex of children

Furthermore, Table 2 depicts the prevalence of underweight across different sociodemographic and health-related characteristics. For example, children from rural areas had a higher prevalence of being underweight than children from urban areas (e.g., moderate underweight: 18.5% vs. 10.2% and severe underweight: 9.3% vs. 3.7%). In addition, children whose mothers were illiterate had a higher prevalence of underweight compared to children whose mothers had higher education (e.g., moderate underweight: 19.6% vs. 5.3% and severe underweight: 10.8% vs. 1.9%). Moreover, children whose fathers were illiterate had a higher prevalence of underweight compared to children whose fathers had a higher education (e.g., moderate underweight: 20.9% vs. 10.8% vs. 5%). Besides, children with more than eight people in the household had a higher prevalence of underweight than those with less than three people (e.g., moderate underweight: 18.5% vs. 15.9% and severe underweight: 8.3% vs. 6.7%). Children older than 48 months had a higher prevalence of underweight than children younger than six months (e.g., moderate underweight: 21.6% vs. 7% and severe underweight: 8.3% vs. 4.7%).

Additionally, children who were second multiple of birth had a higher prevalence of underweight than those who were single at birth (e.g., moderate underweight: 21. % vs. 16.9% and severe underweight:16% vs. 8.1%). Children who were small at birth had a higher prevalence of underweight than those who were large (e.g., moderate underweight: 20.7% vs. 13.8% and severe underweight:12.3% vs. 5.9%).

Furthermore, children with a one to three birth order had a higher prevalence of being underweight than those with a birth order of more than seven (e.g., moderate underweight: 46.5% vs. 18.9% and severe underweight: 43.4% vs. 17.8%). Children who experienced diarrhea had a higher prevalence of underweight than those who had not experienced diarrhea (e.g., moderately underweight: 20.1% vs. 14.4% and severely underweight: 10.8% vs. 6.8%). Children who hadn’t eaten fortified food had a higher prevalence of underweight than those who had fortified food (e.g., moderate underweight: 18.6% vs. 16.16.8% and severe underweight: 10.7% vs. 8%). In addition, children born at home had a higher prevalence of underweight than those born in the government sector (clinics, hospitals or others) (e.g., moderately underweight: 21.1% vs. 15.9% and severely underweight: 13.2% vs. 7.9%). Children who had not been breastfed had a higher prevalence of underweight than those who had been breastfed (e.g., moderate underweight: 65.8% vs. 34.2%, and severe underweight: 68.4% vs. 31.6%).

Table 2 Prevalence of underweight and associated risk factors among under-five children

Proportional odds model assumptions (POM)

POM provides a comprehensive framework for examining the factors of underweight. We used this approach to assess the effect of various predictors on the likelihood of underweight categories. The score test for the POM yielded a p-value of 0.059 > 0.05, indicating that the assumption of parallel lines is valid for our data. Therefore, we used a POM to analyse the underweight categories (normal, moderately, and severely). This model allowed us to interpret the coefficients as ORs, simplifying the calculation of the ORs, confidence intervals (CI), and p-values. The two intercepts of the model (intercept 1 = 2.024 and intercept 2= -3.418) were used to compare the different underweight categories. Table 2 reveals the risk factors associated with being underweight. Sex, place of residence, mother and father’s educational level, the incidence of diarrhea in the last two weeks, employment status of parents, child’s age group, childbirth order, breastfeeding status, size at birth, and fortified baby food consumption were significantly associated with the prevalence of underweight, with a 5% level of significance. For instance, after controlling for all other factors, girls have 16% lower odds of being underweight than boys (OR = 0.84, 95% CI:0.763, 0.928). In addition, children living in rural areas have 43.5% higher odds of being underweight than those living in urban areas (OR = 1.43, 95% CI:1.209, 1.702). Moreover, the age of groups of children is a statistically significant factor for children underweight. For instance, children aged 48 + months are about 3.6 times more likely to be underweight than those under 6 months (OR = 3.639,95% CI:2.943, 4.499). In addition, children aged 12–23 months are about 2.9 times more likely to be underweight than those under 6 months (OR = 2.857, 95% CI:2.321, 3.518). Besides, children aged 24–47 months are about 3.4 times more likely to be underweight than those under 6 months (OR = 3.444, 95% CI:2.826, 4.198). Furthermore, children aged 6–11 months are about 1.6 times more likely to be underweight than those under 6 months (OR = 1.613, 95% CI:1.242, 2.094). This study revealed that children whose mothers have primary education have 25% lower odds of being underweight compared to those whose mothers are illiterate (OR = 0.750, 95% CI = 0.657, 0.855). In addition, children whose mothers have secondary education have 48.6% lower odds of being underweight compared to those whose mothers are illiterate (OR = 0.514, 95% CI: 0.386, 0.683). Moreover, children whose mothers have higher education have 63.9% lower odds of being underweight compared to those whose mothers are illiterate (OR = 0.361, 95% CI: 0.225, 0.579).

