Causal Effects of Antenatal Care (ANC) on Child Malnutrition: A Machine Learning Approach in Ethiopia and Rwanda

Causal Effects of Antenatal Care (ANC) on Child Malnutrition: A Machine Learning Approach in Ethiopia and Rwanda | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Causal Effects of Antenatal Care (ANC) on Child Malnutrition: A Machine Learning Approach in Ethiopia and Rwanda Alehegn Moges Tessema, Temesgen Zewotir, Richard Kabanda This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6131494/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 28 Nov, 2025 Read the published version in BMC Nutrition → Version 1 posted 12 You are reading this latest preprint version Abstract Malnutrition among children under five remains a critical public health challenge in Ethiopia and Rwanda, with stunting rates of 37% and 33%, respectively. This study examines the causal effects of antenatal care (ANC) on child malnutrition, measured using the Composite Index of Anthropometric Failure (CIAF), leveraging data from the Demographic and Health Surveys (DHS). Using machine learning techniques, specifically the Causal Forest model, the study estimates both overall and subgroup-specific impacts of ANC while adjusting for socio-demographic and environmental factors. Results show that ANC significantly reduces child malnutrition, with an overall reduction of 2.63 percentage points. However, the impact varies across groups: urban residents, educated mothers, and wealthier households benefit more, highlighting disparities. For example, urban areas see a 7.05 percentage point reduction compared to 3.06 in rural areas. Complementary factors like breastfeeding, clean water access, and improved sanitation further enhance ANC’s effectiveness. However, the impact of ANC has declined over time, underscoring the need for program improvements. By employing the Causal Forest model and validating results through 5-fold cross-validation, this study provides robust, data-driven insights into ANC’s heterogeneous effects. These findings highlight the importance of addressing structural inequities and integrating ANC with broader health interventions. The use of machine learning offers policymakers precise, actionable evidence to design targeted, equitable strategies for improving child nutrition in Ethiopia, Rwanda, and similar contexts. Antenatal Care (ANC) Child Malnutrition Composite Index of Anthropometric Failure (CIAF) Causal Forest Model Machine Learning Heterogeneous Treatment Effects 5-Fold Cross-Validation Figures Figure 1 Figure 2 1. INTRODUCTION Malnutrition remains a critical public health challenge in Ethiopia and Rwanda, particularly among children under five, contributing to high rates of illness, death, stunted physical growth, impaired cognitive development, and reduced overall well-being [ 1 ] [ 2 ]. Despite ongoing efforts to address malnutrition, progress has been inconsistent. In Ethiopia, nearly 37% of children under five are stunted, while in Rwanda, around 33% face similar issues, reflecting widespread chronic undernutrition [ 3 ] [ 4 ]. These persistent challenges highlight the urgent need for targeted and effective interventions that address both the immediate and underlying causes of malnutrition. Beyond its health impacts, malnutrition has significant economic and social consequences, perpetuating poverty, reducing productivity, and hindering national development [ 5 ] [ 6 ]. To design effective interventions, a clear understanding of the multiple causes of malnutrition is essential. Immediate causes include poor dietary intake and frequent illnesses such as diarrhea, while underlying factors like inadequate maternal health, low household income, and limited access to clean water and sanitation worsen the problem [ 7 ] [ 8 ]. Broader systemic issues, including food insecurity, unequal access to healthcare, and educational gaps, further increase the risk of malnutrition, particularly among vulnerable groups [ 9 ]. Addressing these interconnected factors requires interventions that go beyond immediate needs to tackle the root causes of malnutrition. Existing research has identified associations between certain interventions—such as maternal education, antenatal care, and improved water and sanitation—and better child nutrition outcomes [ 10 ] [ 11 ]. However, these studies often focus on correlations rather than causation. Correlation-based findings do not account for other influencing factors like socioeconomic status, geographic location, or household characteristics, which can distort the observed relationships [ 12 ] [ 13 ]. This limitation makes it difficult for policymakers to design evidence-based strategies, as they lack clear evidence on which interventions directly lead to improved outcomes. This study aims to addresses this gap by leveraging machine learning-based causal inference methods to evaluate the causal effects of specific nutrition and health interventions on child nutrition outcomes in Ethiopia and Rwanda. Unlike traditional approaches that rely on correlation, this research employs machine learning techniques such as causal forests, doubly robust estimation, and propensity score matching to isolate the causal impact of interventions like maternal education, antenatal care, and improved water access. These methods are particularly well-suited for handling complex, high-dimensional data and adjusting for a wide range of confounding variables, providing more accurate estimates of the Average Treatment Effects (ATE) and Conditional Average Treatment Effects (CATE). Causal forests, a machine learning method, are used to identify heterogeneous treatment effects, allowing the study to uncover how the impact of interventions varies across different subgroups, such as rural versus urban populations or varying wealth levels. For example, the analysis might reveal that maternal education has a stronger effect in urban areas with better healthcare infrastructure, while improved water access is more impactful in rural regions with limited sanitation facilities. By quantifying these variations, the study provides policymakers with actionable, context-specific evidence to design targeted interventions. The use of machine learning in causal inference represents a significant advancement in malnutrition research. These methods not only improve the precision of causal estimates but also enable the identification of nuanced patterns in the data that traditional statistical approaches might miss. For instance, machine learning algorithms can automatically detect interactions between variables or non-linear relationships, offering deeper insights into how interventions interact with contextual factors. This level of detail ensures that resources are allocated equitably and effectively, maximizing the impact of interventions. 2. METHODS 2.1. The Data The study utilized Demographic and Health Surveys (DHS) datasets from Ethiopia and Rwanda to analyze the impact of antenatal care (ANC) on child malnutrition, measured using the Composite Index of Anthropometric Failure (CIAF). The datasets included key socio-demographic variables such as residence, maternal education, wealth index, and survey year to enable a comprehensive analysis. Data preprocessing involved encoding categorical variables and standardizing continuous variables to ensure compatibility with machine learning algorithms. Propensity score methods were employed to balance treated and untreated groups, mitigating selection bias and enhancing the validity of causal estimates. Each dataset includes crucial information on household demographics, maternal and child health, socio-economic factors, and access to essential services such as healthcare, water, and sanitation. Independent variables such as wealth index, maternal education, access to healthcare, and environmental conditions were selected for their relevance in predicting child malnutrition based on prior studies. The details of the variable definitions are presented in Table 1 below. Table 1 The description of the response variable and the respective covariates included in the model Variable Categories Description Composite Index of Anthropometric Failure (CIAF) No, Yes \(\:\left\{\begin{array}{c}1\:\:if\:the\:child\:has\:some\:form\:of\:malnutrition\\\:0\:if\:the\:child\:has\:no\:any\:form\:of\:malnutrition\end{array}\right.\) Place of Residence Urban, Rural Urban vs. rural living environments. Mother’s Education No education, Primary, Secondary, Higher Education levels of the mother: no formal education to higher education. Sex of Child Male, Female The biological sex of the child. Vitamin A No, Yes Whether the child received Vitamin A supplementation. Wealth Quintile Poorest, Poorer, Middle, Richer, Richest The economic status of the household, ranked from poorest to richest. Country Ethiopia, Rwanda Country where the data was collected. Year 2005, 2010, 2016 The year of data collection. Source of Drinking Water Unimproved, Improved Whether the household uses unimproved or improved drinking water sources. Toilet Facilities Unimproved, Improved Whether the household has unimproved or improved sanitation facilities. Took Iron Tablets No, Yes Whether the mother took iron tablets during pregnancy. Place of Delivery No, Yes Whether the child was born in a healthcare facility. First Born No, Yes Whether the child is the firstborn. Types of Birth Single, Twins Whether the child was part of a single or multiple birth (e.g., twins). Breastfeeding No, Yes Whether the child was breastfed. ANC (Antenatal Care) No, Yes Whether the mother received antenatal care during pregnancy. Age of Mother 15–24, 25–34, 35–49 Age range of the mother at the time of the child's birth. Size of Child at Birth Above average, Average, Smaller than average The perceived size of the child at birth. BMI of Mother < 18.5 kg/m² (Underweight), 18.5–24.9 kg/m² (Normal weight), ≥ 25 kg/m² (Overweight) Body Mass Index (BMI) of the mother, categorized as underweight, normal, or overweight. Household Size < 4, 5–9, 10+ Number of people living in the household. Number of Under-Five Children 0–1, 2, 3 or more The number of children under the age of five in the household. 2.2. Analytical Framework This study employs a robust analytical framework to assess the impact of Antenatal Care (ANC) on child malnutrition. By integrating Average Treatment Effect (ATE), Conditional Average Treatment Effect (CATE), and a causal forest model, the framework provides a comprehensive and data-driven understanding of ANC’s effectiveness across different populations. The analytical framework outlined in Fig. 1 supports evidence-based decision-making, enabling policymakers to develop targeted interventions for improving maternal and child health. The analysis begins by estimating the Average Treatment Effect (ATE), which measures the overall impact of Antenatal Care (ANC) on reducing child malnutrition at the population level. The ATE is calculated by comparing outcomes between those who received ANC and those who did not, while controlling for confounding factors such as socioeconomic status and healthcare access. This step provides a broad estimate of ANC’s effectiveness, establishing a foundation for understanding its role in reducing child malnutrition. To gain deeper insights, the study conducts a subgroup analysis using Conditional Average Treatment Effects (CATE). This approach examines how ANC’s effectiveness varies across socio-demographic groups, including residence (urban vs. rural), maternal education (educated vs. non-educated), household wealth (richer vs. poorer), country-specific differences, and time trends. Additional factors such as breastfeeding practices, access to clean water, improved sanitation, and vitamin A supplementation are also considered. The CATE analysis identifies which groups benefit the most or least from ANC, revealing disparities that may require targeted interventions. For example, urban residents and educated mothers experience greater improvements in child nutrition compared to rural residents and non-educated mothers. These findings help policymakers design ANC programs tailored to the needs of underserved populations. A key component of the analysis is the use of a causal forest model, a machine learning technique that estimates how ANC’s effects vary across individuals or groups. Unlike traditional methods, this model captures variations based on factors such as location, education, and wealth, providing more detailed and accurate estimates of ANC’s effectiveness. By combining multiple decision trees, the causal forest model detects complex patterns in the data, offering a nuanced understanding of how different groups respond to ANC. This enables policymakers to make informed decisions to improve program outcomes. To ensure reliability, the study employs two validation techniques: 5-fold cross-validation and analysis of treatment effect distribution. In 5-fold cross-validation, the dataset is divided into five parts, with the model trained on four parts and tested on the fifth, repeating the process five times. The results are averaged to provide a stable estimate of ANC’s impact, ensuring the findings are not dependent on a specific dataset split. The second method involves analyzing the distribution of treatment effects, including generating a histogram and calculating descriptive statistics such as mean, standard deviation, and range. A low variation in treatment effects indicates stable and reliable estimates, confirming the model’s robustness. Finally, the study interprets the results of the ATE, CATE, and validation analyses in the context of socio-demographic and environmental factors. This interpretation provides actionable insights for policymakers, highlighting disparities in ANC’s effectiveness and identifying populations that may require additional support. For instance, if rural areas or poorer households benefit less from ANC, efforts can be directed toward improving healthcare access in these communities. By understanding how ANC’s impact varies across groups, stakeholders can design programs that allocate resources where they are most needed, leading to more effective interventions that reduce child malnutrition and improve overall health outcomes. 