Prediction of incomplete vaccination among children aged 12-35 months in sub-Saharan Africa: Application of machine learning algorithm using recent (2021-2024) Demographic and Health Survey Data

Prediction of incomplete vaccination among children aged 12-35 months in sub-Saharan Africa: Application of machine learning algorithm using recent (2021-2024) Demographic and Health Survey Data | 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 Prediction of incomplete vaccination among children aged 12-35 months in sub-Saharan Africa: Application of machine learning algorithm using recent (2021-2024) Demographic and Health Survey Data Abdulkerim Hassen Moloro, Bizunesh Fantahun Kase, Angwach Abrham Asnake, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8945443/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Despite immunization preventing millions of deaths worldwide, sub‑Saharan Africa continues to face a heavy burden of vaccine‑preventable diseases, with nearly one in five children missing essential doses such as the third diphtheria, tetanus, and pertussis vaccine. Routine programs remain constrained by maternal, household, and contextual barriers, leaving the region far from achieving IA2030 targets. Earlier studies relied mainly on descriptive or regression‑based analyses using older datasets, limiting predictive accuracy. This study applies modern machine learning to recent DHS data (2021–2024), offering improved prediction and transparent identification of key determinants of incomplete vaccination. Methods: A secondary analysis was conducted using recent Demographic and Health Survey (DHS) data (2021–2024) from 16 sub‑Saharan African countries. The weighted sample comprised 57,527 children aged 12–35 months. Data cleaning, harmonization, and pooled analysis were performed in STATA 17, with forest plots illustrating pooled and country‑specific incomplete vaccination rates. Eight supervised machine learning algorithms; Naïve Bayes, Decision Trees, K‑Nearest Neighbor, Logistic Regression, Artificial Neural Networks, Extreme Gradient Boosting (XGBoost), CatBoost, and Random Forest were applied for classification and comparison. SHAP analysis enhanced interpretability by ranking maternal, household, and contextual predictors. All analyses were conducted in Python 3.10.2 within Google Colab using scikit‑learn, imblearn, XGBoost, CatBoost, and SHAP packages. Result: The pooled prevalence of incomplete vaccination among children aged 12–35 months in 16 sub‑Saharan African countries was 46.21% (95% CI: 38.58, 53.83%), with the lowest level observed in Ghana (25.21%) and the highest in the Democratic Republic of Congo (73.21%). CatBoost emerged as the best‑performing machine learning algorithm for predicting incomplete childhood vaccination, achieving the highest accuracy (65%) and area under the curve (AUC (70%)) among the models tested. SHAP feature importance analysis revealed that adequate antenatal care visits, maternal media exposure, institutional delivery, being rural residence, health insurance coverage, married marital status, birth order between two and four, and household with high wealth index were the most influential attributes in predicting vaccination outcomes. Conclusion: In conclusion, this study reveals that nearly half of children aged 12–35 months in sub‑Saharan Africa remain incompletely vaccinated, with striking disparities across countries from Ghana’s relatively low 25.21% to the Democratic Republic of Congo’s alarming 73.21%. CatBoost achieved strong predictive accuracy and SHAP feature importance analysis revealed adequate antenatal care visits, maternal media exposure, institutional delivery, being rural residence, health insurance coverage, married marital status, and household with high wealth index were the most influential attributes in predicting vaccination outcomes. These findings underscore the urgent need for targeted interventions that strengthen maternal health services, expand access to facilities, and reduce rural urban inequities. Integrating AI‑driven monitoring into immunization programs offers policymakers actionable tools to accelerate progress toward Immunization Agenda 2030 and safeguard child health. Incomplete vaccination Children aged 12–35 months Sub‑Saharan Africa Machine learning algorithms Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Immunization currently prevents more than 4 million deaths each year worldwide. In the African region alone, vaccines save approximately 800,000 lives annually( 1 ). Despite this progress, in 2023 an estimated 21 million children globally missed lifesaving diphtheria, pertussis, and tetanus (DPT) vaccines, while 6.5 million received only partial vaccination( 2 , 3 ). In Africa, one in five children failed to receive essential immunizations such as the third dose of DPT (DTP3)( 4 ). The basic childhood vaccines include BCG, pentavalent, polio, and measles( 5 ). In Africa, vaccine-preventable diseases (VPDs) such as measles, diphtheria, tetanus, polio, and pertussis conditions nearly eradicated in many high-income countries continue to affect more than 30 million children under five each year. These infections cause over 500,000 deaths annually, representing about 58% of all global child deaths from VPDs( 1 , 6 ). Globally, an estimated 700,000 children under five died from VPDs in 2018, with nearly 99% of these deaths occurring in low- and middle-income countries( 7 ). In 2023, VPDs accounted for almost 30% of all under-five deaths in sub-Saharan Africa, where the mortality rate among children under five was nearly 14 times higher than in Europe and 18 times higher than in Australia and New Zealand( 3 , 8 ). Despite significant efforts to expand immunization coverage across Africa over the past decade, routine vaccination programs continue to face major challenges. As long as children remain vulnerable to outbreaks of vaccine-preventable diseases, most African countries are unlikely to achieve their Sustainable Development Goals (SDGs)( 4 ). The Immunization Agenda 2030 (IA2030) sets ambitious targets of reaching 90% coverage and reducing the number of “zero-dose” children worldwide to fewer than 6.5 million by 2030( 9 ). However, in sub-Saharan Africa, only 54.1% of children receive the full set of basic childhood vaccines, while 36.1% are only partially immunized. Evidence shows that the region continues to struggle in meeting IA2030 targets( 10 – 12 ), with substantial variation in coverage across countries( 13 , 14 ). Factors contributing to incomplete immunization include young maternal age, limited knowledge about vaccination, and negative perceptions of vaccine side effects. Children of single mothers, those born without the assistance of a skilled birth attendant, and those whose mothers did not receive postnatal care are also more likely to be incompletely immunized. Additional determinants include poor maternal awareness of routine immunization, residence in rural areas, low household income, and living more than 30 minutes from the nearest vaccination facility( 15 , 16 ). Previous research on childhood immunization in sub‑Saharan Africa has provided valuable insights into spatial heterogeneity and country‑specific determinants, but methodological limitations remain. Most studies relied on descriptive statistics or regression‑based approaches, such as geographically weighted regression (GWR), which assume linear relationships and cannot fully capture the complex, nonlinear interactions among diverse predictors( 17 , 18 ). While GWR improved local interpretability, it did not provide predictive accuracy benchmarks or systematic feature ranking across multiple determinants. Furthermore, earlier works often used older DHS datasets collected before 2020( 11 , 19 ), rely on specific country( 20 ) and focused narrowly on children aged 12–23 months, leaving broader age ranges and more recent immunization trends underexplored( 13 ). As a result, prior evidence offered limited predictive power and insufficient transparency in explaining the drivers of incomplete vaccination across sub‑regions. This study addresses those gaps by applying ensemble machine learning algorithms including XGBoost, Random Forest, and CatBoost on recent DHS datasets (2021–2024) from 16 sub‑Saharan African countries, covering children aged 12–35 months. Unlike regression models, machine learning captures nonlinear interactions and provides robust predictive accuracy, while SHAP analysis enhances transparency by ranking the most influential maternal, household, and contextual factors. This methodological innovation not only strengthens academic evidence but also equips policymakers with actionable insights to identify high‑risk populations, design targeted interventions, and reduce the burden of “zero‑dose” children. By bridging the limitations of earlier regression‑based studies, this paper offers a recent reproducible, data‑driven framework that supports regional progress toward the Immunization Agenda 2030 goals and improves child health outcomes in sub-Saharan Africa. Methods Study Area, Design, and Period This study conducted secondary analysis using data from the Demographic and Health Surveys (DHS) of 23 countries in Sub-Saharan Africa (SSA). The countries involved were Angola, Burkina Faso, DR Congo, Côte d'Ivoire, Gabon, Ghana, Kenya, Lesotho, Madagascar, Mali, Mauritania, Mozambique, Nigeria, Senegal, Tanzania, and Zambia. The selection of country was based on the recent survey year, availability of a standardized and unrestricted dataset, and presence of observations for the outcome variable in the datasets. The DHS surveys across all countries employed a cross-sectional study design to collect data on basic sociodemographic characteristics and various health indicators, including maternal and health facility related. For the current analysis, we included the countries that have their recent DHS conducted between 2021 and 2024. Population, sampling technique and weight The source population for this study comprised all children aged 12 to 35 months old in 16 selected SSA countries. The study population included all children aged 12 to 35 months in the survey. Across all countries, the surveys used a multistage stratified cluster sampling technique to select the study participants. In the first stage, each country was divided into clusters, and clusters were randomly selected based on the probability proportional to their contribution to overall country’s population. In the second stage, using the housing census as a sampling frame, a representative number of households was selected from each cluster. To account for the complex survey design, non-response, and to ensure comparability across countries while preventing any single country or survey year from disproportionally influencing the pooled estimates, DHS sampling weights were applied as recommended. The weight variable (v005) was rescaled by dividing by 1,000,000 and incorporated into pooled analyses using the svyset command in Stata, accounting for primary sampling units (v021) and strata (v022). This produced a weighted analytic sample of 57, 527 children aged 12–35 months with complete data on the variables of interest (Table 1). Additionally, sensitivity analyses were conducted to assess the influence of individual countries by systematically excluding one country at a time and recalculating the pooled prevalence. This approach adjusts for unequal probabilities of selection and non-response within each survey, ensuring that each country’s contribution to the pooled prevalence is proportional to its population size. Table-1: Survey years and sample sizes of children aged 12–35 months from 16 Sub-Saharan African (SSA) countries included in the study, from 2021 to 2024 Country Survey year Unweighted sample size Weighted sample size Angola 2023–2024 2,911 2,746 Burkina Faso 2021 4,215 4,194 DR Congo 2023–2024 6,020 5,876 Cote d’Ivoire 2021 3,424 3,246 Gabon 2021 1,048 699 Ghana 2022–2023 3,582 3,296 Kenya 2022 6,978 6,390 Lesotho 2023–2024 970 929 Madagascar 2021 3,702 3,719 Mali 2023–2024 4,623 5,056 Mauritania 2021 492 946 Mozambique 2022–2023 3,137 3,198 Nigeria 2023–2024 6,432 6,484 Senegal 2023 3,606 3,498 Tanzania 2022 3,992 4,008 Zambia 2024 3,317 3,242 Total 58,449 57,527 Data source The data for this study were obtained from the DHS women’s questionnaire, focusing on children aged 12–35 months, and from the Kids Record dataset (KR file) across 16 countries. All datasets were sourced directly from the official Demographic and Health Surveys (DHS) program website ( https://dhsprogram.com ). Data extraction and management of missing observations Prior to data extraction, we identified all sub-Saharan African countries with DHS datasets collected between 2021 and 2024, as our analysis focused exclusively on this timeframe. The DHS surveys employed standardized data collection tools and face-to-face interviews to ensure comparability across countries. For this study, we extracted relevant variables from the Kids Record (KR) files of 16 countries and merged them into a pooled dataset. Country identifiers and survey weights were retained to account for sampling design and cross-country differences. Variables were harmonized and recoded consistently across surveys, following the Guide to DHS Statistics, to ensure compatibility. For incomplete vaccination outcomes, children aged 12–35 months were included. Records were excluded if vaccination information was missing, incomplete, or flagged as inconsistent with DHS protocols. Specifically, children were dropped if: (i) vaccination cards were not seen and caregiver recall was unavailable, (ii) vaccination dates were missing or incomplete, or (iii) implausible vaccination sequences were recorded (e.g., doses out of chronological order). These exclusions were applied uniformly to both numerator and denominator, constituting a complete-case analysis. Independent variables were also subjected to complete-case analysis. Observations with missing values for any covariates included in the final machine learning models were excluded. This approach ensured transparency and minimized bias due to incomplete data. While complete-case analysis may reduce sample size, it enhances the validity of cross-country comparisons by maintaining consistency in variable definitions and data quality. All procedures for handling missing data adhered to DHS methodological guidelines. A detailed description of DHS protocols for managing incomplete or implausible observations is available in the Guide to DHS Statistics( 21 ). Variables and measurements Dependent Variable : The dependent variable was defined as incomplete immunization with basic childhood vaccines among children aged 12–35 months. A fully vaccinated child is one who has received all required doses of the basic childhood vaccines, namely BCG, OPV1, OPV2, OPV3, Penta1, Penta2, Penta3, and MCV1. Missing at least one of these vaccines was considered incomplete immunization. The eight recoded variables were then summed, and the outcome was categorized as follows: “0” for children who received all of the recommended vaccines, and “1” for those who missed at least one of the basic vaccines( 17 , 19 , 22 ). Independent Variables A total of 18 variables were incorporated into this study, selected through a systematic process that considered both their availability in the Demographic and Health Survey (DHS) and their documented associations with childhood stunting in prior research( 11 , 17 – 19 , 22 ). Each variable was chosen for its relevance to the study objectives and its role as a known or hypothesized determinants of incomplete vaccination reflecting influences at individual, household, and contextual level. These variables were categorized into individual-level factors, household factors, and contextual factors. The individual-level factors included the sex of the child (male and female), age of the child in month, birth order (1, 2–4, and 5+), perceived size at birth (large, average, and small). Other individual-level factors were maternal educational level (no formal education, primary, secondary, and higher), current maternal working status (yes and no), antenatal care visits during pregnancy (0, 1–3, and 4 or more) and postnatal checks within 2 months (yes and no). Variables such as maternal age ( 15 – 24 and 25 – 49 ), place of delivery (home, health facility, other), and marital status (single and married) were also considered. Household factors included household size (small, medium, and large), exposure to media (yes and no), health insurance coverage, distance to health facility and wealth index (poor, middle, rich). Contextual factors encompassed the place of residence (urban and rural) and geographical sub-regions (West, East, Central, and Southern). Data management and statistical analysis Stata version 17 was utilized for data cleaning, labelling and pooled analysis. Prior to the analysis, the presence of the outcome variable in the DHS dataset for each country was confirmed. All the variables considered in the study were reviewed for missing values. Subsequently, the datasets from 16 SSA countries were appended and weighted to maintain sample representativeness and obtain reliable estimates and standard errors. The pooled prevalence of child incomplete vaccination was calculated using weighted data on the number of affected children with outcome variable and the total number of study participants in each country included in the analysis. The Stata command for meta-analysis “metan” was executed to present the country-specific and pooled estimates with 95% CI in a forest plot. To handle heterogeneity, meta-regressions, sensitivity analysis and sub-group analysis were conducted( 23 ). Before developing a classification model, preprocess the raw data for analysis was crucial to increase model accuracy and performance( 24 ). Yufeng Guo’s steps of machine learning were used to classify knowledge of fertility period, such as data collection, data preparation, model selection, model training, model evaluation, and prediction( 25 ). In addition, interpretation of the model was performed using SHAP. Data analysis was conducted in Google Coolab using the Python 3.10.2 programming language within a Jupiter Notebook environment. The packages imblearn, scikit-learn (sklearn)( 26 ), XGBoost( 27 ), and SHAP( 28 ) were utilized to perform the necessary calculations and analyses (Fig. 1 ). Model selection and development Supervised classifier machine learning algorithms were used to evaluate the predictive power of machine learning techniques in classifying incomplete vaccination among children aged 12–35 months. Naïve Bayes, Decision Trees, K-Nearest Neighbor (KNN), Logistic Regression, Artificial Neural Networks (ANN), Extreme Gradient Boosting (XGBoost), Cat Boosting (CatBoost) and Random Forest were used to select the best model. Those algorithms were chosen based on their interpretability, amount of features, computational efficiency, accuracy, and characteristics of the dataset( 29 ). Data pre-processing Data pre-processing was carried out to address missing and incomplete records, as well as duplicate entries. The dataset contained noise, outliers, and inconsistencies, which were systematically managed to ensure data quality. At this stage, unnecessary values and duplicate variables were removed. In addition, all string and categorical variables were transformed into nominal data types to facilitate processing. Model training The selected supervised machine learning classifiers were applied to analyze the dataset, focusing on binary classification of incomplete vaccination. For the final classification on unseen test data, the best-performing predictive model was identified and trained to ensure optimal performance. Model evaluation Model performance was evaluated using the receiver operating characteristic (ROC) curve, which measures the ability to distinguish between classes, assists in selecting optimal thresholds, and enables comparison across models. Key metrics included the area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Among these, AUC was considered the most reliable indicator of discriminatory power, effectively demonstrating the model’s ability to differentiate between children with incomplete vaccination and complete vaccination. The higher AUC obtained from the training dataset highlighted superior discriminatory performance compared to accuracy, sensitivity, specificity, PPV, and NPV( 30 ). Making classification and model interpretation Classification was performed using selected predictor variables. The best-performing classifier was chosen based on its accuracy. To interpret the model, Shapley Additive Explanations (SHAP) were employed. Because many powerful machine learning models particularly tree-based algorithms are often regarded as “black boxes,” their internal decision-making processes are rarely transparent in research( 31 ). SHAP, a game-theory–based approach, provides a robust framework for explaining model outputs both locally (at the individual level) and globally (across the entire dataset). This technique reduces the limitations of interpreting complex machine learning findings and enhances the transparency of the classification results( 32 ). Ethical considerations For this study, we utilized Demographic and Health Survey (DHS) data from 16 sub-Saharan African countries. The DHS survey procedures were approved by the ICF Institutional Review Board (IRB) and the respective host country IRB; therefore, no additional ethical approval was required for this secondary analysis. The dataset accessed contain no identifiable participant information, ensuring confidentiality and privacy. Access to the data was formally authorized by the DHS program, which serve as the institutional custodian of the datasets, through an online request submitted at https://dhsprogram.com and supported by authorization letter AuthLetter_215093 on dated 01/19/2025. Results Child and mothers’ characteristics The study included and analyzed a weighted sample of 57,527 children aged 12 to 35 months with mothers. Of these