In addition, further education level is a factor in children underweight. Our result found that children whose fathers have primary education have 24% lower odds of being underweight compared to those whose fathers are illiterate (OR = 0.760, 95% CI: 0.676, 0.855). In addition, children whose fathers have secondary education have 18.9% lower odds of being underweight compared to those whose fathers are illiterate (OR = 0.811, 95% CI:0.664,0.991). Moreover, children whose fathers have a higher education have 22.9% lower odds of being underweight compared to those whose fathers are illiterate (OR = 0.771, 95% CI:0.607, 0.979). Besides, parents’ employment status is an associated factor in children underweight. For example, children who had employed parents had 35.7% higher odds of being underweight than children who had unemployed parents (OR = 1.357, 95% CI:1.071, 1.719). Also, children with a birth order of 4–6 have 16.1% higher odds of being underweight than those with a birth order of less than three (OR = 1.161, 95% CI:1.030, 1.309). Children who were never breastfed have 21% higher odds of being underweight than those who were breastfed (OR = 1.210, 95%, CI:1.082, 1.353). Children who are twins are first multiple have 81.7% higher odds of being underweight compared to those who are single at birth (OR = 1.817, 95% CI:0.217, 2.713). Children who are twins are second multiple have 71.5% higher odds of being underweight compared to those who are single at birth (OR = 1.71, 95% CI: 1.118, 2.630). Children of average size at birth have 30.5% higher odds of being underweight than those who were large (OR = 1.305, 95% CI:1.153, 1.476). Children who were small at birth have 95.8% higher odds of being underweight than those who were large (OR = 1.958, 95%CI: 1.718, 2.230). In addition, children who have experienced diarrhea in the last two weeks have 33.9% higher odds of being underweight than those who have not had diarrhea in the previous two weeks (OR = 1.339, 95% CI:1.154,1.553).

Table 3 Associated risk factors associated with underweight among under-five children

Goodness of fit the model

Table 4. revealed that goodness of fit the fitted models using the BIC, AIC, -2LOG L suggests that the intercepts with covariates model were better than the intercept-only model, e.g., intercept with covariates have the smallest BIC (12654.09) and AIC (12412.47). The overall model fitted assesses the contribution of each effect to the model. Furthermore, the score, Wald and Likelihood ratio test results for filtering the model’s goodness fit our data well.

Table 4 Goodness of fit the fitted models

Model diagnostics

To assess the adequacy of the fitted model, such as diagnosing the outliers, results, leverage values, and influential points, we used plots of standardised residuals, Pearson residuals, deviance residuals, Cook’s distance, and leverage values with predicted probability. In Fig. 3a, observations are a little far away from the others; these are not influential since all of Cook’s influence statistics are less than one. Figure 3b shows that few observations lie far away from the rest, but all absolute deviance residuals are less than three. Thus, there is no lack of fit. In Fig. 4b, the standard residuals do not influence the model; standard residuals are less than three. In addition, Fig. 4a shows plots of leverage values versus the predicted probabilities of all observations, revealing that leverage values are less than one. Therefore, there are no outliers.