3. RESULT 3.1. Characteristics of the study participants and prevalence of Composite Index of Anthropometric Failure (CIAF) This analysis included data from 32,099 children aged 0–59 months, with 15,726 (49.01%) experiencing at least one form of undernutrition, as measured by the Composite Index of Anthropometric Failure (CIAF). The prevalence of CIAF was analyzed across various child, maternal, and household-level factors. Higher rates of CIAF were observed among children of mothers with no formal education, those living in rural areas, and households in the lowest wealth quintile. Children from homes with unimproved water sources and sanitation, and whose mothers lacked access to antenatal care or iron supplements, were also more likely to experience undernutrition. Significant covariates identified through chi-square tests were utilized to develop machine learning models on the dataset (Table 2 ). Table 2 Prevalence of CIAF Across Socio-Demographic and Environmental Variables Variables Categories Composite Index of Anthropometric Failure Chi-square Sig. Nourished (%) CIAF (%) Place of residence Urban 66.2% 33.8% 655.498 .000 * Rural 47.1% 52.9% Mother’s education No education 45.4% 54.6% 783.057 .000 * Primary 52.8% 47.2% Secondary 70.6% 29.4% Higher 82.3% 17.7% Sex of child Male 48.3% 51.7% 56.034 .000 * Female 52.5% 47.5% Vitamin A No 49.9% 50.1% 15.361 .000 * Yes 53.1% 46.9% Wealth quintile Poorest 43.4% 56.6% 935.602 .000 * Poorer 45.0% 55.0% Middle 48.5% 51.5% Richer 52.0% 48.0% Richest 67.1% 32.9% Country Ethiopia 50.2% 49.8% 0.654 0.419 Rwanda 50.7% 49.3% Year 2005 41.4% 58.6% 451.776 .000 * 2010 50.2% 49.8% 2016 56.9% 43.1% Source of drinking water Unimproved 45.4% 54.6% 234.443 .000 * Improved 54.1% 45.9% Toilet facilities Unimproved 47.2% 52.8% 256.845 .000 * Improved 56.6% 43.4% Took iron tablets No 47.8% 52.2% 149.896 .000 * Yes 54.8% 45.2% Place of Delivery Category No 45.4% 54.6% 693.336 .000 * Yes 61.3% 38.7% First born No 49.1% 50.9% 80.697 .000 * Yes 55.4% 44.6% Types of birth Single 50.6% 49.4% 43.053 .000 * Twines 38.2% 61.8% Breastfeeding No 51.3% 48.7% 0.280 0.597 Yes 50.3% 49.7% ANC No 43.8% 56.2% 337.177 .000 * Yes 54.4% 45.6% Age of mother 15–24 51.7% 48.3% 24.154 .000 * 25–34 50.9% 49.1% 35–49 48.1% 51.9% Size of the child at birth Above average 55.3% 44.7% 280.695 .000 * Average 50.8% 49.2% Smaller than average 42.9% 57.1% BMI < 18.5 kg/m² (Underweight) 43.9% 56.1% 443.764 .000 * 18.5–24.9 kg/m² (Normal weight) 49.8% 50.2% ≥ 25 kg/m² (Overweight) 66.2% 33.8% Household size < 4 52.6% 47.4% 28.507 .000 * 5–9 49.3% 50.7% 10+ 51.8% 48.2% Number of under five children in the household 0–1 52.9% 47.1% 66.859 .000 * 2 48.0% 52.0% 3 or more 51.7% 48.3% Place of residence, maternal education, wealth, and access to healthcare are significant factors in child nutrition. Urban children show a significantly higher nourishment rate (66.2%) compared to rural children (47.1%) (χ² = 655.498, p < 0.001). Similarly, children of mothers with no education have higher malnutrition rates (54.6%), while those with highly educated mothers show lower rates (17.7%) (χ² = 783.057, p < 0.001). Wealthier households also exhibit better child nutrition, with the poorest having the highest CIAF rate (56.6%) and the richest the lowest (32.9%) (χ² = 935.602, p < 0.001). Male children are slightly more likely to experience malnutrition (51.7%) compared to females (47.5%) (χ² = 56.034, p < 0.001). Vitamin A supplementation is associated with improved nutrition (46.9% vs. 50.1%, χ² = 15.361, p < 0.001), and access to improved water sources and sanitation significantly reduces malnutrition (45.9% for improved water vs. 54.6% for unimproved, χ² = 234.443, p < 0.001; 43.4% for improved sanitation vs. 52.8%, χ² = 256.845, p < 0.001). Children born in healthcare facilities have better nutritional outcomes (38.7%) compared to those born outside such facilities (54.6%) (χ² = 693.336, p < 0.001). First-born children tend to be better nourished (44.6% vs. 50.9% for later-born, χ² = 80.697, p < 0.001), while children from multiple births are more vulnerable to malnutrition (61.8% vs. 49.4%, χ² = 43.053, p < 0.001). Antenatal care (ANC) is a protective factor, with children of mothers who received ANC showing lower malnutrition rates (45.6% vs. 56.2%, χ² = 337.177, p < 0.001). Maternal age and BMI are also important; younger mothers (ages 15–24) have children with better nutrition (48.3% vs. 51.9%, χ² = 24.154, p < 0.001), and underweight mothers’ children face higher malnutrition rates (56.1% vs. 33.8% for overweight mothers, χ² = 443.764, p < 0.001). Generally, the multifaceted nature of CIAF prevalence underscores the importance of addressing socio-economic inequalities, enhancing access to healthcare and sanitation services, and promoting maternal and child health practices to ensure better child health outcomes. The statistically significant associations identified in this analysis offer valuable insights for designing targeted interventions aimed at reducing child illness and malnutrition, ultimately contributing to improved child well-being and development. It is in light of these results that the selected features are chosen for multivariate analysis and inclusion in the machine learning algorithm for further analysis. 3.2. Estimation of Treatment Effects In this section, we present the estimation of treatment effects, including the Average Treatment Effect (ATE) and Conditional Average Treatment Effect (CATE). These estimates provide insight into the overall and subgroup-specific impacts of Antenatal Care (ANC) on child malnutrition. However, before drawing conclusions from these estimates, it is essential to assess the reliability of the model used for estimation. The interpretation of these results will be provided in the Interpretation of Results Section after confirming the model’s robustness and stability through validation. 3.2.1. Average Treatment Effect (ATE) The overall impact of ANC on child malnutrition was estimated using the ATE. The results indicate that ANC significantly reduces the likelihood of child malnutrition. The ATE was found to be -0.0263, meaning that, on average, ANC lowers the probability of malnutrition by 2.63 percentage points across the entire population. This provides a broad measure of ANC’s effectiveness in improving child nutrition outcomes. 3.2.2. Conditional Average Treatment Effect (CATE) To explore how ANC’s effectiveness varies across different population groups, we estimate the Conditional Average Treatment Effect (CATE) across key socio-demographic and contextual factors. These factors include residence (urban vs. rural), maternal education, wealth status, country-specific differences, and temporal trends. Additionally, we assess the influence of complementary health and environmental factors such as breastfeeding, access to improved water, sanitation, and vitamin A supplementation. The results presented in Table 3 shows a detailed and insightful look at how Antenatal Care (ANC) affects child malnutrition across different groups. Using Conditional Average Treatment Effect (CATE) analysis, the study moves beyond average effects to explore how ANC’s effectiveness varies based on factors like where people live (urban vs. rural), their education level, wealth, country, and other health and environmental conditions. These findings are crucial for policymakers and healthcare providers because they reveal disparities in ANC’s impact and pinpoint groups that may need more targeted support. The table not only shows how much ANC reduces malnutrition but also explains why these differences exist, helping to shape fairer and more effective healthcare strategies. These findings will be interpreted in details under result interpretation section of the article. Table 3: Subgroup Analysis Results: A Conditional Average Treatment Effect Analysis Subgroup CATE (Conditional Average Treatment Effect) 1 Cross-Validated ATE Interpretation 2 Residence: Urban -0.0705 -0.0306 Urban residents experienced a significant reduction in malnutrition (7.05%) compared to rural areas. Residence: Rural -0.0306 -0.0301 Rural residents experienced a smaller but consistent reduction in malnutrition (3.06%). Education: Educated -0.0394 -0.0312 Educated mothers benefited more from ANC, reducing malnutrition by 3.94%. Education: Non-Educated -0.0263 -0.0303 Non-educated mothers experienced a reduction of 2.63% in malnutrition likelihood. Wealth Index: Richer -0.0299 -0.0313 Wealthier households saw a reduction of 2.99% in malnutrition, benefiting significantly from ANC. Wealth Index: Poorer -0.0214 -0.0301 Poorer households experienced a smaller reduction (2.14%) due to economic barriers. Country: Ethiopia -0.0487 -0.0309 Ethiopia showed a stronger ANC effect, reducing malnutrition by 4.87%. Country: Rwanda -0.0313 -0.0301 Rwanda's reduction (3.13%) was smaller, possibly due to differences in ANC programs. Year: 2005 -0.0508 -0.0308 ANC had its highest impact in 2005, reducing malnutrition by 5.08%. Year: 2010 -0.0433 -0.0306 In 2010, ANC reduced malnutrition by 4.33%, showing a declining trend over time. Year: 2015/16 -0.0316 -0.0301 By 2015/16, ANC's impact declined to 3.16%, suggesting diminishing marginal returns. Breastfeeding -0.0379 -0.0301 Breastfeeding households experienced a 3.79% reduction, showing synergy with ANC. Improved Water Source -0.0403 -0.0299 Households with access to improved water saw a 4.03% reduction in malnutrition. Improved Sanitation -0.0433 -0.0313 Improved sanitation facilities enhanced ANC’s impact, reducing malnutrition by 4.33%. Vitamin A Supplementation -0.0509 -0.0315 Households with vitamin A supplementation had the largest impact, reducing malnutrition by 5.09%. 1 CATE Estimate: Conditional Average Treatment Effect (the estimated impact of ANC on child malnutrition for each subgroup). 2 Interpretation: Describes the effectiveness of ANC in reducing child malnutrition for the specific subgroup. 3.3. Model Performance Evaluating the performance of the causal forest model is crucial to ensure the reliability and robustness of the findings. This section presents two key assessments of the model's performance: Cross-Validated Results and Distribution of Treatment Effects. These metrics provide insights into the model's ability to consistently and accurately estimate treatment effects across the dataset. 3.3.1. Cross-Validated Results Cross-validation is a widely used technique for evaluating the generalizability of a model’s findings. In this analysis, a 5-fold cross-validation procedure was implemented to assess the reliability of the causal forest model. The dataset was divided into five subsets, where the model was iteratively trained on four subsets and tested on the remaining subset. This process ensured that each observation in the dataset was used for both training and testing. The Average Treatment Effect (ATE) was calculated for each fold, and the cross-validated ATE was obtained by averaging the ATE values across all folds. The cross-validated ATE was estimated at -0.0311, indicating that antenatal care (ANC) reduces the likelihood of child malnutrition by 3.11 percentage points on average across all folds. This result aligns closely with the overall ATE, calculated using the full dataset, which was − 0.0263. The close alignment between the cross-validated ATE and the overall ATE demonstrates that the model provides consistent estimates of the treatment effect across different subsets of the data. The consistency of ATE values across folds highlights the robustness of the causal forest model. The small variations observed between folds suggest that the model is not overfitting and that its findings are generalizable to different subsets of the data. This reliability ensures that the estimated treatment effects are credible and can be confidently applied to policy and programmatic decision-making. By confirming the model’s stability and robustness, cross-validation strengthens the validity of the analysis and the insights derived from it. 3.3.2. Distribution of Treatment Effects The distribution of treatment effects was analyzed to evaluate how the estimated effects vary across individual observations in the dataset. A histogram was generated to visualize the spread of treatment effects, providing a clear picture of their variability and consistency. This analysis is crucial for understanding the model's ability to generate reliable and stable estimates of the impact of antenatal care (ANC) on child malnutrition. The histogram revealed a symmetric, bell-shaped distribution centered around the Average Treatment Effect (ATE) value of -0.0263. Most treatment effects clustered close to the mean, indicating low variability in the model's estimates. Notably, no significant outliers were observed, suggesting that the model does not produce extreme or erratic estimates, further reinforcing its reliability. (Fig. 2 ) The descriptive statistics for treatment effects provide additional insights into the model’s performance. The mean treatment effect (ATE) was calculated to be -0.0263, confirming the overall impact of ANC in reducing malnutrition by 2.63 percentage points. The standard deviation of 0.0047 reflects minimal variability in the estimates. The minimum and maximum treatment effects were − 0.0381 and − 0.0152, respectively, showing that the treatment effects across observations are tightly distributed around the mean. The symmetric, bell-shaped distribution indicates that the model consistently captures the causal relationships between ANC and child malnutrition across observations. The low standard deviation and absence of extreme values confirm that the treatment effect estimates are both stable and reliable. This consistency across observations enhances the credibility of the model's findings and ensures that the results are robust enough to inform policy and programmatic decisions with confidence. The combined results of these metrics—cross-validated ATE and the distribution of treatment effects—validate the causal forest model’s suitability for this analysis. The model’s ability to produce consistent and stable estimates ensures its reliability for policy and programmatic decision-making. These findings provide confidence that the model accurately captures the causal relationship between ANC and child malnutrition, enabling stakeholders to draw actionable insights for targeted interventions. 3.4. Interpretation of Results The estimates presented in Section 3.2 are now interpreted in light of the model validation results from Section 3.3. The validation process confirmed that our causal forest model is robust, stable, and reliable, meaning that the estimated treatment effects can be confidently used for policy recommendations. The key findings from the estimation of treatment effects presented in Table 3 above distinguishes between confounding factors, which provide context for disparities, and intervention-related factors, which can be leveraged to enhance ANC's effectiveness. 