children, 50.68% (29,155) were male, and 55.99% (32,211) were between 12–24 months old. Vaccination incomplete was highly prevalent, affecting 54.48% (14,270) of children between 12–24 months, compared to older age groups. Among the included children, 60.83% (34,993) lived in rural areas, and 40.1% (27,071) were from households in the poor wealth index category. Additionally, 46.45% (26,720) of the children were from West African countries, and 21.31% (12,257) were delivered at home (Table 2 ). Of the 57,527 children’s mothers included in the study, 17,952 (31.21%) were between 15 and 24 years old, and 14,832(25.78%) were currently single. About one-third of the mothers, 16,850(29.31%) had no formal education. Among included children’s mothers, 21,543(37.45%), of the mothers were not working during the time of the interview and only 18,052 (31.38%) had no access to mass media. Regarding household size, the majority of children 51,184(88.97%) lived in large households. More than half of the children 29,688(51.61%) were reported to have an average birth size, and 7,667 (13.33%) children’s mothers did not receive a post-natal checkup within two months after birth. Among mothers, 12,934(22.48%) attended only one to three antenatal care visits, while 2,614(4.54%) had no antenatal care. In terms of birth order, most children 28,528 (49.59%) were born as second to fourth order births. Furthermore, 20,970(36.45%) respondents reported that distance to a health facility was a major problem when seeking medical help, and 51,900(90.22%) households had no health insurance coverage. Table 2 Child and mothers characteristic by vaccine incomplete and their overall background status, in 16 SSA countries, 2021–2024 (n = 57,527 ) Characteristics Weighted Frequency (%) Vaccine incomplete Sex of the child Male 29,155(50.68) 13,174(50.30) Female 28,372(49.32) 13,017(49.70) Age of the child (months) 12–24 months 32,211(55.99) 14,270(54.48) 25–35 months 25,316(44.01) 11,921(45.52) Birth order First 14,166(24.62) 6,079(23.21) 2 to 4 28,528(49.52) 12,800(48.87) 5 or higher 14,833(25.79) 7,312(27.49) Perceived size at birth Large 17,291(30.06) 8,129(31.04) Average 29,688(51.61) 13,539(51.69) Small 10,547(18.33) 10,547(18.33) Mothers’ educational status No education 16,860(29.31) 8,247(31.49) Primary 16,594(28.85) 7,753(29.60) Secondary 19,959(34.69) 8,767(33.47) Higher 4,114(7.15) 1,424(5.44) Mother currently working Yes 35,984(62.55) 16,128(61.58) No 21,543(37.45) 10,063(38.42) Postnatal care visits Yes 49,860(86.67) 22,384(85.46) No 7,667(13.33) 3,807(14.54) Mother’s age (years) 15–24 17,952(31.21) 8,494(32.43) 25–49 39,575(68.79) 17,697(67.57) Number of ANC visits No visit 2,614(4.54) 1,646(6.29) 1 to 3 visits 12,934(22.48) 6,660(25.43) 4 and above visit 41,979(72.97) 17,885(68.29) Place of delivery Home 12,257(21.31) 6,450(24.63) Health facility 45,270(78.69) 19,741(75.37) Marital status Single 14,832(25.78) 7,536(28.77) Married 42,695(74.22) 18,655(71.23) Household size Small 3,572(6.21) 1,687(6.44) Medium 2,770(4.82) 1,388(5.30) Large 51,184(88.97) 23,117(88.26) Mass media exposure No 18,052(31.38) 9,910(37.84) Yes 39,475(68.62) 16,281(62.16) Health insurance coverage No 51,900(90.22) 24,400(93.16) Yes 5,627(9.78) 1,791(6.84) Distance to health facility Big problem 20,969(36.45) 9,858(37.64) Not a big problem 36,557(63.55) 16,333(62.36) Wealth index Poor 23,071(40.10) 11,156(42.59) Middle 11,285(19.62) 5,138(19.62) Rich 23,170(40.28) 9,897(37.79) Place of residence Urban 25,534(39.17) 10,367(39.58) Rural 34,993(60.83) 15,824(60.42) Subregion West African Countries 26,719(46.45) 11,217(42.83) East African Countries 14,116(24.54) 5,316(20.30) Central African Countries 6,575(11.43) 4,655(17.77) Southern African Countries 10,117(17.59) 5,003(19.10) Pooled prevalence of incomplete vaccination among children aged 12–35 months in sub–Saharan African The pooled prevalence of incomplete vaccination among children aged 12–35 months 16 sub-Saharan African (SSA) countries was 46.21% (95% CI: 38.58, 53.83%), with significant variation observed across countries (I 2 = 99.8, P-value = 0.000) (Fig. 2 ). Democratic Republic of Congo reported the highest incomplete vaccination at 73.21% while Ghana had the lowest incomplete vaccination prevalence at 25.21%. Among the 16 countries analyzed, 13 had incomplete vaccination prevalence of 30% or higher among children aged 12–35 months. Handling heterogeneity The random-effects model revealed considerable heterogeneity. To address this, sensitivity analysis, subgroup analysis, and meta-regression were conducted. Sensitivity analysis Sensitivity analysis was performed to evaluate the effect of individual country on the pooled estimated. When individual country was omitted, the pooled prevalence obtained was within the 95% CI of the overall pooled prevalence. This confirms the absence of single study impact on the overall pooled effect size. Therefore, from the random effects model, there were no country that excessively influence the overall pooled estimate of incomplete vaccination (Fig. 3 ). Sub-group analysis Subgroup analyses were carried out by sub-region, and year of publication. Stunting by Sub-Region Subgroup analysis by sub-region indicated that Central African countries had the highest pooled prevalence of incomplete vaccination among children aged 12–35 months (64.61%, 95% CI: 47.70–81.52), followed by Southern African countries (48.12%, 95% CI: 32.79–63.45), West African countries (43.12%, 95% CI: 32.38–53.86) and East African countries (38.59%, 95% CI: 32.81–44.36). Substantial heterogeneity was observed in West and Central Africa (I² = 99.7%, P = 0.00), Central Africa (I² = 99.5%, P = 0.00), East African (I² = 98%, P = 0.00), and Southern Africa (I² = 99.6%, P = 0.00) (Fig. 4 ). Stunting by Year of Survey Subgroup analysis by year of survey showed that the pooled prevalence of incomplete vaccination was highest in 2021–2024 (46.21%, 95% CI: 38.58–53.83) and lowest in 2018–2020 (38.77%, 95% CI: 20.06–57.48). Both periods demonstrated significant heterogeneity (2021–2024: I² = 99.8%, p < 0.00; 2018–2020: I² = 99.9%, p < 0.00) (Fig. 5 ). Meta-regression A meta-regression was performed to assess whether the year of survey, sample size, country, and sub-region could explain the heterogeneity in incomplete vaccination prevalence (Table 3 ). None of these variables were statistically significant predictors of heterogeneity (all p-values > 0.05). The high residual heterogeneity in the pooled prevalence of incomplete vaccination suggests that other unmeasured factors are responsible for the variability across studies. Table 3 Meta-regression for the studies of pooled incomplete vaccination among children aged 12–35 months in 16 sub-Sharan Africa, 2021–2024. Outcome categories Variables Coefficients P-value [95% conf. interval] Pooled incomplete vaccination among 12–35 months children Year of Survey 1.02 0.70 0.88, 1.19 Sample size 1.00 0.86 0.99, 1.00 Countries 0.98 0.41 0.96, 1.01 Sub-region 0.91 0.32 0.76, 1.01 Model performance to predict childhood incomplete vaccination in sub-Saharan Africa Ten machine learning algorithms were applied to predict childhood incomplete vaccination in sub-Saharan Africa. The models evaluated included; Naïve Bayes, Decision Trees, K-Nearest Neighbor (KNN), Logistic Regression, Artificial Neural Networks (ANN), Extreme Gradient Boosting (XGBoost), Cat Boosting (CatBoost) and Random Forest. Their performance was assessed using Receiver Operating Characteristic (ROC) curves, Area Under the Curve (AUC), Accuracy, Precision, Recall (Sensitivity), F1-score, and Ten-Fold Cross-Validation (Table 4 ). Among the evaluated models CatBoost and XGBoost demonstrated the strongest predictive performance for incomplete vaccination in sub-Saharan Africa. CatBoost achieved the highest overall accuracy (65%) and AUC (70%), supported by balanced precision (66%) and recall (61.45%), yielding an F1-score of 63.36%. XGBoost closely followed, with an accuracy of 64.52% and an AUC of 70%, alongside a precision of 65.36% and recall of 61.78%. Overall, the comparative analysis underscores that CatBoost are the most effective algorithms for predicting childhood vaccination outcomes in sub-Saharan Africa, offering reliable and well-balanced classification performance (Fig. 6 ). Table 4 Model performance comparison of the included machine learning algorithm for vaccination incomplete among children aged 12–35 months in 16 sub-Sharan Africa, 2021–2024. Models Accuracy AUC Precision Recall F1-Score Random Forest 63.45 68.02 63.50 63.28 63.39 Decision Tree 60.00 60.03 60.80 57.78 58.71 K-Nearest Neighbor (KNN) 61.59 65.37 61.61 61.51 61.56 Extreme Gradient Boosting (XGBoost) 64.51 70.00 65.36 61.78 63.52 Logistic Regression 63.43 68.98 64.91 58.47 61.52 Naïve Bayes 62.36 67.15 64.16 55.98 59.79 Artificial Neural Networks (ANN) 62.48 67.15 62.80 63.01 62.91 Cat Boosting (CatBoost) 65 70.00 65.95 61.45 63.62 AUC = Area Under the Curve Global feature selection with SHAP SHAP global feature importance was used to identify the top independent variables that determine incomplete vaccination. The CatBoost classifier algorithm was employed to optimize and select important features using SHAP. To determine the top predictors of incomplete vaccination, the mean absolute SHAP value (MASHAPV) was calculated for each independent variable. The top eight independent features were then selected based on their MASHAPV in descending order. Those features were maternal media exposure (media_exposure_1), antenatal care attendance (antenatal_visits_2), being rural residence (Residence_2), being married (marital_status_2), health insurance coverage (Health_insurance_coverage_1), delivery in a health facility (place_of_delivery_2), birth order 2 to 4 (birth_order_category_3), and rich household wealth index (wealth_index_3). These features were determined to be the most important predictors of incomplete vaccination among the study population (Fig. 7 ). Model interpretation with bees warm SHAP plot The SHAP bees warm plot was used to visualize and provide a dense summary of how independent features influence the likelihood of incomplete vaccination among children aged 12–35 months in sub‑Saharan Africa (Fig. 8 ). Each dot represents an individual child’s data point, with color indicating the original feature value: blue for low values (coded “0”) and red for high values (coded “1”). Points located on the left side of the vertical axis (negative SHAP values) indicate that the feature reduces the likelihood of incomplete vaccination, whereas points on the right side (positive SHAP values) indicate that the feature increases the likelihood of incomplete vaccination. The plot highlights the relative contribution of key sociodemographic and health service factors. Antenatal care attendance (antenatal_visits_2) and maternal media exposure (media_exposure_1) emerged as the strongest protective determinants, with higher values shifting predictions toward complete vaccination. Children whose mothers reported antenatal care visits and exposure to health information through media were less likely to experience incomplete vaccination, as indicated by negative SHAP values for higher feature categories. Rural residence (Residence_1) was consistently associated with increased risk of incomplete vaccination. Children of mothers residing in rural areas were consistently at increased risk of incomplete vaccination, as indicated by positive SHAP values for rural residence. Health insurance coverage (Health_insurance_coverage_1) and delivery in a health facility (place_of_delivery_2) were protective, shifting predictions toward complete vaccination. Children whose mothers reported having health insurance and giving birth in a health facility were less likely to experience incomplete vaccination, as these conditions shifted predictions toward complete vaccination. Married marital status (marital_status_2), birth order 2 to 4 (birth_order_category_3), and rich household wealth index (wealth_index_3) also contributed to reduced likelihood of incomplete vaccination. Specifically, children of married mothers, those born with birth order between two and four, and those living in rich households were less likely to experience incomplete vaccination, as reflected by negative SHAP values for these categories. Overall, the plot demonstrates that access to maternal health services, being married, low birth order, socioeconomic advantages, and exposure to health information are critical in lowering the probability of incomplete vaccination, whereas rural residence increase vulnerability to incomplete vaccination. Discussion By 2024, the number of children missing all routine vaccinations (zero-dose children) had declined to approximately 14.3 million, down from the 25 million reported in 2021. However, this figure remains nearly 1.4 million higher than pre-pandemic levels, signaling stalled progress toward achieving the Immunization Agenda 2030 (IA2030) targets( 33 ). In March 2025, the U.S. government ended a $ 2.63 billion grant to Gavi, the Vaccine Alliance, citing concerns about the organization’s focus on “zero-dose” children( 34 ). By ending such a large grant, millions of children specifically 75 million may no longer have access to essential vaccines( 35 ). Crucially, the sudden termination of the Demographic and Health Surveys (DHS) Program in February 2025 has left a critical data vacuum, as DHS previously provided over 50% of child mortality data in sub-Saharan Africa( 36 ). According to this study, the pooled prevalence of incomplete vaccination among children aged 12–35 months was 46.21% (95% CI: 38.58–53.83%), with the Democratic Republic of Congo reporting the highest rate at 73.21% and Ghana the lowest at 25.21%, indicating that nearly one in two children in the region did not receive the full set of WHO-recommended vaccines. The finding is consistent with prevalence reported from Pakistan (46%)( 37 ). However, the finding is higher than previous study conducted in sub-Saharan Africa, which indicated that the pooled prevalence of incomplete vaccine among children aged 12–23 months in 16 countries was 35.5%, with Rwanda recording the lowest rate at 4.21% and Mauritania the highest at 56.3%( 38 ). The findings also higher than study conducted in India (32%)( 39 ), in Africa (35.5%) and Sub-Saharan Africa (35.1%)( 40 , 41 ). These results exceed the WHO’s recommended threshold of 10% for vaccine incompletion and are also higher than estimates reported in studies from Australia (20%) and Myanmar (25.8%)( 42 , 43 ). Furthermore, the result is lower than the findings reported from Nigeria (69.6%)( 44 ), and Indonesia (58.9%)( 45 ). The possible justification for the variation could be disruption of immunization services caused by the COVID-19 pandemic( 46 ), persistent armed conflict and internal displacement making it difficult for mobile health teams to reach "zero-dose" children and for parents to access fixed clinics( 47 ), maternal socioeconomic factors including low educational attainment and lack of media exposure as mothers with limited health literacy are less likely to navigate complex multi-dose schedules( 48 ) and unreliability of vaccine supply chains and frequent "stock-outs" at the facility level( 49 ). Seventy percent of the total observations were allocated for model training and thirty percent for evaluation to assess the performance of machine learning algorithms and identify the most suitable model for predicting childhood vaccination incompleteness in sub-Saharan Africa. Accordingly, eight algorithms were compared: Random Forest, Decision Tree, Logistic Regression, K-Nearest Neighbor, Artificial Neural Network, Naïve Bayes, Extreme Gradient Boosting (XGBoost), and CatBoost. Model performance was evaluated using accuracy, precision, recall, F1-score, and AUC. Among the candidate models, CatBoost achieved the highest overall performance, recording an AUC of 70.00 and highest accuracy (65.00%) and the most promising classifiers algorithm for predicting vaccination incompleteness in the study population. This finding was agreed with studies done about data classification and terms of association( 50 ), application to predict childhood vaccination among children aged 12–23 months in Ethiopia( 20 ), and prediction of incomplete immunization among under-five children in East Africa( 19 ). Additionally, among the evaluated models, Extreme Gradient Boosting (XGBoost) and Random Forest emerged as top-performing algorithms for predicting childhood vaccination incompleteness, following CatBoost. XGBoost achieved an accuracy of 64.51% and shared the highest AUC (70.00) with CatBoost, while Random Forest attained an accuracy of 63.45% and an AUC of 68.02. These findings align with previous studies that have done on model to predict incomplete vaccination among Ethiopian children( 51 ), predict malnutrition among children( 52 ), predict child mortality( 53 ) and prediction of incomplete immunization among under-five children in East Africa( 19 ). A SHAP bee swarm plot was employed to visualize and summarize the relative contribution of independent features to the likelihood of incomplete vaccination among children aged 12–35 months in sub-Saharan Africa. Antenatal care attendance, maternal media exposure, delivery in a health facility, health insurance coverage, married marital status, birth order between two and four, and a high household wealth index emerged as protective determinants, each shifting predictions toward complete vaccination. In contrast, rural residence was consistently associated with increased risk, as indicated by positive SHAP values. Antenatal care attendance emerged as one of the strongest protective determinants of childhood vaccination completeness among children aged 12–35 months in sub-Saharan Africa. The SHAP bee swarm plot demonstrated that higher values of antenatal care visits consistently shifted predictions toward complete vaccination, indicating that children whose mothers attended antenatal care were less likely to experience incomplete vaccination. This finding is consistent with previous studies conducted in Ethiopia( 20 , 54 , 55 ), East Africa( 19 ) and Zimbabwe( 56 ), which highlighted the role of maternal health service utilization in improving child immunization outcomes. The protective effect may be explained by the counseling services and education on child immunization provided during antenatal care visits( 57 – 59 ), as well as the reinforcement of maternal awareness regarding the importance of postnatal follow-up. Children whose mothers had exposure to health information through media were less likely to experience incomplete vaccination, as indicated by negative SHAP values for higher feature categories. This finding is supported by similar studies conducted in sub-Saharan Africa( 60 ), Ethiopia( 61 , 62 ), Cameroon( 63 ), and Vietnam( 64 ) which emphasize the role of maternal access to health information in improving vaccination coverage. Mothers exposed to media are more likely to receive timely information about the importance of immunization, potential side effects, and appropriate vaccine initiation schedules. In addition, media exposure reinforces awareness of basic childhood vaccination services and appointment schedules, thereby enhancing adherence to immunization programs( 65 , 66 ). Being rural residence was the third most important attribute in predicting childhood incomplete vaccination in SSA. Children of mothers residing in rural areas were consistently at increased risk of incomplete vaccination, as indicated by positive SHAP values for rural residence. Similarly previous study conducted in resource limited countries( 18 ), Nigeria( 67 ), Africa( 40 ), and Ethiopia( 68 ) reported that children living in urban areas are more likely to receive complete vaccination compared to those in rural settings. This disparity may be attributed to rural mothers having limited knowledge of vaccination programs, largely due to reduced proximity to health facilities, restricted access to media, and minimal contact with health professionals( 40 ). Health insurance coverage was the fourth most important attribute in predicting childhood incomplete vaccination. Children whose mothers reported having health insurance were less likely to experience incomplete vaccination, as these conditions shifted predictions toward complete vaccination. The finding is similar with previous studies( 69 – 71 ). This association may be explained by the fact that health insurance coverage enables mothers to engage in ongoing consultations regarding childhood immunization programs. Delivery in a health facility was the fifth most important attribute in predicting childhood incomplete vaccination. Children whose mothers reported giving birth in a health facility were less likely to experience incomplete vaccination. This findings is supported by similar studies done in Ethiopia( 20 , 72 , 73 ) and Nigeria( 74 ). This may be explained by the fact that institutional delivery provides an opportunity for mothers to interact with health professionals, allowing them to gain knowledge about the importance of immunization, potential side effects, and appropriate timing for vaccine initiation( 57 ). In addition, mothers who give birth in health facilities are often informed about essential childhood vaccination services and provided with schedules for current and future appointments( 75 ). Married marital status was the sixth most important attribute in predicting childhood incomplete vaccination. Children of married mothers were less likely to experience incomplete vaccination, as reflected by negative SHAP values for these categories. The findings similar with previous study conducted in resource limited countries( 18 , 76 ). This is due to the fact that unmarried mothers are more likely to miss or never attend ANC follow-up visits compared to married mothers. This difference may be attributed to the fact that married couples often make joint decisions, including those related to their children’s healthcare. Such collaborative decision-making increases the likelihood of