Fig. 3

(a) Plot of Cook’s influence distance statistics with predicted probabilities. (b): Plot of deviance residual value with predicted probabilities

Fig. 4

(a) Plot of leverage value with predicted probabilities. (b). Plot of standard residuals with predicted probabilities

Discussion

Of the 9,013 children under five, 74.7% were normal, 17.01% were moderately underweight, and 8.29% were severely underweight, consistent with the 2016 EDHS report [13, 14]. POM is appropriate for this data analysis, as the test of parallel lines is not significant at the 5% level (p-value = 0.0528), validating the model. POM fits well due to the proportionality assumption. We employed the PROC LOGISTIC statement in a statistical analysis system to examine factors underweight among under-five children, analysing sociodemographic and health-related characteristics. This study has identified significant risk factors for underweight such as sex of children, place of residence, parents’ educational level, the incidence of diarrhea incidence in the last two weeks, employment status, child age groups, twin status, size order at birth, breastfeeding status, children size at birth, and fortified baby food. However, delivery of place, marital status, household income, and family size did not significantly affect the status of underweight among under-five children.

The findings revealed sex differences in the prevalence of underweight, with females being 0.269 times less likely to experience underweight than males, evidenced by a β coefficient of -0.17, an OR = 0.84, and a p-value < 0.005. This difference may stem from boys’ higher susceptibility to environmental stress, supported by previous studies [13, 29,30,31,32]. Understanding sex differences in underweight is crucial for prevention and treatment services.

This study also found that rural residence is associated with a higher prevalence of underweight. It also highlights the need for targeted interventions in rural areas. Children from rural areas had a 1.44 times greater likelihood of being underweight compared to children from urban areas, consistent with previous studies [13, 31]. This suggests that rural areas need improved living standards and prospects to enhance children’s health and nutritional outcomes. These studies emphasise the need for targeted strategies to improve children’s nutritional status and overall health in rural areas. Therefore, investigating factors contributing to health access and utilisation disparities beyond merely rural and urban divides is crucial, as this will inform more effective solutions. Moreover, the age of children is a statistically significant factor for being underweight. For instance, children aged 48 + months are about 3.6 times more likely to be underweight than those under 6 months. Similarly, children aged 12–23 months are about 2.9 times more likely to be underweight, and those aged 24–47 months are about 3.4 times more likely to be underweight compared to children under 6 months. Additionally, children aged 6–11 months are about 1.6 times more likely to be underweight than those under 6 months. These findings are in line with previous research [13, 30, 33].

The increased likelihood of being underweight among older children could be attributed to several factors. As children grow older, they may face a higher risk of infections and illnesses, which can affect their nutritional status. Children who are twins or part of multiple births have 71.5% higher odds of being underweight compared to those who are single at birth. This finding is consistent with previous research [13, 32]. The increased risk of underweight status among twins and multiples can be attributed to several factors. Multiple births often result in preterm delivery and lower birth weights, which can predispose children to undernutrition. The current study revealed that children whose mothers have primary education have 25% lower odds of being underweight compared to those whose mothers are illiterate. Furthermore, children whose mothers have secondary education have 48.6% lower odds of being underweight compared to those whose mothers are illiterate. Additionally, children whose mothers have higher education have 63.9% lower odds of being underweight compared to those whose mothers are illiterate. These findings align with previous studies [13, 30, 32, 34, 35]. The results highlight the significant impact of maternal education on child nutrition.

Moreover, children whose fathers have a higher education have 22.9% lower odds of being underweight compared to those whose fathers are illiterate. This result is in line with previous studies [34]. The influence of paternal education on child nutrition is significant. Fathers with higher education levels are likely to have better economic opportunities, which can improve the family’s overall living conditions and access to nutritious food. In addition, children with employed parents have 35.7% higher odds of being underweight compared to those with unemployed parents. These findings are consistent with previous studies [13, 36]. This may be due to time constraints that limit the ability to provide nutritious meals and adequate care. Unemployment, while linked to poverty and food insecurity, also allows more time for childcare. Supporting working parents through these measures can improve children’s nutritional status and overall well-being. Additionally, children with a birth order of 4–6 have 16.1% higher odds of being underweight compared to those with a birth order of less than three. This finding is consistent with a study conducted by Habyarimana et al. It suggests that higher birth order may be linked to reduced access to resources or attention, which can negatively impact nutritional status [13, 32]. Size at birth is a significant predictor of underweight among under-five children. Low birth size is a critical risk factor for child mortality and morbidity but is often overlooked as undernutrition. Children with average size at birth have a 1.31 times higher risk of being underweight than larger-sized children, while small-sized children have almost twice the risk. These findings are consistent with previous studies [13, 32, 37]. This association could be attributed to the impact of maternal malnutrition during pregnancy, which may result in smaller birth sizes and contribute to poorer postnatal growth. Additionally, children born smaller might be more susceptible to infections, have reduced immunity, and have slower growth trajectories, increasing their risk of undernutrition in early childhood. The current study also found that children who consumed fortified baby food were 15% less likely to be underweight compared to those who did not consume fortified baby food. This finding is consistent with the previous studies [13, 33, 37]. Fortified baby food is designed to provide essential nutrients that may be lacking in a child’s diet, particularly in regions where food insecurity or malnutrition is prevalent. The addition of vitamins and minerals to baby food can significantly improve a child’s nutritional status, supporting healthy growth and development.