3.4.1. Confounding Factors and Contextual Disparities Residence: Urban vs. Rural The analysis reveals a significant disparity in ANC’s effectiveness between urban and rural areas. Urban residents experience a 7.05 percentage point reduction in child malnutrition (CATE: -0.0705), while rural residents see a smaller reduction of 3.06 percentage points (CATE: -0.0306). This gap is likely influenced by several factors, including better healthcare access, service quality, and awareness in urban settings. Urban populations are more likely to receive timely ANC services, benefit from skilled health professionals, and have access to necessary medications and diagnostic tools. Education: Educated vs. Non-Educated Mothers Maternal education plays a critical role in ANC’s effectiveness. Educated mothers experience a 3.94 percentage point reduction in child malnutrition (CATE: -0.0394), compared to 2.63 percentage points (CATE: -0.0263) for non-educated mothers. This suggests that education enhances a mother’s ability to understand, seek, and utilize ANC services effectively. Educated mothers are more likely to attend ANC visits regularly, follow healthcare recommendations on nutrition and supplementation, and recognize early signs of malnutrition or pregnancy complications, leading to better health outcomes for their children. Wealth Index: Richer vs. Poorer Households Economic disparities significantly influence ANC’s effectiveness. Wealthier households experience a 2.99 percentage point reduction in malnutrition (CATE: -0.0299), while poorer households see a smaller 2.14 percentage point reduction (CATE: -0.0214). Wealthier families tend to have better healthcare access, improved dietary diversity, and lower financial stress, enabling them to comply with ANC recommendations. They are also more likely to access private healthcare services, which may offer higher-quality ANC. Country: Ethiopia vs. Rwanda The study finds that ANC’s impact on malnutrition reduction differs between Ethiopia (4.87 percentage points, CATE: -0.0487) and Rwanda (3.13 percentage points, CATE: -0.0313). These differences may reflect health system variations, ANC program implementation differences, or baseline malnutrition rates in each country. Ethiopia’s stronger impact could be due to higher initial malnutrition rates, creating greater room for improvement, or more intensive ANC interventions targeted at malnutrition. Trends Over Time: 2005–2015/16 The effectiveness of ANC in reducing malnutrition has declined over time, with the greatest impact observed in 2005 (-5.08 percentage points) and the lowest in 2015/16 (-3.16 percentage points). This trend suggests diminishing marginal returns, where earlier ANC programs reached the most vulnerable populations first, leading to an initially strong effect. As ANC coverage expanded, the remaining population may have been less responsive, either due to lower initial malnutrition risk or barriers that ANC alone cannot fully address. 3.4.2. Health and Environmental Interventions The study finds that ANC’s effectiveness in reducing child malnutrition is significantly enhanced when combined with other health and environmental interventions. This highlights the importance of adopting a holistic approach to maternal and child health rather than viewing ANC as a standalone intervention. The following subsections discuss specific interventions that amplify ANC’s impact and provide policy insights on how they can be effectively integrated into ANC programs. Breastfeeding The results indicate that breastfeeding households experience a 3.79 percentage point reduction in child malnutrition (CATE: -0.0379). This underscores the synergistic effect of ANC and breastfeeding promotion, as both contribute to improved child health outcomes. Breastfeeding provides essential nutrients, boosts immunity, and protects children against infections, all of which help reduce malnutrition. However, many mothers, particularly in low-income and rural communities, may lack the necessary information and support to practice exclusive breastfeeding. Improved Water Source Access to clean drinking water significantly enhances the effectiveness of ANC, as seen in the 4.03 percentage point reduction in child malnutrition (CATE: -0.0403) among households with improved water sources. Contaminated water contributes to frequent infections, especially diarrheal diseases, which impair nutrient absorption and worsen malnutrition. Pregnant mothers who consume unsafe water are also at risk of developing waterborne diseases that affect fetal development and birth outcomes. Improved Sanitation Households with improved sanitation experience a 4.33 percentage point reduction in malnutrition (CATE: -0.0433), highlighting the significant role of hygiene and sanitation in preventing child malnutrition. Poor sanitation leads to frequent infections, including intestinal worm infestations and diarrheal diseases, which increase the risk of malnutrition by reducing nutrient absorption. Pregnant women living in unsanitary conditions are also more susceptible to infections, leading to pregnancy complications that can affect child health. Vitamin A Supplementation Among all the interventions examined, vitamin A supplementation has the largest impact, reducing child malnutrition by 5.09 percentage points (CATE: -0.0509). Vitamin A is essential for immune function, vision, and overall child development, and its deficiency is a major contributor to child morbidity and mortality. ANC programs that incorporate vitamin A supplementation provide mothers with the necessary nutrients to improve fetal growth and postnatal health. 4. DISCUSSION This study examines the role of antenatal care (ANC) in reducing child malnutrition, using machine learning-based causal inference to quantify disparities in ANC effectiveness across residence, maternal education, household wealth, country-specific health policies, and time trends. The results align with existing literature while offering new insights into causal effects and heterogeneity in ANC impact. ANC is more effective in urban areas (CATE = -0.0705) than in rural areas (CATE = -0.0306), consistent with research highlighting better healthcare access and maternal health literacy in urban settings [ 6 ][ 10 ]. In contrast, rural populations face barriers such as limited skilled health workers, transportation difficulties, and cultural constraints [ 7 ] [ 9 ]. Expanding rural ANC access through mobile health services, community-based ANC models, and CHW programs can help bridge this gap. Rwanda’s CHW-driven maternal health model has improved ANC access, and Ethiopia could benefit from scaling similar interventions [ 8 ] [ 13 ]. Maternal education significantly influences ANC effectiveness, with educated mothers experiencing greater reductions in child malnutrition (CATE = -0.0394) than non-educated mothers (CATE = -0.0263). Educated mothers are more likely to seek skilled ANC, follow medical advice, and adopt healthier child-feeding and hygiene practices [ 6 ] [ 14 ]. However, for non-educated mothers, low health literacy and reliance on traditional beliefs limit ANC’s impact [ 7 ][ 10 ]. Integrating maternal health education into ANC services through community outreach and culturally adapted messaging can improve outcomes [ 8 ] [ 9 ]. Economic disparities also play a role, with wealthier households benefiting more (CATE = -0.0299) than poorer households (CATE = -0.0214), as indirect costs like transportation and time off work hinder ANC utilization among low-income women [ 8 ][ 13 ]. Targeted financial and nutrition support programs, such as Conditional Cash Transfers (CCTs) and food supplementation, improve ANC use [ 13 ] [ 15 ]. Rwanda’s Fortified Blended Food (FBF) initiative, which provides cash transfers and meals to vulnerable households, is a strong model for integrating ANC with social protection [ 15 ]. Expanding such nutrition-sensitive ANC interventions in Ethiopia could enhance ANC effectiveness among low-income households. ANC has a stronger impact in Ethiopia (CATE = -0.0487) than Rwanda (CATE = -0.0313), likely due to differences in maternal health policies and baseline malnutrition rates [ 14 ] [ 15 ]. Ethiopia’s greater impact may stem from more intensive community-based ANC programs, while Rwanda could enhance its effectiveness by strengthening ANC-nutrition linkages through targeted supplementation, breastfeeding support, and sanitation improvements [ 7 ] [ 9 ]. Another critical finding is that ANC’s effectiveness has declined over time, with CATE dropping from − 0.0508 in 2005 to -0.0316 in 2015/16, mirroring trends where early health improvements stagnate as coverage expands[ 5 ] [ 15 ]. Sustaining ANC impact requires continuous adaptation, including improving service quality, postnatal care, and integrating nutrition and sanitation support [ 8 ] [ 14 ]. Governments should train healthcare workers, enhance supply chain management, and leverage digital health tracking to improve ANC service delivery [ 9 ][ 13 ]. ANC is most effective when combined with complementary interventions. Households practicing breastfeeding (CATE = -0.0379), improved sanitation (CATE = -0.0433), and vitamin A supplementation (CATE = -0.0509) experience greater reductions in malnutrition, aligning with prior findings on synergistic health interventions [ 13 ] [ 15 ]. Coordinating ANC with breastfeeding education, micronutrient supplementation, and WASH (water, sanitation, and hygiene) programs enhances maternal and child health [ 9 ] [ 14 ]. Strengthening cross-sector collaboration between health, nutrition, and sanitation programs is critical for maximizing ANC’s long-term impact on child malnutrition. The findings carry important policy implications, highlighting the need for context-specific and data-driven approaches to strengthen ANC programs. Expanding rural healthcare infrastructure, deploying mobile ANC services, and integrating maternal education programs can help close the urban-rural gap. Addressing economic barriers through pro-poor policies, such as nutrition assistance and transportation subsidies, can further improve ANC utilization and outcomes. Coordinating ANC with other essential maternal and child health services—such as breastfeeding promotion, vitamin A supplementation, and sanitation improvements—can significantly enhance its impact. The application of machine learning techniques, particularly the Causal Forest model, is a key strength of this study, allowing for precise estimation of treatment effects across subpopulations. However, the reliance on cross-sectional data limits the ability to establish long-term causal relationships. Future research should explore longitudinal data and incorporate qualitative methods to better understand contextual factors influencing ANC's effectiveness. 5. CONCLUSION In conclusion, this study provides robust evidence on the causal effects of antenatal care (ANC) in reducing child malnutrition, as measured by the Composite Index of Anthropometric Failure (CIAF). By leveraging advanced machine learning techniques, particularly the Causal Forest model, the study uncovers significant disparities in ANC’s effectiveness across socio-demographic factors, including residence, maternal education, household wealth, country-specific health policies, and time trends. These findings have important implications for policy and practice, offering a pathway toward more targeted and effective interventions to combat malnutrition. Based on the analysis, countries like Ethiopia and Rwanda should prioritize interventions targeting the most influential factors. Urban residents, educated mothers, and wealthier households consistently benefited more from ANC, highlighting the need to address structural inequities in healthcare access and utilization. Rural populations, non-educated mothers, and poorer households faced greater barriers, underscoring the importance of expanding rural healthcare infrastructure, deploying mobile ANC services, and integrating maternal education programs into ANC services. Addressing economic barriers through pro-poor policies, such as nutrition assistance and transportation subsidies, can further improve ANC utilization and outcomes. The study also emphasizes the importance of integrating ANC with other essential maternal and child health services, such as breastfeeding promotion, vitamin A supplementation, and improved sanitation. Programs like Rwanda’s Fortified Blended Food (FBF) initiative offer valuable models for integrating ANC with social protection programs, ensuring that resources are allocated equitably and effectively. Additionally, the declining trend in ANC’s impact over time signals the need for sustained innovation, quality improvement, and integration with complementary interventions to maintain its long-term effectiveness. The application of machine learning techniques, particularly causal forests, represents a significant methodological advancement, enabling precise, flexible, and robust estimation of treatment effects across subpopulations. Cross-validation techniques enhance the reliability and generalizability of the findings, while feature importance analysis identifies key drivers of child malnutrition, helping policymakers prioritize high-impact interventions. Moving forward, leveraging technology-driven approaches and cross-country learnings can enhance maternal and child health interventions globally, ensuring that ANC continues to be a powerful tool in the fight against child malnutrition. Declarations Ethics approval and consent to participate This study involved secondary analysis of publicly available datasets from the Demographic and Health Surveys (DHS) program, which are fully anonymized with no identifiable information on the participants. The DHS surveys are conducted following ethical guidelines, including informed consent obtained from all respondents, and adhere strictly to principles outlined in the Declaration of Helsinki. DHS protocols and procedures are reviewed and approved by the Institutional Review Board (IRB) of ICF International and by ethical review committees within each country involved. Therefore, the current analysis complies fully with ethical standards consistent with the Declaration of Helsinki. Consent for publication Not Applicable Competing interests Not Applicable Funding Not Applicable Authors' contributions Alehegn Moges Tessema: Conceptualized the study, developed the methodology, conducted data analysis, interpreted the results, and drafted the manuscript. Prof. Temesgen Zewotir: Supervised the study, reviewed and edited the manuscript, provided guidance, and contributed to the interpretation of findings. Dr. Richard Kabanda: Supervised the study, reviewed and edited the manuscript, contributed to the interpretation of findings, and provided valuable feedback on the manuscript drafts. All authors have read and approved the final manuscript and agree to be accountable for all aspects of the work to ensure that questions related to its accuracy or integrity are appropriately investigated and resolved. Acknowledgement I would like to thank the African Center of Excellence for Data Science at the University of Rwanda for their invaluable support and guidance throughout this research. I am also grateful to the editorial team of BMC Public Health for providing the platform to publish this article and for their constructive feedback. My sincere appreciation also goes to the Demographic and Health Surveys (DHS) Program for providing the datasets that were instrumental to this study. This work would not have been possible without the support of these institutions, and I am deeply grateful for their contributions. References Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, Mathers C. Maternal and child undernutrition: global and regional exposures and health consequences. Lancet. 2008;371(9608):243–60. UNICEF. Levels and trends in child malnutrition, UNICEF-WHO-World Bank Group Joint Malnutrition Estimates, 2020. Ethiopian Public Health Institute (EPHI) &, Program DHS. Ethiopia Mini Demographic and Health Survey 2019: Final Report. DHS Program, Addis Ababa, Ethiopia; 2019. National Institute of Statistics of. Rwanda (NISR) & DHS Program, Rwanda Demographic and Health Survey. Rwanda: DHS Program, Kigali; 2020. Martorell R, Horta BL, Adair LS, Stein AD, Richter L, Fall CH, Victora CG. Improved nutrition in the first 1,000 days and its effects on adult human capital and health. Am J Clin Nutr, 98, 5, pp. 1192S-1200S, 2013. Victora CG, Adair L, Fall CH, Hallal PC, Martorell R, Richter L, Sachdev HS. Maternal and child undernutrition: consequences for adult health and human capital. Lancet. 