reaching a mutual agreement to ensure timely completion of essential vaccinations for their child. Being born with a birth order between two and four was identified as the seventh most important attribute in predicting incomplete childhood vaccination. Children born with a birth order between two and four were less likely to experience incomplete vaccination. The findings similar with previous studies( 18 , 19 , 77 , 78 ). A possible explanation is that firstborn children hold particular significance for new parents, who tend to ensure appropriate vaccinations are administered promptly. However, as the number of children in a household increases, limited resources may contribute to incomplete vaccination and a higher risk of childhood illness. Being born into a household with a high wealth index was identified as the least important attribute in predicting incomplete childhood vaccination in this study. Children born into wealthy households were less likely to experience incomplete vaccination. The findings similar with previous studies conducted in East Africa( 78 ), Ethiopia( 57 ), Senegal( 79 ), and Nigeria( 80 ). This might be might be due to an increase in child care practice, better health-seeking behavior, and health care access among wealthier households( 81 ). Additionally, might be due to mothers of the richest wealth status have a higher likelihood of accessing modern healthcare services for their families, particularly their children, which implies greater freedom( 82 ). Conversely, children of mothers in the lowest wealth quintile are less likely to complete their vaccinations, largely due to barriers in accessing health facilities, such as the cost of transportation( 83 ). Strength and limitation of the study This study drew upon a large, nationally representative DHS dataset encompassing 57,527 children aged 12–35 months across 16 sub-Saharan African countries surveyed between 2021 and 2024, which substantially enhanced statistical power and improved the generalizability of findings across diverse settings. By pooling data across countries, the study provided robust estimates of incomplete vaccination prevalence and revealed marked heterogeneity across subregions and countries. The use of advanced analytical approaches including meta-analysis, sensitivity and subgroup analyses, meta-regression, and utilization of eight machine learning algorithm that learn from data rather than relying on prior assumptions as in classical analysis methods strengthened the reliability of the results and allowed for predictive modeling. Furthermore, the integration of SHAP feature importance analysis offered novel insights into the relative contribution of maternal, child, and household-level factors. Nonetheless, the cross-sectional design of DHS surveys precludes establishing causal relationships between predictors and incomplete vaccination outcomes. Considerable heterogeneity persisted across countries and subregions even after subgroup and sensitivity analyses, suggesting that unmeasured contextual factors such as political instability, vaccine supply chain disruptions, and conflict may have contributed to variability. Some potentially important determinants, including cultural beliefs, and health system quality indicators, were not available in DHS data and thus could not be assessed. Additionally, unlike traditional regression models, machine learning algorithms do not produce coefficients such as odds ratios. As a result, the precise strength and direction of associations between predictors and outcomes cannot be directly determined. Conclusion and implications of the study In conclusion, the pooled prevalence of incomplete vaccination among children aged 12–35 months in 16 sub‑Saharan African countries was 46.21%, with the lowest level observed in Ghana (25.21%) and the highest in the Democratic Republic of Congo (73.21%). Considerable heterogeneity was noted across countries, with 16 of the 13 analyzed reporting incomplete vaccination prevalence above 30%. CatBoost, XGBoost, and Random Forest were identified as the first, second, and third best-performing machine learning algorithms to predict incomplete childhood vaccination in sub-Saharan Africa, with CatBoost achieving the highest accuracy (65%) and AUC (70%). SHAP feature importance analysis revealed that adequate antenatal care visits, maternal media exposure, institutional delivery, being rural residence, health insurance coverage, married marital status, birth order between two and four, and household with high wealth index were the most influential attributes in predicting vaccination outcomes. Among these, antenatal care attendance and maternal media exposure emerged as the strongest protective determinants, while rural residence consistently increased the risk of incomplete vaccination. The findings of this study provide critical evidence to support policymakers and stakeholders in developing targeted childcare intervention mechanisms and early preparedness strategies to improve immunization coverage and reduce vaccine dropouts. The predictive rules generated by machine learning models contribute to knowledge creation and representation, offering practical insights for program design. Specifically, stakeholders are recommended to enhance mothers’ antenatal care visits and institutional deliveries by expanding access to nearby health facilities, strengthening health insurance coverage, and improving maternal awareness through media outreach. Creating income opportunities and empowering mothers with health literacy are also essential interventions to ensure timely vaccination. Moreover, this study serves as a reference for future research, highlighting the need for longitudinal designs to establish causal pathways and guide more effective policy strategies. By integrating maternal health services, socioeconomic support, and health system strengthening, countries in sub-Saharan Africa can make significant progress toward achieving Immunization Agenda 2030 targets and reducing the burden of incomplete childhood vaccination. The implications of this study extend beyond academic knowledge to practical policy and programmatic action. First, machine learning models can be integrated into national immunization monitoring systems to identify high-risk populations and guide resource allocation more efficiently. Second, health ministries and partners should prioritize maternal health service utilization and media-based health communication as rapid, scalable strategies to improve vaccination adherence. Third, addressing rural–urban disparities through infrastructure development and transportation support are vital to reducing inequities in vaccine access. Finally, the study highlights the importance of combining individual-level interventions with structural reforms such as poverty alleviation, resilient vaccine supply chains, and political stability to achieve sustainable progress toward Immunization Agenda 2030 targets. Abbreviations and Acronyms ANC Antenatal Care AUC Area Under the Curve CI Confidence Interval DHS Demographic Health Survey DTP Diphtheria Pertussis Tetanus IRB Institutional Review Board ML Machine Learning NPV Negative Predict Value PPV Positive Predict Value ROC Receiver Operating Characteristic SHAP Shapely Additive Explanation SSA Sub-Saharan Africa WHO World Health Organization Declarations Clinical Trial Number Not applicable Consent for publication Not Applicable Competing interests The authors declare no competing interests. Funding No specific fund was received for this work. Author Contribution A.H. conceptualization, formal analysis, investigation, methodology, writing original draft, review and editing. B.F. methodology, review and editing. A.A.A methodology, software, review and editing. A.K. review and editing. E.W. methodology, review and editing. K.G. methodology, software, review and editing. K.U. methodology, review and editing. A.H.S methodology, software, review and editing. A.A.B methodology, software, review and editing. H.A. methodology, software, review and editing. Acknowledgements The authors thank ICF International for granting access to the SSA DHS data set used in this study. Data Availability The datasets analyzed during the current study are publicly available from the Demographic and Health Survey (DHS) program, managed by ICF International. The data can be accessed upon reasonable request through the DHS program website at https://dhsprogram.comResearchers must register and submit a data access request through the site. For inquiries regarding data access, please contact the DHS Data Archivist at [email protected] . The data used in this study are cited as: ICF. Demographic and Health Surveys (various years). Funded by USAID. Rockville, Maryland. Available at: https://dhsprogram.com References Immunization WHO. https://www.afro.who.int/health-topics/immunization . 2024. Vaccination WHO. https://www.who.int/data/gho/data/themes/topics/sdg-target-3_b-development-assistance-and-vaccine-coverage . 2023. WHO. Immunization coverage. 2023. 2023. Africa SA. S. Routine immunization. https://www.speakupafrica.org/wp-content/uploads/2019/05/Routine-Immunization.pdf . 2019. Childhood MK. immunization. 2002. https://media.path.org/documents/CVP_Occ_Paper5.pdf . 2002. A. C. Millions of African children still lack access to life-saving vaccines. 2025. 2025. Frenkel LD, editor. The global burden of vaccine-preventable infectious diseases in children less than 5 years of age: implications for COVID-19 vaccination. How can we do better? Allergy and asthma proceedings; 2021: OceanSide Publications, Inc. CME, UUf. Level and trend in child mortality. 2024. PAHO. Global childhood immunization levels stalled in 2023 leaving many without life-saving protection. PAHO. 2023. 2023. Simon DJ, Tokpovi VCK, Dianou K, Okonji OC, Kiragu A, Olorunsaiye CZ et al. Regional, subregional and country-level full vaccination coverage in children aged 12–23 months for 34 countries in sub-Saharan Africa: a global analysis using demographic and health survey data. BMJ Global Health. 2025;10(3). Tsegaw TK, Alemaw HB, Wale YB, Nigatu SG, Birhan TY, Taddese AA. Incomplete immunization uptake and associated factors among children aged 12–23 months in sub-Saharan African countries; multilevel analysis evidenced from latest demography and health survey data, 2023. Ital J Pediatr. 2024;50(1):96. Haeuser E, Byrne S, Nguyen J, Raggi C, McLaughlin SA, Bisignano C, et al. Global, regional, and national trends in routine childhood vaccination coverage from 1980 to 2023 with forecasts to 2030: a systematic analysis for the Global Burden of Disease Study 2023. Lancet. 2025;406(10500):235–60. Fenta SM, Biresaw HB, Fentaw KD, Gebremichael SG. Determinants of full childhood immunization among children aged 12–23 months in sub-Saharan Africa: a multilevel analysis using Demographic and Health Survey Data. Trop Med health. 2021;49(1):29. Bobo F, Dawson MWAAA, Hayen A, editors. Monitoring Inequalities in Child Vaccination in Sub-Saharan Africa: Analysis of 25 National Health Surveys. 2023 World Congress on Health Economics. IHEA; 2023. Negussie A, Kassahun W, Assegid S, Hagan AK. Factors associated with incomplete childhood immunization in Arbegona district, southern Ethiopia: a case–control study. BMC Public Health. 2016;16:27. Eze P, Agu UJ, Aniebo CL, Agu SA, Lawani LO, Acharya Y. Factors associated with incomplete immunisation in children aged 12–23 months at subnational level, Nigeria: a cross-sectional study. BMJ open. 2021;11(6):e047445. Tsegaw TK, Alemu EA, Arage FG, Tadese ZB, Taye EA, Abate TG. Geographically weighted regression analysis of incomplete basic childhood vaccination in Sub-Saharan Africa: Evidence from DHS, 2019–2024. PLoS ONE. 2025;20(11):e0336498. Getnet M, Jejaw M, Belachew TB, Addis B, Dellie E, Tafere TZ, et al. Incomplete basic vaccination and associated factors among children aged 12–23 months in resource-limited countries: a spatial and multilevel regression analysis of recent DHS data from 48 countries. Front Public Health. 2025;13:1463303. Tadese ZB, Nigatu AM, Yehuala TZ, Sebastian Y. Prediction of incomplete immunization among under-five children in East Africa from recent demographic and health surveys: a machine learning approach. Sci Rep. 2024;14(1):11529. Demsash AW, Chereka AA, Walle AD, Kassie SY, Bekele F, Bekana T. Machine learning algorithms’ application to predict childhood vaccination among children aged 12–23 months in Ethiopia: Evidence 2016 Ethiopian Demographic and Health Survey dataset. PLoS ONE. 2023;18(10):e0288867. USAID. Demographic and health survey(DHS) methodology. https://dhsprogram.com/methodology/Survey-Types/DHS-Methodology.cfm n.d . Available. Melaku MS, Nigatu AM, Mewosha WZ. Spatial distribution of incomplete immunization among under-five children in Ethiopia: evidence from 2005, 2011, and 2016 Ethiopian demographic and health survey data. BMC Public Health. 2020;20(1):1362. Cooper H, Hedges LV, Valentine JC. The handbook of research synthesis and meta-analysis. Russell Sage Foundation; 2019. Brownlee J. Data preparation for machine learning: data cleaning, feature selection, and data transforms in Python. Machine Learning Mastery; 2020. Kebede SD, Mamo DN, Adem JB, Semagn BE, Walle AD. Machine learning modeling for identifying predictors of unmet need for family planning among married/in-union women in Ethiopia: Evidence from performance monitoring and accountability (PMA) survey 2019 dataset. PLOS Digit Health. 2023;2(10):e0000345. Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al. Scikit-learn: Machine learning in Python. J Mach Learn Res. 2011;12:2825–30. Chen T, Guestrin C, editors. Xgboost: A scalable tree boosting system. Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining; 2016. Lundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, et al. From local explanations to global understanding with explainable AI for trees. Nat Mach Intell. 2020;2(1):56–67. Uddin S, Khan A, Hossain ME, Moni MA. Comparing different supervised machine learning algorithms for disease prediction. BMC Med Inf Decis Mak. 2019;19(1):1–16. Luque A, Carrasco A, Martín A, de Las Heras A. The impact of class imbalance in classification performance metrics based on the binary confusion matrix. Pattern Recogn. 2019;91:216–31. Lundberg S. A unified approach to interpreting model predictions. arXiv preprint arXiv:170507874. 2017. Liu Y, Liu Z, Luo X, Zhao H. Diagnosis of Parkinson's disease based on SHAP value feature selection. Biocybernetics Biomedical Eng. 2022;42(3):856–69. World Health Organization. Immunization coverage; Avaliable: https://www.who.int/news-room/fact-sheets/detail/immunization-coverage . 2025. Sulgodu Ramachandra S, Webster P. USA's exit from the WHO and freeze on USAID funds globally: its perils and possible opportunities. J Public Health (Oxf). 2025;47(Supplement1):i16–20. Immunization Economics. US Government intends to end financial support for Gavi, threatening flagship vaccine programs. 2025. United Nations. Assessing the impact of the loss of the Demographic and Health Surveys on global population and family planning data and estimates and on population research 2025. Riaz A, Husain S, Yousafzai MT, Nisar I, Shaheen F, Mahesar W, et al. Reasons for non-vaccination and incomplete vaccinations among children in Pakistan. Vaccine. 2018;36(35):5288–93. Tsegaw TK, Alemu EA, Arage FG, Tadese ZB, Taye EA, Abate TG. Geographically weighted regression analysis of incomplete basic childhood vaccination in Sub-Saharan Africa: Evidence from DHS, 2019–2024. PLoS ONE. 2025;20(11):e0336498. Francis M. Vaccination Coverage and Factors Associated with Routine Childhood Vaccination Uptake in India: Findings from a National Survey and Household Surveys in Vellore, Southern India. 2022. Atnafu Gebeyehu N, Abebe Gelaw K, Asmare Adella G, Dagnaw Tegegne K, Adie Admass B, Mesele Gesese M. Incomplete immunization and its determinants among children in Africa: Systematic review and meta-analysis. Hum vaccines immunotherapeutics. 2023;19(1):2202125. Bobo FT, Asante A, Woldie M, Dawson A, Hayen A. Child vaccination in sub-Saharan Africa: Increasing coverage addresses inequalities. Vaccine. 2022;40(1):141–50. Adedokun ST, Uthman OA, Adekanmbi VT, Wiysonge CS. Incomplete childhood immunization in Nigeria: a multilevel analysis of individual and contextual factors. BMC Public Health. 2017;17(1):236. Than SLMA, Laosee O. Determinants of incomplete immunization among hill tribe children aged under two years in Myanmar. 2016. Ogundele OA, Ogundele T, Fehintola FO, Fagbemi AT, Beloved OO, Osunmakinwa OO. Determinants of incomplete vaccination among children 12–23 months in Nigeria: an analysis of a national sample. Tzu Chi Med J. 2022;34(4):448–55. Setiawan MS, Wijayanto AW. Determinants of immunization status of children under two years old in Sumatera, Indonesia: A multilevel analysis of the 2020 Indonesia National Socio-Economic Survey. Vaccine. 2022;40(12):1821–8. Causey K, Fullman N, Sorensen RJ, Galles NC, Zheng P, Aravkin A, et al. Estimating global and regional disruptions to routine childhood vaccine coverage during the COVID-19 pandemic in 2020: a modelling study. Lancet. 2021;398(10299):522–34. Zhang T, Qi X, He Q, Hee J, Takesue R, Yan Y, et al. editors. The effects of conflicts and self-reported insecurity on maternal healthcare utilisation and children health outcomes in the democratic republic of congo (Drc). Healthcare: MDPI; 2021. Ogundele OA, Ogunwemimo HS, Fehintola FO, Ogundele T, Olorunsola A, Bello OE, et al. Predictors of incomplete childhood vaccination in four West African countries: a population based cross-sectional study. Sci Rep. 2025;15(1):17119. Nigatu T, Abraham L, Willems H, Tilaye M, Tiruneh F, Gebru F, et al. The status of immunization program and challenges in Ethiopia: A mixed method study. Sage Open Med. 2024;12:20503121241237115. Dhar A, Dash NS, Roy K, editors. An innovative method of feature extraction for text classification using part classifier. International Conference on Information, Communication and Computing Technology; 2018: Springer. Mihret BG, Belay BN, Kelkay JM, Amare AM, Tarekegn AD, Asegu AA et al. Design a model to predict incomplete immunization among Ethiopian children using ensemble machine learning algorithms. Sci Rep. 2025. Talukder A, Ahammed B. Machine learning algorithms for predicting malnutrition among under-five children in Bangladesh. Nutrition. 2020;78:110861. Demsash AW. Using best performance machine learning algorithm to predict child death before celebrating their fifth birthday. Inf Med unlocked. 2023;40:101298. Tesfa GA, Demeke AD, Hailegebreal S, Amede ES, Ngusie HS, Kasie MN, et al. Spatial distribution and associated factors of measles vaccination among children aged 12–23 months in Ethiopia. A spatial and multilevel analysis. Hum Vaccines Immunotherapeutics. 2022;18(1):2035558. Dirirsa K, Makuria M, Mulu E, Deriba BS. Assessment of vaccination timeliness and associated factors among children in Toke Kutaye district, central Ethiopia: A Mixed study. PLoS ONE. 2022;17(1):e0262320. Mukungwa T. Factors Associated with full Immunization Coverage amongst children aged 12–23 months in Zimbabwe. Afr Popul Stud. 2015;29(2). Tamirat KS, Sisay MM. Full immunization coverage and its associated factors among children aged 12–23 months in Ethiopia: further analysis from the 2016 Ethiopia demographic and health survey. BMC Public Health. 2019;19(1):1019. Gualu T, Dilie A. Vaccination coverage and associated factors among children aged 12–23 months in debre markos town, Amhara regional state, Ethiopia. Adv Public Health. 2017;2017(1):5352847. Tefera YA, Wagner AL, Mekonen EB, Carlson BF, Boulton ML. Predictors and barriers to full vaccination among children in Ethiopia. Vaccines. 2018;6(2):22. Jejaw M, Demissie KA, Teshale G, Geberu DM, Tiruneh MG, Hagos A, et al. Estimated effect of media use on mothers’ vaccination of their children in Sub-Saharan Africa: a quasi-experimental propensity score matching analysis using DHS data. BMC Public Health. 2025;25(1):2018. Mamo Habtie G, Melaku MS, Aychiluhm SB. A multilevel analysis of factors associated with unimmunized children in Ethiopia. Front Public Health. 2025;13:1433606. Chikako TU, Seidu A-A, Hagan JE Jr, Aboagye RG, Ahinkorah BO. Bayesian analysis of predictors of incomplete vaccination against polio among children aged 12–23 months in Ethiopia. Int J Environ Res Public Health. 2021;18(22):11820. Russo G, Miglietta A, Pezzotti P, Biguioh RM, Bouting Mayaka G, Sobze MS, et al. Vaccine coverage and determinants of incomplete vaccination in children aged 12–23 months in Dschang, West Region, Cameroon: a cross-sectional survey during a polio outbreak. BMC Public Health. 2015;15(1):630. Tran BX, Boggiano VL, Nguyen LH, Latkin CA, Nguyen HLT, Tran TT et al. Media representation of vaccine side effects and its impact on utilization of vaccination services in Vietnam. Patient Prefer Adherence. 2018:1717–28. Endalamew SG, Gebeye E, Assefa SK, Geremew BM. Immunization Coverage and Its Determinants Among Children Aged 12–23 Months in East Africa: A Bayesian Hierarchical Modeling Approach Based on DHS 2019–2022. Biomed Res Int. 2025;2025(1):6645541. Rodrigues F, Ziade N, Jatuworapruk K, Caballero-Uribe CV, Khursheed T, Gupta L. The impact of social media on vaccination: a narrative review. J Korean Med Sci. 2023;38(40). Obanewa OA, Newell ML. The role of place of residency in childhood immunisation coverage in Nigeria: analysis of data from three DHS rounds 2003–2013. BMC Public Health. 2020;20(1):123. Desalew A, Semahegn A, Birhanu S, Tesfaye G. Incomplete vaccination and its predictors among children in Ethiopia: a systematic review and meta-analysis. Global Pediatr health. 2020;7:2333794X20968681. de Cantuária Tauil M, Sato APS, Waldman EA. Factors associated with incomplete or delayed vaccination across countries: a systematic review. Vaccine. 2016;34(24):2635–43. Sakou I-I, Tsitsika AK, Papaevangelou V, Tzavela EC, Greydanus DE, Tsolia MN. Vaccination coverage among adolescents and risk factors associated with incomplete immunization. Eur J Pediatrics. 