Additionally, children who were never breastfed have 21% higher odds of being underweight than those who were breastfed. Breastfeeding provides essential nutrients and antibodies that support healthy growth and development, which may explain this association. These findings are consistent with previous studies [13, 30, 38], reinforcing the importance of breastfeeding and resource allocation within larger families to prevent underweight. Healthcare professionals should develop strategies to raise awareness and improve breastfeeding adherence among mothers. Furthermore, children who have experienced diarrhea in the last two weeks have 33.9% higher odds of being underweight compared to those who have not had diarrhea during the same period. This finding is highly significant (p < 0.001) and aligns with previous studies [13, 31, 39, 40], which has shown that diarrhea negatively impacts children’s nutritional status. Persistent diarrhea can lead to severe nutritional deficiencies and dehydration, necessitating medical attention for affected children. Additionally, undernutrition among children under five varies with age, highlighting the need for age-specific interventions to address this issue effectively.

Limitations and strengths

This study has several limitations and strengths. It used nationally representative data collected from nine regions and two main cities to provide generalised information on children’s underweight status. These results led to significant conclusions that were applied to the samples. This study is limited by its cross-sectional design, which does not allow for causal inferences. Data is collected simultaneously, making establishing temporal relationships between variables challenging. To address this limitation, future research should consider longitudinal follow-ups, which involve collecting data from the same subjects multiple times. This approach can help identify changes over time and provide more substantial evidence for causal relationships. Additionally, reliance on self-reported data may introduce reporting bias. Furthermore, the lack of spatial analysis limits the ability to understand geographic disparities in underweight, which is essential for effective public health planning. Future research should also focus on geospatial and temporal changes to identify areas with a high incidence of underweight in the district in rural areas. Geospatial analysis can reveal spatial patterns and hotspots of underweight, which are crucial for targeted interventions. Temporal analysis can help track changes over time, providing insights into the effectiveness of policies and programs.

Policy implications

In Ethiopia, the government opened the Seqota guidelines (declaration), a great effort to reduce the burden of undernutrition among under-five children by 2030 [41]. Despite this interesting effect, the current study revealed that the prevalence of being underweight is greater and public health consent in the county among under-five children. The associated risk factors can explain this result. The results of this study are essential for policy, program and practice at the regional level, consistent with the second goal of sustainable development dedicated to nutrition [42] and infant, young children and maternal nutrition are the six global targets to be achieved between 2012 and 2025 by the World Health Assembly’s [43]. Therefore, this finding suggests that policymakers and programs should emphasise reducing the burden of undernutrition to achieve the second goal of sustainable development and World Health Assembly targets by implementing specific interventions and nutrition sensitivity that can address the associated risk factors. Furthermore, underweight is a public health concern that requires consistent and targeted interventions in rural areas of district, zonal, and regional health offices and other government and non-government stakeholders to achieve strong commitments to reduce the burden of undernutrition among under-five children. Furthermore, there is a critical need to design and enhance educational programs for parents. These programs should focus on raising awareness about the importance of balanced diets, proper feeding practices, and the long-term impacts of undernutrition on children’s health and development.

Conclusions

Being underweight among children under five age years old is a public health concern in Ethiopia. Therefore, improving maternal educational levels and targeting nutritional interventions in rural areas could significantly reduce undernutrition rates. Efforts should also focus on improving sanitation and personal hygiene to prevent the spreading of diseases such as diarrhea. Integrating maternal nutrition health programs into the community management of chronic undernutrition programs during maternal health services is essential for providing holistic care and improving child health programs. Therefore, there must be a coordinated effort at all levels (e.g., federal, regional, and zonal) to enhance interventions improving parents’ feeding practices for their children.