2008;371(9609):340–57. Bhutta ZA, Ahmed T, Black RE, Cousens S, Dewey K, Giugliani E, Shekar M. What works? Interventions for maternal and child undernutrition and survival. Lancet. 2008;371(9610):417–40. Haddad L, Achadi E, Bendech MA, Ahuja A, de Pee S, Engesveen K, Bhutta ZA. The Global Nutrition Report 2015: Actions and accountability to advance nutrition and sustainable development. International Food Policy Research Institute (IFPRI); 2015. Ruel MT, Alderman H, a. M, Group CNS. Nutrition-sensitive interventions and programmes: How can they help to accelerate progress in improving maternal and child nutrition? Lancet. 2013;382(9891):536–51. Smith LC, Ruel MT, Ndiaye A. Why is child malnutrition lower in urban than in rural areas? Evidence from 36 developing countries. World Dev. 2018;105:274–82. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects, Biometrika , vol. 70, no. 1, pp. 41–55, 1983. Wager S, Athey S. Estimation and inference of heterogeneous treatment effects using random forests. J Name: J Am Stat Association. 2018;113(523):1228–42. Bhutta ZA, Das JK, Rizvi A, Gaffey MF, Walker N, Horton S, Black RE. Evidence-based interventions for improvement of maternal and child nutrition: What can be done and at what cost? Lancet. 2013;382(9890):452–77. Black RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, Mathers C. Maternal and child undernutrition: Global and regional exposures and health consequences. Lancet. 2013;382(9890):427–51. World Health Organization (WHO). (2016). Recommendations on antenatal care for a positive pregnancy experience. World Health Organization, Recommendations on antenatal care for a positive pregnancy experience, World Health Organization, 2016. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 28 Nov, 2025 Read the published version in BMC Nutrition → Version 1 posted Editorial decision: Revision requested 11 Jul, 2025 Reviews received at journal 31 May, 2025 Reviews received at journal 19 May, 2025 Reviews received at journal 09 May, 2025 Reviewers agreed at journal 08 May, 2025 Reviewers agreed at journal 08 May, 2025 Reviewers agreed at journal 03 May, 2025 Reviewers invited by journal 19 Apr, 2025 Editor assigned by journal 15 Apr, 2025 Editor invited by journal 27 Mar, 2025 Submission checks completed at journal 24 Mar, 2025 First submitted to journal 24 Mar, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6131494","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":445339381,"identity":"f10fd309-1b28-4922-b662-b4225fcf5abf","order_by":0,"name":"Alehegn Moges Tessema","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAElEQVRIiWNgGAWjYBADGQYJEFUBxMzMDURp4YFoOQPSwkiKFsY2EJuAFt329osPfzDY8fBLNx9+XTivNpq/HajlR8U2nFrMzpwpNuZhSOaRnHMszXrmtuO5Mw4zNjD2nLmNW8uNnDRpoPt5DG7kmBnzbjuW2wDUwszYhldL+s8fDPU89jfyvxnzzjmWO5+wlvRjQL8f5jGQyGF+zNtQk7uBoJYzZ5ileQyO80jcSDNj5jl2IHcjUMtBvH453v7w44+Kajn+GcmPP/PU1OXOO3/44IMfFbi1AGPEgIHBAMxiA0bNYTDrAB71QMD+AMZi/sDAUIdf8SgYBaNgFIxIAAC2+ll5DAwLKQAAAABJRU5ErkJggg==","orcid":"","institution":"University of Rwanda","correspondingAuthor":true,"prefix":"","firstName":"Alehegn","middleName":"Moges","lastName":"Tessema","suffix":""},{"id":445339382,"identity":"7d3fe940-c251-47a9-b110-d58d34efd7dc","order_by":1,"name":"Temesgen Zewotir","email":"","orcid":"","institution":"University of KwaZulu-Natal","correspondingAuthor":false,"prefix":"","firstName":"Temesgen","middleName":"","lastName":"Zewotir","suffix":""},{"id":445339383,"identity":"678eba2d-53b0-47e2-a38d-ba032c3489eb","order_by":2,"name":"Richard Kabanda","email":"","orcid":"","institution":"University of Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Richard","middleName":"","lastName":"Kabanda","suffix":""}],"badges":[],"createdAt":"2025-02-28 21:53:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6131494/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6131494/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40795-025-01208-w","type":"published","date":"2025-11-28T15:58:44+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":81692398,"identity":"cf4539ed-8454-4763-857f-2b7a4f438abc","added_by":"auto","created_at":"2025-04-30 11:41:48","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":96679,"visible":true,"origin":"","legend":"\u003cp\u003eAnalytical Framework\u003c/p\u003e","description":"","filename":"groupimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6131494/v1/30b462a4756836944319521a.jpeg"},{"id":81692399,"identity":"bbe4c037-cf7f-4342-8e04-9140441554d5","added_by":"auto","created_at":"2025-04-30 11:41:49","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":19371,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of Treatment Effects\u003c/p\u003e","description":"","filename":"groupimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6131494/v1/03d7cea3e0c2b350a5c6b481.jpeg"},{"id":97178642,"identity":"d7be2060-7a3a-4ebe-a5ef-c89dff795812","added_by":"auto","created_at":"2025-12-01 16:12:07","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":945329,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6131494/v1/cbe29738-d5c7-41a9-aa60-b5c843220b2c.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Causal Effects of Antenatal Care (ANC) on Child Malnutrition: A Machine Learning Approach in Ethiopia and Rwanda","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eMalnutrition remains a critical public health challenge in Ethiopia and Rwanda, particularly among children under five, contributing to high rates of illness, death, stunted physical growth, impaired cognitive development, and reduced overall well-being [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Despite ongoing efforts to address malnutrition, progress has been inconsistent. In Ethiopia, nearly 37% of children under five are stunted, while in Rwanda, around 33% face similar issues, reflecting widespread chronic undernutrition [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. These persistent challenges highlight the urgent need for targeted and effective interventions that address both the immediate and underlying causes of malnutrition. Beyond its health impacts, malnutrition has significant economic and social consequences, perpetuating poverty, reducing productivity, and hindering national development [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e] [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eTo design effective interventions, a clear understanding of the multiple causes of malnutrition is essential. Immediate causes include poor dietary intake and frequent illnesses such as diarrhea, while underlying factors like inadequate maternal health, low household income, and limited access to clean water and sanitation worsen the problem [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e] [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Broader systemic issues, including food insecurity, unequal access to healthcare, and educational gaps, further increase the risk of malnutrition, particularly among vulnerable groups [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Addressing these interconnected factors requires interventions that go beyond immediate needs to tackle the root causes of malnutrition.\u003c/p\u003e \u003cp\u003eExisting research has identified associations between certain interventions\u0026mdash;such as maternal education, antenatal care, and improved water and sanitation\u0026mdash;and better child nutrition outcomes [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. However, these studies often focus on correlations rather than causation. Correlation-based findings do not account for other influencing factors like socioeconomic status, geographic location, or household characteristics, which can distort the observed relationships [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. This limitation makes it difficult for policymakers to design evidence-based strategies, as they lack clear evidence on which interventions directly lead to improved outcomes.\u003c/p\u003e \u003cp\u003eThis study aims to addresses this gap by leveraging machine learning-based causal inference methods to evaluate the causal effects of specific nutrition and health interventions on child nutrition outcomes in Ethiopia and Rwanda. Unlike traditional approaches that rely on correlation, this research employs machine learning techniques such as causal forests, doubly robust estimation, and propensity score matching to isolate the causal impact of interventions like maternal education, antenatal care, and improved water access. These methods are particularly well-suited for handling complex, high-dimensional data and adjusting for a wide range of confounding variables, providing more accurate estimates of the Average Treatment Effects (ATE) and Conditional Average Treatment Effects (CATE).\u003c/p\u003e \u003cp\u003eCausal forests, a machine learning method, are used to identify heterogeneous treatment effects, allowing the study to uncover how the impact of interventions varies across different subgroups, such as rural versus urban populations or varying wealth levels. For example, the analysis might reveal that maternal education has a stronger effect in urban areas with better healthcare infrastructure, while improved water access is more impactful in rural regions with limited sanitation facilities. By quantifying these variations, the study provides policymakers with actionable, context-specific evidence to design targeted interventions.\u003c/p\u003e \u003cp\u003eThe use of machine learning in causal inference represents a significant advancement in malnutrition research. These methods not only improve the precision of causal estimates but also enable the identification of nuanced patterns in the data that traditional statistical approaches might miss. For instance, machine learning algorithms can automatically detect interactions between variables or non-linear relationships, offering deeper insights into how interventions interact with contextual factors. This level of detail ensures that resources are allocated equitably and effectively, maximizing the impact of interventions.\u003c/p\u003e"},{"header":"2. METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1. The Data\u003c/h2\u003e \u003cp\u003eThe study utilized Demographic and Health Surveys (DHS) datasets from Ethiopia and Rwanda to analyze the impact of antenatal care (ANC) on child malnutrition, measured using the Composite Index of Anthropometric Failure (CIAF). The datasets included key socio-demographic variables such as residence, maternal education, wealth index, and survey year to enable a comprehensive analysis. Data preprocessing involved encoding categorical variables and standardizing continuous variables to ensure compatibility with machine learning algorithms. Propensity score methods were employed to balance treated and untreated groups, mitigating selection bias and enhancing the validity of causal estimates.\u003c/p\u003e \u003cp\u003eEach dataset includes crucial information on household demographics, maternal and child health, socio-economic factors, and access to essential services such as healthcare, water, and sanitation. Independent variables such as wealth index, maternal education, access to healthcare, and environmental conditions were selected for their relevance in predicting child malnutrition based on prior studies. The details of the variable definitions are presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e below.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe description of the response variable and the respective covariates included in the model\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVariable\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCategories\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDescription\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eComposite Index of Anthropometric Failure (CIAF)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo, Yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\left\\{\\begin{array}{c}1\\:\\:if\\:the\\:child\\:has\\:some\\:form\\:of\\:malnutrition\\\\\\:0\\:if\\:the\\:child\\:has\\:no\\:any\\:form\\:of\\:malnutrition\\end{array}\\right.\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlace of Residence\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUrban, Rural\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUrban vs. rural living environments.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMother\u0026rsquo;s Education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo education, Primary, Secondary, Higher\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEducation levels of the mother: no formal education to higher education.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex of Child\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale, Female\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eThe biological sex of the child.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVitamin A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo, Yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the child received Vitamin A supplementation.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWealth Quintile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePoorest, Poorer, Middle, Richer, Richest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eThe economic status of the household, ranked from poorest to richest.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEthiopia, Rwanda\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCountry where the data was collected.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2005, 2010, 2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eThe year of data collection.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSource of Drinking Water\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnimproved, Improved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the household uses unimproved or improved drinking water sources.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eToilet Facilities\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnimproved, Improved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the household has unimproved or improved sanitation facilities.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTook Iron Tablets\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo, Yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the mother took iron tablets during pregnancy.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePlace of Delivery\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo, Yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the child was born in a healthcare facility.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFirst Born\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo, Yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the child is the firstborn.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTypes of Birth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSingle, Twins\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the child was part of a single or multiple birth (e.g., twins).