2011;170(11):1419–26. Rieck T, Feig M, Eckmanns T, Benzler J, Siedler A, Wichmann O. Vaccination coverage among children in Germany estimated by analysis of health insurance claims data. Hum vaccines immunotherapeutics. 2014;10(2):476–84. Sakelo AN, Assefa N, Oljira L, Assefa ZM. Newborn care practice and associated factors among mothers of one-month‐old infants in Southwest Ethiopia. Int J Pediatr. 2020;2020(1):3897427. Tesfaye TD, Temesgen WA, Kasa AS. Vaccination coverage and associated factors among children aged 12–23 months in Northwest Ethiopia. Hum vaccines immunotherapeutics. 2018;14(10):2348–54. Antai D. Regional inequalities in under-5 mortality in Nigeria: a population-based analysis of individual-and community-level determinants. Popul health metrics. 2011;9(1):6. Mutua MK, Kimani-Murage E, Ettarh RR. Childhood vaccination in informal urban settlements in Nairobi, Kenya: who gets vaccinated? BMC Public Health. 2011;11(1):6. Dana AD, Kebede DL, Betru KT. Prevalence of incomplete vaccination and associated factors among children aged 24–35 months in Dale woreda, Sidama region, Ethiopia. Ethiop J Med Health Sci. 2022;2(2):136–49. Lin A, Pham D, Rosenthal H, Milanaik R. Birth order and up-to-date vaccination status. Pediatrics. 2022;150(4). Tesema GA, Tessema ZT, Tamirat KS, Teshale AB. Complete basic childhood vaccination and associated factors among children aged 12–23 months in East Africa: a multilevel analysis of recent demographic and health surveys. BMC Public Health. 2020;20(1):1837. Mbengue MAS, Sarr M, Faye A, Badiane O, Camara FBN, Mboup S, et al. Determinants of complete immunization among senegalese children aged 12–23 months: evidence from the demographic and health survey. BMC Public Health. 2017;17(1):630. Chidiebere ODI, Uchenna E, Kenechi O. Maternal sociodemographic factors that influence full child immunisation uptake in Nigeria. South Afr J Child Health. 2014;8(4):138–42. Peters DH, Garg A, Bloom G, Walker DG, Brieger WR, Hafizur Rahman M. Poverty and access to health care in developing countries. Ann N Y Acad Sci. 2008;1136(1):161–71. Sabahath F, Lewis P. Study of the immunization status and reasons for incomplete vaccination of children attending an urban hospital. Pediatr Infect Disease. 2022;4(3):92–100. Faisal S, Zahid FM, Kamal S, Shahzad K, Seidu A-A, Ahinkorah BO. Modeling the factors associated with incomplete immunization among children. Math Probl Eng. 2022;2022(1):8460837. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted 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-8945443","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":602718828,"identity":"89de23e2-a0bf-4dcc-9dc0-c39dbbdc6c3f","order_by":0,"name":"Abdulkerim Hassen Moloro","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7klEQVRIiWNgGAWjYBACAwYeBgmGggQGBmYg5wNQhI2dKC0GEC2FM0BamInWAgSfeUAkIS3m7GcP3uYxSJMzb+dO3Gzza5s8HzMD44ePObi1WPbkJVvzGOQYyxzm3Wyc23fbsI2ZgVly5jY8DjuQYybNY1CROIOZd5txbs9tRqAWNiAbj5bzb8Ba6oFatv+27LltT1jLDbAtOQkSzLwbjBl+3E4kQsu7ZMs5BmmGQFs2GPY23E5uY2Zsxu+X87kHb7ypSJaX4D+7weDHn9u289ubD374iEcLKmBsA5MNxKoHgT+kKB4Fo2AUjIKRAgCWM0tEABb/hAAAAABJRU5ErkJggg==","orcid":"","institution":"Samara University","correspondingAuthor":true,"prefix":"","firstName":"Abdulkerim","middleName":"Hassen","lastName":"Moloro","suffix":""},{"id":602718829,"identity":"ed9d7438-cb03-4974-8456-7a94a2960738","order_by":1,"name":"Bizunesh Fantahun Kase","email":"","orcid":"","institution":"Mattu University","correspondingAuthor":false,"prefix":"","firstName":"Bizunesh","middleName":"Fantahun","lastName":"Kase","suffix":""},{"id":602718830,"identity":"8932ac50-6655-4e24-8425-d1df633f5fc3","order_by":2,"name":"Angwach Abrham Asnake","email":"","orcid":"","institution":"Wolaita Sodo University","correspondingAuthor":false,"prefix":"","firstName":"Angwach","middleName":"Abrham","lastName":"Asnake","suffix":""},{"id":602718831,"identity":"25375130-5cd9-40f0-8889-e0237f7d0072","order_by":3,"name":"Alemayehu Kasu Gebrehana","email":"","orcid":"","institution":"Salale University","correspondingAuthor":false,"prefix":"","firstName":"Alemayehu","middleName":"Kasu","lastName":"Gebrehana","suffix":""},{"id":602718832,"identity":"34d92afe-dd14-4c31-8820-0e077653ce3a","order_by":4,"name":"Etsay Woldu Anbesu","email":"","orcid":"","institution":"Samara University","correspondingAuthor":false,"prefix":"","firstName":"Etsay","middleName":"Woldu","lastName":"Anbesu","suffix":""},{"id":602718833,"identity":"035b6364-262d-4d13-9b61-5325105f206a","order_by":5,"name":"Kebede Gemeda Sabo","email":"","orcid":"","institution":"Samara University","correspondingAuthor":false,"prefix":"","firstName":"Kebede","middleName":"Gemeda","lastName":"Sabo","suffix":""},{"id":602718834,"identity":"c19425c6-a3fc-4d57-87ba-2ba19cac910b","order_by":6,"name":"Kusse Urmale Mare","email":"","orcid":"","institution":"Samara University","correspondingAuthor":false,"prefix":"","firstName":"Kusse","middleName":"Urmale","lastName":"Mare","suffix":""},{"id":602718835,"identity":"a320370a-3634-406f-8041-48d53bf26e8f","order_by":7,"name":"Abdu Hailu Shibeshi","email":"","orcid":"","institution":"Samara University","correspondingAuthor":false,"prefix":"","firstName":"Abdu","middleName":"Hailu","lastName":"Shibeshi","suffix":""},{"id":602718836,"identity":"975afc88-9df9-4290-a498-cd8adb585bde","order_by":8,"name":"Andualem Addisu Birlie","email":"","orcid":"","institution":"Samara University","correspondingAuthor":false,"prefix":"","firstName":"Andualem","middleName":"Addisu","lastName":"Birlie","suffix":""},{"id":602718837,"identity":"00433900-be1a-40db-a1cf-b1cf0fc0cf28","order_by":9,"name":"Hiwot Altaye Asebe","email":"","orcid":"","institution":"Samara University","correspondingAuthor":false,"prefix":"","firstName":"Hiwot","middleName":"Altaye","lastName":"Asebe","suffix":""}],"badges":[],"createdAt":"2026-02-23 09:40:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8945443/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8945443/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104322754,"identity":"f1bc35a8-746c-407b-a18e-3a2d91ef54e0","added_by":"auto","created_at":"2026-03-10 13:27:18","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":64599,"visible":true,"origin":"","legend":"\u003cp\u003eData management process of incomplete vaccination among children aged 12-35 months in sub-Saharan Africa from 2021 to 2024.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/df01c0f7243cb63b97d715f7.png"},{"id":104322753,"identity":"d2e1382c-585d-491f-822e-642a7896eeac","added_by":"auto","created_at":"2026-03-10 13:27:18","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":441051,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot of the country-level and pooled prevalence of incomplete vaccination among children aged 12-35 months across 16 SSA countries, 2021–2024\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/aea140b614a965effb23c25a.png"},{"id":104322783,"identity":"8130d339-84ea-41b0-8595-b040731f61f8","added_by":"auto","created_at":"2026-03-10 13:27:21","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":91201,"visible":true,"origin":"","legend":"\u003cp\u003eSensitivity analysis forest plot of the country-level and pooled prevalence of incomplete vaccination among children aged 12-35 months across 16 SSA countries, 2021–2024\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/358467cbc038acea13bf559d.png"},{"id":104322806,"identity":"b4d9133a-c3cb-40ca-ae94-8952534660d2","added_by":"auto","created_at":"2026-03-10 13:27:31","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":663100,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot showing the pooled prevalence of incomplete vaccination among children aged 12-35 months in 16 sub-Saharan Africa, by subregion, based on a random-effects model analysis, 2021-2024.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/aaad8b6cf3d3adb12e19d2b5.png"},{"id":104322791,"identity":"5521d1f3-a160-4df8-911d-4e9ac70ec516","added_by":"auto","created_at":"2026-03-10 13:27:23","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1391582,"visible":true,"origin":"","legend":"\u003cp\u003eForest plot showing the pooled prevalence of incomplete vaccination among children aged 12-35 months in sub-Saharan Africa, by survey year category, based on a random-effects model analysis, 2021-2024.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/c9aced0d53b09f63117c701f.png"},{"id":104322756,"identity":"237d61ee-4027-4ac9-adc1-716539d3e58a","added_by":"auto","created_at":"2026-03-10 13:27:18","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":186106,"visible":true,"origin":"","legend":"\u003cp\u003eROC curve for vaccination incomplete among children aged 12-35 months in 16 sub-Sharan Africa, 2021-2024.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/093a90a46a70b718c2496836.png"},{"id":104322809,"identity":"e8eabcb9-5cf2-413d-834b-d11f0ccc27cc","added_by":"auto","created_at":"2026-03-10 13:27:31","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":46873,"visible":true,"origin":"","legend":"\u003cp\u003eGlobal feature importance for vaccination incomplete among children aged 12-35 months in sub-Saharan Africa, 2021-2024.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/3a960e4c03c769fc49ac1dc2.png"},{"id":104322796,"identity":"882447fc-8471-4fec-a11d-4c2d9ba4811e","added_by":"auto","created_at":"2026-03-10 13:27:29","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":64168,"visible":true,"origin":"","legend":"\u003cp\u003eBees warm plot for vaccination incomplete among children aged 12-35 months in sub-Saharan Africa, 2021-2024.\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/f4e25643a17d5b52f26f6d4b.png"},{"id":108183187,"identity":"15ab6aa7-5407-4769-9555-3a90eef74d34","added_by":"auto","created_at":"2026-04-30 08:59:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3243426,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8945443/v1/8f381593-aa9a-4286-94ed-6864aeeb0159.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Prediction of incomplete vaccination among children aged 12-35 months in sub-Saharan Africa: Application of machine learning algorithm using recent (2021-2024) Demographic and Health Survey Data","fulltext":[{"header":"Introduction","content":"\u003cp\u003eImmunization currently prevents more than 4\u0026nbsp;million deaths each year worldwide. In the African region alone, vaccines save approximately 800,000 lives annually(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). Despite this progress, in 2023 an estimated 21\u0026nbsp;million children globally missed lifesaving diphtheria, pertussis, and tetanus (DPT) vaccines, while 6.5\u0026nbsp;million received only partial vaccination(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). In Africa, one in five children failed to receive essential immunizations such as the third dose of DPT (DTP3)(\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). The basic childhood vaccines include BCG, pentavalent, polio, and measles(\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn Africa, vaccine-preventable diseases (VPDs) such as measles, diphtheria, tetanus, polio, and pertussis conditions nearly eradicated in many high-income countries continue to affect more than 30\u0026nbsp;million children under five each year. These infections cause over 500,000 deaths annually, representing about 58% of all global child deaths from VPDs(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e). Globally, an estimated 700,000 children under five died from VPDs in 2018, with nearly 99% of these deaths occurring in low- and middle-income countries(\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). In 2023, VPDs accounted for almost 30% of all under-five deaths in sub-Saharan Africa, where the mortality rate among children under five was nearly 14 times higher than in Europe and 18 times higher than in Australia and New Zealand(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eDespite significant efforts to expand immunization coverage across Africa over the past decade, routine vaccination programs continue to face major challenges. As long as children remain vulnerable to outbreaks of vaccine-preventable diseases, most African countries are unlikely to achieve their Sustainable Development Goals (SDGs)(\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). The Immunization Agenda 2030 (IA2030) sets ambitious targets of reaching 90% coverage and reducing the number of \u0026ldquo;zero-dose\u0026rdquo; children worldwide to fewer than 6.5\u0026nbsp;million by 2030(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). However, in sub-Saharan Africa, only 54.1% of children receive the full set of basic childhood vaccines, while 36.1% are only partially immunized. Evidence shows that the region continues to struggle in meeting IA2030 targets(\u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e), with substantial variation in coverage across countries(\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFactors contributing to incomplete immunization include young maternal age, limited knowledge about vaccination, and negative perceptions of vaccine side effects. Children of single mothers, those born without the assistance of a skilled birth attendant, and those whose mothers did not receive postnatal care are also more likely to be incompletely immunized. Additional determinants include poor maternal awareness of routine immunization, residence in rural areas, low household income, and living more than 30 minutes from the nearest vaccination facility(\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e).\u003c/p\u003e \u003cp\u003ePrevious research on childhood immunization in sub‑Saharan Africa has provided valuable insights into spatial heterogeneity and country‑specific determinants, but methodological limitations remain. Most studies relied on descriptive statistics or regression‑based approaches, such as geographically weighted regression (GWR), which assume linear relationships and cannot fully capture the complex, nonlinear interactions among diverse predictors(\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). While GWR improved local interpretability, it did not provide predictive accuracy benchmarks or systematic feature ranking across multiple determinants. Furthermore, earlier works often used older DHS datasets collected before 2020(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e), rely on specific country(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e) and focused narrowly on children aged 12\u0026ndash;23 months, leaving broader age ranges and more recent immunization trends underexplored(\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). As a result, prior evidence offered limited predictive power and insufficient transparency in explaining the drivers of incomplete vaccination across sub‑regions.\u003c/p\u003e \u003cp\u003eThis study addresses those gaps by applying ensemble machine learning algorithms including XGBoost, Random Forest, and CatBoost on recent DHS datasets (2021\u0026ndash;2024) from 16 sub‑Saharan African countries, covering children aged 12\u0026ndash;35 months. Unlike regression models, machine learning captures nonlinear interactions and provides robust predictive accuracy, while SHAP analysis enhances transparency by ranking the most influential maternal, household, and contextual factors. This methodological innovation not only strengthens academic evidence but also equips policymakers with actionable insights to identify high‑risk populations, design targeted interventions, and reduce the burden of \u0026ldquo;zero‑dose\u0026rdquo; children. By bridging the limitations of earlier regression‑based studies, this paper offers a recent reproducible, data‑driven framework that supports regional progress toward the Immunization Agenda 2030 goals and improves child health outcomes in sub-Saharan Africa.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Area, Design, and Period\u003c/h2\u003e \u003cp\u003eThis study conducted secondary analysis using data from the Demographic and Health Surveys (DHS) of 23 countries in Sub-Saharan Africa (SSA). The countries involved were Angola, Burkina Faso, DR Congo, C\u0026ocirc;te d'Ivoire, Gabon, Ghana, Kenya, Lesotho, Madagascar, Mali, Mauritania, Mozambique, Nigeria, Senegal, Tanzania, and Zambia. The selection of country was based on the recent survey year, availability of a standardized and unrestricted dataset, and presence of observations for the outcome variable in the datasets.\u003c/p\u003e \u003cp\u003eThe DHS surveys across all countries employed a cross-sectional study design to collect data on basic sociodemographic characteristics and various health indicators, including maternal and health facility related. For the current analysis, we included the countries that have their recent DHS conducted between 2021 and 2024.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePopulation, sampling technique and weight\u003c/h3\u003e\n\u003cp\u003eThe source population for this study comprised all children aged 12 to 35 months old in 16 selected SSA countries. The study population included all children aged 12 to 35 months in the survey. Across all countries, the surveys used a multistage stratified cluster sampling technique to select the study participants. In the first stage, each country was divided into clusters, and clusters were randomly selected based on the probability proportional to their contribution to overall country\u0026rsquo;s population. In the second stage, using the housing census as a sampling frame, a representative number of households was selected from each cluster.\u003c/p\u003e \u003cp\u003eTo account for the complex survey design, non-response, and to ensure comparability across countries while preventing any single country or survey year from disproportionally influencing the pooled estimates, DHS sampling weights were applied as recommended. The weight variable (v005) was rescaled by dividing by 1,000,000 and incorporated into pooled analyses using the \u003cem\u003esvyset\u003c/em\u003e command in Stata, accounting for primary sampling units (v021) and strata (v022). This produced a weighted analytic sample of \u003cb\u003e57, 527\u003c/b\u003e children aged 12\u0026ndash;35 months with complete data on the variables of interest (Table\u0026nbsp;1). Additionally, sensitivity analyses were conducted to assess the influence of individual countries by systematically excluding one country at a time and recalculating the pooled prevalence. This approach adjusts for unequal probabilities of selection and non-response within each survey, ensuring that each country\u0026rsquo;s contribution to the pooled prevalence is proportional to its population size.\u003c/p\u003e \u003cp\u003eTable-1: Survey years and sample sizes of children aged 12\u0026ndash;35 months from 16 Sub-Saharan African (SSA) countries included in the study, from 2021 to 2024\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"4\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSurvey year\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eUnweighted sample size\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eWeighted sample size\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAngola\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2023\u0026ndash;2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2,911\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2,746\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBurkina Faso\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,215\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4,194\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDR Congo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2023\u0026ndash;2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6,020\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5,876\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCote d\u0026rsquo;Ivoire\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,424\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3,246\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGabon\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,048\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e699\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGhana\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2022\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,582\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3,296\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKenya\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6,978\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6,390\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesotho\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2023\u0026ndash;2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e970\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e929\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMadagascar\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,702\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3,719\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMali\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2023\u0026ndash;2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,623\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e5,056\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMauritania\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2021\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e492\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e946\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMozambique\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2022\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,137\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3,198\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNigeria\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2023\u0026ndash;2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6,432\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6,484\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSenegal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,606\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3,498\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTanzania\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,992\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4,008\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eZambia\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2024\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,317\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3,242\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e58,449\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e57,527\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eData source\u003c/h3\u003e\n\u003cp\u003eThe data for this study were obtained from the DHS women\u0026rsquo;s questionnaire, focusing on children aged 12\u0026ndash;35 months, and from the Kids Record dataset (KR file) across 16 countries. All datasets were sourced directly from the official Demographic and Health Surveys (DHS) program website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://dhsprogram.com\u003c/span\u003e\u003cspan address=\"https://dhsprogram.com\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eData extraction and management of missing observations\u003c/h3\u003e\n\u003cp\u003ePrior to data extraction, we identified all sub-Saharan African countries with DHS datasets collected between 2021 and 2024, as our analysis focused exclusively on this timeframe. The DHS surveys employed standardized data collection tools and face-to-face interviews to ensure comparability across countries. For this study, we extracted relevant variables from the Kids Record (KR) files of 16 countries and merged them into a pooled dataset. Country identifiers and survey weights were retained to account for sampling design and cross-country differences. Variables were harmonized and recoded consistently across surveys, following the Guide to DHS Statistics, to ensure compatibility.\u003c/p\u003e \u003cp\u003eFor incomplete vaccination outcomes, children aged 12\u0026ndash;35 months were included. Records were excluded if vaccination information was missing, incomplete, or flagged as inconsistent with DHS protocols. Specifically, children were dropped if: (i) vaccination cards were not seen and caregiver recall was unavailable, (ii) vaccination dates were missing or incomplete, or (iii) implausible vaccination sequences were recorded (e.g., doses out of chronological order). These exclusions were applied uniformly to both numerator and denominator, constituting a complete-case analysis.\u003c/p\u003e \u003cp\u003eIndependent variables were also subjected to complete-case analysis. Observations with missing values for any covariates included in the final machine learning models were excluded. This approach ensured transparency and minimized bias due to incomplete data. While complete-case analysis may reduce sample size, it enhances the validity of cross-country comparisons by maintaining consistency in variable definitions and data quality. All procedures for handling missing data adhered to DHS methodological guidelines. A detailed description of DHS protocols for managing incomplete or implausible observations is available in the Guide to DHS Statistics(\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eVariables and measurements\u003c/h3\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003eDependent Variable\u003c/b\u003e:\u003c/h2\u003e \u003cp\u003eThe dependent variable was defined as incomplete immunization with basic childhood vaccines among children aged 12\u0026ndash;35 months. A fully vaccinated child is one who has received all required doses of the basic childhood vaccines, namely BCG, OPV1, OPV2, OPV3, Penta1, Penta2, Penta3, and MCV1. Missing at least one of these vaccines was considered incomplete immunization. The eight recoded variables were then summed, and the outcome was categorized as follows: \u0026ldquo;0\u0026rdquo; for children who received all of the recommended vaccines, and \u0026ldquo;1\u0026rdquo; for those who missed at least one of the basic vaccines(\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eIndependent Variables\u003c/strong\u003e \u003cp\u003eA total of 18 variables were incorporated into this study, selected through a systematic process that considered both their availability in the Demographic and Health Survey (DHS) and their documented associations with childhood stunting in prior research(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan additionalcitationids=\"CR18\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e). Each variable was chosen for its relevance to the study objectives and its role as a known or hypothesized determinants of incomplete vaccination reflecting influences at individual, household, and contextual level.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eThese variables were categorized into individual-level factors, household factors, and contextual factors. The individual-level factors included the sex of the child (male and female), age of the child in month, birth order (1, 2\u0026ndash;4, and 5+), perceived size at birth (large, average, and small). Other individual-level factors were maternal educational level (no formal education, primary, secondary, and higher), current maternal working status (yes and no), antenatal care visits during pregnancy (0, 1\u0026ndash;3, and 4 or more) and postnatal checks within 2 months (yes and no). Variables such as maternal age (\u003cspan additionalcitationids=\"CR16 CR17 CR18 CR19 CR20 CR21 CR22 CR23\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e and \u003cspan additionalcitationids=\"CR26 CR27 CR28 CR29 CR30 CR31 CR32 CR33 CR34 CR35 CR36 CR37 CR38 CR39 CR40 CR41 CR42 CR43 CR44 CR45 CR46 CR47 CR48\" citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e), place of delivery (home, health facility, other), and marital status (single and married) were also considered.\u003c/p\u003e \u003cp\u003eHousehold factors included household size (small, medium, and large), exposure to media (yes and no), health insurance coverage, distance to health facility and wealth index (poor, middle, rich). Contextual factors encompassed the place of residence (urban and rural) and geographical sub-regions (West, East, Central, and Southern).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eData management and statistical analysis\u003c/h3\u003e\n\u003cp\u003eStata version 17 was utilized for data cleaning, labelling and pooled analysis. Prior to the analysis, the presence of the outcome variable in the DHS dataset for each country was confirmed. All the variables considered in the study were reviewed for missing values. Subsequently, the datasets from 16 SSA countries were appended and weighted to maintain sample representativeness and obtain reliable estimates and standard errors. The pooled prevalence of child incomplete vaccination was calculated using weighted data on the number of affected children with outcome variable and the total number of study participants in each country included in the analysis. The Stata command for meta-analysis \u0026ldquo;metan\u0026rdquo; was executed to present the country-specific and pooled estimates with 95% CI in a forest plot. To handle heterogeneity, meta-regressions, sensitivity analysis and sub-group analysis were conducted(\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBefore developing a classification model, preprocess the raw data for analysis was crucial to increase model accuracy and performance(\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). Yufeng Guo\u0026rsquo;s steps of machine learning were used to classify knowledge of fertility period, such as data collection, data preparation, model selection, model training, model evaluation, and prediction(\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). In addition, interpretation of the model was performed using SHAP. Data analysis was conducted in Google Coolab using the Python 3.10.2 programming language within a Jupiter Notebook environment. The packages imblearn, scikit-learn (sklearn)(\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e), XGBoost(\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e), and SHAP(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e) were utilized to perform the necessary calculations and analyses (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eModel selection and development\u003c/h3\u003e\n\u003cp\u003eSupervised classifier machine learning algorithms were used to evaluate the predictive power of machine learning techniques in classifying incomplete vaccination among children aged 12\u0026ndash;35 months. Na\u0026iuml;ve Bayes, Decision Trees, K-Nearest Neighbor (KNN), Logistic Regression, Artificial Neural Networks (ANN), Extreme Gradient Boosting (XGBoost), Cat Boosting (CatBoost) and Random Forest were used to select the best model. Those algorithms were chosen based on their interpretability, amount of features, computational efficiency, accuracy, and characteristics of the dataset(\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eData pre-processing\u003c/h2\u003e \u003cp\u003eData pre-processing was carried out to address missing and incomplete records, as well as duplicate entries. The dataset contained noise, outliers, and inconsistencies, which were systematically managed to ensure data quality. At this stage, unnecessary values and duplicate variables were removed. In addition, all string and categorical variables were transformed into nominal data types to facilitate processing.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eModel training\u003c/h2\u003e \u003cp\u003eThe selected supervised machine learning classifiers were applied to analyze the dataset, focusing on binary classification of incomplete vaccination. For the final classification on unseen test data, the best-performing predictive model was identified and trained to ensure optimal performance.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eModel evaluation\u003c/h2\u003e \u003cp\u003eModel performance was evaluated using the receiver operating characteristic (ROC) curve, which measures the ability to distinguish between classes, assists in selecting optimal thresholds, and enables comparison across models. Key metrics included the area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Among these, AUC was considered the most reliable indicator of discriminatory power, effectively demonstrating the model\u0026rsquo;s ability to differentiate between children with incomplete vaccination and complete vaccination. The higher AUC obtained from the training dataset highlighted superior discriminatory performance compared to accuracy, sensitivity, specificity, PPV, and NPV(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eMaking classification and model interpretation\u003c/h2\u003e \u003cp\u003eClassification was performed using selected predictor variables. The best-performing classifier was chosen based on its accuracy. To interpret the model, Shapley Additive Explanations (SHAP) were employed. Because many powerful machine learning models particularly tree-based algorithms are often regarded as \u0026ldquo;black boxes,\u0026rdquo; their internal decision-making processes are rarely transparent in research(\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). SHAP, a game-theory\u0026ndash;based approach, provides a robust framework for explaining model outputs both locally (at the individual level) and globally (across the entire dataset). This technique reduces the limitations of interpreting complex machine learning findings and enhances the transparency of the classification results(\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eEthical considerations\u003c/h2\u003e \u003cp\u003eFor this study, we utilized Demographic and Health Survey (DHS) data from 16 sub-Saharan African countries. The DHS survey procedures were approved by the ICF Institutional Review Board (IRB) and the respective host country IRB; therefore, no additional ethical approval was required for this secondary analysis. The dataset accessed contain no identifiable participant information, ensuring confidentiality and privacy. Access to the data was formally authorized by the DHS program, which serve as the institutional custodian of the datasets, through an online request submitted at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://dhsprogram.com\u003c/span\u003e\u003cspan address=\"https://dhsprogram.com\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e and supported by authorization letter AuthLetter_215093 on dated 01/19/2025.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eChild and mothers\u0026rsquo; characteristics\u003c/h2\u003e \u003cp\u003eThe study included and analyzed a weighted sample of 57,527 children aged 12 to 35 months with mothers. Of these children, 50.68% (29,155) were male, and 55.99% (32,211) were between 12\u0026ndash;24 months old. Vaccination incomplete was highly prevalent, affecting 54.48% (14,270) of children between 12\u0026ndash;24 months, compared to older age groups. Among the included children, 60.83% (34,993) lived in rural areas, and 40.1% (27,071) were from households in the poor wealth index category. Additionally, 46.45% (26,720) of the children were from West African countries, and 21.31% (12,257) were delivered at home (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eOf the 57,527 children\u0026rsquo;s mothers included in the study, 17,952 (31.21%) were between 15 and 24 years old, and 14,832(25.78%) were currently single. About one-third of the mothers, 16,850(29.31%) had no formal education. Among included children\u0026rsquo;s mothers, 21,543(37.45%), of the mothers were not working during the time of the interview and only 18,052 (31.38%) had no access to mass media.\u003c/p\u003e \u003cp\u003eRegarding household size, the majority of children 51,184(88.97%) lived in large households. More than half of the children 29,688(51.61%) were reported to have an average birth size, and 7,667 (13.33%) children\u0026rsquo;s mothers did not receive a post-natal checkup within two months after birth. Among mothers, 12,934(22.48%) attended only one to three antenatal care visits, while 2,614(4.54%) had no antenatal care. In terms of birth order, most children 28,528 (49.59%) were born as second to fourth order births. Furthermore, 20,970(36.45%) respondents reported that distance to a health facility was a major problem when seeking medical help, and 51,900(90.22%) households had no health insurance coverage.\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 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eChild and mothers characteristic by vaccine incomplete and their overall background status, in 16 SSA countries, 2021\u0026ndash;2024 (n\u0026thinsp;=\u0026thinsp;\u003cb\u003e57,527\u003c/b\u003e)\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=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eWeighted Frequency (%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eVaccine incomplete\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSex of the child\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e29,155(50.68)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13,174(50.30)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFemale\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e28,372(49.32)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13,017(49.70)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge of the child (months)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e12\u0026ndash;24 months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e32,211(55.99)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e14,270(54.48)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e25\u0026ndash;35 months\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e25,316(44.01)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e11,921(45.52)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBirth order\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFirst\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14,166(24.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6,079(23.21)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2 to 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e28,528(49.52)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e12,800(48.87)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5 or higher\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14,833(25.79)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7,312(27.49)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived size at birth\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLarge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e17,291(30.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8,129(31.04)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAverage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e29,688(51.61)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e13,539(51.69)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e10,547(18.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10,547(18.33)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMothers\u0026rsquo; educational status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo education\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e16,860(29.31)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8,247(31.49)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePrimary\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e16,594(28.85)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7,753(29.60)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSecondary\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e19,959(34.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8,767(33.47)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHigher\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,114(7.15)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,424(5.44)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMother currently working\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e35,984(62.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16,128(61.58)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e21,543(37.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10,063(38.42)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePostnatal care visits\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e49,860(86.67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e22,384(85.46)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e7,667(13.33)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3,807(14.54)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMother\u0026rsquo;s age (years)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e15\u0026ndash;24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e17,952(31.21)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8,494(32.43)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e25\u0026ndash;49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e39,575(68.79)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e17,697(67.57)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNumber of ANC visits\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo visit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2,614(4.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,646(6.29)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1 to 3 visits\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12,934(22.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6,660(25.43)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4 and above visit\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e41,979(72.97)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e17,885(68.29)\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\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHome\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e12,257(21.31)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e6,450(24.63)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHealth facility\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e45,270(78.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e19,741(75.37)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarital status\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSingle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14,832(25.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7,536(28.77)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMarried\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e42,695(74.22)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e18,655(71.23)\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\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3,572(6.21)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,687(6.44)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2,770(4.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,388(5.30)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLarge\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e51,184(88.97)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e23,117(88.26)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMass media exposure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e18,052(31.38)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9,910(37.84)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e39,475(68.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16,281(62.16)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHealth insurance coverage\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e51,900(90.22)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e24,400(93.16)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eYes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e5,627(9.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1,791(6.84)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDistance to health facility\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBig problem\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e20,969(36.