\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBreastfeeding\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo, Yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the child was breastfed.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eANC (Antenatal Care)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo, Yes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhether the mother received antenatal care during pregnancy.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge of Mother\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15\u0026ndash;24, 25\u0026ndash;34, 35\u0026ndash;49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAge range of the mother at the time of the child's birth.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize of Child at Birth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAbove average, Average, Smaller than average\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eThe perceived size of the child at birth.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBMI of Mother\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;18.5 kg/m\u0026sup2; (Underweight), 18.5\u0026ndash;24.9 kg/m\u0026sup2; (Normal weight), \u0026ge;\u0026thinsp;25 kg/m\u0026sup2; (Overweight)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eBody Mass Index (BMI) of the mother, categorized as underweight, normal, or overweight.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHousehold Size\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;4, 5\u0026ndash;9, 10+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNumber of people living in the household.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber of Under-Five Children\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u0026ndash;1, 2, 3 or more\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eThe number of children under the age of five in the household.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2. Analytical Framework\u003c/h2\u003e \u003cp\u003eThis study employs a robust analytical framework to assess the impact of Antenatal Care (ANC) on child malnutrition. By integrating Average Treatment Effect (ATE), Conditional Average Treatment Effect (CATE), and a causal forest model, the framework provides a comprehensive and data-driven understanding of ANC\u0026rsquo;s effectiveness across different populations.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe analytical framework outlined in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e supports evidence-based decision-making, enabling policymakers to develop targeted interventions for improving maternal and child health. The analysis begins by estimating the Average Treatment Effect (ATE), which measures the overall impact of Antenatal Care (ANC) on reducing child malnutrition at the population level. The ATE is calculated by comparing outcomes between those who received ANC and those who did not, while controlling for confounding factors such as socioeconomic status and healthcare access. This step provides a broad estimate of ANC\u0026rsquo;s effectiveness, establishing a foundation for understanding its role in reducing child malnutrition.\u003c/p\u003e \u003cp\u003eTo gain deeper insights, the study conducts a subgroup analysis using Conditional Average Treatment Effects (CATE). This approach examines how ANC\u0026rsquo;s effectiveness varies across socio-demographic groups, including residence (urban vs. rural), maternal education (educated vs. non-educated), household wealth (richer vs. poorer), country-specific differences, and time trends. Additional factors such as breastfeeding practices, access to clean water, improved sanitation, and vitamin A supplementation are also considered. The CATE analysis identifies which groups benefit the most or least from ANC, revealing disparities that may require targeted interventions. For example, urban residents and educated mothers experience greater improvements in child nutrition compared to rural residents and non-educated mothers. These findings help policymakers design ANC programs tailored to the needs of underserved populations.\u003c/p\u003e \u003cp\u003eA key component of the analysis is the use of a causal forest model, a machine learning technique that estimates how ANC\u0026rsquo;s effects vary across individuals or groups. Unlike traditional methods, this model captures variations based on factors such as location, education, and wealth, providing more detailed and accurate estimates of ANC\u0026rsquo;s effectiveness. By combining multiple decision trees, the causal forest model detects complex patterns in the data, offering a nuanced understanding of how different groups respond to ANC. This enables policymakers to make informed decisions to improve program outcomes.\u003c/p\u003e \u003cp\u003eTo ensure reliability, the study employs two validation techniques: 5-fold cross-validation and analysis of treatment effect distribution. In 5-fold cross-validation, the dataset is divided into five parts, with the model trained on four parts and tested on the fifth, repeating the process five times. The results are averaged to provide a stable estimate of ANC\u0026rsquo;s impact, ensuring the findings are not dependent on a specific dataset split. The second method involves analyzing the distribution of treatment effects, including generating a histogram and calculating descriptive statistics such as mean, standard deviation, and range. A low variation in treatment effects indicates stable and reliable estimates, confirming the model\u0026rsquo;s robustness.\u003c/p\u003e \u003cp\u003eFinally, the study interprets the results of the ATE, CATE, and validation analyses in the context of socio-demographic and environmental factors. This interpretation provides actionable insights for policymakers, highlighting disparities in ANC\u0026rsquo;s effectiveness and identifying populations that may require additional support. For instance, if rural areas or poorer households benefit less from ANC, efforts can be directed toward improving healthcare access in these communities. By understanding how ANC\u0026rsquo;s impact varies across groups, stakeholders can design programs that allocate resources where they are most needed, leading to more effective interventions that reduce child malnutrition and improve overall health outcomes.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. RESULT","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Characteristics of the study participants and prevalence of Composite Index of Anthropometric Failure (CIAF)\u003c/h2\u003e \u003cp\u003eThis analysis included data from 32,099 children aged 0\u0026ndash;59 months, with 15,726 (49.01%) experiencing at least one form of undernutrition, as measured by the Composite Index of Anthropometric Failure (CIAF). The prevalence of CIAF was analyzed across various child, maternal, and household-level factors. Higher rates of CIAF were observed among children of mothers with no formal education, those living in rural areas, and households in the lowest wealth quintile. Children from homes with unimproved water sources and sanitation, and whose mothers lacked access to antenatal care or iron supplements, were also more likely to experience undernutrition. Significant covariates identified through chi-square tests were utilized to develop machine learning models on the dataset (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePrevalence of CIAF Across Socio-Demographic and Environmental Variables\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCategories\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003eComposite Index of Anthropometric Failure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eChi-square\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSig.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNourished (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCIAF (%)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ePlace of residence\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUrban\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e33.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e655.498\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRural\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e52.9%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003eMother\u0026rsquo;s education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e54.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e783.057\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePrimary\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e47.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSecondary\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e70.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29.4%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHigher\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e82.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e17.7%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSex of child\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e48.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e51.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e56.034\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e47.5%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVitamin A\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e15.361\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e53.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e46.9%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"4\" rowspan=\"5\"\u003e \u003cp\u003eWealth quintile\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePoorest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e56.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"4\" rowspan=\"5\"\u003e \u003cp\u003e935.602\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"4\" rowspan=\"5\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePoorer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e55.0%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMiddle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e48.5%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e51.5%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRicher\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48.0%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRichest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e67.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e32.9%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEthiopia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e0.654\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e0.419\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eRwanda\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.3%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eYear\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2005\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e41.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e58.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e451.776\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e56.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e43.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSource of drinking water\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnimproved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e54.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e234.443\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eImproved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e45.9%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eToilet facilities\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUnimproved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e52.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e256.845\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eImproved\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e56.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e43.4%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eTook iron tablets\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e47.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e52.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e149.896\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e45.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003ePlace of Delivery Category\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e54.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e693.336\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e61.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e38.7%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eFirst born\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e80.697\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e55.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44.6%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eTypes of birth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSingle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e43.053\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTwines\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e61.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eBreastfeeding\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e0.280\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e0.597\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.7%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eANC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e56.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e337.177\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e45.6%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eAge of mother\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15\u0026ndash;24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e24.154\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25\u0026ndash;34\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e35\u0026ndash;49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e48.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e51.9%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eSize of the child at birth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAbove average\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e55.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e44.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e280.695\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAverage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e50.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e49.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSmaller than average\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e42.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e57.1%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eBMI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;18.5 kg/m\u0026sup2; (Underweight)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e56.