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9,858(37.64)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNot a big problem\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e36,557(63.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16,333(62.36)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWealth index\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoor\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e23,071(40.10)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e11,156(42.59)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMiddle\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e11,285(19.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,138(19.62)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRich\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e23,170(40.28)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9,897(37.79)\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\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrban\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e25,534(39.17)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10,367(39.58)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRural\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e34,993(60.83)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e15,824(60.42)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSubregion\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eWest African Countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e26,719(46.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e11,217(42.83)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEast African Countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e14,116(24.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,316(20.30)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCentral African Countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e6,575(11.43)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4,655(17.77)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSouthern African Countries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e10,117(17.59)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5,003(19.10)\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=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003ePooled prevalence of incomplete vaccination among children aged 12\u0026ndash;35 months in sub\u0026ndash;Saharan African\u003c/h2\u003e \u003cp\u003eThe pooled prevalence of incomplete vaccination among children aged 12\u0026ndash;35 months 16 sub-Saharan African (SSA) countries was 46.21% (95% CI: 38.58, 53.83%), with significant variation observed across countries (I\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;99.8, P-value\u0026thinsp;=\u0026thinsp;0.000) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Democratic Republic of Congo reported the highest incomplete vaccination at 73.21% while Ghana had the lowest incomplete vaccination prevalence at 25.21%. Among the 16 countries analyzed, 13 had incomplete vaccination prevalence of 30% or higher among children aged 12\u0026ndash;35 months.\u003c/p\u003e \u003cp\u003e \u003cb\u003eHandling heterogeneity\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe random-effects model revealed considerable heterogeneity. To address this, sensitivity analysis, subgroup analysis, and meta-regression were conducted.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eSensitivity analysis\u003c/h2\u003e \u003cp\u003eSensitivity analysis was performed to evaluate the effect of individual country on the pooled estimated. When individual country was omitted, the pooled prevalence obtained was within the 95% CI of the overall pooled prevalence. This confirms the absence of single study impact on the overall pooled effect size. Therefore, from the random effects model, there were no country that excessively influence the overall pooled estimate of incomplete vaccination (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eSub-group analysis\u003c/h2\u003e \u003cp\u003eSubgroup analyses were carried out by sub-region, and year of publication.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eStunting by Sub-Region\u003c/h2\u003e \u003cp\u003eSubgroup analysis by sub-region indicated that Central African countries had the highest pooled prevalence of incomplete vaccination among children aged 12\u0026ndash;35 months (64.61%, 95% CI: 47.70\u0026ndash;81.52), followed by Southern African countries (48.12%, 95% CI: 32.79\u0026ndash;63.45), West African countries (43.12%, 95% CI: 32.38\u0026ndash;53.86) and East African countries (38.59%, 95% CI: 32.81\u0026ndash;44.36). Substantial heterogeneity was observed in West and Central Africa (I\u0026sup2; = 99.7%, P\u0026thinsp;=\u0026thinsp;0.00), Central Africa (I\u0026sup2; = 99.5%, P\u0026thinsp;=\u0026thinsp;0.00), East African (I\u0026sup2; = 98%, P\u0026thinsp;=\u0026thinsp;0.00), and Southern Africa (I\u0026sup2; = 99.6%, P\u0026thinsp;=\u0026thinsp;0.00) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eStunting by Year of Survey\u003c/h2\u003e \u003cp\u003eSubgroup analysis by year of survey showed that the pooled prevalence of incomplete vaccination was highest in 2021\u0026ndash;2024 (46.21%, 95% CI: 38.58\u0026ndash;53.83) and lowest in 2018\u0026ndash;2020 (38.77%, 95% CI: 20.06\u0026ndash;57.48). Both periods demonstrated significant heterogeneity (2021\u0026ndash;2024: I\u0026sup2; = 99.8%, p\u0026thinsp;\u0026lt;\u0026thinsp;0.00; 2018\u0026ndash;2020: I\u0026sup2; = 99.9%, p\u0026thinsp;\u0026lt;\u0026thinsp;0.00) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eMeta-regression\u003c/h2\u003e \u003cp\u003eA meta-regression was performed to assess whether the year of survey, sample size, country, and sub-region could explain the heterogeneity in incomplete vaccination prevalence (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e3\u003c/span\u003e). None of these variables were statistically significant predictors of heterogeneity (all p-values\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The high residual heterogeneity in the pooled prevalence of incomplete vaccination suggests that other unmeasured factors are responsible for the variability across studies.\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 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMeta-regression for the studies of pooled incomplete vaccination among children aged 12\u0026ndash;35 months in 16 sub-Sharan Africa, 2021\u0026ndash;2024.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOutcome categories\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCoefficients\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eP-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e[95% conf. interval]\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e \u003cp\u003e\u003cem\u003ePooled incomplete vaccination among 12\u0026ndash;35 months children\u003c/em\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eYear of Survey\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.70\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.88, 1.19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSample size\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.99, 1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCountries\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.96, 1.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSub-region\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.32\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.76, 1.01\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 \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eModel performance to predict childhood incomplete vaccination in sub-Saharan Africa\u003c/h2\u003e \u003cp\u003eTen machine learning algorithms were applied to predict childhood incomplete vaccination in sub-Saharan Africa. The models evaluated included; Na\u0026iuml;ve Bayes, Decision Trees, K-Nearest Neighbor (KNN), Logistic Regression, Artificial Neural Networks (ANN), Extreme Gradient Boosting (XGBoost), Cat Boosting (CatBoost) and Random Forest. Their performance was assessed using Receiver Operating Characteristic (ROC) curves, Area Under the Curve (AUC), Accuracy, Precision, Recall (Sensitivity), F1-score, and Ten-Fold Cross-Validation (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAmong the evaluated models CatBoost and XGBoost demonstrated the strongest predictive performance for incomplete vaccination in sub-Saharan Africa. CatBoost achieved the highest overall accuracy (65%) and AUC (70%), supported by balanced precision (66%) and recall (61.45%), yielding an F1-score of 63.36%. XGBoost closely followed, with an accuracy of 64.52% and an AUC of 70%, alongside a precision of 65.36% and recall of 61.78%. Overall, the comparative analysis underscores that CatBoost are the most effective algorithms for predicting childhood vaccination outcomes in sub-Saharan Africa, offering reliable and well-balanced classification performance (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModel performance comparison of the included machine learning algorithm for vaccination incomplete among children aged 12\u0026ndash;35 months in 16 sub-Sharan Africa, 2021\u0026ndash;2024.\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModels\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAccuracy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePrecision\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eRecall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eF1-Score\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRandom Forest\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e63.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e68.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e63.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e63.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e63.39\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDecision Tree\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e60.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e60.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e60.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e57.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e58.71\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eK-Nearest Neighbor (KNN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e65.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e61.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e61.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e61.56\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eExtreme Gradient Boosting (XGBoost)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e70.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e65.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e61.78\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e63.52\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLogistic Regression\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e63.43\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e68.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e64.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e58.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e61.52\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNa\u0026iuml;ve Bayes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e62.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e67.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e64.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e55.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e59.79\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eArtificial Neural Networks (ANN)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e62.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e67.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e62.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e63.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e62.91\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCat Boosting (CatBoost)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003e65\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003e70.00\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003e65.95\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003e61.45\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003e63.62\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003cb\u003eAUC\u003c/b\u003e\u0026thinsp;\u003cem\u003e=\u0026thinsp;Area Under the Curve\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003eGlobal feature selection with SHAP\u003c/h2\u003e \u003cp\u003eSHAP global feature importance was used to identify the top independent variables that determine incomplete vaccination. The CatBoost classifier algorithm was employed to optimize and select important features using SHAP. To determine the top predictors of incomplete vaccination, the mean absolute SHAP value (MASHAPV) was calculated for each independent variable. The top eight independent features were then selected based on their MASHAPV in descending order.\u003c/p\u003e \u003cp\u003eThose features were maternal media exposure (media_exposure_1), antenatal care attendance (antenatal_visits_2), being rural residence (Residence_2), being married (marital_status_2), health insurance coverage (Health_insurance_coverage_1), delivery in a health facility (place_of_delivery_2), birth order 2 to 4 (birth_order_category_3), and rich household wealth index (wealth_index_3). These features were determined to be the most important predictors of incomplete vaccination among the study population (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e \u003ch2\u003eModel interpretation with bees warm SHAP plot\u003c/h2\u003e \u003cp\u003eThe SHAP bees warm plot was used to visualize and provide a dense summary of how independent features influence the likelihood of incomplete vaccination among children aged 12\u0026ndash;35 months in sub‑Saharan Africa (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e). Each dot represents an individual child\u0026rsquo;s data point, with color indicating the original feature value: blue for low values (coded \u0026ldquo;0\u0026rdquo;) and red for high values (coded \u0026ldquo;1\u0026rdquo;). Points located on the left side of the vertical axis (negative SHAP values) indicate that the feature reduces the likelihood of incomplete vaccination, whereas points on the right side (positive SHAP values) indicate that the feature increases the likelihood of incomplete vaccination.\u003c/p\u003e \u003cp\u003eThe plot highlights the relative contribution of key sociodemographic and health service factors. Antenatal care attendance (antenatal_visits_2) and maternal media exposure (media_exposure_1) emerged as the strongest protective determinants, with higher values shifting predictions toward complete vaccination. Children whose mothers reported antenatal care visits and exposure to health information through media were less likely to experience incomplete vaccination, as indicated by negative SHAP values for higher feature categories.\u003c/p\u003e \u003cp\u003eRural residence (Residence_1) was consistently associated with increased risk of incomplete vaccination. Children of mothers residing in rural areas were consistently at increased risk of incomplete vaccination, as indicated by positive SHAP values for rural residence. Health insurance coverage (Health_insurance_coverage_1) and delivery in a health facility (place_of_delivery_2) were protective, shifting predictions toward complete vaccination. Children whose mothers reported having health insurance and giving birth in a health facility were less likely to experience incomplete vaccination, as these conditions shifted predictions toward complete vaccination.\u003c/p\u003e \u003cp\u003eMarried marital status (marital_status_2), birth order 2 to 4 (birth_order_category_3), and rich household wealth index (wealth_index_3) also contributed to reduced likelihood of incomplete vaccination. Specifically, children of married mothers, those born with birth order between two and four, and those living in rich households were less likely to experience incomplete vaccination, as reflected by negative SHAP values for these categories. Overall, the plot demonstrates that access to maternal health services, being married, low birth order, socioeconomic advantages, and exposure to health information are critical in lowering the probability of incomplete vaccination, whereas rural residence increase vulnerability to incomplete vaccination.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eBy 2024, the number of children missing all routine vaccinations (zero-dose children) had declined to approximately 14.3\u0026nbsp;million, down from the 25\u0026nbsp;million reported in 2021. However, this figure remains nearly 1.4\u0026nbsp;million higher than pre-pandemic levels, signaling stalled progress toward achieving the Immunization Agenda 2030 (IA2030) targets(\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). In March 2025, the U.S. government ended a \u003cspan\u003e$\u003c/span\u003e2.63\u0026nbsp;billion grant to Gavi, the Vaccine Alliance, citing concerns about the organization\u0026rsquo;s focus on \u0026ldquo;zero-dose\u0026rdquo; children(\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). By ending such a large grant, millions of children specifically 75\u0026nbsp;million may no longer have access to essential vaccines(\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e). Crucially, the sudden termination of the Demographic and Health Surveys (DHS) Program in February 2025 has left a critical data vacuum, as DHS previously provided over 50% of child mortality data in sub-Saharan Africa(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAccording to this study, the pooled prevalence of incomplete vaccination among children aged 12\u0026ndash;35 months was 46.21% (95% CI: 38.58\u0026ndash;53.83%), with the Democratic Republic of Congo reporting the highest rate at 73.21% and Ghana the lowest at 25.21%, indicating that nearly one in two children in the region did not receive the full set of WHO-recommended vaccines. The finding is consistent with prevalence reported from Pakistan (46%)(\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e). However, the finding is higher than previous study conducted in sub-Saharan Africa, which indicated that the pooled prevalence of incomplete vaccine among children aged 12\u0026ndash;23 months in 16 countries was 35.5%, with Rwanda recording the lowest rate at 4.21% and Mauritania the highest at 56.3%(\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe findings also higher than study conducted in India (32%)(\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e), in Africa (35.5%) and Sub-Saharan Africa (35.1%)(\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). These results exceed the WHO\u0026rsquo;s recommended threshold of 10% for vaccine incompletion and are also higher than estimates reported in studies from Australia (20%) and Myanmar (25.8%)(\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). Furthermore, the result is lower than the findings reported from Nigeria (69.6%)(\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e), and Indonesia (58.9%)(\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe possible justification for the variation could be disruption of immunization services caused by the COVID-19 pandemic(\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e), persistent armed conflict and internal displacement making it difficult for mobile health teams to reach \"zero-dose\" children and for parents to access fixed clinics(\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e), maternal socioeconomic factors including low educational attainment and lack of media exposure as mothers with limited health literacy are less likely to navigate complex multi-dose schedules(\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e) and unreliability of vaccine supply chains and frequent \"stock-outs\" at the facility level(\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSeventy percent of the total observations were allocated for model training and thirty percent for evaluation to assess the performance of machine learning algorithms and identify the most suitable model for predicting childhood vaccination incompleteness in sub-Saharan Africa. Accordingly, eight algorithms were compared: Random Forest, Decision Tree, Logistic Regression, K-Nearest Neighbor, Artificial Neural Network, Na\u0026iuml;ve Bayes, Extreme Gradient Boosting (XGBoost), and CatBoost. Model performance was evaluated using accuracy, precision, recall, F1-score, and AUC. Among the candidate models, CatBoost achieved the highest overall performance, recording an AUC of 70.00 and highest accuracy (65.00%) and the most promising classifiers algorithm for predicting vaccination incompleteness in the study population. This finding was agreed with studies done about data classification and terms of association(\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e), application to predict childhood vaccination among children aged 12\u0026ndash;23 months in Ethiopia(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e), and prediction of incomplete immunization among under-five children in East Africa(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAdditionally, among the evaluated models, Extreme Gradient Boosting (XGBoost) and Random Forest emerged as top-performing algorithms for predicting childhood vaccination incompleteness, following CatBoost. XGBoost achieved an accuracy of 64.51% and shared the highest AUC (70.00) with CatBoost, while Random Forest attained an accuracy of 63.45% and an AUC of 68.02. These findings align with previous studies that have done on model to predict incomplete vaccination among Ethiopian children(\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e), predict malnutrition among children(\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e), predict child mortality(\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e) and prediction of incomplete immunization among under-five children in East Africa(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eA SHAP bee swarm plot was employed to visualize and summarize the relative contribution of independent features to the likelihood of incomplete vaccination among children aged 12\u0026ndash;35 months in sub-Saharan Africa. Antenatal care attendance, maternal media exposure, delivery in a health facility, health insurance coverage, married marital status, birth order between two and four, and a high household wealth index emerged as protective determinants, each shifting predictions toward complete vaccination. In contrast, rural residence was consistently associated with increased risk, as indicated by positive SHAP values.\u003c/p\u003e \u003cp\u003eAntenatal care attendance emerged as one of the strongest protective determinants of childhood vaccination completeness among children aged 12\u0026ndash;35 months in sub-Saharan Africa. The SHAP bee swarm plot demonstrated that higher values of antenatal care visits consistently shifted predictions toward complete vaccination, indicating that children whose mothers attended antenatal care were less likely to experience incomplete vaccination. This finding is consistent with previous studies conducted in Ethiopia(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e), East Africa(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e) and Zimbabwe(\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e), which highlighted the role of maternal health service utilization in improving child immunization outcomes. The protective effect may be explained by the counseling services and education on child immunization provided during antenatal care visits(\u003cspan additionalcitationids=\"CR58\" citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e), as well as the reinforcement of maternal awareness regarding the importance of postnatal follow-up.\u003c/p\u003e \u003cp\u003eChildren whose mothers had exposure to health information through media were less likely to experience incomplete vaccination, as indicated by negative SHAP values for higher feature categories. This finding is supported by similar studies conducted in sub-Saharan Africa(\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e), Ethiopia(\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e, \u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e), Cameroon(\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e), and Vietnam(\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e) which emphasize the role of maternal access to health information in improving vaccination coverage. Mothers exposed to media are more likely to receive timely information about the importance of immunization, potential side effects, and appropriate vaccine initiation schedules. In addition, media exposure reinforces awareness of basic childhood vaccination services and appointment schedules, thereby enhancing adherence to immunization programs(\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e, \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eBeing rural residence was the third most important attribute in predicting childhood incomplete vaccination in SSA. Children of mothers residing in rural areas were consistently at increased risk of incomplete vaccination, as indicated by positive SHAP values for rural residence. Similarly previous study conducted in resource limited countries(\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e), Nigeria(\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e), Africa(\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e), and Ethiopia(\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e) reported that children living in urban areas are more likely to receive complete vaccination compared to those in rural settings. This disparity may be attributed to rural mothers having limited knowledge of vaccination programs, largely due to reduced proximity to health facilities, restricted access to media, and minimal contact with health professionals(\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHealth insurance coverage was the fourth most important attribute in predicting childhood incomplete vaccination. Children whose mothers reported having health insurance were less likely to experience incomplete vaccination, as these conditions shifted predictions toward complete vaccination. The finding is similar with previous studies(\u003cspan additionalcitationids=\"CR70\" citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e). This association may be explained by the fact that health insurance coverage enables mothers to engage in ongoing consultations regarding childhood immunization programs.\u003c/p\u003e \u003cp\u003eDelivery in a health facility was the fifth most important attribute in predicting childhood incomplete vaccination. Children whose mothers reported giving birth in a health facility were less likely to experience incomplete vaccination. This findings is supported by similar studies done in Ethiopia(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e, \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e) and Nigeria(\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e). This may be explained by the fact that institutional delivery provides an opportunity for mothers to interact with health professionals, allowing them to gain knowledge about the importance of immunization, potential side effects, and appropriate timing for vaccine initiation(\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). In addition, mothers who give birth in health facilities are often informed about essential childhood vaccination services and provided with schedules for current and future appointments(\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMarried marital status was the sixth most important attribute in predicting childhood incomplete vaccination. Children of married mothers were less likely to experience incomplete vaccination, as reflected by negative SHAP values for these categories. The findings similar with previous study conducted in resource limited countries(\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e). This is due to the fact that unmarried mothers are more likely to miss or never attend ANC follow-up visits compared to married mothers. This difference may be attributed to the fact that married couples often make joint decisions, including those related to their children\u0026rsquo;s healthcare. Such collaborative decision-making increases the likelihood of reaching a mutual agreement to ensure timely completion of essential vaccinations for their child.\u003c/p\u003e \u003cp\u003eBeing born with a birth order between two and four was identified as the seventh most important attribute in predicting incomplete childhood vaccination. Children born with a birth order between two and four were less likely to experience incomplete vaccination. The findings similar with previous studies(\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e, \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e). A possible explanation is that firstborn children hold particular significance for new parents, who tend to ensure appropriate vaccinations are administered promptly. However, as the number of children in a household increases, limited resources may contribute to incomplete vaccination and a higher risk of childhood illness.\u003c/p\u003e \u003cp\u003eBeing born into a household with a high wealth index was identified as the least important attribute in predicting incomplete childhood vaccination in this study. Children born into wealthy households were less likely to experience incomplete vaccination. The findings similar with previous studies conducted in East Africa(\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e), Ethiopia(\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e), Senegal(\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e), and Nigeria(\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e). This might be might be due to an increase in child care practice, better health-seeking behavior, and health care access among wealthier households(\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e). Additionally, might be due to mothers of the richest wealth status have a higher likelihood of accessing modern healthcare services for their families, particularly their children, which implies greater freedom(\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e). Conversely, children of mothers in the lowest wealth quintile are less likely to complete their vaccinations, largely due to barriers in accessing health facilities, such as the cost of transportation(\u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003eStrength and limitation of the study\u003c/h2\u003e \u003cp\u003eThis study drew upon a large, nationally representative DHS dataset encompassing 57,527 children aged 12\u0026ndash;35 months across 16 sub-Saharan African countries surveyed between 2021 and 2024, which substantially enhanced statistical power and improved the generalizability of findings across diverse settings. By pooling data across countries, the study provided robust estimates of incomplete vaccination prevalence and revealed marked heterogeneity across subregions and countries.\u003c/p\u003e \u003cp\u003eThe use of advanced analytical approaches including meta-analysis, sensitivity and subgroup analyses, meta-regression, and utilization of eight machine learning algorithm that learn from data rather than relying on prior assumptions as in classical analysis methods strengthened the reliability of the results and allowed for predictive modeling. Furthermore, the integration of SHAP feature importance analysis offered novel insights into the relative contribution of maternal, child, and household-level factors.\u003c/p\u003e \u003cp\u003eNonetheless, the cross-sectional design of DHS surveys precludes establishing causal relationships between predictors and incomplete vaccination outcomes. Considerable heterogeneity persisted across countries and subregions even after subgroup and sensitivity analyses, suggesting that unmeasured contextual factors such as political instability, vaccine supply chain disruptions, and conflict may have contributed to variability. Some potentially important determinants, including cultural beliefs, and health system quality indicators, were not available in DHS data and thus could not be assessed. Additionally, unlike traditional regression models, machine learning algorithms do not produce coefficients such as odds ratios. As a result, the precise strength and direction of associations between predictors and outcomes cannot be directly determined.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003eConclusion and implications of the study\u003c/h2\u003e \u003cp\u003eIn conclusion, the pooled prevalence of incomplete vaccination among children aged 12\u0026ndash;35 months in 16 sub‑Saharan African countries was 46.21%, with the lowest level observed in Ghana (25.21%) and the highest in the Democratic Republic of Congo (73.21%). Considerable heterogeneity was noted across countries, with 16 of the 13 analyzed reporting incomplete vaccination prevalence above 30%. CatBoost, XGBoost, and Random Forest were identified as the first, second, and third best-performing machine learning algorithms to predict incomplete childhood vaccination in sub-Saharan Africa, with CatBoost achieving the highest accuracy (65%) and AUC (70%). SHAP feature importance analysis revealed that adequate antenatal care visits, maternal media exposure, institutional delivery, being rural residence, health insurance coverage, married marital status, birth order between two and four, and household with high wealth index were the most influential attributes in predicting vaccination outcomes. Among these, antenatal care attendance and maternal media exposure emerged as the strongest protective determinants, while rural residence consistently increased the risk of incomplete vaccination.\u003c/p\u003e \u003cp\u003eThe findings of this study provide critical evidence to support policymakers and stakeholders in developing targeted childcare intervention mechanisms and early preparedness strategies to improve immunization coverage and reduce vaccine dropouts. The predictive rules generated by machine learning models contribute to knowledge creation and representation, offering practical insights for program design. Specifically, stakeholders are recommended to enhance mothers\u0026rsquo; antenatal care visits and institutional deliveries by expanding access to nearby health facilities, strengthening health insurance coverage, and improving maternal awareness through media outreach. Creating income opportunities and empowering mothers with health literacy are also essential interventions to ensure timely vaccination.\u003c/p\u003e \u003cp\u003eMoreover, this study serves as a reference for future research, highlighting the need for longitudinal designs to establish causal pathways and guide more effective policy strategies. By integrating maternal health services, socioeconomic support, and health system strengthening, countries in sub-Saharan Africa can make significant progress toward achieving Immunization Agenda 2030 targets and reducing the burden of incomplete childhood vaccination.\u003c/p\u003e \u003cp\u003eThe implications of this study extend beyond academic knowledge to practical policy and programmatic action. First, machine learning models can be integrated into national immunization monitoring systems to identify high-risk populations and guide resource allocation more efficiently. Second, health ministries and partners should prioritize maternal health service utilization and media-based health communication as rapid, scalable strategies to improve vaccination adherence. Third, addressing rural\u0026ndash;urban disparities through infrastructure development and transportation support are vital to reducing inequities in vaccine access. Finally, the study highlights the importance of combining individual-level interventions with structural reforms such as poverty alleviation, resilient vaccine supply chains, and political stability to achieve sustainable progress toward Immunization Agenda 2030 targets.\u003c/p\u003e \u003c/div\u003e"},{"header":"Abbreviations","content":"\u003cdiv id=\"AGS1\" class=\"AbbreviationGroupSection\"\u003e \u003cdiv class=\"Heading\"\u003eand Acronyms\u003c/div\u003e \u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eANC\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eAntenatal Care\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eAUC\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eArea Under the Curve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eCI\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eConfidence Interval\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eDHS\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDemographic Health Survey\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eDTP\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDiphtheria Pertussis Tetanus\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eIRB\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eInstitutional Review Board\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eML\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMachine Learning\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eNPV\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eNegative Predict Value\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003ePPV\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePositive Predict Value\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eROC\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eReceiver Operating Characteristic\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eSHAP\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eShapely Additive Explanation\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eSSA\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eSub-Saharan Africa\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003e\u003cb\u003eWHO\u003c/b\u003e\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eWorld Health Organization\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eClinical Trial Number\u003c/h2\u003e \u003cp\u003eNot applicable\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eConsent for publication\u003c/strong\u003e \u003cp\u003eNot Applicable\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eCompeting interests\u003c/strong\u003e \u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eNo specific fund was received for this work.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eA.H. conceptualization, formal analysis, investigation, methodology, writing original draft, review and editing. B.F. methodology, review and editing. A.A.A methodology, software, review and editing. A.K. review and editing. E.W. methodology, review and editing. K.G. methodology, software, review and editing. K.U. methodology, review and editing. A.H.S methodology, software, review and editing. A.A.B methodology, software, review and editing. H.A. methodology, software, review and editing.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eThe authors thank ICF International for granting access to the SSA DHS data set used in this study.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe datasets analyzed during the current study are publicly available from the Demographic and Health Survey (DHS) program, managed by ICF International. The data can be accessed upon reasonable request through the DHS program website at https://dhsprogram.comResearchers must register and submit a data access request through the site. For inquiries regarding data access, please contact the DHS Data Archivist at [email protected]. The data used in this study are cited as: ICF. Demographic and Health Surveys (various years). Funded by USAID. Rockville, Maryland. Available at: https://dhsprogram.com\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eImmunization WHO. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.afro.who.int/health-topics/immunization\u003c/span\u003e\u003cspan address=\"https://www.afro.who.int/health-topics/immunization\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 2024.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVaccination WHO. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/data/gho/data/themes/topics/sdg-target-3_b-development-assistance-and-vaccine-coverage\u003c/span\u003e\u003cspan address=\"https://www.who.int/data/gho/data/themes/topics/sdg-target-3_b-development-assistance-and-vaccine-coverage\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWHO. Immunization coverage. 2023. 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAfrica SA. S. Routine immunization. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.speakupafrica.org/wp-content/uploads/2019/05/Routine-Immunization.pdf\u003c/span\u003e\u003cspan address=\"https://www.speakupafrica.org/wp-content/uploads/2019/05/Routine-Immunization.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 2019.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChildhood MK. immunization. 2002. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://media.path.org/documents/CVP_Occ_Paper5.pdf\u003c/span\u003e\u003cspan address=\"https://media.path.org/documents/CVP_Occ_Paper5.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 2002.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA. C. Millions of African children still lack access to life-saving vaccines. 2025. 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFrenkel LD, editor. The global burden of vaccine-preventable infectious diseases in children less than 5 years of age: implications for COVID-19 vaccination. How can we do better? Allergy and asthma proceedings; 2021: OceanSide Publications, Inc.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCME, UUf. Level and trend in child mortality. 2024.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePAHO. Global childhood immunization levels stalled in 2023 leaving many without life-saving protection. PAHO. 2023. 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimon DJ, Tokpovi VCK, Dianou K, Okonji OC, Kiragu A, Olorunsaiye CZ et al. Regional, subregional and country-level full vaccination coverage in children aged 12\u0026ndash;23 months for 34 countries in sub-Saharan Africa: a global analysis using demographic and health survey data. BMJ Global Health. 2025;10(3).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTsegaw TK, Alemaw HB, Wale YB, Nigatu SG, Birhan TY, Taddese AA. Incomplete immunization uptake and associated factors among children aged 12\u0026ndash;23 months in sub-Saharan African countries; multilevel analysis evidenced from latest demography and health survey data, 2023. Ital J Pediatr. 2024;50(1):96.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHaeuser E, Byrne S, Nguyen J, Raggi C, McLaughlin SA, Bisignano C, et al. Global, regional, and national trends in routine childhood vaccination coverage from 1980 to 2023 with forecasts to 2030: a systematic analysis for the Global Burden of Disease Study 2023. Lancet. 2025;406(10500):235\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFenta SM, Biresaw HB, Fentaw KD, Gebremichael SG. Determinants of full childhood immunization among children aged 12\u0026ndash;23 months in sub-Saharan Africa: a multilevel analysis using Demographic and Health Survey Data. Trop Med health. 2021;49(1):29.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBobo F, Dawson MWAAA, Hayen A, editors. Monitoring Inequalities in Child Vaccination in Sub-Saharan Africa: Analysis of 25 National Health Surveys. 2023 World Congress on Health Economics. IHEA; 2023.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNegussie A, Kassahun W, Assegid S, Hagan AK. Factors associated with incomplete childhood immunization in Arbegona district, southern Ethiopia: a case\u0026ndash;control study. BMC Public Health. 