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e443.764\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e18.5\u0026ndash;24.9 kg/m\u0026sup2; (Normal weight)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026ge;\u0026thinsp;25 kg/m\u0026sup2; (Overweight)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e66.2%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e33.8%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eHousehold size\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.6%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e47.4%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e28.507\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u0026ndash;9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49.3%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e50.7%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e10+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51.8%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48.2%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eNumber of under five children in the household\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u0026ndash;1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.9%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e47.1%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e66.859\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e.000\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e48.0%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e52.0%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 or more\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e51.7%\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e48.3%\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003ePlace of residence, maternal education, wealth, and access to healthcare are significant factors in child nutrition. Urban children show a significantly higher nourishment rate (66.2%) compared to rural children (47.1%) (χ\u0026sup2; = 655.498, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Similarly, children of mothers with no education have higher malnutrition rates (54.6%), while those with highly educated mothers show lower rates (17.7%) (χ\u0026sup2; = 783.057, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Wealthier households also exhibit better child nutrition, with the poorest having the highest CIAF rate (56.6%) and the richest the lowest (32.9%) (χ\u0026sup2; = 935.602, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eMale children are slightly more likely to experience malnutrition (51.7%) compared to females (47.5%) (χ\u0026sup2; = 56.034, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Vitamin A supplementation is associated with improved nutrition (46.9% vs. 50.1%, χ\u0026sup2; = 15.361, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and access to improved water sources and sanitation significantly reduces malnutrition (45.9% for improved water vs. 54.6% for unimproved, χ\u0026sup2; = 234.443, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; 43.4% for improved sanitation vs. 52.8%, χ\u0026sup2; = 256.845, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eChildren born in healthcare facilities have better nutritional outcomes (38.7%) compared to those born outside such facilities (54.6%) (χ\u0026sup2; = 693.336, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). First-born children tend to be better nourished (44.6% vs. 50.9% for later-born, χ\u0026sup2; = 80.697, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while children from multiple births are more vulnerable to malnutrition (61.8% vs. 49.4%, χ\u0026sup2; = 43.053, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eAntenatal care (ANC) is a protective factor, with children of mothers who received ANC showing lower malnutrition rates (45.6% vs. 56.2%, χ\u0026sup2; = 337.177, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Maternal age and BMI are also important; younger mothers (ages 15\u0026ndash;24) have children with better nutrition (48.3% vs. 51.9%, χ\u0026sup2; = 24.154, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and underweight mothers\u0026rsquo; children face higher malnutrition rates (56.1% vs. 33.8% for overweight mothers, χ\u0026sup2; = 443.764, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eGenerally, the multifaceted nature of CIAF prevalence underscores the importance of addressing socio-economic inequalities, enhancing access to healthcare and sanitation services, and promoting maternal and child health practices to ensure better child health outcomes. The statistically significant associations identified in this analysis offer valuable insights for designing targeted interventions aimed at reducing child illness and malnutrition, ultimately contributing to improved child well-being and development. It is in light of these results that the selected features are chosen for multivariate analysis and inclusion in the machine learning algorithm for further analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Estimation of Treatment Effects\u003c/h2\u003e \u003cp\u003eIn this section, we present the estimation of treatment effects, including the Average Treatment Effect (ATE) and Conditional Average Treatment Effect (CATE). These estimates provide insight into the overall and subgroup-specific impacts of Antenatal Care (ANC) on child malnutrition. However, before drawing conclusions from these estimates, it is essential to assess the reliability of the model used for estimation. The interpretation of these results will be provided in the Interpretation of Results Section after confirming the model\u0026rsquo;s robustness and stability through validation.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e3.2.1. Average Treatment Effect (ATE)\u003c/h2\u003e \u003cp\u003eThe overall impact of ANC on child malnutrition was estimated using the ATE. The results indicate that ANC significantly reduces the likelihood of child malnutrition. The ATE was found to be -0.0263, meaning that, on average, ANC lowers the probability of malnutrition by 2.63 percentage points across the entire population. This provides a broad measure of ANC\u0026rsquo;s effectiveness in improving child nutrition outcomes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e3.2.2. Conditional Average Treatment Effect (CATE)\u003c/h2\u003e \u003cp\u003eTo explore how ANC\u0026rsquo;s effectiveness varies across different population groups, we estimate the Conditional Average Treatment Effect (CATE) across key socio-demographic and contextual factors. These factors include residence (urban vs. rural), maternal education, wealth status, country-specific differences, and temporal trends. Additionally, we assess the influence of complementary health and environmental factors such as breastfeeding, access to improved water, sanitation, and vitamin A supplementation.\u003c/p\u003e \u003cp\u003eThe results presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e shows a detailed and insightful look at how Antenatal Care (ANC) affects child malnutrition across different groups. Using Conditional Average Treatment Effect (CATE) analysis, the study moves beyond average effects to explore how ANC\u0026rsquo;s effectiveness varies based on factors like where people live (urban vs. rural), their education level, wealth, country, and other health and environmental conditions. These findings are crucial for policymakers and healthcare providers because they reveal disparities in ANC\u0026rsquo;s impact and pinpoint groups that may need more targeted support. The table not only shows how much ANC reduces malnutrition but also explains why these differences exist, helping to shape fairer and more effective healthcare strategies. These findings will be interpreted in details under result interpretation section of the article.\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;3: Subgroup Analysis Results: A Conditional Average Treatment Effect Analysis\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"614\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eSubgroup\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003eCATE (Conditional Average Treatment Effect)\u003csup\u003e1\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003eCross-Validated ATE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eInterpretation\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eResidence: Urban\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0705\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0306\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eUrban residents experienced a significant reduction in malnutrition (7.05%) compared to rural areas.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eResidence: Rural\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0306\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eRural residents experienced a smaller but consistent reduction in malnutrition (3.06%).\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eEducation: Educated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0394\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0312\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eEducated mothers benefited more from ANC, reducing malnutrition by 3.94%.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eEducation: Non-Educated\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0263\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0303\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eNon-educated mothers experienced a reduction of 2.63% in malnutrition likelihood.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eWealth Index: Richer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0299\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0313\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eWealthier households saw a reduction of 2.99% in malnutrition, benefiting significantly from ANC.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eWealth Index: Poorer\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0214\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003ePoorer households experienced a smaller reduction (2.14%) due to economic barriers.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eCountry: Ethiopia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0487\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0309\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eEthiopia showed a stronger ANC effect, reducing malnutrition by 4.87%.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eCountry: Rwanda\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0313\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eRwanda\u0026apos;s reduction (3.13%) was smaller, possibly due to differences in ANC programs.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eYear: 2005\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0508\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0308\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eANC had its highest impact in 2005, reducing malnutrition by 5.08%.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eYear: 2010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0433\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0306\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eIn 2010, ANC reduced malnutrition by 4.33%, showing a declining trend over time.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eYear: 2015/16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0316\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eBy 2015/16, ANC\u0026apos;s impact declined to 3.16%, suggesting diminishing marginal returns.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eBreastfeeding\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0379\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eBreastfeeding households experienced a 3.79% reduction, showing synergy with ANC.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eImproved Water Source\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0403\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0299\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eHouseholds with access to improved water saw a 4.03% reduction in malnutrition.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eImproved Sanitation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0433\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0313\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eImproved sanitation facilities enhanced ANC\u0026rsquo;s impact, reducing malnutrition by 4.33%.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 114px;\"\u003e\n \u003cp\u003eVitamin A Supplementation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 117px;\"\u003e\n \u003cp\u003e-0.0509\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e-0.0315\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 300px;\"\u003e\n \u003cp\u003eHouseholds with vitamin A supplementation had the largest impact, reducing malnutrition by 5.09%.\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003csup\u003e1\u003c/sup\u003e\u003cem\u003e\u0026nbsp;CATE Estimate: Conditional Average Treatment Effect (the estimated impact of ANC on child malnutrition for each subgroup).\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e2\u003c/sup\u003e\u003cem\u003e\u0026nbsp;Interpretation: Describes the effectiveness of ANC in reducing child malnutrition for the specific subgroup.\u003c/em\u003e\u003c/p\u003e\n \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.3. Model Performance\u003c/h2\u003e \u003cp\u003eEvaluating the performance of the causal forest model is crucial to ensure the reliability and robustness of the findings. This section presents two key assessments of the model's performance: Cross-Validated Results and Distribution of Treatment Effects. These metrics provide insights into the model's ability to consistently and accurately estimate treatment effects across the dataset.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section3\"\u003e \u003ch2\u003e3.3.1. Cross-Validated Results\u003c/h2\u003e \u003cp\u003eCross-validation is a widely used technique for evaluating the generalizability of a model\u0026rsquo;s findings. In this analysis, a 5-fold cross-validation procedure was implemented to assess the reliability of the causal forest model. The dataset was divided into five subsets, where the model was iteratively trained on four subsets and tested on the remaining subset. This process ensured that each observation in the dataset was used for both training and testing. The Average Treatment Effect (ATE) was calculated for each fold, and the cross-validated ATE was obtained by averaging the ATE values across all folds.\u003c/p\u003e \u003cp\u003eThe cross-validated ATE was estimated at -0.0311, indicating that antenatal care (ANC) reduces the likelihood of child malnutrition by 3.11 percentage points on average across all folds. This result aligns closely with the overall ATE, calculated using the full dataset, which was \u0026minus;\u0026thinsp;0.0263. The close alignment between the cross-validated ATE and the overall ATE demonstrates that the model provides consistent estimates of the treatment effect across different subsets of the data.