2016;16:27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEze P, Agu UJ, Aniebo CL, Agu SA, Lawani LO, Acharya Y. Factors associated with incomplete immunisation in children aged 12\u0026ndash;23 months at subnational level, Nigeria: a cross-sectional study. BMJ open. 2021;11(6):e047445.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTsegaw TK, Alemu EA, Arage FG, Tadese ZB, Taye EA, Abate TG. Geographically weighted regression analysis of incomplete basic childhood vaccination in Sub-Saharan Africa: Evidence from DHS, 2019\u0026ndash;2024. PLoS ONE. 2025;20(11):e0336498.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGetnet M, Jejaw M, Belachew TB, Addis B, Dellie E, Tafere TZ, et al. Incomplete basic vaccination and associated factors among children aged 12\u0026ndash;23 months in resource-limited countries: a spatial and multilevel regression analysis of recent DHS data from 48 countries. Front Public Health. 2025;13:1463303.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTadese ZB, Nigatu AM, Yehuala TZ, Sebastian Y. Prediction of incomplete immunization among under-five children in East Africa from recent demographic and health surveys: a machine learning approach. Sci Rep. 2024;14(1):11529.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemsash AW, Chereka AA, Walle AD, Kassie SY, Bekele F, Bekana T. Machine learning algorithms\u0026rsquo; application to predict childhood vaccination among children aged 12\u0026ndash;23 months in Ethiopia: Evidence 2016 Ethiopian Demographic and Health Survey dataset. PLoS ONE. 2023;18(10):e0288867.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUSAID. Demographic and health survey(DHS) methodology. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://dhsprogram.com/methodology/Survey-Types/DHS-Methodology.cfm n.d\u003c/span\u003e\u003cspan address=\"https://dhsprogram.com/methodology/Survey-Types/DHS-Methodology.cfm n.d\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Available.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMelaku MS, Nigatu AM, Mewosha WZ. Spatial distribution of incomplete immunization among under-five children in Ethiopia: evidence from 2005, 2011, and 2016 Ethiopian demographic and health survey data. BMC Public Health. 2020;20(1):1362.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCooper H, Hedges LV, Valentine JC. The handbook of research synthesis and meta-analysis. Russell Sage Foundation; 2019.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrownlee J. Data preparation for machine learning: data cleaning, feature selection, and data transforms in Python. Machine Learning Mastery; 2020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKebede SD, Mamo DN, Adem JB, Semagn BE, Walle AD. Machine learning modeling for identifying predictors of unmet need for family planning among married/in-union women in Ethiopia: Evidence from performance monitoring and accountability (PMA) survey 2019 dataset. PLOS Digit Health. 2023;2(10):e0000345.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al. Scikit-learn: Machine learning in Python. J Mach Learn Res. 2011;12:2825\u0026ndash;30.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen T, Guestrin C, editors. Xgboost: A scalable tree boosting system. Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining; 2016.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLundberg SM, Erion G, Chen H, DeGrave A, Prutkin JM, Nair B, et al. From local explanations to global understanding with explainable AI for trees. Nat Mach Intell. 2020;2(1):56\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUddin S, Khan A, Hossain ME, Moni MA. Comparing different supervised machine learning algorithms for disease prediction. BMC Med Inf Decis Mak. 2019;19(1):1\u0026ndash;16.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuque A, Carrasco A, Mart\u0026iacute;n A, de Las Heras A. The impact of class imbalance in classification performance metrics based on the binary confusion matrix. Pattern Recogn. 2019;91:216\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLundberg S. A unified approach to interpreting model predictions. arXiv preprint arXiv:170507874. 2017.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Y, Liu Z, Luo X, Zhao H. Diagnosis of Parkinson's disease based on SHAP value feature selection. Biocybernetics Biomedical Eng. 2022;42(3):856\u0026ndash;69.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWorld Health Organization. Immunization coverage; Avaliable:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/news-room/fact-sheets/detail/immunization-coverage\u003c/span\u003e\u003cspan address=\"https://www.who.int/news-room/fact-sheets/detail/immunization-coverage\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSulgodu Ramachandra S, Webster P. USA's exit from the WHO and freeze on USAID funds globally: its perils and possible opportunities. J Public Health (Oxf). 2025;47(Supplement1):i16\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eImmunization Economics. US Government intends to end financial support for Gavi, threatening flagship vaccine programs. 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUnited Nations. Assessing the impact of the loss of the Demographic and Health Surveys on global population and family planning data and estimates and on population research 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRiaz A, Husain S, Yousafzai MT, Nisar I, Shaheen F, Mahesar W, et al. Reasons for non-vaccination and incomplete vaccinations among children in Pakistan. Vaccine. 2018;36(35):5288\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTsegaw TK, Alemu EA, Arage FG, Tadese ZB, Taye EA, Abate TG. Geographically weighted regression analysis of incomplete basic childhood vaccination in Sub-Saharan Africa: Evidence from DHS, 2019\u0026ndash;2024. PLoS ONE. 2025;20(11):e0336498.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFrancis M. Vaccination Coverage and Factors Associated with Routine Childhood Vaccination Uptake in India: Findings from a National Survey and Household Surveys in Vellore, Southern India. 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAtnafu Gebeyehu N, Abebe Gelaw K, Asmare Adella G, Dagnaw Tegegne K, Adie Admass B, Mesele Gesese M. Incomplete immunization and its determinants among children in Africa: Systematic review and meta-analysis. Hum vaccines immunotherapeutics. 2023;19(1):2202125.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBobo FT, Asante A, Woldie M, Dawson A, Hayen A. Child vaccination in sub-Saharan Africa: Increasing coverage addresses inequalities. Vaccine. 2022;40(1):141\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAdedokun ST, Uthman OA, Adekanmbi VT, Wiysonge CS. Incomplete childhood immunization in Nigeria: a multilevel analysis of individual and contextual factors. BMC Public Health. 2017;17(1):236.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThan SLMA, Laosee O. Determinants of incomplete immunization among hill tribe children aged under two years in Myanmar. 2016.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOgundele OA, Ogundele T, Fehintola FO, Fagbemi AT, Beloved OO, Osunmakinwa OO. Determinants of incomplete vaccination among children 12\u0026ndash;23 months in Nigeria: an analysis of a national sample. Tzu Chi Med J. 2022;34(4):448\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSetiawan MS, Wijayanto AW. Determinants of immunization status of children under two years old in Sumatera, Indonesia: A multilevel analysis of the 2020 Indonesia National Socio-Economic Survey. Vaccine. 2022;40(12):1821\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCausey K, Fullman N, Sorensen RJ, Galles NC, Zheng P, Aravkin A, et al. Estimating global and regional disruptions to routine childhood vaccine coverage during the COVID-19 pandemic in 2020: a modelling study. Lancet. 2021;398(10299):522\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang T, Qi X, He Q, Hee J, Takesue R, Yan Y, et al. editors. The effects of conflicts and self-reported insecurity on maternal healthcare utilisation and children health outcomes in the democratic republic of congo (Drc). Healthcare: MDPI; 2021.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOgundele OA, Ogunwemimo HS, Fehintola FO, Ogundele T, Olorunsola A, Bello OE, et al. Predictors of incomplete childhood vaccination in four West African countries: a population based cross-sectional study. Sci Rep. 2025;15(1):17119.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNigatu T, Abraham L, Willems H, Tilaye M, Tiruneh F, Gebru F, et al. The status of immunization program and challenges in Ethiopia: A mixed method study. Sage Open Med. 2024;12:20503121241237115.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDhar A, Dash NS, Roy K, editors. An innovative method of feature extraction for text classification using part classifier. International Conference on Information, Communication and Computing Technology; 2018: Springer.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMihret BG, Belay BN, Kelkay JM, Amare AM, Tarekegn AD, Asegu AA et al. Design a model to predict incomplete immunization among Ethiopian children using ensemble machine learning algorithms. Sci Rep. 2025.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTalukder A, Ahammed B. Machine learning algorithms for predicting malnutrition among under-five children in Bangladesh. Nutrition. 2020;78:110861.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemsash AW. Using best performance machine learning algorithm to predict child death before celebrating their fifth birthday. Inf Med unlocked. 2023;40:101298.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTesfa GA, Demeke AD, Hailegebreal S, Amede ES, Ngusie HS, Kasie MN, et al. Spatial distribution and associated factors of measles vaccination among children aged 12\u0026ndash;23 months in Ethiopia. A spatial and multilevel analysis. Hum Vaccines Immunotherapeutics. 2022;18(1):2035558.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDirirsa K, Makuria M, Mulu E, Deriba BS. Assessment of vaccination timeliness and associated factors among children in Toke Kutaye district, central Ethiopia: A Mixed study. PLoS ONE. 2022;17(1):e0262320.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMukungwa T. Factors Associated with full Immunization Coverage amongst children aged 12\u0026ndash;23 months in Zimbabwe. Afr Popul Stud. 2015;29(2).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTamirat KS, Sisay MM. Full immunization coverage and its associated factors among children aged 12\u0026ndash;23 months in Ethiopia: further analysis from the 2016 Ethiopia demographic and health survey. BMC Public Health. 2019;19(1):1019.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGualu T, Dilie A. Vaccination coverage and associated factors among children aged 12\u0026ndash;23 months in debre markos town, Amhara regional state, Ethiopia. Adv Public Health. 2017;2017(1):5352847.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTefera YA, Wagner AL, Mekonen EB, Carlson BF, Boulton ML. Predictors and barriers to full vaccination among children in Ethiopia. Vaccines. 2018;6(2):22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJejaw M, Demissie KA, Teshale G, Geberu DM, Tiruneh MG, Hagos A, et al. Estimated effect of media use on mothers\u0026rsquo; vaccination of their children in Sub-Saharan Africa: a quasi-experimental propensity score matching analysis using DHS data. BMC Public Health. 2025;25(1):2018.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMamo Habtie G, Melaku MS, Aychiluhm SB. A multilevel analysis of factors associated with unimmunized children in Ethiopia. Front Public Health. 2025;13:1433606.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChikako TU, Seidu A-A, Hagan JE Jr, Aboagye RG, Ahinkorah BO. Bayesian analysis of predictors of incomplete vaccination against polio among children aged 12\u0026ndash;23 months in Ethiopia. Int J Environ Res Public Health. 2021;18(22):11820.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRusso G, Miglietta A, Pezzotti P, Biguioh RM, Bouting Mayaka G, Sobze MS, et al. Vaccine coverage and determinants of incomplete vaccination in children aged 12\u0026ndash;23 months in Dschang, West Region, Cameroon: a cross-sectional survey during a polio outbreak. BMC Public Health. 2015;15(1):630.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTran BX, Boggiano VL, Nguyen LH, Latkin CA, Nguyen HLT, Tran TT et al. Media representation of vaccine side effects and its impact on utilization of vaccination services in Vietnam. Patient Prefer Adherence. 2018:1717\u0026ndash;28.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEndalamew SG, Gebeye E, Assefa SK, Geremew BM. Immunization Coverage and Its Determinants Among Children Aged 12\u0026ndash;23 Months in East Africa: A Bayesian Hierarchical Modeling Approach Based on DHS 2019\u0026ndash;2022. Biomed Res Int. 2025;2025(1):6645541.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRodrigues F, Ziade N, Jatuworapruk K, Caballero-Uribe CV, Khursheed T, Gupta L. The impact of social media on vaccination: a narrative review. J Korean Med Sci. 2023;38(40).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eObanewa OA, Newell ML. The role of place of residency in childhood immunisation coverage in Nigeria: analysis of data from three DHS rounds 2003\u0026ndash;2013. BMC Public Health. 2020;20(1):123.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDesalew A, Semahegn A, Birhanu S, Tesfaye G. Incomplete vaccination and its predictors among children in Ethiopia: a systematic review and meta-analysis. Global Pediatr health. 2020;7:2333794X20968681.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede Cantu\u0026aacute;ria Tauil M, Sato APS, Waldman EA. Factors associated with incomplete or delayed vaccination across countries: a systematic review. Vaccine. 2016;34(24):2635\u0026ndash;43.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSakou I-I, Tsitsika AK, Papaevangelou V, Tzavela EC, Greydanus DE, Tsolia MN. Vaccination coverage among adolescents and risk factors associated with incomplete immunization. Eur J Pediatrics. 2011;170(11):1419\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRieck T, Feig M, Eckmanns T, Benzler J, Siedler A, Wichmann O. Vaccination coverage among children in Germany estimated by analysis of health insurance claims data. Hum vaccines immunotherapeutics. 2014;10(2):476\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSakelo AN, Assefa N, Oljira L, Assefa ZM. Newborn care practice and associated factors among mothers of one-month‐old infants in Southwest Ethiopia. Int J Pediatr. 2020;2020(1):3897427.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTesfaye TD, Temesgen WA, Kasa AS. Vaccination coverage and associated factors among children aged 12\u0026ndash;23 months in Northwest Ethiopia. Hum vaccines immunotherapeutics. 2018;14(10):2348\u0026ndash;54.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAntai D. Regional inequalities in under-5 mortality in Nigeria: a population-based analysis of individual-and community-level determinants. Popul health metrics. 2011;9(1):6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMutua MK, Kimani-Murage E, Ettarh RR. Childhood vaccination in informal urban settlements in Nairobi, Kenya: who gets vaccinated? BMC Public Health. 2011;11(1):6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDana AD, Kebede DL, Betru KT. Prevalence of incomplete vaccination and associated factors among children aged 24\u0026ndash;35 months in Dale woreda, Sidama region, Ethiopia. Ethiop J Med Health Sci. 2022;2(2):136\u0026ndash;49.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLin A, Pham D, Rosenthal H, Milanaik R. Birth order and up-to-date vaccination status. Pediatrics. 2022;150(4).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTesema GA, Tessema ZT, Tamirat KS, Teshale AB. Complete basic childhood vaccination and associated factors among children aged 12\u0026ndash;23 months in East Africa: a multilevel analysis of recent demographic and health surveys. BMC Public Health. 2020;20(1):1837.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMbengue MAS, Sarr M, Faye A, Badiane O, Camara FBN, Mboup S, et al. Determinants of complete immunization among senegalese children aged 12\u0026ndash;23 months: evidence from the demographic and health survey. BMC Public Health. 2017;17(1):630.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChidiebere ODI, Uchenna E, Kenechi O. Maternal sociodemographic factors that influence full child immunisation uptake in Nigeria. South Afr J Child Health. 2014;8(4):138\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeters DH, Garg A, Bloom G, Walker DG, Brieger WR, Hafizur Rahman M. Poverty and access to health care in developing countries. Ann N Y Acad Sci. 2008;1136(1):161\u0026ndash;71.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSabahath F, Lewis P. Study of the immunization status and reasons for incomplete vaccination of children attending an urban hospital. Pediatr Infect Disease. 2022;4(3):92\u0026ndash;100.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFaisal S, Zahid FM, Kamal S, Shahzad K, Seidu A-A, Ahinkorah BO. Modeling the factors associated with incomplete immunization among children. Math Probl Eng. 2022;2022(1):8460837.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Incomplete vaccination, Children aged 12–35 months, Sub‑Saharan Africa, Machine learning algorithms","lastPublishedDoi":"10.21203/rs.3.rs-8945443/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8945443/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eDespite immunization preventing millions of deaths worldwide, sub‑Saharan Africa continues to face a heavy burden of vaccine‑preventable diseases, with nearly one in five children missing essential doses such as the third diphtheria, tetanus, and pertussis vaccine. Routine programs remain constrained by maternal, household, and contextual barriers, leaving the region far from achieving IA2030 targets. Earlier studies relied mainly on descriptive or regression‑based analyses using older datasets, limiting predictive accuracy. This study applies modern machine learning to recent DHS data (2021–2024), offering improved prediction and transparent identification of key determinants of incomplete vaccination.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eA secondary analysis was conducted using recent Demographic and Health Survey (DHS) data (2021–2024) from 16 sub‑Saharan African countries. The weighted sample comprised 57,527 children aged 12–35 months. Data cleaning, harmonization, and pooled analysis were performed in STATA 17, with forest plots illustrating pooled and country‑specific incomplete vaccination rates. Eight supervised machine learning algorithms; Naïve Bayes, Decision Trees, K‑Nearest Neighbor, Logistic Regression, Artificial Neural Networks, Extreme Gradient Boosting (XGBoost), CatBoost, and Random Forest were applied for classification and comparison. SHAP analysis enhanced interpretability by ranking maternal, household, and contextual predictors. All analyses were conducted in Python 3.10.2 within Google Colab using scikit‑learn, imblearn, XGBoost, CatBoost, and SHAP packages.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResult: \u003c/strong\u003eThe pooled prevalence of incomplete vaccination among children aged 12–35 months in 16 sub‑Saharan African countries was 46.21% (95% CI: 38.58, 53.83%), with the lowest level observed in Ghana (25.21%) and the highest in the Democratic Republic of Congo (73.21%). CatBoost emerged as the best‑performing machine learning algorithm for predicting incomplete childhood vaccination, achieving the highest accuracy (65%) and area under the curve (AUC (70%)) among the models tested. SHAP feature importance analysis revealed that adequate antenatal care visits, maternal media exposure, institutional delivery, being rural residence, health insurance coverage, married marital status, birth order between two and four, and household with high wealth index were the most influential attributes in predicting vaccination outcomes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion: \u003c/strong\u003eIn conclusion, this study reveals that nearly half of children aged 12–35 months in sub‑Saharan Africa remain incompletely vaccinated, with striking disparities across countries from Ghana’s relatively low 25.21% to the Democratic Republic of Congo’s alarming 73.21%. CatBoost achieved strong predictive accuracy and SHAP feature importance analysis revealed adequate antenatal care visits, maternal media exposure, institutional delivery, being rural residence, health insurance coverage, married marital status, and household with high wealth index were the most influential attributes in predicting vaccination outcomes. These findings underscore the urgent need for targeted interventions that strengthen maternal health services, expand access to facilities, and reduce rural urban inequities. Integrating AI‑driven monitoring into immunization programs offers policymakers actionable tools to accelerate progress toward Immunization Agenda 2030 and safeguard child health.\u003c/p\u003e","manuscriptTitle":"Prediction of incomplete vaccination among children aged 12-35 months in sub-Saharan Africa: Application of machine learning algorithm using recent (2021-2024) Demographic and Health Survey Data","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-10 13:24:35","doi":"10.21203/rs.3.rs-8945443/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"3b28fb9c-b48c-49ca-91b6-6960bddf2d43","owner":[],"postedDate":"March 10th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Rejected","date":"2026-04-30T06:41:08+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-30T06:56:20+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-10 13:24:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8945443","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8945443","identity":"rs-8945443","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}