\u003c/p\u003e \u003cp\u003eThe consistency of ATE values across folds highlights the robustness of the causal forest model. The small variations observed between folds suggest that the model is not overfitting and that its findings are generalizable to different subsets of the data. This reliability ensures that the estimated treatment effects are credible and can be confidently applied to policy and programmatic decision-making. By confirming the model\u0026rsquo;s stability and robustness, cross-validation strengthens the validity of the analysis and the insights derived from it.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e \u003ch2\u003e3.3.2. Distribution of Treatment Effects\u003c/h2\u003e \u003cp\u003eThe distribution of treatment effects was analyzed to evaluate how the estimated effects vary across individual observations in the dataset. A histogram was generated to visualize the spread of treatment effects, providing a clear picture of their variability and consistency. This analysis is crucial for understanding the model's ability to generate reliable and stable estimates of the impact of antenatal care (ANC) on child malnutrition.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe histogram revealed a symmetric, bell-shaped distribution centered around the Average Treatment Effect (ATE) value of -0.0263. Most treatment effects clustered close to the mean, indicating low variability in the model's estimates. Notably, no significant outliers were observed, suggesting that the model does not produce extreme or erratic estimates, further reinforcing its reliability. (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e)\u003c/p\u003e \u003cp\u003eThe descriptive statistics for treatment effects provide additional insights into the model\u0026rsquo;s performance. The mean treatment effect (ATE) was calculated to be -0.0263, confirming the overall impact of ANC in reducing malnutrition by 2.63 percentage points. The standard deviation of 0.0047 reflects minimal variability in the estimates. The minimum and maximum treatment effects were \u0026minus;\u0026thinsp;0.0381 and \u0026minus;\u0026thinsp;0.0152, respectively, showing that the treatment effects across observations are tightly distributed around the mean.\u003c/p\u003e \u003cp\u003eThe symmetric, bell-shaped distribution indicates that the model consistently captures the causal relationships between ANC and child malnutrition across observations. The low standard deviation and absence of extreme values confirm that the treatment effect estimates are both stable and reliable. This consistency across observations enhances the credibility of the model's findings and ensures that the results are robust enough to inform policy and programmatic decisions with confidence.\u003c/p\u003e \u003cp\u003eThe combined results of these metrics\u0026mdash;cross-validated ATE and the distribution of treatment effects\u0026mdash;validate the causal forest model\u0026rsquo;s suitability for this analysis. The model\u0026rsquo;s ability to produce consistent and stable estimates ensures its reliability for policy and programmatic decision-making. These findings provide confidence that the model accurately captures the causal relationship between ANC and child malnutrition, enabling stakeholders to draw actionable insights for targeted interventions.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.4. Interpretation of Results\u003c/h2\u003e \u003cp\u003eThe estimates presented in Section 3.2 are now interpreted in light of the model validation results from Section 3.3. The validation process confirmed that our causal forest model is robust, stable, and reliable, meaning that the estimated treatment effects can be confidently used for policy recommendations. The key findings from the estimation of treatment effects presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e above distinguishes between confounding factors, which provide context for disparities, and intervention-related factors, which can be leveraged to enhance ANC's effectiveness.\u003c/p\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003e3.4.1. Confounding Factors and Contextual Disparities\u003c/h2\u003e \u003cp\u003eResidence: Urban vs. Rural\u003c/p\u003e \u003cp\u003eThe analysis reveals a significant disparity in ANC\u0026rsquo;s effectiveness between urban and rural areas. Urban residents experience a 7.05 percentage point reduction in child malnutrition (CATE: -0.0705), while rural residents see a smaller reduction of 3.06 percentage points (CATE: -0.0306). This gap is likely influenced by several factors, including better healthcare access, service quality, and awareness in urban settings. Urban populations are more likely to receive timely ANC services, benefit from skilled health professionals, and have access to necessary medications and diagnostic tools.\u003c/p\u003e \u003cp\u003eEducation: Educated vs. Non-Educated Mothers\u003c/p\u003e \u003cp\u003eMaternal education plays a critical role in ANC\u0026rsquo;s effectiveness. Educated mothers experience a 3.94 percentage point reduction in child malnutrition (CATE: -0.0394), compared to 2.63 percentage points (CATE: -0.0263) for non-educated mothers. This suggests that education enhances a mother\u0026rsquo;s ability to understand, seek, and utilize ANC services effectively. Educated mothers are more likely to attend ANC visits regularly, follow healthcare recommendations on nutrition and supplementation, and recognize early signs of malnutrition or pregnancy complications, leading to better health outcomes for their children.\u003c/p\u003e \u003cp\u003eWealth Index: Richer vs. Poorer Households\u003c/p\u003e \u003cp\u003eEconomic disparities significantly influence ANC\u0026rsquo;s effectiveness. Wealthier households experience a 2.99 percentage point reduction in malnutrition (CATE: -0.0299), while poorer households see a smaller 2.14 percentage point reduction (CATE: -0.0214). Wealthier families tend to have better healthcare access, improved dietary diversity, and lower financial stress, enabling them to comply with ANC recommendations. They are also more likely to access private healthcare services, which may offer higher-quality ANC.\u003c/p\u003e \u003cp\u003eCountry: Ethiopia vs. Rwanda\u003c/p\u003e \u003cp\u003eThe study finds that ANC\u0026rsquo;s impact on malnutrition reduction differs between Ethiopia (4.87 percentage points, CATE: -0.0487) and Rwanda (3.13 percentage points, CATE: -0.0313). These differences may reflect health system variations, ANC program implementation differences, or baseline malnutrition rates in each country. Ethiopia\u0026rsquo;s stronger impact could be due to higher initial malnutrition rates, creating greater room for improvement, or more intensive ANC interventions targeted at malnutrition.\u003c/p\u003e \u003cp\u003eTrends Over Time: 2005\u0026ndash;2015/16\u003c/p\u003e \u003cp\u003eThe effectiveness of ANC in reducing malnutrition has declined over time, with the greatest impact observed in 2005 (-5.08 percentage points) and the lowest in 2015/16 (-3.16 percentage points). This trend suggests diminishing marginal returns, where earlier ANC programs reached the most vulnerable populations first, leading to an initially strong effect. As ANC coverage expanded, the remaining population may have been less responsive, either due to lower initial malnutrition risk or barriers that ANC alone cannot fully address.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section3\"\u003e \u003ch2\u003e3.4.2. Health and Environmental Interventions\u003c/h2\u003e \u003cp\u003eThe study finds that ANC\u0026rsquo;s effectiveness in reducing child malnutrition is significantly enhanced when combined with other health and environmental interventions. This highlights the importance of adopting a holistic approach to maternal and child health rather than viewing ANC as a standalone intervention. The following subsections discuss specific interventions that amplify ANC\u0026rsquo;s impact and provide policy insights on how they can be effectively integrated into ANC programs.\u003c/p\u003e \u003cp\u003eBreastfeeding\u003c/p\u003e \u003cp\u003eThe results indicate that breastfeeding households experience a 3.79 percentage point reduction in child malnutrition (CATE: -0.0379). This underscores the synergistic effect of ANC and breastfeeding promotion, as both contribute to improved child health outcomes. Breastfeeding provides essential nutrients, boosts immunity, and protects children against infections, all of which help reduce malnutrition. However, many mothers, particularly in low-income and rural communities, may lack the necessary information and support to practice exclusive breastfeeding.\u003c/p\u003e \u003cp\u003eImproved Water Source\u003c/p\u003e \u003cp\u003eAccess to clean drinking water significantly enhances the effectiveness of ANC, as seen in the 4.03 percentage point reduction in child malnutrition (CATE: -0.0403) among households with improved water sources. Contaminated water contributes to frequent infections, especially diarrheal diseases, which impair nutrient absorption and worsen malnutrition. Pregnant mothers who consume unsafe water are also at risk of developing waterborne diseases that affect fetal development and birth outcomes.\u003c/p\u003e \u003cp\u003eImproved Sanitation\u003c/p\u003e \u003cp\u003eHouseholds with improved sanitation experience a 4.33 percentage point reduction in malnutrition (CATE: -0.0433), highlighting the significant role of hygiene and sanitation in preventing child malnutrition. Poor sanitation leads to frequent infections, including intestinal worm infestations and diarrheal diseases, which increase the risk of malnutrition by reducing nutrient absorption. Pregnant women living in unsanitary conditions are also more susceptible to infections, leading to pregnancy complications that can affect child health.\u003c/p\u003e \u003cp\u003eVitamin A Supplementation\u003c/p\u003e \u003cp\u003eAmong all the interventions examined, vitamin A supplementation has the largest impact, reducing child malnutrition by 5.09 percentage points (CATE: -0.0509). Vitamin A is essential for immune function, vision, and overall child development, and its deficiency is a major contributor to child morbidity and mortality. ANC programs that incorporate vitamin A supplementation provide mothers with the necessary nutrients to improve fetal growth and postnatal health.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"4. DISCUSSION","content":"\u003cp\u003eThis study examines the role of antenatal care (ANC) in reducing child malnutrition, using machine learning-based causal inference to quantify disparities in ANC effectiveness across residence, maternal education, household wealth, country-specific health policies, and time trends. The results align with existing literature while offering new insights into causal effects and heterogeneity in ANC impact.\u003c/p\u003e \u003cp\u003eANC is more effective in urban areas (CATE = -0.0705) than in rural areas (CATE = -0.0306), consistent with research highlighting better healthcare access and maternal health literacy in urban settings [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e][\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. In contrast, rural populations face barriers such as limited skilled health workers, transportation difficulties, and cultural constraints [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e] [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Expanding rural ANC access through mobile health services, community-based ANC models, and CHW programs can help bridge this gap. Rwanda\u0026rsquo;s CHW-driven maternal health model has improved ANC access, and Ethiopia could benefit from scaling similar interventions [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eMaternal education significantly influences ANC effectiveness, with educated mothers experiencing greater reductions in child malnutrition (CATE = -0.0394) than non-educated mothers (CATE = -0.0263). Educated mothers are more likely to seek skilled ANC, follow medical advice, and adopt healthier child-feeding and hygiene practices [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e] [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. However, for non-educated mothers, low health literacy and reliance on traditional beliefs limit ANC\u0026rsquo;s impact [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e][\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Integrating maternal health education into ANC services through community outreach and culturally adapted messaging can improve outcomes [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eEconomic disparities also play a role, with wealthier households benefiting more (CATE = -0.0299) than poorer households (CATE = -0.0214), as indirect costs like transportation and time off work hinder ANC utilization among low-income women [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e][\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Targeted financial and nutrition support programs, such as Conditional Cash Transfers (CCTs) and food supplementation, improve ANC use [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Rwanda\u0026rsquo;s Fortified Blended Food (FBF) initiative, which provides cash transfers and meals to vulnerable households, is a strong model for integrating ANC with social protection [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Expanding such nutrition-sensitive ANC interventions in Ethiopia could enhance ANC effectiveness among low-income households.\u003c/p\u003e \u003cp\u003eANC has a stronger impact in Ethiopia (CATE = -0.0487) than Rwanda (CATE = -0.0313), likely due to differences in maternal health policies and baseline malnutrition rates [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Ethiopia\u0026rsquo;s greater impact may stem from more intensive community-based ANC programs, while Rwanda could enhance its effectiveness by strengthening ANC-nutrition linkages through targeted supplementation, breastfeeding support, and sanitation improvements [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e] [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAnother critical finding is that ANC\u0026rsquo;s effectiveness has declined over time, with CATE dropping from \u0026minus;\u0026thinsp;0.0508 in 2005 to -0.0316 in 2015/16, mirroring trends where early health improvements stagnate as coverage expands[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e] [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Sustaining ANC impact requires continuous adaptation, including improving service quality, postnatal care, and integrating nutrition and sanitation support [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Governments should train healthcare workers, enhance supply chain management, and leverage digital health tracking to improve ANC service delivery [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e][\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eANC is most effective when combined with complementary interventions. Households practicing breastfeeding (CATE = -0.0379), improved sanitation (CATE = -0.0433), and vitamin A supplementation (CATE = -0.0509) experience greater reductions in malnutrition, aligning with prior findings on synergistic health interventions [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Coordinating ANC with breastfeeding education, micronutrient supplementation, and WASH (water, sanitation, and hygiene) programs enhances maternal and child health [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e] [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Strengthening cross-sector collaboration between health, nutrition, and sanitation programs is critical for maximizing ANC\u0026rsquo;s long-term impact on child malnutrition.\u003c/p\u003e \u003cp\u003eThe findings carry important policy implications, highlighting the need for context-specific and data-driven approaches to strengthen ANC programs. Expanding rural healthcare infrastructure, deploying mobile ANC services, and integrating maternal education programs can help close the urban-rural gap. Addressing economic barriers through pro-poor policies, such as nutrition assistance and transportation subsidies, can further improve ANC utilization and outcomes. Coordinating ANC with other essential maternal and child health services\u0026mdash;such as breastfeeding promotion, vitamin A supplementation, and sanitation improvements\u0026mdash;can significantly enhance its impact. The application of machine learning techniques, particularly the Causal Forest model, is a key strength of this study, allowing for precise estimation of treatment effects across subpopulations. However, the reliance on cross-sectional data limits the ability to establish long-term causal relationships. Future research should explore longitudinal data and incorporate qualitative methods to better understand contextual factors influencing ANC's effectiveness.\u003c/p\u003e"},{"header":"5. CONCLUSION","content":"\u003cp\u003eIn conclusion, this study provides robust evidence on the causal effects of antenatal care (ANC) in reducing child malnutrition, as measured by the Composite Index of Anthropometric Failure (CIAF). By leveraging advanced machine learning techniques, particularly the Causal Forest model, the study uncovers significant disparities in ANC\u0026rsquo;s effectiveness across socio-demographic factors, including residence, maternal education, household wealth, country-specific health policies, and time trends. These findings have important implications for policy and practice, offering a pathway toward more targeted and effective interventions to combat malnutrition.\u003c/p\u003e \u003cp\u003eBased on the analysis, countries like Ethiopia and Rwanda should prioritize interventions targeting the most influential factors. Urban residents, educated mothers, and wealthier households consistently benefited more from ANC, highlighting the need to address structural inequities in healthcare access and utilization. Rural populations, non-educated mothers, and poorer households faced greater barriers, underscoring the importance of expanding rural healthcare infrastructure, deploying mobile ANC services, and integrating maternal education programs into ANC services. Addressing economic barriers through pro-poor policies, such as nutrition assistance and transportation subsidies, can further improve ANC utilization and outcomes.\u003c/p\u003e \u003cp\u003eThe study also emphasizes the importance of integrating ANC with other essential maternal and child health services, such as breastfeeding promotion, vitamin A supplementation, and improved sanitation. Programs like Rwanda\u0026rsquo;s Fortified Blended Food (FBF) initiative offer valuable models for integrating ANC with social protection programs, ensuring that resources are allocated equitably and effectively. Additionally, the declining trend in ANC\u0026rsquo;s impact over time signals the need for sustained innovation, quality improvement, and integration with complementary interventions to maintain its long-term effectiveness.\u003c/p\u003e \u003cp\u003eThe application of machine learning techniques, particularly causal forests, represents a significant methodological advancement, enabling precise, flexible, and robust estimation of treatment effects across subpopulations. Cross-validation techniques enhance the reliability and generalizability of the findings, while feature importance analysis identifies key drivers of child malnutrition, helping policymakers prioritize high-impact interventions. Moving forward, leveraging technology-driven approaches and cross-country learnings can enhance maternal and child health interventions globally, ensuring that ANC continues to be a powerful tool in the fight against child malnutrition.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eEthics approval and consent to participate\u003c/p\u003e\n\u003cp\u003eThis study involved secondary analysis of publicly available datasets from the Demographic and Health Surveys (DHS) program, which are fully anonymized with no identifiable information on the participants. The DHS surveys are conducted following ethical guidelines, including informed consent obtained from all respondents, and adhere strictly to principles outlined in the Declaration of Helsinki. DHS protocols and procedures are reviewed and approved by the Institutional Review Board (IRB) of ICF International and by ethical review committees within each country involved. Therefore, the current analysis complies fully with ethical standards consistent with the Declaration of Helsinki.\u003c/p\u003e\n\u003cp\u003eConsent for publication\u003c/p\u003e\n\u003cp\u003eNot Applicable\u003c/p\u003e\n\u003cp\u003eCompeting interests\u003c/p\u003e\n\u003cp\u003eNot Applicable\u003c/p\u003e\n\u003cp\u003eFunding\u003c/p\u003e\n\u003cp\u003eNot Applicable\u003c/p\u003e\n\u003cp\u003eAuthors\u0026apos; contributions\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAlehegn Moges Tessema: Conceptualized the study, developed the methodology, conducted data analysis, interpreted the results, and drafted the manuscript.\u003c/p\u003e\n\u003cp\u003eProf. Temesgen Zewotir: Supervised the study, reviewed and edited the manuscript, provided guidance, and contributed to the interpretation of findings.\u003c/p\u003e\n\u003cp\u003eDr. Richard Kabanda: Supervised the study, reviewed and edited the manuscript, contributed to the interpretation of findings, and provided valuable feedback on the manuscript drafts.\u003c/p\u003e\n\u003cp\u003eAll authors have read and approved the final manuscript and agree to be accountable for all aspects of the work to ensure that questions related to its accuracy or integrity are appropriately investigated and resolved.\u003c/p\u003e\n\u003cp\u003eAcknowledgement\u003c/p\u003e\n\u003cp\u003eI would like to thank the African Center of Excellence for Data Science at the University of Rwanda for their invaluable support and guidance throughout this research. I am also grateful to the editorial team of BMC Public Health for providing the platform to publish this article and for their constructive feedback.\u003c/p\u003e\n\u003cp\u003eMy sincere appreciation also goes to the Demographic and Health Surveys (DHS) Program for providing the datasets that were instrumental to this study.\u003c/p\u003e\n\u003cp\u003eThis work would not have been possible without the support of these institutions, and I am deeply grateful for their contributions.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBlack RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, Mathers C. Maternal and child undernutrition: global and regional exposures and health consequences. Lancet. 2008;371(9608):243\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUNICEF. Levels and trends in child malnutrition, UNICEF-WHO-World Bank Group Joint Malnutrition Estimates, 2020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEthiopian Public Health Institute (EPHI) \u0026amp;, Program DHS. Ethiopia Mini Demographic and Health Survey 2019: Final Report. DHS Program, Addis Ababa, Ethiopia; 2019.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Institute of Statistics of. Rwanda (NISR) \u0026amp; DHS Program, Rwanda Demographic and Health Survey. Rwanda: DHS Program, Kigali; 2020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartorell R, Horta BL, Adair LS, Stein AD, Richter L, Fall CH, Victora CG. Improved nutrition in the first 1,000 days and its effects on adult human capital and health. Am J Clin Nutr, 98, 5, pp. 1192S-1200S, 2013.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVictora CG, Adair L, Fall CH, Hallal PC, Martorell R, Richter L, Sachdev HS. Maternal and child undernutrition: consequences for adult health and human capital. Lancet. 2008;371(9609):340\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBhutta ZA, Ahmed T, Black RE, Cousens S, Dewey K, Giugliani E, Shekar M. What works? Interventions for maternal and child undernutrition and survival. Lancet. 2008;371(9610):417\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHaddad L, Achadi E, Bendech MA, Ahuja A, de Pee S, Engesveen K, Bhutta ZA. The Global Nutrition Report 2015: Actions and accountability to advance nutrition and sustainable development. International Food Policy Research Institute (IFPRI); 2015.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRuel MT, Alderman H, a. M, Group CNS. Nutrition-sensitive interventions and programmes: How can they help to accelerate progress in improving maternal and child nutrition? Lancet. 2013;382(9891):536\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSmith LC, Ruel MT, Ndiaye A. Why is child malnutrition lower in urban than in rural areas? Evidence from 36 developing countries. World Dev. 2018;105:274\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects, \u003cem\u003eBiometrika\u003c/em\u003e, vol. 70, no. 1, pp. 41\u0026ndash;55, 1983.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWager S, Athey S. Estimation and inference of heterogeneous treatment effects using random forests. J Name: J Am Stat Association. 2018;113(523):1228\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBhutta ZA, Das JK, Rizvi A, Gaffey MF, Walker N, Horton S, Black RE. Evidence-based interventions for improvement of maternal and child nutrition: What can be done and at what cost? Lancet. 2013;382(9890):452\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlack RE, Allen LH, Bhutta ZA, Caulfield LE, de Onis M, Ezzati M, Mathers C. Maternal and child undernutrition: Global and regional exposures and health consequences. Lancet. 2013;382(9890):427\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Health Organization (WHO). (2016). Recommendations on antenatal care for a positive pregnancy experience. World Health Organization, Recommendations on antenatal care for a positive pregnancy experience, World Health Organization, 2016.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-nutrition","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nutn","sideBox":"Learn more about [BMC Nutrition](http://bmcnutr.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/nutn/default.aspx","title":"BMC Nutrition","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Antenatal Care (ANC), Child Malnutrition, Composite Index of Anthropometric Failure (CIAF), Causal Forest Model, Machine Learning, Heterogeneous Treatment Effects, 5-Fold Cross-Validation","lastPublishedDoi":"10.21203/rs.3.rs-6131494/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6131494/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMalnutrition among children under five remains a critical public health challenge in Ethiopia and Rwanda, with stunting rates of 37% and 33%, respectively. This study examines the causal effects of antenatal care (ANC) on child malnutrition, measured using the Composite Index of Anthropometric Failure (CIAF), leveraging data from the Demographic and Health Surveys (DHS). Using machine learning techniques, specifically the Causal Forest model, the study estimates both overall and subgroup-specific impacts of ANC while adjusting for socio-demographic and environmental factors.\u003c/p\u003e \u003cp\u003eResults show that ANC significantly reduces child malnutrition, with an overall reduction of 2.63 percentage points. However, the impact varies across groups: urban residents, educated mothers, and wealthier households benefit more, highlighting disparities. For example, urban areas see a 7.05 percentage point reduction compared to 3.06 in rural areas. Complementary factors like breastfeeding, clean water access, and improved sanitation further enhance ANC\u0026rsquo;s effectiveness. However, the impact of ANC has declined over time, underscoring the need for program improvements.\u003c/p\u003e \u003cp\u003eBy employing the Causal Forest model and validating results through 5-fold cross-validation, this study provides robust, data-driven insights into ANC\u0026rsquo;s heterogeneous effects. These findings highlight the importance of addressing structural inequities and integrating ANC with broader health interventions. The use of machine learning offers policymakers precise, actionable evidence to design targeted, equitable strategies for improving child nutrition in Ethiopia, Rwanda, and similar contexts.\u003c/p\u003e","manuscriptTitle":"Causal Effects of Antenatal Care (ANC) on Child Malnutrition: A Machine Learning Approach in Ethiopia and Rwanda","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-30 11:33:44","doi":"10.21203/rs.3.rs-6131494/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-07-11T11:15:26+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-05-31T15:31:23+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-05-19T07:04:12+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-05-09T08:16:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"177000371059789355709418351378043656999","date":"2025-05-08T13:39:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"289721515272399237349184595858204437892","date":"2025-05-08T11:18:55+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"277974525961360254684695053986902659167","date":"2025-05-03T09:30:49+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-19T17:41:03+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-15T08:08:43+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-03-27T11:52:06+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-03-24T09:19:59+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Nutrition","date":"2025-03-24T09:18:48+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-nutrition","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"nutn","sideBox":"Learn more about [BMC Nutrition](http://bmcnutr.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/nutn/default.aspx","title":"BMC Nutrition","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d6d358bb-ad13-4b5c-aaed-bfd21038403e","owner":[],"postedDate":"April 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-12-01T16:05:30+00:00","versionOfRecord":{"articleIdentity":"rs-6131494","link":"https://doi.org/10.1186/s40795-025-01208-w","journal":{"identity":"bmc-nutrition","isVorOnly":false,"title":"BMC Nutrition"},"publishedOn":"2025-11-28 15:58:44","publishedOnDateReadable":"November 28th, 2025"},"versionCreatedAt":"2025-04-30 11:33:44","video":"","vorDoi":"10.1186/s40795-025-01208-w","vorDoiUrl":"https://doi.org/10.1186/s40795-025-01208-w","workflowStages":[]},"version":"v1","identity":"rs-6131494","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6131494","identity":"rs-6131494","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}